Objective and subjective predictors of cancer-related stress symptoms in testicular cancer survivors

Quality of life of

testicular cancer survivors

Joke Fleer

FYN Werk-Fleer proefschrift.indb1 1 11-1-2006 23:43:44

Joke FleerQuality of life of testicular cancer survivorsThesis University of GroningenISBN 90-771-1340-1

Cover design STUDIO FRANK & LISA (www.studiofrank-lisa.nl)

Lay out FYN Werk (www.fynwerk.nl)

Printed by Drukkerij C. Regenboog Groningen

This research was supported by a grant from The Dutch Cancer Society, no. RUG 99-2130

Financial support for this thesis was kindly given byAstraZenecaBV; Fleremaheerd Erfgoedlogies (www.fleremaheerd.nl); Integraal Kanker-centrum Noord-Nederland; KWF Kankerbestrijding; Northern Centre for Healthcare Research Groningen; Stichting Werkgroep Interne Oncologie Groningen

©2006, Joke Fleer, Groningen, The NetherlandsAll rights reserved


Rijksuniversiteit Groningen

Quality of Life of

Testicular Cancer Survivors

Proefschrift

ter verkrijging van het doctoraat in de Medische Wetenschappen aan de Rijksuniversiteit Groningen op gezag van de Rector Magnificus, dr. F. Zwarts,

in het openbaar te verdedigen op woensdag 1 maart 2006 om 13.15 uur

door

Joke Fleergeboren op 10 september 1974 te Niehove


PromotoresProf. dr. H.J. Hoekstra

Prof. dr. D.Th. Sleijfer

Prof. dr. E.C. Klip

CopromotorDr. J.E.H.M. Hoekstra-Weebers

BeoordelingscommissieProf. dr. J.W. Groothoff

Prof. dr. J.M. Nijman

Prof. dr. M.A.G. Sprangers


Contents

1 Introduction 7

2 Quality of life of survivors of testicular germ cell cancer: a review of the literature 17

3 Quality of life of testicular cancer survivors and the relationship with sociodemographics, cancer-related variables and life events 41

4 The role of meaning in the prediction of psychosocial well-being of testicular cancer survivors 59

5 Objective and subjective predictors of cancer-related stress symptoms in testicular cancer survivors 81

6 Prevalence, changes in, and correlates of fatigue the first year after diagnosis of testicular cancer 103

7 Different recruitment methods identify different individuals 119

8 General discussion 133

9 Summary 143

10 Samenvatting 151

Dankwoord 159


1Introduction


1

9

Introduction

Only 25 years ago, disseminated testicular cancer was the leading cause of cancer death in men between 15 and 45 years of age. After the introduction of cisplatinum into the chemo-therapeutic regimen in the late 1970s, testicular cancer has become a curable disease for approximately 90% of the patients [1;2]. Because of the young age at diagnosis, testicular cancer survivors have an additional life expectancy of up to 50 years and, consequently, they have to face possible sequel of diagnosis and treatment for the rest of their lives. At the Uni-versity Medical Center Groningen (UMCG), in The Netherlands, a tertiary referral center for patients with testicular cancer, considerable institutional research has been performed into the epidemiology, genetic susceptibility, and short- and long-term medical and sexual sequel of testicular cancer [3-10]. However, the possible impact of the experience with tes-ticular cancer on the quality of life of survivors has not received empirical attention yet. This thesis addresses the quality of life of testicular cancer survivors treated at the UMCG between 1977 and 2003. This introductory chapter starts with an elaboration on the con-cept of quality of life and the possible effects of cancer on the quality of life of patients and survivors. After that, factors that may influence quality of life will be discussed. Lastly, the aim of the thesis and an overview of the chapters will be presented.

Quality of life

Calman [11] defines quality of life as the gap between the hopes, expectations, and desires of persons during a particular period of time and their present life experiences. In the same tenor, Cella [12] refers to quality of life as the patient’s appraisals of and satisfaction with their current level of functioning compared with what they perceive to be possible or ideal. The greater the gap between the actual and the ideal situation, the lower a person’s quality of life will be. The perception of a person’s quality of life varies between individuals. This means that people with different expectations will report a different quality of life, even when they have the same objective health status. Therefore, insight into a patient’s quality of life can only be obtained by asking a patient’s perspective. Quality of life encompasses several life domains, usually physical, psychological and social well-being [13-15]. Recently, investigators have started to include spirituality as a separate domain, because they have recognized that the experience of life as meaningful and the need to find meaning in life events may be important in determining a person’s quality of life [16;17].Ferrell and Dow [18] have defined the domains for cancer survivors as follows:

• Physical well-being is the control or relief of symptoms and the maintenance of function and independence.

• Psychological well-being is the attempt to maintain a sense of control in the face of life-threatening illness characterized by emotional distress, altered life priorities, and fear of the unknown as well as positive life changes.

• Social well-being is the effort to deal with the impact of cancer on individuals, their roles and relationships.


10

1

• Spiritual well-being is the ability to maintain hope and derive meaning from the cancer experience which is characterized by uncertainty.’

Research has found that sequel from cancer diagnosis and treatment may affect all these domains in survivors, months or even years after successful completion of treatment. For example, the cancer experience has been found to induce fatigue [19-21], infertility [22-25], cardiovascular diseases [26;27], posttraumatic stress disorder, and fear for recurrence and death [28-30], and to negatively affect social support systems [31], marital relationships [32], job performance [33-35], and outlook on life [36]. The specific combination of clinical features of testicular cancer may make the population of testicular cancer survivors especially vulnerable for an impaired quality of life. Testicular cancer strikes at an age when health is taken for granted and life-threatening illnesses and the possibility of dying do not fit their outlook on life [2;37;38]. It assaults an organ associated with sexuality and reproduction at a time of life when sexual desire and performance, sense of masculinity, body image, and fertility are central themes [5;39]. And it mostly occurs in a period of life characterized by major social life changes. Between the ages of 15 and 45, impor-tant decisions about marriage, starting a family and a professional career are generally made [14]. Thus, the life stage in which testicular cancer occurs has its specific developmental tasks and social roles, and therefore it may make testicular cancer a particularly distressing experi-ence. The confrontation with testicular cancer may continue to influence the well-being of the survivors months or even years after treatment completion, when sequel of the cancer experience continue to interfere with involvements in valued activities, interests, and desires.

Factors that may affect quality of life

The effect of a disease on people’s lives is generally examined by measuring the outcomes as well as the predictors of outcomes [17]. Quality of life as an outcome has been studied quite extensively in psychosocial oncology. Increasingly, research focuses on risk and resilience factors associated with quality of life. It is important to identify factors predictive of an impaired qua-lity of life, because it may help health care providers to detect distressed individuals at an early stage and refer them for psychosocial care [40]. Based on the stress-coping literature, Holland and colleagues have developed a research model for the field of psycho-oncology [16], in which they have identified variables that may influence quality of life. In this model, cancer is the stressful life event and the disease- and treatment-re-lated variables are the measurable characteristics of the event that may affect the patient’s func-tioning. The personal variables and life stresses represent the resources of a patient that may be relevant for the outcome of the adjustment process (e.g., quality of life). Figure 1 is a schematic representation of the model of Holland [16] and summarizes the specific variables that were identified as possibly relevant predictors of (domains of ) quality of life in the cancer popula-tion. The effect of these variables on quality of life will be investigated in the current thesis.

Disease and treatment variablesTime since treatment, type of treatment, and the experience of a second cancer event will be studied in the present thesis. There is some evidence that cancer patients who have been diagnosed more recently, who have been treated with more extensive treatment, or who


1

11

have experienced a recurrence are at a higher risk for an impaired quality of life and cancer-related stress [28;40]. Furthermore, hemoglobin and testosterone levels will be examined in relation to fatigue during the first year after the diagnosis of testicular cancer. Chemo-therapy can lower the levels of hemoglobin (mainly a short-term effect) and testosterone (mainly a long-term effect) which can cause excessive tiredness [27;41-43]. Associations between hemoglobin and testosterone levels and fatigue in testicular cancer patients have never received attention before.

Personal variablesSociodemographic characteristics (e.g., age, educational level, marital status, employment status) have been found to be associated to some extent with how well people adjust to ha-ving (had) cancer [40], and should thus be included in studies that aim to elucidate factors that affect quality of life. In general, having a sense of meaning has been identified as an important contributor to a person’s quality of life [16;17]. Stressful life events, such as testicular cancer, may threaten the belief that life is meaningful and this may have a negative effect on well-being. There-fore, it may be relevant to assess meaning and its relationship with psychosocial well-being and cancer-related distress in testicular cancer survivors. Trait anxiety will be investigated in relation to cancer-related fatigue. Little research has focused on this relationship, but there are indications that cancer patients with an anxious disposition pay more attention to physical sensations than less anxious patients [44;45]. So, it might be that this personality trait increases the sensitivity to fatigue.

Life stressesThe main focus of this thesis is on the quality of life of testicular cancer survivors who have been treated over a period of 25 years (1977-2003). This large range in time since treatment is associated with a methodological problem, because in long-term survivors it is more dif-ficult to distinguish effects caused by cancer and its treatment, from those attributable to other factors, such as additionally experienced life events and comorbidities [46]. The prob-ability of experiencing additional major life events and of developing a functional limitation or chronic disease increases with advancing age, and this may influence the quality of life more than the earlier experience with cancer.

Aims of the thesis

The general aims of the thesis are: 1. To study the quality of life of testicular cancer survivors; 2. To investigate the extent to which characteristics of the disease and treatment, personal

variables, and additional life stresses affect (domains of ) quality of life of testicular can-cer survivors.

Chapter 2 offers a review of the literature on the quality of life of testicular cancer survivors. The aims of this review are (1) to gain insight into the current state of knowledge on the physical, psychological, and social well-being of testicular cancer survivors, and (2) to assess the impact of disease and treatment characteristics on these domains of quality of life. The


12

1

study that is presented in chapter 3, describes the quality of life (physical, psychological and social domains) of testicular cancer survivors in comparison to a reference group of Dutch men. Furthermore, relationships between quality of life and disease-and treatment-related variables, sociodemographic characteristics, concurrent chronic diseases and recently expe-rienced life events are examined to identify testicular cancer survivors at risk for an impaired quality of life. In chapter 4, a study is presented that examines meaning among testicular survivors and describes the relative contribution of meaning in the prediction of psychoso-cial well-being and cancer-related distress, in addition to sociodemographics, disease- and treatment-related variables, concurrent chronic disease and recently experienced life events. The primary goal of the study described in chapter 5, is to investigate the prevalence of can-cer-related stress symptoms among survivors of testicular cancer. The second goal is to gain insight into the relationships between such stress symptoms and disease- and treatment-re-lated variables and sociodemographics, and to examine whether objective and subjective as-pects of cancer diagnosis and treatment are associated with posttraumatic stress symptoms. Chapter 6 describes the levels of and changes in fatigue among testicular cancer patients at three time points during the first year after diagnosis. Also, relationships between fatigue on the one hand and sociodemographics, hemoglobin- and testosterone levels, and trait anxiety on the other hand are examined. In chapter 7 testicular cancer survivors recruited through the UMCG database are compared with survivors recruited through a patient association on sociodemographic and disease- and treatment-related characteristics and quality of life to investigate heterogeneity and differences in the quality of life between indi-viduals recruited through different sources. Finally, in chapter 8, a discussion of the findings of the present study is presented.

Personal variablessociodemographic characteristics meaningtrait anxiety

Life stressesconcurrent chronic diseaserecently experienced life events

---

--

Disease and treatment variablestime since treatmenttype of treatmentsecond cancer eventhemoglobin leveltestosterone level

-----

Quality of life (multidimensional)Cancer-related stress symptoms MeaningSatisfaction with lifeFatigue (multidimensional)

Testicular cancer Outcome variablesRelated variables

Figure 1. Schematic representation of the model of research in psycho-oncology of Holland and colleagues [16]


1

13

Reference List

1. Bosl GJ, Motzer RJ (1997) Testicular germ-cell cancer. N.Engl.J Med. 337:242-253

2. Dearnaley D, Huddart R, Horwich A (2001) Regular review: Managing testicular cancer. BMJ 322:1583-1588

3. Nuver J (2005) Chemotherapy-related vascular toxicity - studies in germ cell tumour patients. Disserta-tion Rijksuniversiteit Groningen, available from URL http://irs.ub.rug.nl/ppn/275149285.

4. Pool G (2003) Surviving testicular cancer - sexuality and other existential issues. Dissertation Rijksuniver-siteit Groningen, available from URL http://irs.ub.rug.nl/ppn/255787200.

5. Van Basten JPA (1999) The sexual sequelae of testicular cancer. Dissertation Rijksuniversiteit Groningen, available from URL http://irs.ub.rug.nl/ppn/181455765.

6. Holzik MFL, Rapley EA, Hoekstra HJ, Sleijfer DT, Nolte IM, Sijmons RH (2004) Genetic predisposi-tion to testicular germ-cell tumours. Lancet Oncology 5:363-371

7. Holzik MFL, Sijmons RH, Sleijfer DT, Sonneveld DJA, Hoekstra-Weebers JEHM, Echten-Arends J, Hoekstra HJ (2003) Syndromic aspects of testicular carcinoma. Cancer 97:984-992

8. Sonneveld DJA (2002) Testicular germ cell tumours: new insights in epidemiology, genetic susceptibi-lity and outcome. Dissertation Rijksuniversiteit Groningen, available from URL http://irs.ub.rug.nl/ppn/238686795.

9. Holzik MFL, Nolte IM, Sijmons RH, van der Steege G, Sleijfer DT, Bruinenberg M, Hoekstra HJ (2005) Association analysis fails to confirm Xq27 as candidate region for a testicular germ cell tumour susceptibi-lity gene. Journal of Clinical Oncology 23:397S

10. Holzik MFL, Storm K, Sijmons RH, D’Hollander M, Arts EGJM, Verstraaten ML, Sleijfer DT, Hoek-stra HJ (2005) Absence of constitutional Y chromosome AZF deletions in patients with testicular germ cell tumors. Urology 65:196-201

11. Calman KC (1984) Quality of life in cancer patients - an hypothesis. Journal of Medical Ethics 10:124-127

12. Cella DF (1992) Quality of life: the concept. Journal of Palliative Care 8:8-13

13. Van Knippenberg FC, de Haes JC (1988) Measuring the quality of life of cancer patients: psychometric properties of instruments. Journal of Clinical Epidemiology 41:1043-1053

14. Vidrine DJ, Gritz ER (2003) Psychosocial outcomes after testicular cancer treatment. In: Steele jr. GD, Phillips TL, Chabner BA, Gansler TS (eds) Germ cell tumors - American Cancer Society atlas of clinical oncology. BC Decker Inc., Hamilton, London, pp 295-305

15. Ferrell BR, Dow KH, Grant M (1995) Measurement of the quality of life in cancer survivors. Quality of Life Research 4:523-531

16. Holland JC (2002) History of psycho-oncology: overcoming attitudinal and conceptual barriers. Psycho-somatic Medicine 64:206-221

17. Zebrack B (2000) Cancer survivors and quality of life: a critical review of the literature. Oncology Nursing Forum 27:1395-1401

18. Ferrell BR, Hassey Dow K (1997) Quality of life among long-term cancer survivors. Oncology 11:565-8, 571

19. Servaes P, Verhagen C, Bleijenberg G (2002) Fatigue in cancer patients during and after treatment: preva-lence, correlates and interventions. European Journal of Cancer 38:27-43

20. Tavio M, Milan I, Tirelli U (2002) Cancer-related fatigue (Review). International Journal of Oncology 21:1093-1099

21. Stasi R, Abriani L, Beccaglia P, Terzoli E, Amadori S (2003) Cancer-related fatigue - Evolving concepts in evaluation and treatment. Cancer 98:1786-1801


14

1

22. Schover LR (1999) Psychosocial aspects of infertility and decisions about reproduction in young cancer survivors: A review. Medical and Pediatric Oncology 33:53-59

23. Schover LR (1997) Sexuality and infertility after cancer. John Wiley and sons, Inc., New York

24. Schover LR, Rybicki LA, Martin BA, Bringelsen KA (1999) Having children after cancer. A pilot survey of survivors’ attitudes and experiences. Cancer 86:697-709

25. Wallace WH, Anderson RA, Irvine DS (2005) Fertility preservation for young patients with cancer: who is at risk and what can be offered? Lancet Oncology 6:209-218

26. Meinardi MT, Gietema JA, van der Graaf WT, van Veldhuisen DJ, Runne MA, Sluiter WJ, de Vries EG, Willemse PB, Mulder NH, van den Berg MP, Koops HS, Sleijfer DT (2000) Cardiovascular morbidity in long-term survivors of metastatic testicular cancer. Journal of Clinical Oncology 18:1725-1732

27. Nuver J, Smit AJ, Sleijfer DT, van Gessel AI, van Roon AM, van der Meer J, van den Berg MP, Hoekstra HJ, Sluiter WJ, Gietema JA (2005) Left ventricular and cardiac autonomic function in survivors of testicu-lar cancer. European Journal of Clinical Investigation 35:99-103

28. Kangas M, Henry JL, Bryant RA (2002) Posttraumatic stress disorder following cancer - A conceptual and empirical review. Clinical Psychology Review 22:499-524

29. Smith MY, Redd WH, Peyser C, Vogl D (1999) Post-traumatic stress disorder in cancer: a review. Psycho Oncology 8:521-537

30. Taieb O, Moro MR, Baubet T, Revah-Levy A, Flament MF (2003) Posttraumatic stress symptoms after childhood cancer. European Child & Adolescent Psychiatry 12:255-264

31. Helgeson VS, Cohen S (1996) Social support and adjustment to cancer: reconciling descriptive, correla-tional, and intervention research. Health Psychology 15:135-148

32. Manne S (1998) Cancer in the marital context: a review of the literature. Cancer Investigations 16:188-202

33. Abrahamsen AF, Loge JH, Hannisdal E, Holte H, Kvaloy S (1998) Socio-medical situation for long-term survivors of Hodgkin’s disease: a survey of 459 patients treated at one institution. European Journal of Cancer 34:1865-1870

34. Bradley CJ, Bednarek HL (2002) Employment patterns of long-term cancer survivors. Psycho Oncology 11:188-198

35. Short PF, Vasey JJ, Tunceli K (2005) Employment pathways in a large cohort of adult cancer survivors. Cancer 103:1292-1301

36. White CA (2004) Meaning and its measurement in psychosocial oncology. Psycho Oncology 13:468-481

37. Jones GY, Payne S (2000) Searching for safety signals: The experience of medical surveillance amongst men with testicular teratomas. Psycho Oncology 9:385-394

38. Van ’t Spijker A, Trijsburg RW, Duivenvoorden HJ (1997) Psychological sequelae of cancer diagnosis: a meta-analytical review of 58 studies after 1980. Psychosomatic Medicine 59:280-293

39. Tross S (1990) Psychological adjustment in testicular cancer. In: Holland JC, Rowland JH (eds) Hand-book of psychooncology. Oxford University Press, New York, pp 240-245

40. Parker PA, Baile WF, Moor CC, Cohen L (2003) Psychosocial and demographic predictors of quality of life in a large sample of cancer patients. Psycho Oncology 12:183-193

41. Pujade-Lauraine E, Gascon P (2004) The burden of anaemia in patients with cancer. Oncology 67:1-4

42. Nord C, Bjoro T, Ellingsen D, Mykletun A, Dahl O, Klepp O, Bremnes RM, Wist E, Fossa SD (2003) Go-nadal hormones in long-term survivors 10 years after treatment for unilateral testicular cancer. European Urology 44:322-328

43. Brennemann W, StoffelWagner B, Helmers A, Mezger J, Jager N, Klingmuller D (1997) Gonadal function of patients treated with cisplatin based chemotherapy for germ cell cancer. Journal of Urology 158:844-850


1

15

44. Stark D, Kiely M, Smith A, Morley S, Selby P, House A (2004) Reassurance and the anxious cancer patient. British Journal of Cancer 91:893-899

45. Cameron LD, Leventhal H, Love RR (1998) Trait anxiety, symptom perceptions, and illness-related re-sponses among women with breast cancer in remission during a tamoxifen clinical trial. Health Psychology 17:459-469

46. Gotay CC, Muraoka MY (1998) Quality of life in long-term survivors of adult-onset cancers. Journal of the National Cancer Institute 90:656-667


2Quality of life of survivors of

testicular germ cell cancer:

a review of the literature

Joke Fleer

Harald J Hoekstra

Dirk Th Sleijfer

Josette EHM Hoekstra-Weebers

Supportive Care in Cancer (2004) 12: 476-486


2

19

Introduction

Men between 15 and 45 years are in the prime of life. Major issues that concern them are career, interpersonal relationships and starting a family. Life-threatening illnesses and the possibility of dying do not fit their outlook in this period of life [1;2]. Nevertheless, a small percentage of these men will develop the most common neoplasm in young men: testicular germ cell cancer (TC). Despite increases in incidence, TC is an uncommon di-sease. It accounts for approximately 1% of all malignancies in men, although the incidence varies according to geographical area and race. Compared to other malignancies, the age distribution of TC is unusual in that the incidence declines with advancing age. Since the introduction of cisplatin-based polychemotherapy in the late 1970s, (disseminated) TC has become one of the most curable malignancies [3;4]. There are two histological types of TC: seminomatous and nonseminomatous tumors. Treatment for TC depends on histological type and stage of disease. Nowadays, treat-ment for stage I nonseminomatous disease consists of surgical removal of the affected testis (orchidectomy) and surveillance or nerve-sparing retroperitoneal lymph node dis-section. Initial treatment for disseminated disease (stages II-IV) is cisplatin-based poly-chemotherapy followed by resection of residual retroperitoneal or pulmonary tumor mass if necessary. Patients with a stage I and IIA/B seminoma are treated with orchidectomy and radiotherapy, while those with stages IIC-IV are treated with chemotherapy [4]. At present, almost 90% of testicular cancer patients can be cured with existing treatment modalities. Owing to the excellent prognosis, the young age and an increasing incidence, the number of testicular cancer survivors (TCSs) is growing. These men may have an additional life expectancy of perhaps 50 years after treatment and consequently they will have to face possible sequel of diagnosis and treatment for the rest of their lives [5]. The first reports that dealt with the quality of life (QoL, an umbrella term for physical, psy-chological and social well-being) of TCSs appeared a few years after the breakthrough in medical treatment. Since that time, studies have been performed on the consequences of the experience with TC for a variety of QoL domains. Unfortunately, a critical overview of the current state of knowledge on the QoL of survivors has not been published at this point in time, while the group of survivors is growing and the literature continuous to expand. A thorough review of the literature may guide clinicians when providing patients with information on possible short- and long term effects of the experience with TC on their lives, and it may provide insight into areas that need more research. The aims of this study were: 1) to review the literature on the QoL (physical, psychological and social well-being) of TCSs, and 2) to assess the impact of treatment-related characteristics, such as time since diagnosis and type of treatment, on the QoL of survivors.

Patients and methods

MEDLINE, EMBASE, PsycINFO and CancerLIT databases were used to identify relevant publications, as were the references of these papers. Keywords were ‘testicular cancer’, ‘testicular neoplasm’, ‘neoplasm’, ‘cancer’, ‘cancer survivors’, ‘survivors’ and ‘quality of life’. To encompass a wide range of QoL outcomes, we also used the following descrip-


2

20

tor terms: ‘depression’, ‘anxiety’, ‘distress’, ‘marital functioning’, ‘social functioning’ and ‘work-related problems’. Selection criteria were: 1) papers published in English between 1980 and October 2003, 2) studies on patients with TC who were in complete remission, 3) studies on physical, psychological and social well-being, and 4) papers in which sub-groups of TCSs could be clearly identified. Operationalisations of physical, psychological and social well-being were based on the literature on QoL in cancer survivors [6-11]. For physical well-being, the focal points were subjective perceptions of general health, fatigue, and body image. Physical side-effects of the treatment for TC were not included, because these aspects have been addressed in previous studies [4;12-14]. For psychological well-being, distress (including anxiety and depression), health worries and psychological well-being were chosen from the variety of possible operationalisations. In addition, distress about infertility due to treatment was evaluated. For social well-being, the focal points were marital functioning, social support and functional life. Although sexual well-being is also an important outcome variable, it was not included, because this issue has received extensive attention recently [15;16]. Methodological aspects were not used as an inclusion criterion, because the aim was to make an inventory of the literature on the QoL of TCSs. Studies with methodologi-cal shortcomings, such as non-standardized questionnaires and small sample sizes, may have detected relevant and valuable (site-specific) information that strong methodologi-cal studies overlooked. These studies may be of value in providing directions for future research. However, this decision not to use methodological aspects as selection criteria may decrease the power of the findings in the current study [17]. Therefore, it was decided to assess the quality of the studies first. There is no golden standard to determine which criteria should be used in quality assessments [18], but one critical issue is the internal validity of the studies, which is derived from procedures and design of the study. Fur-thermore, with regard to the research questions of this review, we considered treatment-related aspects to be important criteria as well. Therefore, the following methodological and treatment-related aspects were used for the quality assessment: design, sample size, use of comparison groups, measurement instruments, type of treatment, and time since diagnosis. It needs to be pointed out that results of individual studies involving comparison of groups or correlational analyses are only reported when they were statistically significant at a level of p ≤ .05 in the analyses.

Results

A total of 23 studies met the inclusion criteria. Table 1 shows the data on the aspects that were considered in the quality assessment. Quality assessment revealed that studies dif-fered greatly in quality. Seven studies were considered as stronger [5;19-24], because on the whole these studies used suitable designs, had adequate sample sizes, used validated questionnaires and comparison groups (with exception of the two prospective studies) [19;20], and had large ranges in time since treatment (with exception of the two prospec-tive studies). Furthermore, these studies considered time since treatment in their analyses, whereas in total, only nine studies investigated whether time since diagnosis was related to


2

21

one of the outcome variables [21-29]. None of these studies found a relationship between time since diagnosis and any of the dimensions of QoL. Five of the qualitatively stronger studies were published recently (between 2000 and 2003). However, the remaining two stronger studies were published in 1989, indicating that the poorer-quality studies were not necessarily the earlier studies.Because it is probable that the findings from the stronger-quality studies will have more power and would best answer the research questions of this review, it was decided to discuss the results of these studies first. The results of the remaining studies will be re-lated to these seven studies, and possible differences in results will be explained using the quality criteria. One last remark should be made about type of treatment. Eight retrospective/cross-sec-tional studies included both TCSs that had been treated before and after 1978 (when cis-platin-based chemotherapy was introduced) [24;25;27-32]. However, it was not possible to compare the two time periods, because the studies did not consider changes due to developments in treatment in their analyses. Furthermore, most studies only considered treatment-related characteristics as independent variables. Other factors that could pos-sibly influence the outcome, such as sociodemographic characteristics and the patient’s support system, were usually not considered. All factors reported on by the studies that have a significant effect on the outcome variable are reported in the current review.

Physical well-being

Health perceptionFour of the qualitatively stronger studies reported on health perception. Prospectively, it was shown that physical functioning of TCSs treated with chemotherapy recovered to baseline levels during the 2 years following diagnosis [20]. At time of study, over 70% of the TCSs assessed their general health as good [23], and perceived their health and physical functioning as equal to that of age-matched healthy men [22;23]. None of these qualitatively stronger studies compared treatment groups. The poorer-quality studies reported comparable findings. Several studies reported that TCSs experienced no change in health status compared to the pretreatment situation and between 63 and 90% assessed their general health as good [25;26;30;33;34]. Two studies found no differences in perceived health and physical functioning between TCSs and age-matched healthy men or men treated for Hodgkin’s disease [26;35]. However, one study that did not use a validated questionnaire reported that TCSs assessed their own health as better than that of controls [36]. No differences in physical well-being were found between men treated with different treatment modalities [30;34;36].

FatigueFatigue was studied relatively comprehensively by the qualitatively stronger studies. Pro-spectively, an improvement of fatigue was reported in TCSs treated with chemotherapy, although 19% experienced deterioration in fatigue 2 years after treatment [23]. Eighty-four percent experienced no chronic fatigue at time of study [21]. Two studies found that fatigue scores did not differ from those of age-matched healthy men or the general popu-lation [5;22], but Fossa et al. [21] reported that TCSs younger than 30 years of age expe-


2

22

rienced more fatigue than the same age cohort of the general population. Furthermore, it was found that TCSs who experienced more fatigue reported a poorer overall QoL [22]. Fatigue was associated with pretreatment distress, morbidity, lower educational level, old-er age, comorbidity, and higher levels of depression and anxiety [21]. Differences between treatment groups were not found [21, 22].Overall, results of the poorer-quality studies were in line with those of the qualitatively stronger studies. Two studies reported that most TCSs felt energetic (75% [33] and 82% [25]), although it took an average of nine months for energy levels to return to normal [33]. Compared to men treated for Hodgkin’s disease, energy levels returned to normal in more TCSs, and they reported less fatigue at the time of data collection [26;32]. Fatigue scores of TCSs did not differ from those of norm groups or age-matched healthy men [28;33;37]. In contrast with the qualitatively stronger studies, one study reported that men treated for seminoma experienced more fatigue than men treated for non-seminoma [33], but this last study had a much smaller sample size than the stronger studies [21;22].

Body imageThe loss of a testicle, which is an organ associated with masculinity and sexuality, may result in an impaired body image [38]. The body image of TCSs returned to normal in the course of time, and at the time of data collection, most TCSs in the qualitatively strong studies did not report feeling less attractive (77.3% [43] and 84.8% [23]) or less masculine than before TC (79%) [20]. Men who felt less attractive had lower scores in most health-related QoL domains, reported more fatigue, and were less satisfied with their family life than those who did not feel less attractive [5;23]. These results were supported by the poorer-quality studies, which found that the majority of TCSs had no changed body image (56-94%) [25;26;29;39]. In addition, these studies reported that the body image of TCSs did not differ from that of men treated for Hodgkin’s disease and their feelings of masculinity were equal to those of sociodemographically matched men [26;31]. None of the reviewed studies compared treatment modalities.

Psychological well-being

Psychological distressIn the qualitatively stronger studies, the levels of psychological distress experienced by TCSs varied between 9 and 27% [21;23;24]. One of these studies reported that TCSs as a group experienced significantly less distress than controls [24]. However, Fossa et al. [20] found that TCSs experienced more anxiety but less depression than the general popula-tion. One study found that a quarter of the TCSs became more anxious after diagnosis and treatment [24]. The contradictory results may be explained by the different aspects of distress addressed in these studies and the use of different validated questionnaires. None of the qualitatively stronger studies compared treatment modalities.A study of poorer quality reported that 13% of the TCSs experienced depression and 16% tension [25]. In retrospect, TCSs reported more anxiety, depression and distress dur-ing the first 6 months after diagnosis than male college students, but levels decreased over time [29;33;35;36]. One study stated that scores did not return to baseline levels, another study found that 12 months after diagnosis, TCSs treated with radiotherapy ex-


2

23

perienced less distress than male college students [25;35]. A significant number of TCSs experienced more depression since treatment [30]. Partly in contrast with the results of Fossa et al. [20], two poorer-quality studies using self-developed questionnaires, reported that TCSs experienced more anxiety, depression and psychosocial problems than age-matched men and controls [35;36]. Furthermore, three studies using validated question-naires but small sample sizes, found no differences in levels of anxiety and depression between TCSs and age-matched men, psychiatric patients, male college students, and other cancer patients [28;33;37]. In line with Rieker et al. [24], two studies reported that TCSs experienced less distress than healthy sociodemographically matched men, male college students and psychotherapy patients [26;33]. In addition, one study found that men treated with radiotherapy reported more depression, but two other studies did not show any differences between treatment groups [30;34;36]. The small sample sizes of the different treatment groups and the use of self-developed questionnaires may explain these contradictory results.

Health worriesSeventeen percent of the TCSs in the qualitatively stronger study of Fossa et al. [20] re-ported an increase of fear of recurrence 2 years after baseline measurement, whereas anxi-ety had decreased in 36%. Since their experience with TC, 19% of the TCSs reported more fear of dying, while 32% reported less fear [24]. None of the stronger studies reported on differences between treatment modalities in health worries.In concurrence with Fossa et al., [20] decreases in fear of recurrence (from 45 to 12%) and in fear of developing a second primary malignancy (from 50 to 21%) were reported in the poorer-quality studies, although health worries were still present in 54-76% of the TCSs [29;30;32-34]. TCSs who reported more anxiety and depression were more afraid of a tumor recurrence, spent more time worrying about their health and had more con-cerns about their health [29]. One study found no differences in health worries between men who received different treatment modalities; a second study reported that radio-therapy patients worried more about their health after treatment; a third study showed that men treated with chemotherapy experienced greater fear of a tumor recurrence than men treated with radiotherapy [30;33;34]. As mentioned earlier, a likely explanation for the contradictory results between studies that compared treatment groups are the small sample sizes and the use of nonvalidated questionnaires to assess health worries.

Fertility distressTC strikes men at an age when fertility is a major concern. Cancer treatment modalities are known to result in infertility [26]. Prospectively, it was found that fertility distress de-creased during the 2 years after treatment in 28% of the TCSs, but it increased in 11% [20]. Compared to sociodemographically matched healthy men, TCSs reported more overall problems with being infertile (5% compared to 22%) [24]. Infertility distress was reported most by TCSs with posttreatment ejaculatory dysfunction, those without children, of younger age, with a lower income, and those treated with chemotherapy and retroperito-neal lymph-node dissection (RPLND) [24;28]. Across the treatment groups, percentages of fertility distress varied from 11% in the radiotherapy group to 33% in the men treated with chemotherapy and RPLND [24].


2

24

The poorer-quality studies reported that during and after treatment, TCSs were more anxious about infertility than before diagnosis, but men who were infertile and those who were uncertain about their fertility status were not significantly more depressed and anxious than those who were fertile [29;40]. The majority of the 11 TCSs that were in-terviewed reported discomfort and regret about possible infertility [39]. Only one study reported on fertility distress in men treated for bilateral TC. Surprisingly, these men re-ported few problems with their fertility status, but the sample size was small and the study used one question only to assess fertility distress [32].

Psychological well-beingAlthough many QoL studies have shown that cancer and its treatment have detrimental effects on patient’s lives, there is a growing awareness that there may also be positive ef-fects. The prospective studies showed that the QoL of TCSs improved progressively after orchidectomy or chemotherapy [19;20]. At time of study, TCSs and age-matched men re-ported a similar QoL, and there were no differences between treatment modalities [22]. The results of the poorer-quality studies were rather positive as well. One study found that TCSs reported being more satisfied with their lives than age-matched men [36], and two studies reported a similar QoL in TCSs, and norm-group men, and controls [31;32]. Over 73% of the TCSs experienced good QoL after treatment [25]. A positive rather than negative impact was experienced by over half of the TCSs (exact percentages unknown), while 76% considered that surviving cancer was a worthwhile achievement [28]. Several positive effects were mentioned, including the ability to enjoy oneself, re-newed appreciation of life, emotional growth and the resetting of priorities and values [28;39]. The experience of cancer benefited the psychological well-being of most patients treated with chemotherapy (77%) and radiotherapy (82%) [34]. In one study men treated with radiotherapy or chemotherapy reported greater satisfaction with life than men in the surveillance group, but this result was not confirmed by another study [30;36]. Again, a probably explanation for the contradictory results are the small samples and the use of self-developed questionnaires.

Social well-being

Marital functioningTreatment for TC is known to have repercussions on marital functioning [41]. The quali-tatively stronger studies reported that most TCSs did not experience a change in the relationship with their partner after their experience with TC [23;24] and that treatment groups did not differ in their responses to questions about changes in the partner rela-tionship since treatment [19]. The subject of marital functioning has received much more attention in the poorer-qua-lity studies. Quite similar to the results of the stronger studies, most of the TCSs in these studies reported that the relationship with their partner had not changed or had become stronger after their experience with TC [29;30;34;39]. Anxiety about separation from the spouse was not a concern before, during, or after treatment in most TCSs (84-94%), while scores on spouse importance and quality of communication did not differ over the course of time. The degree of support from the spouse increased during treatment as compared


2

25

to the situation prior to diagnosis and decreased again afterwards [40]. In TCSs, of pa-tients who experienced a change in their partner relationship after TC, most married men indicated that it had strengthened (68%), whereas 74% of unmarried men reported that it had become strained [28]. TCSs appeared to be more satisfied with their partner relationship than sociodemographically matched men [31]. Treatment groups did not dif-fer in their responses to questions about changes in the partner relationship since treat-ment, although men treated with the most extensive treatment modality (radiotherapy, chemotherapy and/or RPLND) reported less satisfaction with family life than the other treatment groups [30]. This finding was in line with the qualitatively stronger study.Single TCSs may have difficulties starting a relationship not only on account of the above-mentioned problems with infertility, sexual functioning, and altered body image, but also due to the feeling of estrangement from their peers without a history of cancer. No qualitatively stronger studies considered this issue, and only one poorer-quality study reported on TCSs who were not involved with a partner or were not married at the time of their diagnosis and treatment. This study showed that two out of the ten patients who got married after treatment found that having had TC caused difficulties in the marital process, while 35% of the 28 men who were single at the time of data collection thought that the cancer experience would form a problem in planning a marriage [40].

Social supportOnly one qualitatively stronger study focussed on social functioning after TC. This study reported that TCSs experienced fewer changes in relationships with friends than socio-demographically matched healthy men [22]. The qualitatively poorer studies showed that the experience of TC did not result in changed relationships with family (66-93%) and friends (77-95%) [25;30;34]. Furthermore it was reported that 92% of the TCSs were satisfied with the support they received during treatment [29]. They mentioned more ‘enriching’ relationships with family and friends, although they had become more selec-tive [39]. One study showed no differences between treatment groups, but another study reported improved relationships with family and friends in patients treated with chemo-therapy [30;34].

Functional lifeLong-term side-effects of treatment (such as fatigue, psychosocial problems and fear of tumor recurrence), may cause difficulties in resuming work or study after treatment [42]. The qualitatively stronger studies reported that at the time of data collection, 76-90% of the TCSs were employed [22-24] and that their employment status was no different from that of age-matched healthy men [22]. Furthermore, Joly et al. [22] reported that TCSs and controls had similar problems in their professional lives, although TCSs were less ambitious than controls.The poorer-quality studies reported more extensively on functional life. These studies also showed that most TCSs were employed at the time of data collection (82-98%) [26-29;43]. Their employment status after treatment appeared to be the same as before [25;27;28;34] and it was better than that of men treated for Hodgkin’s disease [43]. Over 90% of the TCSs who returned to work perceived no effect on their career mobility, 45% experienced no effect on their ambition and career plans (while 26% reported a positive


2

26

effect), and 52% experienced no effect on their satisfaction with work (while 32% reported more satisfaction) [28]. In line with the results of Joly et al. [22], two studies found that work and ambition had become less important since TC, because men had reorganized their priorities and had adopted a new philosophy of working to live rather than living to work [27;39]. Compared to men treated for Hodgkin’s disease, the TCSs were better able to work at their former pace [26]. Compared to age-matched controls, TCSs felt less physically exhausted after a working day, and could maintain significantly better concen-tration and attention at work [36].Small percentages of TCSs reported negative effects of the cancer treatment on their work satisfaction (16%), career mobility (10%), ambitions and career planning (29%), and relationships with supervisors and co-workers (5-14%), while 12-25% were unable to work at their former pace, finish tasks and had problems with concentration [26-28;43]. In terms of treatment groups, between 12-31% of men treated with radiotherapy or chemo-therapy reported job loss and loss of job prospects because of their illness [30;34]. De-terioration in professional performance was reported by 17% of the radiotherapy group and 3% of the other treatment modalities [40]. Men in the surveillance group were less comfortable with their work than TCSs treated with chemotherapy or radiotherapy and had more problems with concentration and attention at work than TCSs treated with chemotherapy. However, another study found no differences between treatment groups [30;36]. Men treated with radiotherapy were more satisfied with their work than those treated with chemotherapy or those in the surveillance group and they felt that work was more worthwhile than those in the surveillance group [30]. There was no difference between treatment groups with respect to physical exhaustion after work [30;36]. All of these poorer quality studies had serious methodological shortcomings: the sample sizes of the total groups (only two studies included more than 100 TCSs), and the treatment groups were small, all studies used nonvalidated questionnaires or questionnaires of which the psychometric qualities were not reported, and no study compared TCSs with sociodemographically matched healthy men.


2

27

Tab

le 1.

Stu

dies

incl

uded

in th

is re

view

(alp

habe

tica

l ord

er) a

nd in

form

atio

n on

cri

teri

a us

ed fo

r qu

alit

y as

sess

men

t

Stud

y +

Des

ign

Subj

ects

Age

at t

ime

of

data

col

lect

ion

Mea

sure

men

t ins

trum

ents

Stag

e of

dis

ease

T

reat

men

t in

addi

tion

to

orc

hide

ctom

y T

ime

sinc

e di

agno

sis

Nye

ars

%%

year

s

Ara

i et a

l. (1

996)

[3

0]R

etro

spec

tive

8318

.0-6

0.0,

M=

33.8

SWL

(val

idat

ed);

self-

deve

lope

d an

d tr

ansla

ted

ques

tions

, bas

ed o

n St

uart

et

al.,

1990

[34]

; Kaa

sa e

t al.,

1991

[36]

; A

aron

son

et a

l., 19

87 [5

0] (p

sych

omet

ric

qual

ities

n.r.

)

I=57

.8II

=19

.3II

I=22

.9

Surv

eilla

nce=

8.4

RT

=50

.6C

T±

RR

TM

=41

.01.0

-21.8

, M=

8.0

Bloo

m e

t al.

(198

8) [4

3], (

1993

) [2

6]R

etro

spec

tive

Test

icul

ar: s

ee

Grit

z et

al.

[33]

Con

trol

: 85

See

Grit

z et

al.

[33]

Con

trol

: m

edia

n=28

.0

Bloo

m e

t al.

(198

8, 19

93):

wor

k-re

late

d qu

estio

ns d

evel

oped

by

Baro

fsky

(197

8)

[51]

(psy

chom

etric

qua

litie

s n.r.

)Bl

oom

et a

l. (1

993)

: PO

MS

and

CE

S-D

(s

tand

ardi

zed

ques

tionn

aire

s, pr

oven

ps

ycho

met

ric q

ualit

ies)

, SA

S (p

sych

o-m

etric

qua

litie

s n.r.

), se

lf-de

velo

ped

ques

tions

(psy

chom

etric

qua

litie

s not

as

sess

ed)

See

Grit

z et

al.

[33]

See

Grit

z et

al.

[33]

See

Grit

z et

al.

[33]

Con

trol

: med

ian=

3.0

Brod

sky

(199

5)

[39]

Ret

rosp

ectiv

e

Non

-sem

inom

a:

11 M

=35

.0Q

ualit

ativ

e in

terv

iew

with

ope

n-en

ded

ques

tions

n.r.

n.r.

>3

afte

r tre

atm

ent


2

28

Stud

y +

Des

ign

Subj

ects

Age

at t

ime

of

data

col

lect

ion

Mea

sure

men

t ins

trum

ents

Stag

e of

dis

ease

T

reat

men

t in

addi

tion

to

orc

hide

ctom

y T

ime

sinc

e di

agno

sis

Nye

ars

%%

year

s

Caff

o et

al.

(200

1)

[25]

Ret

rosp

ectiv

eSe

min

oma:

98

26.0

-85.

0,

med

ian=

48.0

Self-

deve

lope

d qu

estio

nnai

re (v

alid

ated

) n.

r.R

T=

100

1.25-

36.0

, med

ian=

10.3

afte

r tre

atm

ent

Cas

sile

th &

Ste

in-

feld

(198

6) [3

1]C

ross

-sec

tiona

l

Sem

inom

a: 3

9C

ontr

ol: 3

9

Sem

inom

a:

22.0

-67.0

, M=

41.0

Con

trol

: 25

.0-7

5.0,

M=

42.9

PO

MS;

self-

deve

lope

d qu

estio

ns (p

sy-

chom

etric

qua

litie

s n.r.

)I=

65.0

>I=

35.0

RT

=10

01.0

-20.

0, M

=7.3

Dou

chez

et a

l. (1

993)

[35]

Ret

rosp

ectiv

e

Non

-sem

inom

a:

109

Con

trol

: 107

Non

-sem

inom

a:

24.0

-66.

0, M

=39

.4C

ontr

ol: a

ge-

mat

ched

Self-

deve

lope

d qu

estio

ns (p

sych

omet

ric

qual

ities

n.r.

)

I=53

.2II

=39

.5II

I=7.3

RR

TM

=3.

7R

T+

CT

=9.

2R

T±

RR

TM

=4.

6C

T±

RR

TM

=65

.1R

T+

CT

+R

RT

M=

17.4

M=

9.1

Edb

ril &

Rie

ker

(198

9) [2

7], R

ieke

r et

al.

(198

5) [2

8]C

ross

-sec

tiona

l

7419

.0-5

9.0,

m

edia

n=30

PO

MS;

CP

BS

(abb

revi

ated

; psy

chom

et-

ric q

ualit

ies n

.r.);

BM

S; se

lf-de

velo

ped

ques

tions

(psy

chom

etric

qua

litie

s n.r.

)n.

r.

RT

=8.

0C

T=

27.0

RR

TM

=16

.0R

T+

CT

+R

RT

M=

50.0

2.0-

10.0

, med

ian=

4.0

afte

r tre

atm

ent

Tab

le 1

cont

inue

d


2

29

Stud

y +

Des

ign

Subj

ects

Age

at t

ime

of

data

col

lect

ion

Mea

sure

men

t ins

trum

ents

Stag

e of

dis

ease

T

reat

men

t in

addi

tion

to

orc

hide

ctom

y T

ime

sinc

e di

agno

sis

Nye

ars

%%

year

s

Foss

a et

al.

(199

9)

[32]

Cro

ss-s

ectio

nal

Bila

tera

l: 43

31.0

-75

med

ian=

41.0

EO

RT

C Q

LQ C

-30;

IES;

GH

Q-2

8 (s

tand

ardi

zed

ques

tionn

aire

s, pr

oven

ps

ycho

met

ric q

ualit

ies)

1st-

2nd

orch

idec

tom

yI=

74.4

-81.5

II=

9.3-

11.6

III=

7.0-2

.3IV

=9.

3-4.

7

1st-

2nd

orch

idec

tom

ysu

rvei

llanc

e=16

.3-6

7.8R

T=

58.1-

18.6

RR

TM

=11.

6-0.

0C

T±

othe

rs=

14.0

-18.

6

n.r.

Foss

a et

al.

(200

3)

[20]

Pros

pect

ive

666

16.0

-63.

0, M

=31

.0E

OR

TC

QLQ

C-3

0 +

TC

mod

ule

>I

CT

=10

0A

fter o

rchi

dect

omy,

3 m

onth

s, 6

mon

ths,

1 yea

r, 2

year

s

Foss

a et

al.

(200

3)

[21]

Cro

ss-s

ectio

nal

Test

icul

ar: 7

91C

ontr

ol 1:

706

Con

trol

2: 2

7083

Test

icul

ar: 2

3.0-

75.0

, m

edia

n=44

.0C

ontr

ol 1:

<75

.0C

ontr

ol 2

: 19-

74

HA

DS;

FQ

(sta

ndar

dize

d qu

estio

n-na

ires,

prov

en p

sych

omet

ric q

ualit

ies)

n.r.

Surv

eilla

nce

±R

RT

M=

20.0

RT

=41

.0C

T±

othe

rs=

39.0

4.0-

21.0

, m

edia

n=12

.0

Grit

z et

al.

(198

8)

[33]

Cro

ss-s

ectio

nal

8818

.0-6

0.0,

M=

35.2

PO

MS;

CE

S-D

; sel

f-de

velo

ped

ques

-tio

ns (p

sych

omet

ric q

ualit

ies n

.r.)

I=48

.0II

=37

.0II

I=15

.0

RT

=42

.0C

T=

47.0

RR

TM

=50

.0

1.0-7

.5, M

=3.

8 af

ter t

reat

men

t

Grit

z et

al.

(199

0)

[37]

Cro

ss-s

ectio

nal

34M

=37

.7P

OM

S; C

ES-

D; O

LQ (s

ense

of c

oher

-en

ce);

FES

(sta

ndar

dize

d qu

estio

nnai

res,

prov

en p

sych

omet

ric q

ualit

ies)

n.r.

Surv

eilla

nce=

2.9

RT

=44

.1C

T=

8.8

RR

TM

=20

.6R

T+

CT

=2.

9C

T+

RR

TM

=20

.6

1-7.5

, M=

4 af

ter

trea

tmen

t

Tab

le 1

cont

inue

d


2

30

Stud

y +

Des

ign

Subj

ects

Age

at t

ime

of

data

col

lect

ion

Mea

sure

men

t ins

trum

ents

Stag

e of

dis

ease

T

reat

men

t in

addi

tion

to

orc

hide

ctom

y T

ime

sinc

e di

agno

sis

Nye

ars

%%

year

s

Joly

et a

l. (2

002)

[2

2]C

ase-

cont

rol

Test

icul

ar: 7

1C

ontr

ol: 1

19

Test

icul

ar:

29.0

-67.0

, M=

47.0

Con

trol

: 29

.0-6

7.0, M

=48

EO

RT

C Q

LQ C

-30;

SF-

36 (s

tand

ard-

ized

que

stio

nnai

re, p

rove

n ps

ycho

met

ric

qual

ities

)

I=64

.8>

I=35

.2

Surv

eilla

nce=

9.0

RT

=48

.2C

T±

RR

TM

=36

.8R

T+

CT

=5.

9R

RT

M=

19.7

5.0-

20.0

, M=

11.0

Kaa

sa e

t al.

(199

1)

[36]

Cro

ss-s

ectio

nal

Test

icul

ar: 1

49C

ontr

ol 1:

627

7C

ontr

ol 2

: 163

8C

ontr

ol 3

: 520

Test

icul

ar: 1

7-64

, M

=34

Con

trol

gro

ups:

age-

mat

ched

HSC

L (s

tand

ardi

zed

ques

tionn

aire

, pr

oven

psy

chom

etric

qua

litie

s); s

elf-

deve

lope

d qu

estio

ns (p

sych

omet

ric

qual

ities

n.r.

)

M=

2.0

I=51

.0II

=32

.2II

I=2.

7IV

=12

.1

RT

=26

.2R

RT

M=

21.5

CT

+R

RT

M=

30.8

CT

+R

T±

RR

TM

=21

.5

Tota

l: 3.

0-9.

0, M

=5.

1

Moy

niha

n (1

987)

[2

9]R

etro

spec

tive

9715

.0-≥

40.0

PSE

(sta

ndar

dize

d in

stru

men

t)I=

43.1

>I=

56.9

Surv

eilla

nce=

19.0

RT

=4.

0C

T=

5.0

RT

+C

T=

31.0

CT

+su

rger

y=29

.0R

T+

CT

+su

rger

y=12

.0

Up

to 5

.0

Tab

le 1

cont

inue

d


2

31

Stud

y +

Des

ign

Subj

ects

Age

at t

ime

of

data

col

lect

ion

Mea

sure

men

t ins

trum

ents

Stag

e of

dis

ease

T

reat

men

t in

addi

tion

to

orc

hide

ctom

y T

ime

sinc

e di

agno

sis

Nye

ars

%%

year

s

Oze

n et

al.

(199

8)

[40]

Cro

ss-s

ectio

nal

140

18.0

-51.0

, m

edia

n=32

.0G

HQ

-28;

self-

deve

lope

d qu

estio

ns

(psy

chom

etric

qua

litie

s n.r.

)

I=43

.6II

=40

.0II

I=3.

6IV

=12

.8

Surv

eilla

nce=

15.7

RT

=16

.4C

T=

48.6

RT

+C

T=

2.9

CT

+po

st-C

T su

r-ge

ry=

16.4

1.2-8

.0, m

edia

n=4.

5

Rie

ker e

t al.

(198

9)

[24]

Ret

rosp

ectiv

e

Test

icul

ar: 2

23C

ontr

ol: 1

20

Test

icul

ar:

17.0

-65.

0,

med

ian=

33.0

Con

trol

: 19.

0-60

.0,

med

ian=

30.5

See

Edb

ril a

nd R

ieke

r (19

89) a

nd R

ieke

r et

al.

(198

5)I=

32.0

>I=

68.0

RT

=32

.0C

T=

9.0

RR

TM

=17

.0C

T+

RR

TM

=33

.0R

T+

surg

ery±

CT

=9.

0

<2:

18.0

2-5:

51.0

≥6:

32.

0

Rud

berg

et a

l. (2

000)

[23]

Ret

rosp

ectiv

e

Test

icul

ar: 2

77C

ontr

ol: 1

22

Test

icul

ar: 1

8-83

, M

=42

.2C

ontr

ol: 2

5-54

, M

=39

SWE

DQ

UA

L (s

tand

ardi

zed

ques

tion-

naire

, pro

ven

psyc

hom

etric

qua

litie

s);

self-

deve

lope

d qu

estio

ns (p

sych

omet

ric

qual

ities

n.r.

)

I=45

.1>

I=54

.9

Surv

eilla

nce=

8.0

RT

=37

.1C

T=

9.8

RR

TM

=11.

9C

T+

RR

TM

=27

.8R

T+

CT

±R

RT

M=

5.4

3-13

, M=

7.8

Tab

le 1

cont

inue

d


2

32

Pts =

pat

ients;

CT

= ch

emot

hera

py; R

T =

radi

othe

rapy

; RRT

M =

rese

ction

of r

esid

ual t

umor

mas

s (in

cludi

ng re

trope

riton

eal l

ymph

nod

e diss

ectio

n an

d re

secti

on o

f pul

mon

ary

tum

or

mas

s); M

= m

ean,

n.r.

= n

ot re

porte

d in

pub

licat

ion.

Abb

revi

atio

ns o

f ins

trum

ents:

SW

L: S

atisf

actio

n wi

th L

ife; P

OM

S: P

rofil

e of M

ood

Stat

es; C

ES-D

: Cen

tre fo

r Epi

dem

iolog

ical S

tudi

es-D

epre

ssion

Sca

le; S

AS:

Soc

ial A

ctivi

ties S

cale;

C

PBS:

Can

cer P

atien

t Beh

avio

r Sca

le; B

MS:

Bra

nnon

Mas

culin

ity S

cale;

EO

RTC

QLQ

-C30

: The E

urop

ean

Org

aniz

atio

n for

Res

earc

h an

d T

reat

men

t Qua

lity

of L

ife Q

uesti

onna

ire-

C30

: IES

: Im

pact

of E

vent

s Sca

le; G

HQ

-28:

Gen

eral

Hea

lth Q

uesti

onna

ire-2

8; H

AD

S: H

ospi

tal A

nxiet

y an

d D

epre

ssion

Sca

le; F

Q: F

atigu

e Que

stion

naire

; OLQ

: Orie

ntat

ion

to L

ife

Que

stion

naire

; FES

: Fam

ily E

nviro

nmen

t Sca

le; S

F-36

: Med

ical O

utco

me S

tudy

36-it

em S

hort

Form

Hea

lth S

urve

y; H

SCL:

Hop

kin

Sym

ptom

Che

cklis

t; PS

E: P

rese

nt S

tate

Exam

inat

ion;

SW

EDQ

UA

L: S

wedi

sh H

ealth

Rela

ted Q

ualit

y of L

ife Q

uesti

onna

ire; G

QL:

Got

henb

urg Q

ualit

y of L

ife In

strum

ent

Stud

y +

Des

ign

Subj

ects

Age

at t

ime

of

data

col

lect

ion

Mea

sure

men

t ins

trum

ents

Stag

e of

dis

ease

T

reat

men

t in

addi

tion

to

orc

hide

ctom

y T

ime

sinc

e di

agno

sis

Nye

ars

%%

year

s

Rud

berg

et a

l. (2

002)

[5]

Ret

rosp

ectiv

e

Test

icul

ar: 2

77C

ontr

ol: 3

92

Test

icul

ar: 1

8-83

, M

=42

.2C

ontr

ol: M

=45

GQ

L (o

nly

sym

ptom

che

cklis

t; re

liabl

e, va

lidity

n.r.

); se

lf-de

velo

ped

ques

tions

(p

sych

omet

ric q

ualit

ies n

.r.)

See

Rud

berg

et a

l. [2

3]Se

e R

udbe

rg e

t al.

[23]

See

Rud

berg

et a

l. [2

3]

Stua

rt e

t al.

(199

0)

[34]

Ret

rosp

ectiv

e62

CT

: 19-

48, m

e-di

an=

29R

T: 2

0-48

, me-

dian

=33

Self-

deve

lope

d qu

estio

ns (p

sych

omet

ric

qual

ities

n.r.

)

I/II

A=

74.2

IIB

/C=

11.3

III=

3.2

IV=

11.3

RT

=54

.8C

T=

45.2

CT

: 0.5

-2.7,

M=

2.7

afte

r tre

atm

ent

RT

: 0.6

-5.6

, M=

2.4

afte

r tre

atm

ent

Tam

burin

i et a

l. (1

989)

[19]

Pros

pect

ive

3118

-50,

M=

29.3

Spitz

er Q

L in

dex

(sta

ndar

dize

d qu

es-

tionn

aire

, pro

ven

psyc

hom

etric

qua

litie

s);

self-

deve

lope

d qu

estio

ns (p

sych

omet

ric

qual

ities

n.r.

)

I=16

.1>

I=83

.9

RT

=16

.1C

T=

41.9

RR

TM

=19

.4C

T+

surg

ery=

22.6

0.25

, 0.5

and

1.5

Tab

le 1

cont

inue

d


2

33

Discussion

The goal of this study was to give a comprehensive review of the literature of primary studies examining physical, psychological and social well-being of TCSs. A total of 23 studies, published between 1985 and 2003, met the inclusion criteria. Because these studies used such a broad variety of methodologies and research questions, this re-view was by necessity descriptive. However, a quality assessment was performed to interpret and explain contradictory results and to increase the power of this review. Based on methodological and treatment-related criteria, seven studies appeared to be qualitatively stronger.Results of both stronger- and poorer-quality studies appeared quite similar. Prospec-tive and retrospective studies showed that QoL after completion of treatment increased and that the negative consequences of TC on life decreased compared to the situation directly after diagnosis when a poorer QoL was observed. For example, energy levels returned to normal, fatigue decreased and the TCSs became used to their changed body image. Decreases were found in psychological distress, anxiety, depression, and fear of tumor recurrence. The stronger-quality studies paid little attention to the social dimension (marital functioning, social support, and functional life) and positive effects of the experience with TC. Poorer-quality studies reported that, besides a decrease in negative consequences of TC on the patients’ lives, the experience of TC seemed to have positive effects. A considerable proportion of the TCSs reported that they had experienced emotional growth, were appreciating life more, and had stronger relation-ships with family and friends. Results on functional life were mainly contradictory.Based on this summary of the results one could assume that in general TCSs expe-rience a good QoL. However, this could be an arguable assumption because of the methodological shortcomings of most studies and because the stronger-quality studies did not investigate all dimensions of QoL with the same extensiveness. Below we dis-cuss various methodological issues.Firstly, many studies used a retrospective design to evaluate the well-being of TCSs and to compare their well-being before and/or after treatment. Data obtained in re-trospect are less reliable for two reasons: [2] the chance is great that recall was distorted (selective memory bias), because 1-36 years had passed between data collection and diagnosis [17;23;28;44] under the influence of certain life events (such as experience with cancer) internal standards can change (response shift) and cause people to evalu-ate new situations according to other standards [45]. For example, people can become accustomed to a higher level of fatigue and after a certain amount of time, report the same QoL as that prior to the illness [46;47]. Such changes in internal standards cannot be reproduced accurately in retrospective studies. The only way that conclu-sions can be drawn from cross-sectional or retrospective studies about the relationship between TC and QoL is to compare the current QoL of TCSs to that of a sociodemo-graphically-matched group.Secondly, a considerable number of studies had no comparison group. Some studies used norm groups, but these often differed from TC patients in sociodemographic areas. In studies that compared TCSs to other cancer patients, it often appeared that


2

34

the other cancer patients had a poorer prognosis and had received other forms of treatment.Thirdly, a wide variety of both validated and non-validated measurement instruments were used to evaluate the same aspects of QoL, which made it difficult to compare results. Therefore, some differences in findings may be due to the different instruments that were used, even if they were validated measures. In the studies that used non-vali-dated instruments, some variables (e.g., body image) were measured with only one or two items. This affected the reliability of study results, and it is not clear whether their scope is of sufficient depth [48]. Fourthly, descriptive and comparison analysis methods were generally used, whereas techniques such as regression analyses would have enabled the identification of pos-sible risk groups. Fifthly, in a large proportion of the studies, reliability and generalizability were affected by small numbers (fewer than 100 TCSs). A small study group means that there is insufficient power to determine the possible influence of demographic and treatment-related variables [22]. Sixthly, the majority of studies did not correct for confounding variables such as age, marital status, education level, time since diagnosis, treatment modality and disease stage, although they aimed to measure the negative impact of illness or treatment on QoL. To identify possible groups of men, who are at risk of developing problems, it is essential to control for demographic and treatment-related variables.In addition to these methodological issues, the studies showed limitations in how they applied the treatment-related variables. Only nine studies considered time since di-agnosis in their analyses, and the categorizations of treatment modalities were rough. None of the studies reflected on changes in treatment over time, which may have caused a threat to the internal validity. From a historical point of view the differences in methodologies between studies and the methodological shortcomings of most studies are well explainable. The earlier stu-dies had to use explorative designs, self-developed and non-validated questionnaires and could only focus on short-term sequel of survivorship. In contrast, the more recent studies could make use of well-validated and reliable instruments developed over the years. Furthermore, the number of survivors has greatly increased due to advances in medical treatment, making methodological stronger studies possible. Therefore, it was not surprising that five of the seven methodological stronger studies were published between 2000 and 2003, although the poorer-quality studies were not necessarily the earlier studies. A problem of studies with methodological shortcomings is that their results may not always concur with reality. This can be illustrated with three examples from this re-view. Firstly, it would have been reasonable to expect that more extensive treatment would cause more short- and long-term physical and psychosocial side effects. It may be that the lack of documented effect of treatment reflects inadequate methodology, lack of adequately powered and comprehensive studies and the rough categorizations of treatment modalities. A second example that illustrates the distorting effect of methodological shortcomings on the outcomes of this review is the wide variety of


2

35

percentages of TCSs reporting negative effects between studies. For example, the level of psychological distress varied between 8% and 27%, while 5% to 29% were experi-encing work-related problems since TC. A third example is the result that men with bilateral TC have less fertility issues. There is only one study that states this, and given the lack of biological rationale for this observation, it may be regarded as casuistry and not as a reflection of reality. As regard to content, more research is needed as well. The qualitatively stronger stu-dies did not exhaustively delve into any of the outcome variables, indicating that more research is needed on all dimensions of QoL but most of all in the area of social well-being and the impact of treatment-related characteristics. The poorer-quality studies provided some directions for future research. In particular, issues that may be of major concern for young men dealing with sequel of cancer diagnosis and treatment, such as body image, fertility issues, intimate relationships, social support and work-related is-sues need more attention. Surprisingly few studies attacked these issues, and all these studies suffered from methodological shortcomings (non-validated questionnaires and small sample sizes). For example, the effects of TC on casual relationships and on men who did not have a partner at the time of TC have received very little attention. One study showed that TC had a negative effect on casual relationships, whereas the relationship of many married men with their spouse became stronger. The only study that investigated whether men who were single at the time of diagnosis experienced problems with starting a new relationship found that 20% of men reported problems. It would seem worthwhile to conduct more research into the effects of TC on casual relationships and on single men, because TC particularly affects young men who are more likely to be single or have a relationship without being married.

Future directives

In conclusion, because of the above-mentioned methodological shortcomings, it would not be justified to provide a general conclusion on the QoL of TCSs. In order to ob-tain such data and to gain clearer insight into the impact of TC on later QoL, more methodological strong, cross-sectional, and prospective research is needed. The studies should include large groups of patients and sociodemographically matched compari-son groups, and the analyses should aim to identify subgroups that are at risk of de-veloping psychosocial problems. To identify risk groups, it is necessary to establish the predictive value of, for example, demographic variables (age, marital status, education level), treatment-related variables (treatment modalities, time since diagnosis, disease stage, experience of a second cancer event), social support, and personality factors. In addition, on a content level, future research should pay more attention to body image, fertility distress, issues concerning intimate relationships, social support and work-related problems. Furthermore, more attention should be paid to possible long-term side-effects of treatment with radiotherapy or chemotherapy, such as an increased risk of secondary leukemia, second primary malignancies, decreased renal functioning, hearing deficits and cardiovascular disease [14;49]. As knowledge of the long-term


2

36

side-effects of cancer treatment has only started to increase fairly recently, very little research has been performed as yet into their possible impact on QoL.


2

37

Reference List


2. Van ’t Spijker A, Trijsburg RW, Duivenvoorden HJ (1997) Psychological sequelae of cancer diagnosis: a meta-analytical review of 58 studies after 1980. Psychosomatic Medicine 59:280-293


4. Bosl GJ, Motzer RJ (1997) Testicular germ-cell cancer. New England Journal of Medicine 337:242-253

5. Rudberg L, Carlsson M, Nilsson S, Wikblad K (2002) Self-perceived physical, psychologic, and general symptoms in survivors of testicular cancer 3 to 13 years after treatment. Cancer Nursing 25:187-195

6. Wyatt G, Friedman LL (1996) Long-term female cancer survivors: quality of life issues and clinical implica-tions. Cancer Nursing 19:1-7

7. Bush NE, Haberman M, Donaldson G, Sullivan KM (1995) Quality of life of 125 adults surviving 6-18 years after bone marrow transplantation. Social Science & Medicine 40:479-490

8. Holzner B, Kemmler G, Kopp M, Moschen R, Schweigkofler H, Dunser M, Margreiter R, Fleischhacker WW, Sperner-Unterweger B (2001) Quality of life in breast cancer patients--not enough attention for long-term survivors? Psychosomatics 42:117-123

9. Joly F, Henry-Amar M, Arveux P, Reman O, Tanguy A, Peny AM, Lebailly P, Mace-Lesec’h J, Vie B, Genot JY, Busson A, Troussard X, Leporrier M (1996) Late psychosocial sequelae in Hodgkin’s disease survivors: a French population-based case-control study. Journal of Clinical Oncology 14:2444-2453

10. Lasry JC, Margolese RG, Poisson R, Shibata H, Fleischer D, Lafleur D, Legault S, Taillefer S (1987) De-pression and body image following mastectomy and lumpectomy. Journal of Chronic Diseases 40:529-534

11. Alter CL, Pelcovitz D, Axelrod A, Goldenberg B, Harris H, Meyers B, Grobois B, Mandel F, Septimus A, Kaplan S (1996) Identification of PTSD in cancer survivors. Psychosomatics 37:137-143

12. Sonneveld DJ, Hoekstra HJ, van der Graaf WT, Sluiter WJ, Mulder NH, Willemse PH, Koops HS, Sleij-fer DT (2001) Improved long term survival of patients with metastatic nonseminomatous testicular germ cell carcinoma in relation to prognostic classification systems during the cisplatin era. Cancer 91:1304-1315

13. Horwich A, Huddart R, Dearnaley D (1998) Markers and management of germ-cell tumours of the testes. Lancet 352:1535-1538


15. Jonker-Pool G, van de Wiel HB, Hoekstra HJ, Sleijfer DT, van Driel MF, van Basten JP, Schraffordt Koops HS (2001) Sexual functioning after treatment for testicular cancer--review and meta-analysis of 36 empiri-cal studies between 1975-2000. Archives of Sexual Behavior 30:55-74

16. Nazareth I, Lewin J, King M (2001) Sexual dysfunction after treatment for testicular cancer. A systematic review. Journal of Psychosomatic Research 51:735-743

17. Kiebert GM, de Haes JC, van de Velde CJ (1991) The impact of breast-conserving treatment and mas-tectomy on the quality of life of early-stage breast cancer patients: a review. Journal of Clinical Oncology 9:1059-1070

18. Alderson P, Green Se (2003) Cochrane Collaboration open learning material for reviewers. Available from URL http://www.cochrane-net.org/openlearning/ Cited Nov 1, 2003

19. Tamburini M, Filiberti A, Barbieri A, Zanoni F, Pizzocaro G, Barletta L, Ventafridda V (1989) Psychologi-cal aspects of testis cancer therapy: a prospective study. Journal of Urology 142:1487-1489


2

38

20. Fossa SD, de Wit R, Roberts JT, Wilkinson PM, de Mulder PH, Mead GM, Cook P, de Prijck L, Stenning S, Aaronson NK, Bottomley A, Collette L (2003) Quality of life in good prognosis patients with metastatic germ cell cancer: a prospective study of the European Organization for Research and Treatment of Cancer Genitourinary Group/Medical Research Council Testicular Cancer Study Group (30941/TE20). Journal of Clinical Oncology 21:1107-1118

21. Fossa SD, Dahl AA, Loge JH (2003) Fatigue, anxiety, and depression in long-term survivors of testicular cancer. Journal of Clinical Oncology 21:1249-1254

22. Joly F, Heron JF, Kalusinski L, Bottet P, Brune D, Allouache N, Mace-Lesec’h J, Couette JE, Peny J, Henry-Amar M (2002) Quality of life in long-term survivors of testicular cancer: a population-based case-control study. Journal of Clinical Oncology 20:73-80

23. Rudberg L, Nilsson S, Wikblad K (2000) Health-related quality of life in survivors of testicular cancer 3 to 13 years after treatment. Journal of Psychosocial Oncology 18:19-31

24. Rieker PP, Fitzgerald EM, Kalish LA, Richie JP, Lederman GS, Edbril SD, Garnick MB (1989) Psycho-social factors, curative therapies, and behavioral outcomes. A comparison of testis cancer survivors and a control group of healthy men. Cancer 64:2399-2407

25. Caffo O, Amichetti M, Tomio L, Galligioni E (2001) Quality of life after radiotherapy for early-stage tes-ticular seminoma. Radiotherapy and Oncology 59:13-20

26. Bloom JR, Fobair P, Gritz E, Wellisch D, Spiegel D, Varghese A, Hoppe R (1993) Psychosocial outcomes of cancer: a comparative analysis of Hodgkin’s disease and testicular cancer. Journal of Clinical Oncology 11:979-988

27. Edbril SD, Rieker PR (1989) The impact of testicular cancer on the work lives of survivors. Journal of Psychosocial Oncology 7:17-29

28. Rieker PP, Edbril SD, Garnick MB (1985) Curative testis cancer therapy: psychosocial sequelae. Journal of Clinical Oncology 3:1117-1126

29. Moynihan C (1987) Testicular cancer: the psychosocial problems of patients and their relatives. Cancer Surveys 6:477-510

30. Arai Y, Kawakita M, Hida S, Terachi T, Okada Y, Yoshida O (1996) Psychosocial aspects in long-term survivors of testicular cancer. Journal of Urology 155:574-578

31. Cassileth BR, Steinfeld AD (1987) Psychological preparation of the patient and family. Cancer 60:547-552

32. Fossa SD, Opjordsmoen S, Haug E (1999) Androgen replacement and quality of life in patients treated for bilateral testicular cancer. European Journal of Cancer 35:1220-1225

33. Gritz ER, Wellisch DK, Landsverk JA (1988) Psychosocial sequelae in long-term survivors of testicular cancer. Journal of Psychosocial Oncology 6:41-63

34. Stuart NS, Grundy R, Woodroffe CM, Cullen MH (1990) Quality of life after treatment for testicular cancer--the patient’s view. European Journal of Cancer 26:291-294

35. Douchez J, Droz JP, Desclaux B, Allain Y, Fargeot P, Caty A, Charrot P (1993) Quality of life in long-term survivors of nonseminomatous germ cell testicular tumors. Journal of Urology 149:498-501

36. Kaasa S, Aass N, Mastekaasa A, Lund E, Fossa SD (1991) Psychosocial well-being in testicular cancer patients. European Journal of Cancer 27:1091-1095

37. Gritz ER, Wellisch DK, Siau J, Wang HJ (1990) Long-term effects of testicular cancer on marital relation-ships. Psychosomatics 31:301-312

38. van Basten JP, Jonker-Pool G, van Driel MF, Sleijfer DT, van de Wiel HB, Mensink HJ, Schraffordt-Koops H, Hoekstra HJ (1996) Fantasies and facts of the testes. British Journal of Urology 78:756-762

39. Brodsky MS (1995) Testicular cancer survivors’ impressions of the impact of the disease on their lives. Quality of Life Research 5:78-96

40. Ozen H, Sahin A, Toklu C, Rastadoskouee M, Kilic C, Gogus A, Kendi S (1998) Psychosocial adjustment after testicular cancer treatment. Journal of Urology 159:1947-1950


2

39

41. Hannah MT, Gritz ER, Wellisch DK, Fobair P, Hoppe RT, Bloom JR, Sun G, Varghese A, Cosgrove MD, Spiegel D (1992) Changes in marital and sexual functioning in long-term survivors and their spouses: testicular cancer versus Hodgkin’s disease. Psycho Oncology 1:89-103

42. Fobair P, Hoppe RT, Bloom J, Cox R, Varghese A, Spiegel D (1986) Psychosocial problems among survi-vors of Hodgkin’s disease. Journal of Clinical Oncology 4:805-814

43. Bloom JR, Hoppe RT, Fobair P, Cox RS, Varghese A, Spiegel D (1988) Effects of treatment on the work experience on long-term survivors of Hodgkin’s disease. Journal of Psychosocial Oncology 6:65-80

44. de Haes J, Curran D, Young T, Bottomley A, Flechtner H, Aaronson N, Blazeby J, Bjordal K, Brandberg Y, Greimel E, Maher J, Sprangers M, Cull A (2000) Quality of life evaluation in oncological clinical trials. European Journal of Cancer 36:821-825

45. Sprangers MA, Schwartz CE (1999) Integrating response shift into health-related quality of life research: a theoretical model. Social Science & Medicine 48:1507-1515

46. Richardson A (1998) Measuring fatigue in patients with cancer. Supportive Care in Cancer 6:94-100

47. Stone P, Richards M, A’Hern R, Hardy J (2001) Fatigue in patients with cancers of the breast or prostate undergoing radical radiotherapy. Journal of Pain and Symptom Management 22:1007-1015

48. Maguire P (1992) Psychosocial well-being in testicular cancer patients. European Journal of Cancer 28A:722

49. Travis LB, Curtis RE, Storm H, Hall P, Holowaty E, Van Leeuwen FE, Kohler BA, Pukkala E, Lynch CF, Andersson M, Bergfeldt K, Clarke EA, Wiklund T, Stoter G, Gospodarowicz M, Sturgeon J, Fraumeni JF, Jr., Boice JD, Jr. (1997) Risk of second malignant neoplasms among long-term survivors of testicular cancer. Journal of the National Cancer Institute 89:1429-1439

50. Aaronson NK, Bakker W, Stewart AL, van Dam FS, van Zandwijk N, Yarnold JR, Kirkpatrick A (1987) Multi-dimensional approach to the measurement of quality of life in cancer clinical trials. In: Aaronson NK, Beckman J (eds) The quality of life of cancer patients. Raven Press, New York, p 63

51. Barofsky I (1978) Compliance, adherence and the therapeutic alliance: Steps in the development of self-care. Social Science & Medicine 12:369-376


3Quality of life of testicular cancer

survivors and the relationship with

sociodemographics, cancer-related

variables and life events

Joke Fleer

Harald J Hoekstra

Dirk Th Sleijfer

Marrit A Tuinman

Ed C Klip


Supportive Care in Cancer (2005): published online Sep 17


3

43

Introduction

Testicular cancer (TC) affects men in a period of life when intimate relationships, star-ting a family, and career are major concerns. Since the introduction of cisplatin-based chemotherapy in the late 1970s, TC has become one of the most curable malignancies, with cure rates between 85 and 90% [1;2]. As a consequence, the majority of men diag-nosed with TC will live the largest part of their lives as cancer survivors and they may have to face sequel of diagnosis and treatment on different domains of their lives. The term “quality of life” (QoL) is generally used to refer to an individual’s perception of his well-being on, among others, the physical, psychological, and social domains [3]. In the research on testicular cancer survivors (TCSs), these domains already have received attention. For example, research on the physical domain shows that TCSs who were treated for metastetic disease have an increased risk of infertility, fatigue, second primary malignancies, decreased renal functioning, hearing deficits, and cardiovascular disease [4;5]. Studies that focused on the psychological and social domains showed that the majority of TCSs experiences good levels of functioning, although subsamples report psychosocial problems such as anxiety, depression, fertility distress, and work-related problems [6].However, a review of the literature on the QoL of TCSs revealed that most of the exis-ting literature on the QoL of TCSs suffers from methodological shortcomings, such as small sample sizes and the use of nonvalidated questionnaires [6]. Furthermore, there is a lack of research into the QoL of the growing group of long-term TCSs. Currently, there are only a few studies that did include long-term survivors [7-10], whereas at this point in time, approximately 25 years after the medical breakthrough in treatment, the group of long-term survivors has grown large enough to provide adequate statistical power to examine their QoL. Then again, the inclusion of long-term survivors does confront researchers with specific methodological concerns. In long-term survivors, it is more difficult to distinguish effects caused by cancer and its treatment, from those attributable to other factors, such as aging, comorbidities [11], and additionally expe-rienced life events. The probability of having a functional limitation or chronic disease increases with advancing age [11], and this may affect QoL. In addition, it is likely that, over time, other life events may influence the current QoL more than the experience with cancer. Therefore, to be able to draw reliable conclusions about the QoL of TCSs, it is essential to investigate the impact of age, comorbidities, and recently experienced life events as well as that of cancer-related variables. Lastly, based on the available data, it is not possible to identify TCSs at risk for an impaired QoL. Because the endpoint of research should always be to bridge the gap with clinical practice [12], the identifica-tion of risk groups should be a primary focus of QoL studies. Risk profiles may help clinicians to detect distressed TCSs in an early stage. This may lead to earlier referral to psychosocial services and as a consequence more patients may be spared from protracted distress [13]. The present study aims to attack the above-mentioned issues by (1) examining the QoL of a large sample of TCSs by comparing them to a reference group of Dutch men; (2) investigating the relationship of sociodemographics, cancer-related variables, and recent-ly experienced life events with QoL and (3) identifying TCSs with an impaired QoL,


3

44

based on sociodemographic variables, cancer-related variables, and recently experienced life events.


Procedure

All men treated for TC between 1977 and 2003 at the University Medical Center Gro-ningen (UMCG) in The Netherlands were approached in writing for the present study. In 1977, cisplatin-based chemotherapy, which led to the improved survival rates, was introduced into the treatment of metastetic TC at the UMCG. Exclusion criteria were age younger than 18 years at study entry, insufficient command of the Dutch language, and time since completion of treatment shorter than 3 months. The decision to use the criterion of 3 months after completion of the last treatment was based on the clinical consideration that TC patients have a very good prognosis. A letter with information about the objectives of the study and an informed consent form were sent to the 702 eligible TCSs. The study was approved of by the Medical Ethics Committee of the Uni-versity Medical Center Groningen.

Measurements

Data on the following sociodemographic variables were collected: age, educational level, marital status, employment status, and chronic disease. Highest educational level com-pleted was measured on a seven-point scale: primary school (1), lower vocational degree (2), lower secondary (3), middle secondary (4), high secondary (5), higher vocational (6) and university (7). We define chronic disease as an illness marked by long duration or frequent recurrence and used two questions to measure the prevalence of a chronic disease. TCSs first responded to a yes/no question: “Do you have a chronic disease (e.g., asthma, multiple sclerosis, rheumatoid arthritis)?” If a TCS reported that he had a chronic disease, he was asked to describe the disease. In addition, the following cancer-related data were collected: age at diagnosis, time since completion of treatment, type of treatment, and whether they had experienced a second cancer event (either a tumor relapse or a second primary malignancy). Type of treatment could comprise: orchidectomy (surgical removal of the affected testicle), orchidectomy and a retroperitoneal lymph node dissection (RPLND), orchidectomy and radiothera-py, orchidectomy and chemotherapy, or orchidectomy and chemotherapy and resection of residual tumor mass (RRTM).

The RAND-36 [14] is an internationally used valid and reliable generic self-report ques-tionnaire to assess QoL. The Dutch version of the RAND-36 [15] was used. It contains eight subscales: physical functioning (ten items), social functioning (two items), role limitations in work or other activities due to physical problems (four items), role limita-tions in work or other activities due to emotional problems (three items), mental health (five items), vitality (two items), pain (two items), and general health perception (five


3

45

items). After recoding and transforming, scores of the subscales could range from 0 to 100, with a higher score indicating better functioning. In the present study, internal con-sistency (Cronbach alpha) of the subscales for the total group of TCSs varied between 0.80 and 0.92. We used reference scores from the Dutch manual for the RAND-36 as comparison to the TCSs. These comprised the mean scores from a group of 372 nonselected men from a random sample of 1,063 persons aged 18 years and older from the population register of a municipality in the north of the Netherlands (number of inhabitants = 108,000). The mean age of the persons in the total random sample was 44.1 years (range 18-89 years) [15].

Life events were measured with the Vragenlijst Recent Meegemaakte Gebeurtenissen (VRMG), a Dutch questionnaire to measure recently experienced life events [16]. This questionnaire contains 25 events around five themes: health, illness and death; preg-nancy/birth; work; relationships, and miscellaneous (e.g., financial gain/loss, moving house, passing an exam). TCSs were asked to report whether they had experienced any of the 11 positive events and 14 negative events during the past year. Sum scores were calculated for the two subscales.

For TCSs, one self-constructed question was added to relate their experience with TC to their current QoL. TCSs responded to the question: “Do you think that your experi-ence with testicular cancer affects your current QoL?” The following answers could be given: 1) “TC affects my current QoL very negatively”, 2) “TC affects my current QoL negatively”, 3) “TC affects my current QoL positively as well as negatively”, 4) “TC affects my current QoL positively”, 5) “TC affects my current QoL very positively”, 6) “TC does not affect my current QoL”.

Statistical analyses

To investigate differences between the TCSs and the reference group, independent T-tests were performed. Effect sizes were calculated with Cohen’s d to assess the clinical significance of differences between TCSs and the reference group on the RAND-36. The interpretation of effect sizes is that a difference greater than 0.8 is large, between 0.5 and 0.8 is moderate, between 0.2 and 0.5 is small, and a difference smaller than 0.2 is insignificant [17]. Pearson correlations, T-tests, ANOVAs, and Mann-Whitney U tests were conducted to examine which sociodemographic variables, life events, and cancer-related variables were significantly related to the RAND-36 subscales. For employment status, a dichoto-mous variable was created with the categories “not employed for wages” (consisting of students, being unemployed, being unable to work and being retired) and “employed for wages” (consisting of being employed for wages). The effect of time since completion of treatment was evaluated in two ways. First, correlational analyses were performed, and secondly, ANOVAs were conducted to compare survivors divided into five groups accor-ding to time elapsed since completion of treatment. Group I, 3 months-2 years; Group II, 2-5 years; Group III, 5-10 years; Group IV, 10-15 years; Group V, more than 15 years. The


3

46

effect of type of treatment was investigated on the basis of two classifications. Firstly, the four treatment groups were compared. Secondly, two treatment groups were compared: “surgical treatment” (consisting of the categories orchidectomy and orchidectomy plus RPLND) and “combined treatment” (consisting of the categories orchidectomy plus radiotherapy, orchidectomy plus chemotherapy and orchidectomy plus chemotherapy plus RRTM). The variables that were significantly related to the dependent variables in the univariate analyses were included in forward regression analyses. Results were considered statistically significant if the probability of occurrence was 0.05 or less. Inasmuch as this is a descriptive study, no formal adjustments were employed to correct for multiple testing. Therefore, caution is warranted in interpretation of findings of p > 0.01.

Results

Descriptives

Fifty percent (n = 354) of the TCSs approached agreed to participate. Of the TCSs who did not participate, 74 (21%) indicated their reason for refusal. Main reasons mentioned were lack of interest, the disease period was considered to be a closed book, and too much of a burden. Nonparticipants did not differ significantly on age, time since diag-nosis, and type of treatment from the participants. The descriptives of the participants are summarized in Table 1. Mean age was 43.7 years, ranging from 18 to 78 years. Highest educational level completed varied from primary school to university degree, but the mean educational level was high secondary educa-tion. Furthermore, most survivors were married or cohabiting (85%), employed for wa-ges (78%), and did not have a chronic disease (85%). The TCSs who did have a chronic disease mainly reported having rheumatoid and degenerative diseases (26%), pulmonary diseases (23%) and cardiovascular diseases (19%). Noticeable, only one TCS reported that his lungs were damaged due to treatment with bleomycin. The mean time since completion of treatment was 10 years, ranging from 3 months to 24 years. Because age at diagnosis correlated highly with age (r = 0.83, p < 0.001), it was decided to leave age at diagnosis out of further analyses. The correlation between age and time since completion of treatment was lower (r = 0.43, p = 0.001), and there-fore, both variables were taken into account in further analyses. The minority of TCSs (34.5%) was treated with surgical treatment (orchidectomy ± RPLND) and almost 9% of the survivors had experienced a second cancer event (recurrence, second TC, or other cancer tumor) (Table 1).

QoL of TCSs: comparison with a reference group

Independent T-tests showed that TCSs reported better physical functioning (t = 2.4, p < .05) and less pain (t = 3.3, p = .001) than the reference group of men, but a worse mental health (t = -2.03, p < 0.05) and less vitality (t = -3.5, p < 0.001). However, the effect sizes of the differences were small for the subscales vitality and pain and insignifi-


3

47

*N is slightly variable because of missing data. aeducational level was measured on a seven-point scale, ranging from primary school (1) to university degree (7); bRPLND = retroperitoneal lymph node dissection; cRRTM = resection of residual tumor mass

TCSs (N = 354*)

Age (in years) mean, range 43.7 18.4-78.5

Educational levela mean, SD 4.2 1.7

Marital status N, %

Single/divorced/separated 53 15

Married/cohabiting/LAT 300 85

Employment status N, %

Employed for wages 275 77.9

Student 12 3.4

Out of work 6 1.7

(Partly) unable to work 28 7.9

Retired 32 9.1

Children N, %

No 127 36

Yes 226 64

Chronic disease N, %

No 299 84.9

Yes 53 15.1

Life events mean, SD

Positive 1.6 2.0

Negative 1.3 1.6

Time since treatment (in years) mean, SD 10.0 6.7

Treatment N,%

Surgical treatment 121 34.5

Orchidectomy 99 28.2

Orchidectomy + RPLNDb 22 6.3

Combined treatment 230 65.6

Orchidectomy + radiotherapy 68 19.4

Orchidectomy + chemotherapy 53 15.1

Orchidectomy + chemotherapy + RRTMc 109 31.1

Second cancer event N, %

Yes 31 8.8

No 322 91.2

Table 1. Sample characteristics


3

48

cant for the other subscales. This indicates that the statistically significant differences are not clinically relevant (Table 2).

Associations between sociodemographic and cancer-related variables and life events and QoL of TCSs

A significant negative correlation was found between age and physical functioning, pain, and general health perception (all p < 0.001). Furthermore, a significant positive re-lationship was found between educational level and physical functioning (p < 0.001). Positive life events was not significantly correlated to any of the RAND-36 subscales, but a significant negative association was found between the experience of more negative life events and social functioning (p < 0.05), role limitations (emotional) (p < 0.01) and mental health (p < 0.05). Time since treatment correlated significantly negatively with physical functioning (p < 0.05) (Table 3). When TCSs were divided in time-cohorts based on time since treatment, it appeared that no significant differences were found between the cohorts on any of the RAND-36 subscales.

Married TCSs reported significantly better mental health (t = -2.23, p < 0.05). TCSs who were employed for wages reported significantly better physical functioning (t = -4.42, p = 0.001), social functioning (t = -3.46, p = 0.001), fewer role limitations due to physi-cal problems (t = -2.86, p < 0.01), better mental health (t = -3.28, p = 0.001), more vita-lity (t = -3.60, p < 0.001), less pain (t = -3.53, p = 0.001), and a better general health perception (t = -5.09, p < 0.001) than did those who were not employed for wages. TCSs without a chronic disease had significantly higher mean scores on the subscales physical functioning (t = -3.78, p < 0.001), social functioning (t = -3.00, p < 0.05), role limitations due to physical problems (t = -2.49, p < 0.05), role limitations due to emotional problems

Table 2. Means and standard deviations for the RAND-36 subscales for TCSs and the reference group and T-tests between the two groups

TCSs Reference group

Mean SD Mean SD t p-valueEffect

size

Physical functioning 88 19 85 22 2.41 0.02 0.18

Social functioning 86 20 88 20 -1.84 0.07 -0.14

Role limitations (physical) 84 31 82 34 1.05 0.29 0.08

Role limitations (emotional) 85 31 87 29 -0.85 0.40 -0.06

Mental health 77 16 79 17 -2.03 0.04 -0.15

Vitality 64 20 70 21 -3.50 <0.001 -0.26

Pain 88 19 83 24 3.28 0.001 0.24

General health perception 71 20 71 23 -0.02 0.95 0.00


3

49

(t = -2.07, p < 0.05), mental health (t = -2.60, p < 0.05), vitality (t = -2.74, p < 0.01), pain (t = -2.74, p < 0.01), and general health perception (t = -5.18, p < 0.001) than did those with a chronic disease.

ANOVAs with type of treatment in four cohorts did not show significant effects of type of treatment received on any of the QoL subscales (results not reported in table). However, when type of treatment was divided in two cohorts, results showed that TCSs who had received combined treatment had significantly lower mean scores on vitality (t = 2.00, p < .05) and general health perception (t = 2.55 , p = 0.01) than those who had received surgical treatment only. Finally, TCSs who had experienced a second cancer event reported worse mental health (t = -2.12 , p < 0.01) and general health perception (t = -3.14, p < 0.01) than TCSs who had not experienced a second cancer event (Table 4).

Univariate analyses showed that both employment status and chronic disease signifi-cantly related to all QoL subscales, except for role limitations due to emotional pro-blems. Because not being employed for wages may be a result of having a chronic disease, it was decided to create a composite variable, which resulted in the following groups:

Group 1, not employed for wages and with a chronic disease (n = 24); Group 2, employed for wages and with a chronic disease (n = 29); Group 3, not employed for wages and without a chronic disease (n = 54); Group 4, employed for wages and without a chronic disease (n = 245).

ANOVA, followed by a Scheffé-test, showed that group 1 differed significantly from the remaining three groups on physical functioning (F = 20.41, p < 0.001), social functioning

Variable AgeEducational

levelPositive

life eventsNegative

life eventsTime since treatment

r r r r r

Physical functioning -0.29*** 0.20*** 0.10 -0.06 -0.12*

Social functioning -0.06 0.01 0.03 -0.12* -0.02

Role limitations (physical) -0.12* 0.08 -0.02 -0.08 0.05

Role limitations (emotional) -0.002 -0.07 -0.006 -0.15** -0.01

Mental health 0.009 -0.02 0.01 -0.14* 0.01

Vitality -0.05 0.03 0.07 -0.09 0.004

Pain -0.25*** 0.09 0.08 -0.05 -0.09

General health perception -0.19*** 0.08 0.08 -0.08 -0.02

Table 3. Pearson correlations between continuous sociodemographic variables, life events, and cancer-related variables with the RAND-36 subscales

*p≤0.05; **p≤0.01; ***p≤0.001


3

50

Tab

le 4

. Mea

ns o

n th

e R

AN

D-3

6 su

bsca

les f

or g

roup

s on

cate

gori

cal s

ocio

dem

ogra

phic

and

can

cer-

rela

ted

vari

able

s

Var

iabl

ePa

rtne

rE

mpl

oyed

for

wag

esC

hron

ic d

isea

seTy

pe o

f tre

atm

ent

Seco

nd c

ance

r ev

ent

Yes

No

Yes

No

Yes

No

Surg

ical

Com

bine

dYe

sN

o

Phys

ical

func

tioni

ng89

8691

78**

*75

90**

*89

8884

89

Soci

al fu

nctio

ning

8683

8878

***

7688

**87

8584

86

Rol

e lim

itatio

ns (p

hysic

al)

8580

8773

**72

86*

8782

8084

Rol

e lim

itatio

ns (e

mot

iona

l)86

8287

8075

87*

8486

8386

Men

tal h

ealth

7872

*79

71**

*70

78*

7877

7078

*

Vita

lity

6560

6657

***

5766

**68

62*

6065

Pain

8986

9181

***

8090

**89

8886

88

Gen

eral

hea

lth p

erce

ptio

n72

7074

61**

*57

74**

*75

69**

6172

**

*p≤

0.05;

**p≤

0.01;

***p

≤0.0

01 si

gnifi

canc

e lev

el of

diff

eren

ces b

etwee

n gr

oups


3

51

(F = 11.62, p < 0.001), role limitations due to physical problems (F = 9.78, p < 0.001), pain (F = 9.50, p < .001), and general health perception (F = 20.02, p < 0.001). Group 1 differed significantly from group 4 on mental health (F = 7.96, p < 0.001) and vitality (F = 6.62, p < 0.001). Therefore, a dummy variable of this composite variable was created for which group 1 was separated from the remaining three groups (dummy variable, 0 = TCSs who were not employed for wages and had a chronic disease; 1 = the rest). This dummy variable was included in the stepwise forward regression analyses instead of the individual variables of employment status and chronic disease for all subscales, except for role limitations due to emotional problems.

Predictors of QoL

Because the third aim of this study was to identify TCSs with an impaired QoL, regres-sion analyses were performed to determine which variables had the strongest predictive power. Only the variables that were significantly related to the dependent variables in the univariate analyses were included in the forward regression analyses. The dummy vari-able of employment status and chronic disease appeared to be the strongest predictor, indicating that the combination of not being employed for wages and having a chronic disease had a particularly negative effect on the QoL of TCSs. Age and educational level had significant independent predictive power in physical functioning, age had also an independent effect on pain and general health perception. The experience of negative life events had a unique negative effect on social functioning, role limitations (emotional) and mental health. With respect to cancer-related variables, type of treatment appeared to have some explanatory power in vitality and general health perception and the ex-perience of a second cancer event in mental health and general health perception. The univariate effects of marital status and time since completion of treatment disappeared in the regression analyses.

The selected variables in the regression models explained significant, although small, proportions of the variance of the subscales of the RAND-36 (ranging from 4% to 22%) (Table 5).

Evaluation of the impact of TC on current QoL

As a response to the question “do you think that the experience with TC affects your current QoL?” 52.6% of the TCSs reported that it had a negative as well as a positive impact on their current QoL. A small percentage (11.2%) of the TCSs reported that the experience with TC still had a negative or very negative impact, 26.6% reported that it had a (very) positive impact, and 9.5% reported that it had no impact at all. TCSs who were not employed for wages and who had a chronic disease differed from the rest of the TCSs in their answers to this question: a significantly larger proportion of this group reported that TC still had a negative impact on their present QoL (33.3% vs. 9.7%, Chi2 = 12.90, p = 0.005).


3

52

Tab

le 5

. For

war

d re

gres

sion

ana

lyse

s of s

ocio

dem

ogra

phic

s, c

ance

r-re

late

d va

riab

les a

nd li

fe e

vent

s on

the

RA

ND

-36

subs

cale

s

Pred

ictor

var

iabl

eP

hysi

cal

Func

tion

ing

Soci

alFu

ncti

onin

g

Rol

e lim

ita-

tion

s(p

hysi

cal)

Rol

e lim

ita-

tion

s(e

mot

iona

l)

Men

tal h

ealth

Vit

alit

yPa

inG

ener

al h

ealth

pe

rcep

tion

beta

beta

beta

beta

beta

beta

beta

beta

Age

-0.2

1***

-0.0

5-0

.20*

**-0

.12*

Edu

catio

n le

vel

0.17

***

Mar

ital s

tatu

s0.

10

Em

ploy

men

t sta

tus +

chr

onic

dis

ease

a0.

31**

*0.

30**

*0.

28**

*0.

22**

*0.

20**

*0.

20**

*0.

29**

*

Chr

onic

dis

ease

b0.

14**

Neg

ativ

e lif

e ev

ents

-0.13

*-0

.16**

-0.14

**

Tim

e sin

ce tr

eatm

ent

-0.0

1

Type

of t

reat

men

tc-0

.12*

-0.11

*

Seco

nd c

ance

r eve

ntd

0.12

*0.

11*

R2

0.22

***

0.10

***

0.08

***

0.04

***

0.08

***

0.07

***

0.10

***

0.16

***

*p≤

0.05;

**p≤

0.01;

***p

≤0.0

01a0

= n

ot em

ploy

ed fo

r wag

es +

chro

nic d

iseas

e, 1 =

rest;

b0

= y

es, 1

= n

o; c0

= su

rgica

l, 1=

com

bine

d; d

0 =

yes

, 1=

no;

empt

y cell

= re

latio

nshi

p wa

s not

sign

ifica

nt in

the

univ

aria

te an

alys

es


3

53

Discussion

Examination of the quality of life (QoL) of testicular cancer survivors (TCSs) was the first objective of this study. Overall, the TCSs in our study reported relatively high levels of QoL. However, some statistically significant differences between TCSs and refer-ence men were found. TCSs reported a better physical functioning and less bodily pain, but worse mental functioning and less vitality. Yet, the effect sizes of the statistical diffe-rences were small to negligible. These results are consistent with those of other studies reporting no differences between TCSs and controls in QoL [9;18] and more generally consistent with the results of studies on other types of cancers [19-21]. Parker et al. [21] explained this lack of difference by arguing that “patients may be taking their cancer diagnosis in consideration when they subjectively evaluate their health status and func-tioning”. Indeed, it may very well be that the experience with cancer changes their expec-tations about life or internal standards and causes one to evaluate ones QoL according to new expectations or different standards [22].

Secondly, we were interested in the relationship between the QoL of TCSs and socio-demographics, cancer-related variables, and life events. Consistent with findings in the general population [15], our results showed an aging effect in the physical domain, with older TCSs reporting a decline in functioning. We also expected to find more QoL problems in TCSs who were more recently treated. Contrary to our expectations, we did not find an effect of time since treatment. QoL of TCSs was comparable among the short-, median-, and long-term survivors, indicating that survivors quickly adapt to the circumstances created by the experience of cancer treatment. Employment status and chronic disease were found to be the most important predictors of QoL of TCSs. More precisely, the combination of employment status and chronic disease appeared to be the strongest predictor. The mean scores of the group of unem-ployed TCSs with a chronic disease deviated significantly from those of men who only had a chronic disease, were only unemployed, or who neither suffered from a chronic disease or unemployment. This joint burden of unemployment and chronic disease on QoL has not been studied before in TCSs, but its negative effect on well-being has been found in the general population [23]. The relationship between these two variables is understandable when it is considered that chronic diseases can cause impairments in social activities and work, which may lead to lower levels of QoL [24-27]. If and how the experience with TC relates to these findings cannot be determined on the basis of the available data, but responses to an open question indicate that unemployed TCSs with a chronic disease significantly more often experience a negative impact of TC on their current QoL than their counterparts. So, it might be that the experience with TC has impacted through chronic disease and employment status on the well-being of these men, but prospective research is needed to verify this result.Having or not having a partner did not affect functioning in TCSs, whereas having a partner has been identified as an predictor of better well-being in the general population [21]. One line of reasoning to explain this result might be that TCSs without a partner may attribute their being single to the experience with TC (“external attribution”) [28] and this might have led to a reevaluation of having a partner as being less important for


3

54

their well-being than, for example, being employed. More research is needed to deter-mine if this is the case. TCSs who were treated with combined treatment reported less vitality and a worse general health perception than those who had received surgery only. Furthermore, TCSs who had received treatment for a second cancer event reported a worse mental health and a worse general health perception than those who did not encounter a second cancer event. Although these relationships make sense, the impact of these variables was low. This lack of effect of objective illness-related variables on QoL has been reported previ-ously [21;29]. It may be that the subjective perception of the seriousness or threat of the disease is a more important determinant of QoL [30]. The impact of positive life events was an insignificant predictor, but the experience of more negative life events during the past year was related to more limitations in emo-tional and social well-being. This is consistent with the general assumption that expo-sure to a cumulative number of negative life events is associated with more psychosocial problems [31]. Recently experienced negative life events influenced the current QoL of the TCSs to a greater extent than objective cancer-related variables.

Limitations

The inclusion of a validated generic QoL questionnaire, the large sample size, and the large range in time since treatment are strengths in the design of this study. However, this study has some limitations as well. First of all, the response rate of 50% may have induced bias, although the participating TCSs did not differ in age, time since diagnosis, and type of treatment from those who did not participate. Matsuda et al. [32] have sum-marized several characteristics of nonresponders and gender was one of them. Men tend to have a lower response rate than women, so this might partly explain the relatively low response rate in the present study. A further limitation was that chronic disease was measured somewhat crudely: respondents were not asked to indicate their perception of the severity of the chronic disease and whether they thought that the chronic disease was a consequence of the treatment for TC. Finally, based on a previous report [33], we decided to dichotomizes employment status into employment for wages and not employed for wages. The last group included students as well as retirees, and this may have resulted in a biased picture of the results. However, we also performed analyses with other dichotomizations (for example, dichotomizations in which students and/or retirees were left out), and this did not give different results (not reported in the results section).

Summary and conclusions

In conclusion, the results of the present study show that, on the whole, TCSs experience a QoL similar to that of men who did not experience cancer. Age, employment status, chronic disease, and negative life events were more important determinants of QoL than objective disease characteristics and should be taken into account when considering the QoL of cancer survivors over time. TCSs who are not employed for wages and who have a chronic disease seemed to be at risk for an impaired functioning. Finally, it should be


3

55

noted that the variance explained by sociodemographic variables, life events, and can-cer-related variables was low, indicating that more important predictors remain to be identified.


3

56

Reference List



3. Vidrine DJ, Gritz ER (2003) Psychosocial outcomes after testicular cancer treatment. In: Steele jr. GD, Phillips TL, Chabner BA, Gansler TS (eds) Germ cell tumors - American Cancer Society atlas of clinical oncology. BC Decker Inc., Hamilton, London, pp 295-305



6. Fleer J, Hoekstra HJ, Sleijfer DT, Hoekstra-Weebers JEHM (2004) Quality of life of survivors of testicu-lar germ cell cancer: a review of the literature. Supportive Care in Cancer 12:476-486

7. Arai Y, Kawakita M, Hida S, Terachi T, Okada Y, Yoshida O (1996) Psychosocial aspects in long-term survivors of testicular cancer. Journal of Urology 155:574-578

8. Caffo O, Amichetti M, Tomio L, Galligioni E (2001) Quality of life after radiotherapy for early-stage tes-ticular seminoma. Radiotherapy and Oncology 59:13-20


10. Fossa SD, Dahl AA, Loge JH (2003) Fatigue, anxiety, and depression in long-term survivors of testicular cancer. Journal of Clinical Oncology 21:1249-1254

11. Gotay CC, Muraoka MY (1998) Quality of life in long-term survivors of adult-onset cancers. Journal of the National Cancer Institute. 90:656-667

12. Efficace F, Bottomley A, Osoba D, Gotay C, Flechtner H, D’haese S, Zurlo A (2003) Beyond the develop-ment of health-related quality-of-life (HRQOL) measures: A checklist for evaluating HRQOL outcomes in cancer clinical trials - Does HRQOL evaluation in prostate cancer research inform clinical decision making? Journal of Clinical Oncology 21:3502-3511

13. Schnoll RA, Harlow LL (2001) Using disease-related and demographic variables to form cancer-distress risk groups. Journal of Behavioral Medicine 24:57-74

14. Hays RD, Donald Sherbourne C, Mazel RM (1993) The RAND 36-item health survey 1.0. Health Eco-nomics 2:217-227

15. Van der Zee, K. I. and Sanderman, R. The measurement of generic health with the RAND-36. Northern Centre for Healthcare Research, available from URL http://www.coo.med.rug.nl. Cited April 14 2005

16. Van de Willige G, Schreurs P, Tellegen B, Zwart F (1985) Het meten van ‘life events’: de Vragenlijst Recent Meegemaakte Gebeurtenissen (VRMG). Nederlands Tijdschrift voor de Psychologie 40:1-19

17. Cohen J (1988) Statistical power analysis for the behavioral sciences. Erlbaum, Hillsdale, NJ

18. Rudberg L, Nilsson S, Wikblad K (2000) Health-related quality of life in survivors of testicular cancer 3 to 13 years after treatment. Journal of Psychosocial Oncology 18:19-31

19. Bardwell WA, Major JM, Rock CL, Newman VA, Thomson CA, Chilton JA, Dimsdale JE, Pierce JP (2004) Health-related quality of life in women previously treated for early-stage breast cancer. Psycho Oncology 13:595-664


3

57

20. Ganz PA, Rowland JH, Desmond K, Meyerowitz BE, Wyatt GE (1998) Life after breast cancer: Under-standing women’s health-related quality of life and sexual functioning. Journal of Clinical Oncology 16:501-514



23. Michelson H, Bolund C, Brandberg Y (2000) Multiple chronic health problems are negatively associated with health related quality of life (HRQoL) irrespective of age. Quality of Life Research 9:1093-1104

24. Michelson H, Bolund C, Nilsson B, Brandberg Y (2000) Health-related quality of life measured by the EORTC QLQ-C30 - Reference values from a large sample of the Swedish population. Acta Oncologica 39:477-484

25. Schlenk EA, Erlen JA, Dunbar-Jacob J, McDowell J, Engberg S, Sereika SM, Rohay JM, Bernier MJ (1998) Health-related quality of life in chronic disorders: a comparison across studies using the MOS SF-36. Quality of Life Research 7:57-65

26. Wingard JR, Curbow B, Baker F, Piantadosi S (1991) Health, Functional Status, and Employment of Adult Survivors of Bone-Marrow Transplantation. Annals of Internal Medicine 114:113-118

27. Greaves-Otte JG, Greaves J, Kruyt PM, van Leeuwen O, van der Wouden JC, van der DE (1991) Problems at social re-integration of long-term cancer survivors. European Journal of Cancer 27:178-181

28. Festinger L (1957) A theory of cognitive dissonance. New York

29. Bjordal K, Mastekaasa A, Kaasa S (1995) Self-reported satisfaction with life and physical health in long-term cancer survivors and a matched control group. European Journal of Cancer. Part B, Oral Oncology 31B:340-345

30. Zebrack BJ, Chesler M (2001) Health-related worries, self-image, and life outlooks of long-term survivors of childhood cancer. Health & Social Work 26:245-256

31. Forehand R, Biggar H, Kotchick BA (1998) Cumulative risk across family stressors: Short- and long-term effects for adolescents. Journal of Abnormal Child Psychology 26:119-128

32. Matsuda T, Marche H, Grosclaude P, Clement S (2004) Participation behavior of bladder cancer survivors in a medical follow-up survey on quality of life in France. European Journal of Epidemiology 19:313-321

33. Brazier JE, Harper R, Jones NMB, O’Cathain A, Thomas KJ, Usherwood T, Westlake L (1992) Validating the SF-36 health survey questionnaire: new outcome measure for primary care. BMJ 305:160-164


4The role of meaning in the prediction

of psychosocial well-being of


Joke Fleer

Harald J Hoekstra

Dirk Th Sleijfer

Marrit A Tuinman


Quality of Life Research, in press


61

4

Introduction

All people have a framework from which they view the world; they have basic beliefs about the world, themselves and themselves in the world. In important psychological theories, such as cognitive dissonance theory [1], cognitive adaptation theory [2-4], as-sumptive world theory [5] and self-regulation theory [6], it is assumed that a central belief of most people is that their life is meaningful. Having a sense of meaning refers to ‘the cognizance of order, coherence, and purpose in one’s existence, the pursuit and attainment of worthwhile goals, and an accompanying sense of fulfillment’ [7]. This belief influences people’s understanding of the past and the present and it determines what is important in life. As such, it generates expectations about the future and forms a guide from which priorities and goals are directed [8]. So, events are experienced as meaningful when they logically follow from people’s framework of beliefs and when they do not disturb the fulfillment of important life goals. It is generally acknowledged that having a sense of meaning is relevant to psychosocial well-being and life satisfaction [9]. Some people encounter a life event that is so aversive (e.g., holocaust, natural disasters, bereavement) that it threatens the basic assumption that life is meaningful (e.g., [5;8;10]). Fortunately, most people do regain a sense of meaning by using a variety of coping ac-tivities, such as construing positive meaning from the experience or adapting their belief systems and redirecting goals and priorities (for an overview of these coping activities see [8]). Persons who find meaning in the trauma (i.e. the belief that the event makes sense in their life), generally experience better psychosocial functioning than those who do not (e.g., [11-13]). Therefore, it is assumed that the process of maintaining or reestablishing meaning is an important task of coping with a negative life event [5;8;10;14]. People who are unable to make sense of their experience may become caught in a cycle of rumination (e.g., continuously trying to find a cause for the event), which is maladaptive [8]. The concept of meaning is relevant within the context of psycho-oncology, because cancer also threatens people’s sense of meaning [9]. For example, most people will not be able to discover the cause for their cancer, which may make it difficult to find meaning in the experience [15]. It may undermine other assumptions as well, such as people’s sense of invulnerability and safety [16]. Prior to the cancer experience, people may feel that they have control over the events that happen to them. A cancer diagnosis may replace this sense of mastery by feelings of vulnerability and fear of death. Cancer patients will have to integrate these threatening cognitions and feelings into their belief system to regain a good quality of life and reduce their cancer-related distress. It has been suggested that the process of meaning-making may actually start only after completion of cancer treatment, because shortly after diagnosis people may be too engaged in coping with the immediate physical and psychosocial consequences of diagnosis and treatment [15]. Researchers increasingly recognize the possible relevance of meaning for the well-being of cancer survivors, which is reflected in the recent publications on meaning and related concepts, such as posttraumatic growth and benefit finding in survivor populations [15;17-19]. Yet, these studies have mainly focused on female (breast cancer) survivors, whereas research into meaning in male cancer survivors is scarce. Furthermore, there are still many theorically and clinically relevant questions that remain to be answered. For instance, there is a lack of research that focuses on the identification of survivors who experi-


62

4

ence less meaning in life after cancer. Different personal (e.g., sociodemographic charac-teristics) and medical variables (e.g., characteristics of disease and treatment) may relate to meaning. Also, there is a lack of research that investigates the relationship between meaning and specific domains of well-being. There are indications that meaning may be related more strongly to some aspects of functioning (e.g., mental health) than to others (e.g., social functioning) [20]. Knowledge on meaning and its impact on functioning may help health care workers to identify survivors who might benefit from an intervention based on the meaning-making process. The purpose of the present study is to investigate meaning and psychosocial well-being in a large group of testicular cancer survivors (TCSs). Testicular cancer (TC) has some typical clinical features that make it an illness capable of shattering basic beliefs and the attainment of valued goals. TC mainly affects men in the prime of their life (15-45 years). It is curable for approximately 90% of the patients [21;22]. Young people generally take their good health for granted, whereas life-threatening illnesses and the possibility of dy-ing do not fit their outlook on life [22-24]. Therefore, the experience of TC may forever undermine this sense of invulnerability and safety and replace it with fear for recurrence, concerns about other cancers, and death. Furthermore, it strikes in a period of life charac-terized by major social life changes. Between the ages of 20 and 40, important decisions about marriage, starting a family and establishing a professional career are generally made. Physical and psychological after-effects of treatment may interfere with plans TCSs had made before cancer (e.g., the plan to pursue a career at a high level may be disturbed by cancer-related fatigue), forcing the survivors to reorganize their planned life trajectories. It also threatens social roles, such as the work role that provides feedback about one’s value and competence in the world [25]. When the cancer experience continues to interfere with planned life trajectories and social roles, the testicular cancer survivor may find it difficult to attain meaning to his cancer experience, and as a consequence experience an impaired psychosocial functioning and more cancer-related distress. Because meaning can be considered as outcome (it may be changed by the cancer experi-ence), as well as a predictor of well-being, the focus of this paper is twofold. Firstly, it is aimed at examining meaning in TCSs and its relationship with sociodemographics, can-cer-related variables, and recently experienced life events. The sociodemographic charac-teristics of TCSs are of interest, because it might help to identify subgroups of men who experience more or less meaning in life after cancer. The relationship with cancer-related variables may be relevant, because type of treatment, time since treatment and the experi-ence of a second cancer event may relate to meaning. For instance, it may be that short-term survivors are more occupied with meaning-making than those who are long-term survivors. We think that it is relevant to examine the relationship between meaning and recent life events, because the effect of recent life events on one’s existential belief system may vary depending on the amount of time that has passed since the cancer experience. Secondly, this paper aims at assessing changes in outlook on life after the experience with TC. Thirdly, it aims at assessing the contribution of meaning in the prediction of psycho-social well-being and cancer-related distress, in addition to sociodemographic and cancer -related variables and recently experienced life events.


4

63


Selection of patients

All men treated for TC between 1977 and 2003 at the University Medical Center Gro-ningen in The Netherlands were approached in writing for the present study. Exclusion criteria were age younger than 18 years at study entry, insufficient command of the Dutch language, and time since completion of treatment less than three months. The decision to use the criterion of three months after completion of the last treatment was based on the clinical consideration that TC patients have a very good prognosis; their estimated chance of survival is approximately 90% [21;22]. A letter with information about the objectives of the study and an informed consent form were sent to the 702 eligible TCSs. The TCSs who agreed to participate received a questionnaire and a prepaid return envelope.The study was approved of by the Medical Ethics Committee of the University Medical Center Groningen.

Measures

Sociodemographic and cancer-related variablesThe following sociodemographic variables were collected: age, educational level, marital status, employment status, prevalence of chronic disease and parenthood. In addition, the following cancer-related data were collected: time since completion of treatment, type of treatment, and experience of a second cancer event.

Meaning in lifeThe construct of meaning in life was assessed with the Dutch version of the Life Regard Index (LRI) [26], a questionnaire which was originally developed by Battista and Al-mond [27]. It is a generic 28-item questionnaire, with two subscales, each comprising of 14 items. The framework subscale was designed to assess the degree into which individuals envi-sion their lives within some meaningful perspective or have derived a set of life-goals, purpose in life or life view from them. Examples of items of the framework subscale: ‘I feel like I have found a really significant meaning for leading my life’, ‘I really don’t have much purpose for living, even for myself ’, ‘I have really come to terms with what is important in my life’, ‘Other people seem to have a better idea of what they want to do with their lives than I do’, ‘I have a philosophy of life that really gives my living significance’, ‘I need to find something that I can really be committed to’, ‘I have some aims and goals that would personally give me a great deal of satisfaction if I could accomplish them’. The fulfillment subscale was designed to assess the degree into which persons see them-selves as having fulfilled or being in the process of fulfilling their framework or life-goals. Examples of items of this subscale: ‘Living is deeply fulfilling’, ‘I spend most of my time doing things that really aren’t important for me in life’, ‘I don’t seem to be able to accom-plish those things that are really important to me’, ‘I feel that I am really going to attain what I want in life’, ‘Something seems to stop me from doing what I really want to do’,


64

4

’When I look at my life I feel the satisfaction of really having worked to accomplish some-thing’, ‘I don’t really value what I’m doing’.Answers were rated on a three-point Likert scale (‘I agree’, ‘I have no opinion’, ‘I disagree’) and scores on both subscales could vary between 14 and 42. In case of the framework subscale, a higher score indicates a higher degree of positive beliefs and goals. In case of the fulfillment subscale, a higher score indicates that persons see themselves in the process of fulfilling their beliefs and goals. In the present study, internal consistency (Cronbach alpha) of the framework subscale was 0.77 and of the fulfillment subscale it was 0.86. This is in line with findings of Debats [26].

Two questions were asked following the LRI. TCSs first responded to the question: ‘Did the experience with TC change your outlook on life?’ The question was measured on a 5-point scale: ‘yes, I think I have a much more positive outlook on life since my experi-ence with TC’ (1); ‘yes, I think I have a somewhat more positive outlook on life since my experience with TC’ (2); ‘no change’ (3); ‘yes, I think I have a somewhat more negative outlook on life since my experience with TC’ (4) ‘yes, I think I have a much more negative outlook on life since my experience with TC’ (5). If a TCS reported that his outlook had changed, a second open question was asked: ‘could you describe how your outlook on life has changed since the testicular cancer?’

Psychosocial well-beingThe following instruments were used to assess psychological well-being:1. The mental health subscale of the Dutch version of the RAND-36[28]. The RAND-

36 [29] is an internationally used valid and reliable generic self-report questionnaire to assess QoL. The mental health subscale contains five items to assess feelings of depression and nervousness. After recoding and transforming, scores of the subscale could range from 0 to 100, with a higher score indicating better mental health. In the present study, internal consistency (Cronbach alpha) of the mental health subscale for the total group of TCSs was 0.84.

2. The Happiness Index [30]. This index consists of two questions: (1) ‘To what extent do you regard yourself as a happy person?’ and (2) ‘How satisfied are you with your present life?’. Responses could be given on a four-point scale (1 = not satisfied or hap-py, 2 = doubtful satisfied or happy, 3 = satisfied or happy, 4 = very satisfied or happy). The sumscores of the answers to these two questions comprise the Happiness Index, with a higher score indicating a higher level of life satisfaction. Previous studies have not provided reliability coefficients of this scale [26;31;32], but in the present study internal consistency (Cronbach alpha) of the happiness index was 0.88.

The following instrument was used to assess social well-being:3. The social functioning subscale of the Dutch version of the RAND-36 [28]. This

subscale contains two items to assess limitations in social activities due to health problems. After recoding and transforming, scores of the subscale could range from 0 to 100, with a higher score indicating better social functioning. In the present study, internal consistency (Cronbach alpha) of the social functioning subscale for the total group of TCSs was 0.80.


4

65

Cancer-related distressCancer-related distress was measured with the Dutch version of the Impact of Event Scale [33;34]. This scale (15 items) makes an inventory of the extent to which a subject is currently occupied with the coping process after a major event by measuring intrusion (intrusively experienced ideas, images, feelings or bad dreams about the event) and avoi-dance of unpleasant feelings or memories of the event. The IES is often used in studies on cancer patients [35]. With this questionnaire, information was obtained about the degree to which confrontation with testicular cancer was influencing the current daily life of the respondent. Higher scores indicate more stress response symptoms. The internal consis-tency of this questionnaire was good (alpha = 0.88).

Life eventsThe questionnaire recently experienced life events contains 25 questions around five themes: health, illness and death; pregnancy/birth; work; relationships and miscellane-ous. Respondents were asked to report whether they had experienced any of the 11 posi-tive events (e.g., promotion at work or important financial improvement) and 14 negative events (e.g., death of a significant other) during the past year [36]. Sumscores were calcu-lated for the two subscales.


Statistical analyses were performed using SPSS 10.0. Descriptive statistics were used to calculate means, frequencies and percentages. Correlational analyses, T-tests and ANO-VAs were performed to assess relationships between the LRI and sociodemographic vari-ables, cancer-related variables, and life events. For these analyses, a dichotomous variable was created for employment status with the categories ‘not employed for wages’ (consis-ting of being a student, being unemployed, being unable to work and being retired) and ‘employed for wages’ (consisting of being employed for wages). Furthermore, the effect of type of treatment was investigated on the basis of two classifications. Firstly, TCSs were divided into five groups: 1) ‘orchidectomy’, survivors had undergone a surgical removal of the affected testicle only; 2) ‘orchidectomy + RPLND’, survivors were treated with an orchidectomy followed by a retroperitoneal lymph node dissection; 3) ‘radiotherapy’, survivors were treated with an orchidectomy and radiotherapy; 4) ‘chemotherapy’, survi-vors were treated with an orchidectomy and chemotherapy; 5) ‘chemotherapy + RRTM’, survivors were treated with an orchidectomy, additional chemotherapy and resection of residual tumor mass. Secondly, TCSs were divided into two groups: ‘surgical treatment’ (consisting of the categories orchidectomy and orchidectomy plus RPLND) and ‘com-bined treatment’ (consisting of the categories orchidectomy plus radiotherapy, orchidec-tomy plus chemotherapy and orchidectomy plus chemotherapy plus RRTM). The answers to the open question were categorized. A respondent could raise several is-sues. Each issue was categorized separately. The predictive power of the independent variables on psychosocial well-being (mental health, the happiness index, and social functioning) and cancer-related distress was as-sessed in two steps. Firstly, correlational analyses, T-tests or ANOVAs were conducted to examine bivariate relationships between potential predictors (sociodemographic vari-


66

4

ables, cancer-related variables, life events, and the LRI) and the outcome variables. Then hierarchical regression analyses were performed, entering only those predictor variables that were significantly associated with the outcome variables. Potential predictors were firstly the antecedent variables (sociodemographic variables, cancer-related variables and life events), and secondly the framework and fulfillment subscales. The incremental con-tribution of the subscales of the LRI on the outcome variables can be examined by statis-tically controlling for the antecedent variables.

Results

Sample characteristics

Fifty percent (n = 354) of the TCSs approached agreed to participate. Of the TCSs who did not participate, 74 (21%) indicated their reason for refusal. Main reasons mentioned were lack of interest, the disease period was considered to be a closed book and too much of a burden. Non-participants did not differ significantly on age, time since diagnosis and type of treatment from the participants. Descriptives of the participants are summarized in Table 1. Mean age was 44 years, rang-ing from 18 to 78 years. Highest educational level varied from primary school to university degree, the mean educational level was comparable to middle secondary education. Fur-thermore, most survivors were married or cohabiting (85%), employed for wages (78%) and did not have a chronic disease (85%). TCSs who did have a chronic disease, mainly reported having rheumatoid and degenerative diseases (26%), pulmonary diseases (23%), or cardiovascular diseases (19%). Mean time since completion of treatment was ten years, ranging from three months to 24 years. Approximately one-third of TCSs (34.5%) was treated surgically (orchidectomy ± RPLND), and 9% of the survivors had experienced a second cancer event (recurrence, second TC, or other malignancy).

Univariate relationships between meaning and sociodemographic variables, cancer-related variables, and recently experienced life events

Sociodemographic variablesCorrelational analyses showed that age and educational level were not significantly re-lated to the LRI (comprising of the framework and the fulfillment subscale) (Table 2). T-tests showed that TCSs who had a partner reported significantly higher scores on the LRI than those who did not have a partner (framework: t = -2.63, p = 0.009; fulfillment: t = -3.52, p = 0.001). Differences were found between those working for wages compared to those not working for wages (framework: t = -4.14, p < 0.001; fulfillment: t = -4.26, p < 0.001) and between those with children compared to those without children (frame-work: t = -2.05, p = 0.04; fulfillment: t = -1.94, p = 0.05). TCSs with a chronic disease re-ported significantly less fulfillment than their counterparts (t = 2.42, p = 0.02) (Table 3).


67

4

Table 1. Sample characteristics

TCSs (N = 354*)



Marital status N, %





Student 12 3.4

Out of work 6 1.7


Retired 32 9.1

Children N, %

No 127 36

Yes 226 64


No 299 84.9

Yes 53 15.1


Positive 1.6 2.0

Negative 1.3 1.6


Treatment N,%









Yes 31 8.8

No 322 91.2



68

4 Var

iabl

eA

geE

duca

tion

al

leve

lT

ime

sinc

e

trea

tmen

tPo

sitiv

e lif

e ev

ents

Neg

ativ

e lif

e ev

ents

Fram

ewor

kFu

lfillm

ent

rr

rr

rr

r

Fram

ewor

k0.

002

-0.0

1-0

.02

-0.0

1-0

.14**

Fulfi

llmen

t

0.01

-0.0

1-0

.007

-0.0

5-0

.16**

0.67

***

RA

ND

36

- Men

tal h

ealth

0.

01-0

.02

-0.0

6-0

.01

-0.14

*0.

29**

*0.

46**

*

Hap

pine

ss In

dex

-0.0

3-0

.09

-0.0

60.

12*

-0.13

*0.

47**

*0.

68**

*

RA

ND

36

- Soc

ial f

unct

ioni

ng-0

.06

0.00

9-0

.04

-0.0

3-0

.12*

0.21

***

0.36

***

IES

- tot

al d

istr

ess

-0.0

5-0

.10-0

.13**

0.07

0.04

-0.10

-0.2

0***

Tab

le 2

. Pea

rson

cor

rela

tion

s bet

wee

n co

ntin

uous

soci

odem

ogra

phic

var

iabl

es, c

ance

r-re

late

d va

riab

les,

and

life

eve

nts

and

the

LRI (

fram

ewor

k an

d fu

lfillm

ent)

, men

tal h

ealth

, hap

pine

ss in

dex

and

soci

al fu

ncti

onin

g

*p≤

0.05;

**p≤

0.01;*

**p≤

0.001


69

4

Var

iabl

ePa

rtne

rE

mpl

oyed

for

wag

esC

hild

ren

Chr

onic

dis

ease

Type

of t

reat

men

t Se

cond

can

cer

even

t

Yes

No

Yes

No

Yes

No

Yes

No

Surg

ical

Com

bine

dYe

sN

o

LRI -

Fra

mew

ork

35.6

33.4

**35

.833

.4**

*35

.734

.6*

34.3

35.4

34.9

35.4

35.0

35.3

LRI -

Ful

fillm

ent

35

.833

.0**

*36

.033

.0**

*35

.834

.533

.635

.6*

35.2

35.4

33.6

35.5

RA

ND

36

- Men

tal h

ealth

77

.772

.3*

78.7

70.8

***

78.2

74.5

*70

.278

.0*

77.5

76.5

69.7

77.6

*

Hap

pine

ss In

dex

6.2

5.5*

**6.

25.

7**

6.2

5.9

5.6

6.1*

*6.

16.

05.

86.

1

RA

ND

36

- Soc

ial f

unct

ioni

ng86

.183

.388

.177

.7**

*86

.384

.975

.587

.5**

87.0

85.0

84.2

85.9

IES

- tot

al d

istr

ess

11.0

15.2

*10

.316

.1***

10.5

13.6

*13

.411.

311.

212

.412

.011.

5

Tab

le 3

. Mea

ns o

n th

e R

AN

D-3

6 su

bsca

les a

nd T

-tes

ts to

exa

min

e di

ffere

nces

bet

wee

n gr

oups

on

cate

gori

cal s

ocio

dem

ogra

phic

an

d ca

ncer

-rel

ated

var

iabl

es

*p≤

.05; *

*p≤

.01; *

**p≤

.001


70

4

Cancer-related variablesCorrelational analyses revealed no significant relationships between time since treat-ment and the framework and fulfillment subscales (Table 2). Additional ANOVAs in which time since treatment was divided into five time cohorts did not yield different re-sults from those obtained with the correlational analyses (results not reported in table). T-tests with two treatment groups and ANOVAs with five treatment groups did not show significant effects on framework and fulfillment (only the results of two treatment groups are reported in Table 3). Furthermore, TCSs treated for a second cancer event and those who did not experience a second cancer event had comparable scores on both subscales of the LRI (Table 3).

Life eventsPositive life events were not significantly related to the LRI, but TCSs who had experi-enced more negative life events reported significantly lower mean scores on framework (p = 0.01) and fulfillment (p = 0.003) (Table 2).

Descriptions of changes in outlook on life

Of the 343 TCSs (96.9%) who answered the question whether their outlook on life had changed since their experience with TC, 58.3% (n = 200) reported that their outlook on life had become somewhat or much more positive. Of the TCSs who reported a some-what or much more positive outlook on life, 169 described the way into which it had changed. The most frequently mentioned changes were: 1) more appreciation of life, enjoy-ing life more, a more conscious experience and appreciation of the here and now (n = 93); 2) seeing things in a different perspective: everything in life has become relative (n = 57); 3) reevaluation of priorities, reprioritization of what is important in life (n = 54).Of the 8.2% (n = 28) of the TCSs who reported that their outlook on life had become somewhat or much more negative, 26 described the way into which it had changed. The most frequently mentioned changes were: 1) more anxiety and worries about the future (n = 12); 2) less satisfaction with life (n = 10); 3) feeling more vulnerable and less in con-trol (n = 5).

Correlates and predictors of psychosocial well-being

Mental healthCorrelational analyses demonstrated significant relationships between mental health and negative life events (p = 0.01), framework (p < 0.001), and fulfillment (p < 0.001) (Table 2). Furthermore, TCSs who had a partner, who were employed for wages, who had children and who did not have a chronic disease reported better mental health than their counter-parts (t = -2.30, p = 0.02; t = -3.30, p = 0.001; t = -2.10, p = 0.04; t = -2.60, p = 0.012 respectively). Also, TCSs who had not experienced a second cancer event reported better mental health than TCSs who had experienced a second cancer event (t = -2.12, p = 0.04) (Table 3). The hierarchical regression analysis (Table 4) showed that the sociodemographics, can-cer-related variables, and life events that were entered in the first step, accounted for 11%


4

71

of the variance explained. Employment status, chronic disease and negative life events had independent significant effects. The LRI accounted for an increment of 14% in the second step.

Happiness indexCorrelational analyses demonstrated significant relationships between the happiness index and positive life events (p = 0.03), negative life events (p = 0.01), framework (p < 0.001), and fulfillment ( p < 0.001) (Table 2). Furthermore, TCSs who had a partner, who were employed for wages, and who did not have a chronic disease reported a higher level of life satisfaction than their counterparts (t = -3.84, p < 0.001; t = -2.71, p = 0.007; t = -2.77, p = 0.006 respectively) (Table 3). The sociodemographics, cancer-related variables, and life events that were entered in the first step of the hierarchical regression analyses, accounted for 11% of the variance ex-plained. Marital status, chronic disease, positive life events and negative life events had significant independent effects. The LRI accounted for an increment of 37% in the second step (Table 4).

Social functioningCorrelational analyses showed significant relationships between social functioning and negative life events (p = 0.03), framework (p < 0.001), and fulfillment ( p < 0.001) (Table 2). Furthermore, TCSs who were employed for wages and who did not have a chronic disease reported better social functioning than their counterparts (t = -3.46, p < 0.001; t = -3.00, p = 0.004 respectively). In the hierarchical regression analyses, employment status, chronic disease and negative life events accounted for 9% of the variance explained. The framework and fulfillment subscales were entered in the second step and accounted for a significant increment of 8% in the variance explained.

Correlates and predictors of cancer-related distress

Correlational analyses demonstrated significant relationships between cancer-related di-stress and time since treatment (p = 0.02), and fulfillment (p < 0.001) (Table 2). Fur-thermore, TCSs who had a partner, who were employed for wages, and who had children reported less cancer-related distress than their counterparts (t = -2.05, p = 0.04; t = -2.71, p = 0.007; t = -2.25, p = 0.03 respectively) (Table 3). The hierarchical regression analysis (Table 4) showed that the sociodemographics, can-cer-related variables, and life events that were entered in the first step, accounted for 7% of the variance explained. Employment status and time since treatment had independent significant effects. The fulfillment subscale accounted for an increment of 2% in the second step.


72

4

Dependent variableIndependent variable

ß R2 R2change Fchange

RAND 36 - Mental health

Step 1: Sociodemographics, cancer-related variables, and life events

0.11 0.11 6.98***

Partnera 0.06

Employed for wagesa 0.18***

Childrena 0.10

Chronic diseaseb 0.14**

Second cancer eventc 0.10

Negative life events -0.13*

Step 2: LRI 0.25 0.14 30.58***

Framework 0.21***

Fulfillment 0.39***

Happiness index


0.11 0.11 8.00***

Partnera 0.17*

Employed for wagesa 0.08

Chronic diseaseb 0.13*

Positive life events 0.16**

Negative life events -0.18***

Step 2: LRI 0.48 0.37 121.73***

Framework 0.43***

Fulfillment 0.65***

RAND 36 - Social functioning


0.09 0.09 11.34***

Employed for wagesa 0.18***

Chronic diseaseb 0.18***

Negative life events -0.12*

Step 2: LRI 0.17 0.08 17.16***

Framework 0.14*

Fulfillment 0.29***

Table 4. Standardized regression coefficients from hierarchical regres-sion analyses of significant sociodemographics, cancer-related variables and life events (step 1) and the LRI (step 2) on mental health, happiness index, social functioning, and cancer-related distress


73

4

Discussion

Stressful life events, such as cancer, may threaten the basic assumption that life is mea-ningful and this may have a negative effect on psychosocial well-being. The concept of meaning has not often been studied in male cancer survivors in general, and never in the population of testicular cancer survivors (TCSs). The first aim of the present study was to identify which TCSs experience less meaning. Therefore, we first explored rela-tionships between meaning and sociodemographics, cancer-related variables and recently experience life events. First of all, TCSs with a partner, work or children had a stronger sense of meaning and saw themselves as having fulfilled or being in the process of ful-filling their life goals more often than those without a partner, work or children. These results subscribe to findings of previous research, which identified having an intimate relationship [37-39], children [38] and work [38] as important sources for the experience of a meaningful life. Furthermore, TCSs without a chronic disease saw themselves more often in the process of fulfilling their goals than those with a chronic disease. This finding makes sense, when it is considered that chronic diseases can cause impairments in social activities and work and that they may affect the attainment of goals negatively. Similarly, it has been found that life outlooks of childhood cancer survivors were significantly worse when physical long-term toxicity resulted in substantial limitations in daily activities [40]. Finally, the sense of meaningfulness was less strong in TCSs who had experienced more negative life events during the previous year than in those who had experienced less nega-tive life events. This is consistent with previous research, in which it was found that the experience of more negative life events are related to more negative beliefs about the world (e.g., less belief in the benevolence and meaningfulness of the world) [8]. It has also been assumed that positive life events have a cumulative effect on meaning [8]. However, we did not find evidence for that supposition. A possible explanation for this finding is that

*p≤0.05; **p≤0.01;***p≤0.001; a0 = no, 1 = yes; b0 = yes, 1= no

Dependent variableIndependent variable

ß R2 R2change Fchange

IES - total distress


0.07 0.07 6.68***

Partnera -0.05

Employed for wagesa -0.22***

Childrena -0.07

Time since treatment -0.10

Step 2: LRI 0.09 0.02 5.64**

Fulfillment -0.13*

Tabel 4 continued


74

4

the TCSs had relatively high mean scores on the questionnaire (mean score = 35, maxi-mum score = 42), also compared with means reported in previous research [37;39], and this may have made it difficult to find an effect of positive life events (ceiling effect).In line with earlier research [31;37], age and educational level were not related to meaning, which confirms the idea that meaning can be considered as a universal phenomenon, independent of how old or how well educated one is. Objective cancer-related characteristics were not related to meaning either, which might indicate that the disease period did not threaten their sense of meaning. However, this contradicts with the assumption that cancer can be a life event that disrupts life trajec-tories [41]. One explanation for this finding is that people do not easily abandon their beliefs, but are more inclined to interpret their experiences in ways that fit into their exis-ting belief system [8]. On the other hand, it is also possible that the experience of TC did threaten the sense of meaningfulness of the survivors and that they have adapted to a new life situation. It might be that a response shift has occurred and that they now evaluate life according to new expectations and standards [42], or that they have disengaged from some goals and taken up others [43]. Based on the cross-sectional design of this study we cannot determine which of these mechanisms are responsible for this outcome. Pros-pective longitudinal research may give us more insight into causal relationships between TC, cancer-related characteristics, adaptation processes and meaning, and into changes over time. What we can deduce from these data is that our sample had higher mean scores on the questionnaire than smaller samples of distressed and non-distressed young adults [37] and adults not confronted with cancer [39], signifying that TCSs experience a meaningful life and see themselves in the process of fulfilling their goals. Even the short-term survivors, the more extensively treated TCSs and the TCSs who were treated for a second cancer event have a distinct sense of meaning and feel that they are in the process of fulfilling their goals.

The second aim of the study was to assess changes in outlook on life after the experience with TC. Approximately two-third of the TCSs experienced a change in their outlook on life. These findings are in line with previous research showing that life-threatening ill-ness can change meaning (e.g., [18;44;45]). By far most of the TCSs reported that their outlook on life had become more positive. Research shows that people tend to reappraise a negative event in a way that they find positive benefit from a situation [2;8]. Park and Folkman [8] call this the meaning-making process and consider it to be a coping strategy useful for reducing discrepancies between the appraised meaning of an event and the pre-existing belief system. Other researchers used terms such as posttraumatic growth (as opposed to posttraumatic stress), and benefit finding to describe positive outcomes of a potential traumatic event [18]. This change in outlook was not evident from the results from the questionnaire, because the questionnaire did not focus on changes in meaning. Perhaps the above-mentioned lack of effect of the cancer-related variables on meaning is reducible to the type of questionnaire used in this study. It is also imaginable that the above-mentioned cancer-variables act more as catalysts for the development of a new sense of meaning than that they are related to the state of a belief system at one particular point in time.


4

75

Thirdly, we investigated the relationship between meaning and specific domains of func-tioning, since research has shown that meaning may be related more strongly to some domains (e.g., mental health) than to others (e.g., social functioning) [20]. In agreement with the suggestion of Vickberg et al. [20], we found that meaning related more strongly to some domains than to others. However, we did not find that the relation of meaning with mental health was much stronger than with social functioning. We found that meaning had the strongest effect on satisfaction with life and the weakest effect on cancer-related distress. Meaning accounted for a significant increment in the prediction of functioning, in addition to sociodemographics, cancer-related variables and life events. An interesting finding was that the fulfillment subscale was a more important predictor than the frame-work subscale. This indicates that the subjective feeling of meaningfulness is a more im-portant determinant of well-being than just the presence of a framework [32;39]. A further important finding was that approximately the same sociodemographics, can-cer-related variables and life events correlated with both psychosocial well-being and mea-ning. These variables were more strongly associated with meaning than with psychosocial well-being, which might indicate that people need a source of meaning to feel happy [46]. Furthermore, this finding indicates the importance of investigating meaning when one wants to understand factors important for well-being [39]. To clarify this, we can use the data of this specific study. The variables marital status, employment status, chronic disease and negative life events correlated significantly with psychological well-being, but more strongly with meaning. So, the experience of not being married, unemployment, having a chronic disease, and negative life events may have impeded on the fulfillment of life goals of TCSs, which contiguously affected their psychological well-being. As mentioned before, having a sense of meaning appeared to have a stronger effect on psychosocial well-being than on cancer-related distress. This finding may be related to the type of questions and the instructions given for the questionnaires being used. The Impact of Event Scale, used to assess cancer-related distress, focusses on cancer-related posttraumatic stress symptoms. The measures used to assess psychosocial well-being con-cern the perception of well-being in general. Therefore, and in line with others [15;20], one may conclude that a sense of purpose and fulfillment is more important to the experience of positive quality of life than to coping with cancer-related post traumatic stress symp-toms. Does this mean that an intervention based on meaning-making would not be useful for the reduction of cancer-related distress (i.e., posttraumatic stress symptoms)? No, not necessarily. Perhaps these findings merely indicate that the underlying mechanism through which meaning and cancer-related distress associate, is more complex. Vickberg et al. [47] suggest, for example, that it may very well be that meaning serves as a buffer against psychological distress. They found no direct relationship between meaning and intrusive thoughts, but they did find that these thoughts were less distressing for the cancer survivors with a stronger sense of meaning than for those with a lower sense of meaning. They reasoned that, although having a strong sense of meaning may not protect against recurrent thoughts, it may be that a cancer survivor who does find meaning in life is better able to put these distressing thoughts into perspective. Alternatively, they sug-gested that it may be that the content of the intrusive thoughts of cancer survivors with a stronger sense of meaning is different (i.e., less threatening) from that of survivors with a lower sense of meaning. If future research confirms that meaning does indeed function as


76

4

a protective factor, than it may be relevant for health care workers to focus their interven-tions on both the reduction of cancer-related distress and the encouragement of creating a stronger sense of meaning in distressed cancer survivors [47].

Although this study has fostered our insights into some of the contextual variables that are related to meaning, there are several limitations that have prevented us from fully understanding the impact of the cancer-experience on meaning. Because of the cross-sec-tional design of our study, we did not assess meaning before the onset of TC. Therefore, we do not know if and how the cancer-experience may change meaning. Yet, both the responses to our open question and results from interview studies have shown that the confrontation with cancer can raise existential issues. Survivors have reported that they experienced their cancer event as life altering; it had changed the way they looked at the world and altered their goals and priorities (e.g., [48-50]). Furthermore, our way of inves-tigating the relationship between TC and meaning, by focusing on objective disease cha-racteristics, might have been insufficient. As noted previously, Zebrack and Chesler [40] reported that life outlooks were influenced by the extent to which after-effects interfered with daily activities. This suggests that concrete physical and psychosocial after-effects of cancer need to be assessed and related to meaning. The cross-sectional nature of our study has also prevented us from interpreting the directionality of the established ef-fects between the antecent variables, meaning, and psychosocial well-being, which means that it is also possible that psychosocial well-being influences the relations between the antecedent variables and meaning. The questionnaire that we have used might have li-mited the results as well. Although it is a psychometrically sound instrument, the items are somewhat abstract [39], whereas beliefs and goals are generally rather concrete and personal [8]. Finally, the response rate of TCSs was 50%, and this may have induced bias, even though analyses showed that responders and non-responders did not differ in age, time since diagnosis and type of treatment received.

Overall, it can be concluded that, despite the experience with an illness capable of shat-tering basic beliefs and the attainment of valued goals, TCSs do have a strong sense of meaning. Single TCSs, as well as childless TCSs, those without employment, those with a chronic disease and those who have experienced more negative life events experienced somewhat less meaning than their counterparts. Furthermore and in accordance with theory and research, this study showed that having a sense of meaning was strongly re-lated to the psychosocial well-being of TCSs. The effect of meaning on cancer-related distress was less strong. However, because previous research has shown that this relation-ship may be more complex, it would be relevant to further investigate the underlying mechanisms through which meaning and cancer-related distress relate.


4

77

Reference List

1. Festinger L (1957) A theory of cognitive dissonance. New York

2. Taylor SE (1983) Adjustment to threatening events - A theory of cognitive adaptation. American Psycholo-gist 38:1161-1173

3. Lepore SJ (2001) A social-cognitive processing model of emotional adjustment to cancer. In: Baum A, Anderson B (eds) Psychosocial interventions for cancer. APA, Washington DC, pp 99-118

4. Redd WH, DuHamel KN, Johnson Vickberg SM, Ostroff JL, Smith MY, Jacobsen PB, Manne SL (2001) Long-term adjustment in cancer survivors: Integration of classical conditioning and cognitive-processing models. In: Baum A, Anderson B (eds) Psychosocial interventions for cancer. APA, Washington DC, pp 77-98

5. Janoff-Bulman R (1989) Assumptive worlds and the stress of traumatic events: applications of the schema construct. Social Cognition 7:113-136

6. Carver CS, Scheier MF (1998) On the self-regulation of behavior. Cambridge University Press, New York

7. Reker GT, Wong PT (1988) Aging as an individual process: Towards a theory of personal meaning. In: Birren JE, Bengston VL (eds) Emergent theories of aging. Springer, New York, pp 214-246

8. Park CL, Folkman S (1997) Meaning in the context of stress and coping. Review of General Psychology 1:115-144

9. White CA (2004) Meaning and its measurement in psychosocial oncology. Psycho Oncology 13:468-481

10. Tedeschi RG, Calhoun LG (1995) Trauma and transformation - growing in the aftermath of suffering. SAGE Publications, London

11. Janoff-Bulman R (1992) Shattered assumptions: Towards a new psychology of trauma. Free Press, New York

12. Silver RL, Boon C, Stones MH (1983) Searching for meaning in misfortune - Making sense of incest. Journal of Social Issues 39:81-101

13. Thompson SC (1991) The search for meaning following a stroke. Basic and Applied Social Psychology 12:81-96

14. Frankl VE (1959) Man’s search for meaning. Hodder & Stoughton, London

15. Tomich PL, Helgeson VS (2002) Five years later: A cross-sectional comparison of breast cancer survivors with healthy women. Psycho Oncology 11:154-169

16. Green BL, Rowland JH, Krupnick JL, Epstein SA, Stockton P, Stern NM, Spertus IL, Steakley C (1998) Prevalence of posttraumatic stress disorder in women with breast cancer [see comments]. Psychosomatics 39:102-111

17. Tomich PL, Helgeson VS (2004) Is finding something good in the bad always good? Benefit finding among women with breast cancer. Health Psychology 23:16-23

18. Lechner SC, Zakowski SG, Antoni MH, Greenhawt M, Block K, Block P (2003) Do sociodemographic and disease-related variables influence benefit-finding in cancer patients? Psycho Oncology 12:491-499

19. Cordova MJ, Cunningham LLC, Carlson CR, Andrykowski MA (2001) Posttraumatic growth following breast cancer: A controlled comparison study. Health Psychology 20:176-185

20. Vickberg SMJ, DuHamel KN, Smith MY, Manne SL, Winkel G, Papadopoulos EB, Redd WH (2001) Global meaning and psychological adjustment among survivors of bone marrow transplant. Psycho Onco-logy 10:29-39




78

4


24. van’t Spijker A, Trijsburg RW, Duivenvoorden HJ (1997) Psychological sequelae of cancer diagnosis: a meta-analytical review of 58 studies after 1980. Psychosomatic Medicine 59:280-293

25. Brennan J (2001) Adjustment to cancer - Coping or personal transition? Psycho Oncology 10:1-18

26. Debats DL (1990) The life regard index: reliability and validity. Psychological Reports 67:27-34

27. Battista J, Almond R (1973) The development of meaning in life. Psychiatry 36:409-427



30. Veenhoven, R. Conditions of happiness. 1984. Erasmus University, Rotterdam.

31. Debats DL, Vanderlubbe PM, Wezeman FRA (1993) On the psychometric properties of the Life Regard Index (LRI) - A measure of meaningful life - An evaluation in 3 independent samples based on the Dutch version. Personality and Individual Differences 14:337-345

32. Debats DL (1996) Meaning in life: clinical relevance and predictive power. British Journal of Clinical Psy-chology 35 ( Pt 4):503-516

33. Horowitz MJ, Wilner N, Alvarez W (1979) Impact of Event Scale: A measure of subjective stress. Psycho-somatic Medicine 41:209-218

34. Van der Ploeg E, Mooren TTM, Kleber RJ, van der Velden PG, Brom D (2004) Construct validation of the Dutch version of the impact of event scale. Psychological Assessment 16:16-26


36. Van de Willige G, Schreurs P, Tellegen B, Zwart F (1985) Het meten van ‘life events’: de Vragenlijst Recent Meegemaakte Gebeurtenissen (VRMG). Nederlands Tijdschrift voor de Psychologie 40:1-19

37. Debats DL (1999) Sources of meaning an investigation of significant commitments in life. Journal of Hu-manistic Psychology 39:30-57

38. Ryff CD, Singer B (1998) The contours of positive human health. Psychological Inquiry 9:1-28

39. Scannell ED, Allen FCL, Burton J (2002) Meaning in life and positive and negative well-being. North American Journal of Psychology 4:93-112


41. Zebrack BJ (2000) Cancer survivor identity and quality of life. Cancer Practice 8:238-242


43. Carver CS, Scheier MF (2000) Scaling back goals and recalibration of the affect system are processes in normal adaptive self-regulation: understanding ‘response shift’ phenomena. Social Science & Medicine 50:1715-1722

44. Dirksen SR (1995) Search for Meaning in Long-Term Cancer Survivors. Journal of Advanced Nursing 21:628-633

45. Fromm K, Andrykowski MA, Hunt J (1996) Positive and negative psychosocial sequelae of bone marrow transplantation: implications for quality of life assessment. Journal of Behavioral Medicine 19:221-240

46. Zika S, Chamberlain K (1992) On the relation between meaning in life and psychological well-being. Brit-ish Journal of Psychology 83:133-145


4

79

47. Vickberg SM, Bovbjerg DH, DuHamel KN, Currie V, Redd WH (2000) Intrusive thoughts and psycho-logical distress among breast cancer survivors: global meaning as a possible protective factor. Behavioral Medicine 25:152-160

48. Brodsky MS (1995) Testicular cancer survivors’ impressions of the impact of the disease on their lives. Quality of Life Research 5:78-96

49. Ferrell BR, Grant M, Funk B, Otis-Green S, Garcia N (1997) Quality of life in breast cancer. Part I: Physi-cal and social well- being. Cancer Nursing 20:398-408

50. O’Connor AP, Wicker CA, Germino BB (1990) Understanding the cancer patient’s search for meaning. Cancer Nursing 13:167-175


5Objective and subjective predictors

of cancer-related stress symptoms in


Joke Fleer

Dirk Th Sleijfer

Harald J Hoekstra

Marrit A Tuinman

Ed C Klip


Patient Education and Counseling, in press (slightly adapted version)


5

83

Introduction

Testicular cancer is most prevalent among men between the ages of 15 and 45, a phase of life during which people tend to enter into intimate relationships, have children, and build careers. Almost 90% survives after treatment and will live the largest part of their life as a cancer survivor [1;2]. Research into the physical sequel of treatment for testicu-lar cancer shows that testicular cancer survivors (TCSs) who were treated for metasta-tic disease have an increased risk of infertility, fatigue, second primary malignancies, decreased renal functioning, hearing deficits and cardiovascular disease [1;2]. Studies that focus on psychosocial sequel show that most TCSs report a good quality of life, but a small proportion shows evidence of anxiety, depression, work-related problems, and sexual problems [3;4]. The presence of these problems suggests that the experience with testicular cancer may cause ongoing distress after cure has been achieved. Yet, the prevalence of cancer-related stress symptoms (e.g., intrusively experienced thoughts and feelings about the event) has never been studied in this population and little is known about the factors that are associated with it, whereas this knowledge is essential for an early detection of distressed TCSs. In general, only a minority (10-20%) of the cancer patients exhibits problems following treatment that indicate cancer-related stress for which professional psychological care is required. The majority adapts well [5;6]. Several studies have shown that sociode-mographics are associated with increased levels of cancer-related stress symptoms in survivors of cancer, such as younger age, lower education, and lower socio-economic status. Poor social support and reduced physical functioning have also been found to predict distress after cancer [6]. The evidence with respect to relationships between objective cancer-related characteristics (e.g., recency of treatment, extensiveness of treatment) and cancer-related stress is mixed [6], but if associations are found they are mainly weak [7-10]. Kazak et al. [9;11] argued that concrete events surrounding diagnosis and treatment may also contribute to the onset of cancer-related stress. The experience of cancer consists of a series of successive potential stressors, including hearing the diagnosis, beginning and ending treatment, the appearance of treatment side effects, follow-up care, and possible relapse. Long-term consequences of treat-ment, including infertility, cognitive deficits (e.g., concentration disorders), and fatigue can subsequently contribute to the persistence of cancer-related stress [6;12]. Other studies have shown that cancer survivors who believe that the cancer experience still has a harmful effect on their daily lives (e.g., as a result of physical after-effects of treatment or a continuous fear for recurrence) experience more distress [7;8;13] and a worse quality of life [14]. It has been suggested that these subjective appraisals are more important contributors than objective variables [7-10]. The goal of this study is to investigate the prevalence of cancer-related stress symptoms after testicular cancer, and relationships between these symptoms and objective and sub-jective aspects of diagnosis and treatment. The study consists of two parts. In the first part the prevalence of cancer-related stress symptoms and its relationship with objective so-ciodemographic and cancer-related variables is investigated among 354 TCSs who com-pleted a self-report questionnaire. Based on findings of prior research, it is expected that a minority (<20%) of the TCSs will experience cancer-related distress. Younger TCSs,


84

5

those with a lower educational level, without a partner, without paid employment, or with a chronic disease are at risk of higher levels of cancer-related distress. Furthermore, it is hypothesized that recency of treatment, extensiveness of treatment and the experience of a second cancer event are only weakly related to higher levels of cancer-related distress.Because the associations between objective cancer-related variables and cancer-related distress tend to be weak, the second part of the study aims to gain insight into the relationship between cancer-related stress symptoms and subjective aspects of diagno-sis and treatment, using interviews. The aim of the interview is to explore if concrete events (e.g., hearing the diagnosis) and the subjective perceptions of these events (e.g., emotions elicited by hearing the diagnosis) may explain why some TCSs experience ongoing stress problems while others do not. Because the relationship between con-crete events and cancer-related stress has received little empirical attention until now, we do not have a specific hypothesis regarding which events are related to cancer-re-lated stress. However, based on the previous finding that distressed cancer patients do more often perceive a harmful effect of the cancer experience on their lives than those without such stress, we do hypothesize that the concrete events elicit more emotions in TCSs with clinically elevated levels of cancer-related stress symptoms.


Procedure and patients

QuestionnaireAll of the men who had been treated for testicular cancer at the University Medi-cal Center Groningen (UMCG; the Netherlands) between 1977 and 2003 were ap-proached. In 1977 cisplatin-based chemotherapy, which led to the improved survival rates, was introduced into the treatment for metastatic testicular cancer at the UMCG. Exclusion criteria were as follows: age younger than 18 years at study entry, insufficient command of Dutch, and time since completing treatment of less than three months. A letter containing information about the objectives of the study and an informed consent form were sent to each of the 702 eligible TCSs. This form was used to obtain informed consent for both the questionnaire and the interview. Fifty percent (n = 354) of the TCSs who were approached agreed to participate and subsequently received questionnaires that were accompanied by prepaid return envelopes. Analyses showed that participating TCSs did not differ from non-participants in age, time since diag-nosis and type of treatment.

InterviewAfter the questionnaires were returned and scored, two groups were selected for inter-views according to their answer pattern on the Dutch version of the Impact of Event Scale (IES) [15-17]: TCSs with low IES scores (0-7; n = 171) and those with high scores (26 or higher; n = 46). From each group, a random selection of 15 TCSs was drawn and these men were invited for interviews. Based on the preference of the par-


5

85

ticipants, the interviews were conducted at the participant’s house or at the UMCG. The study was approved by the Medical Ethics Committee of the UMCG.

Measurement instruments

QuestionnaireThe following sociodemographic variables were collected: age, education level, marital status, employment status, and the prevalence of chronic disease. The highest level of completed education was measured along the following seven-point scale: primary school [1], lower vocational [2], lower secondary [3], middle secondary [4], high se-condary [5], higher vocational [6], and advanced university [7]. In addition, the follow-ing cancer-related data were collected from medical records: time since completion of treatment, type of treatment, and the presence of a second cancer event.

The Dutch version of the Impact of Event Scale (IES) [15-17] was used to assess stress symptoms that were specifically related to the experience of testicular cancer. This questionnaire gathers information concerning the degree to which confrontation with testicular cancer was influencing the respondent’s current daily life. Two dimensions were measured with 15 items: intrusion (intrusively experienced thoughts, images, feelings, or bad dreams about the event) was measured with 7 items, using the fol-lowing response categories: “not at all” (0), “rarely” (1), “sometimes” (3), and “often” (5). Avoidance of unpleasant feelings or recollections of the event were measured with 8 items, using the same response categories. Total scores (intrusion + avoidance) higher than 26 strongly indicated severe cancer-related stress, for which psychological help is recommended. Scores between 8 and 26 are cause for extra attention, while scores below 8 indicate the absence of cancer-related stress. The internal consistency of this questionnaire was good (alpha = 0.88).

InterviewThe Impact of Traumatic Stressors Interview Schedule (ITSIS) is a semi-structured interview, surrounding seven content areas [9]. The following aspects are considered within these content areas:

1. Memory for treatment-related experiences. Questions concerning (a) recollections of the treatment period (b) subjects that family members continue to discuss.

2. Re-experiencing. Questions concerning (a) the prevalence of unwanted memories and (b) the extent to which these intrusions influence daily life.

3. Criteria for DSM-IV PTSD diagnosis. Questions concerning (a) intense fear, (b) horrifying events, and (c) feelings of helplessness during the period of diagnosis and treatment.

4. Thoughts of respondents concerning why they had cancer.


86

5

5. Medical problems. Questions concerning (a) existing problems that are related to cancer and treatment, (b) the extent to which these problems upset the respon-dents, (c) the extent to which these problems make life more difficult, (d) the ex-tent to which these problems are visible to others, and (e) the prevalence of fear for future medical problems.

6. Perceptions of personal changes due to cancer and treatment. Questions concern-ing (a) whether and to what extent respondents were changed by cancer and (b) if so, in what ways they were changed.

7. Communication. Questions concerning (a) the extent to which respondents con-tinue to talk about the period of diagnosis and treatment with others, (b) with whom they discuss this period, (c) the topics they discuss, and (d) how respon-dents feel when talking about these topics.

The ITSIS incorporates both unstructured and structured components. The present study concentrates on the structured part of the interview and describes the responses of the TCSs to: (a) a five-point Likert scale, which was used to assess the subjective percep-tions that cause some TCSs to experience ongoing stress problems while others do not (e.g., “Can you use this scale to indicate how afraid you felt?” Response categories: 1 = not at all, 2 = slightly, 3 = somewhat, 4 = considerably, 5 = very much); (b) a Q-sort of cards containing descriptions of events to examine which specific events cause some TCSs to experience ongoing cancer-related stress problems while others do not. The cards of the Q-sort were sorted into “yes” (experienced) or “no” (not experienced), according to specific questions, (e.g., “Which of the following events made you feel afraid?”). No Q-sort was developed for content area 4.ITSIS was originally developed for children and adolescents who have had cancer, and for their parents. The events experienced by men with testicular tumors are somewhat differ-ent. Some events were not applicable and were therefore omitted. In addition, a number of cards were adapted (e.g., “having an operation” was replaced by “surgical removal of a testicle”). A number of additional events were also included, because these were described in the literature as possibly stressful. The Q-sorts that were used are described in Ap-pendices 1-5, which also indicates whether the texts were translated literally, adapted, or added. The inter-rater reliability of the original version of the ITSIS is high [9]. The validity of the original version of the ITSIS is determined by correlating the total scores of the Q-sorts from each content area with the scores on the self-report scales, including the IES. Some of the correlations were significant [9]. The validity of the original ITSIS was reconfirmed in later analyses [11]. The reliability of the translated version, used in the present study, has yet to determined, as a single interviewer conducted all of the interviews. This interviewer was informed and trained by the original authors.


5

87


Questionnaire. Statistical analyses were performed using SPSS 12.0. Descriptive statis-tics were used to calculate means, frequencies, and percentages. Pearson correlations, T-tests and ANOVAs were conducted to determine which sociodemographic and cancer-related variables were significantly related to the subscales and total score of the IES. For employment status, a dichotomous variable was created with the catego-ries of “not employed for wages” (consisting of students, unemployed persons, those who were unable to work, and retirees) and “employed for wages” (consisting of those who were employed for wages). The effect of type of treatment was investigated accor-ding to two classifications. First, TCSs were divided into the following five groups: (1) “orchidectomy” survivors had undergone the surgical removal of the affected testicle only; (2) “orchidectomy + RPLND” survivors had undergone orchidectomy, followed by retroperitoneal lymph node dissection; (3) “radiotherapy” survivors had undergone orchidectomy + radiotherapy; (4) “chemotherapy” survivors had undergone orchidec-tomy + chemotherapy; (5) “chemotherapy + RRTM” survivors had undergone orchi-dectomy, additional chemotherapy, and resection of residual tumor mass. Second, the following two treatment groups were compared: “surgical treatment” (consisting of the categories orchidectomy and orchidectomy + RPLND) and “combined treatment” (consisting of the categories orchidectomy + radiotherapy, orchidectomy + chemo-therapy, and orchidectomy + chemotherapy + RRTM).Interview. For the interview, a random selection was made of the 15 survivors with a score of 26 or higher on the IES and 15 with a score of ≤ 7 through SPSS 12. Descriptive statistical procedures were used to calculate means, medians, frequencies, and percentages. T-tests and chi2 tests were used to compare the sociodemographic and cancer-related characteris-tics of the interviewed TCSs (n = 30) to those of the total population (n = 354). Because of the small sample sizes, non-parametric tests were used to investigate differences between the TCSs with cancer-related distress (n = 15) and those with no distress (n = 15). Mann-Whitney U tests and chi2 tests were used to compare sociodemographic and cancer-related characteristics of the two groups. Furthermore, Mann-Whitney U tests were used to com-pare the scores of the groups on the Likert scales and chi2 tests on the Q-sort. In this study, results were considered statistically significant if the probability of oc-currence was 0.05 or less. Inasmuch as this is a descriptive and explorative study, no formal adjustments were employed to correct for multiple testing. Therefore, caution is warranted in interpretation of findings of p > 0.01.

Results

Questionnaire

Descriptive statistics concerning the participants that completed the questionnaire (n = 354) are summarized in Table 1. As predicted, a minority of TCSs experienced elevated levels of cancer-related distress. Forty-six (13.1%) of 354 participants reported clinically elevated levels of cancer-related


5

88

Table 1. Sample characteristics total group


TCSs (N = 354*)



Marital status N, %





Student 12 3.4

Out of work 6 1.7


Retired 32 9.1

Children N, %

No 127 36

Yes 226 64


No 299 84.9

Yes 53 15.1


Positive 1.6 2.0

Negative 1.3 1.6


Treatment N,%









Yes 31 8.8

No 322 91.2


5

89

stress (IES total ≥ 26), whereas 48.3% reported no distress concerning their experience with TC (IES total ≤ 7). A further 38.6% had scores between 8 and 25. The mean score on the intrusion subscale was 6.4 (SD = 6.7); the mean score on the avoidance sub-scale was 5.2 (SD = 6.8), and the mean score for total distress was 11.6 (SD = 12.1).Age and chronic disease were unrelated to the IES, but single TCSs and those who were not employed for wages reported higher levels of avoidance (resp. t = 2.10, p = 0.04; t = 2.85, p = 0.005) and total distress (resp. t = 2.05, p = 0.04; t = 2.71, p = 0.007) than did other categories of respondents. Furthermore, lower levels of educa-tion were associated with higher levels of avoidance (r = -0.12, p < 0.05). This means that our prediction of higher levels of cancer-related distress in younger TCSs, as well as in those with a lower educational level, without a partner, without paid employ-ment, and with a chronic disease was partly supported.The hypothesis that objective cancer-related variables are only weakly related to higher levels of cancer-related distress was confirmed. Time since treatment was inversely associated with intrusion (r = -0.14, p = 0.012), avoidance (r = -0.11, p = 0.04), and total distress (r = -0.13, p = 0.013). Type of treatment was not related to the subscales or total score of the IES, neither in the analyses of the five treatment groups nor in those that considered only two groups. Second cancer event was unrelated to the IES as well.

Interview

Based on the scores of the IES, a random sample of 15 of the 46 TCSs with clinically elevated levels of cancer-related stress symptoms (hereafter referred to as ‘group 1’) and 15 of the 171 TCSs with no cancer-related stress symptoms (hereafter referred to as ‘group 2’) were asked for an interview. These 30 TCSs did not differ from the total group of TCSs (n = 354) in sociodemographic and cancer-related characteristics. However, the TCSs in group 1 were significantly younger than were those in group 2 (median = 36.5, range = 24.1 - 72.3 as compared to median = 48.6, range = 31.6-60.9; Z = -2.18 , p = 0.03), more likely to be single (N = 4/15 as compared to N = 0/15), not employed for wages (N = 6/15 as compared to N = 0/15), and had been treated more recently (median = 3.4 years, range = 7 months - 23.5 years as compared to median = 16.4 years, range = 2.8-22.1 years; Z = -2.76, p = 0.006).

1. Memory of treatment-related experiences. Results from the Q-sort (Appendix 1) showed that both groups recalled approximately the same number of events sur-rounding the period of diagnosis and treatment of TC. However, group 1 was sig-nificantly more likely to recall the experience of undergoing a testicular ultrasound and of feeling ashamed that they had testicular cancer. The families of TCSs with clinically elevated levels of cancer-related stress symptoms talked about events sur-rounding diagnosis and treatment significantly more often than did the families of those with no cancer-related stress symptoms. Anxiety about side effects, the fact that the experience of having testicular cancer revealed who their true friends were, and the surgical removal of the testicle were significantly more frequent topics of discussion (Table 2).


90

5

TC

Ss w

ith

SRS

TC

Ss w

ith

no S

RS

Com

pari

son

of g

roup

s on

tota

l num

ber o

f ite

ms,

by Q

-sort

Mea

n (S

D)

Mea

n (S

D)

Zp

Item

-leve

l com

pari

son

of g

roup

s, b

y Q

-sor

tn

nC

hi2

p

Con

tent

are

a 1 (

Q-s

ort o

f 49

item

s, se

e A

ppen

dix

1)

Rec

ollec

tions

of t

reat

men

t per

iod

29.9

(8.6

) 27

.8 (6

.4)

-0.6

90.

49U

nder

goin

g te

stic

ular

ultr

asou

nd13

75.

400.

02B

eing

ash

amed

that

I ha

ve te

stic

ular

can

cer

61

4.66

0.03

Subj

ects

disc

usse

d by

fam

ily m

embe

rs10

.7 (7

.8)

4.1 (

5.0)

-2.33

0.02

Bei

ng w

orrie

d of

get

ting

sick

from

side

effe

cts o

f tre

atm

ent

50

6.00

0.01

That

I di

scov

ered

who

my

real

frie

nds a

re9

35.

000.

03Su

rgic

al re

mov

al o

f the

test

icle

72

3.97

0.05

Con

tent

are

a 2

(Q-s

ort o

f 46

item

s, se

e A

ppen

dix

1)

Even

ts ab

out w

hich

TC

Ss h

ave u

nwan

ted

mem

ories

9.9

(7.8

)4.

2 (4

.3)-2

.25

0.02

Hur

ting

or h

avin

g pa

in7

16.

140.

01G

oing

to th

e em

erge

ncy

room

40

4.62

0.03

Wai

ting

for t

he re

sults

of t

umor

mar

ker t

ests

72

3.97

0.03

Con

tent

are

a 3

(Q-s

ort o

f 46

item

s, se

e A

ppen

dix

1)

Crit

eria

for D

SM-I

V P

TSD

dia

gnos

is: in

tens

e fea

r 9.

0 (3

.9)

5.6 (5

.2)

2.40

0.02

Surg

ical

rem

oval

of t

he te

stic

le9

35.

000.

03

Wai

ting

for X

-ray

resu

lts11

54.

820.

03

Tab

le 2

. Res

pons

e pa

tter

ns fo

r th

e Q

-sor

ts


5

91

TC

Ss w

ith

SRS

TC

Ss w

ith

no S

RS

Com

pari

son

of g

roup

s on

tota

l num

ber o

f ite

ms,

by Q

-sort

Mea

n (S

D)

Mea

n (S

D)

Zp

Item

-leve

l com

paris

on o

f gro

ups,

by Q

-sor

tn

nC

hi2

p

Crit

eria

for D

SM-I

V P

TSD

dia

gnos

is: h

orro

r1.9

(3.5)

0.6

(0.9

)-0

.27

0.79

Crit

eria

for D

SM-I

V P

TSD

dia

gnos

is: h

elples

snes

s5.

(5.7

)2.

1 (2.

6)-1

.78

0.08

Feel

ing

sick

from

che

mot

hera

py4

0.0

4.62

0.03

Con

tent

are

a 5

(Q-s

ort o

f 20

item

s, se

e A

ppen

dix

2)

Perc

eptio

n of

med

ical p

robl

ems

4.2

(4.0

)2.

9 (2

.8)

-0.6

50.

52C

once

ntra

tion

prob

lem

s6

14.

660.

03H

eada

ches

40

4.62

0.03

Con

tent

are

a 6

(Q-s

ort o

f 14

item

s, se

e A

ppen

dix

3)

Idea

s abo

ut p

erso

nal c

hang

es d

ue to

testi

cula

r can

cer

6.3 (

4.0)

3.5 (2

.9)

-2.30

0.02

How

I m

ake

frie

nds

70

9.13

0.00

3M

y pl

ans f

or th

e fu

ture

115

4.82

0.03

Con

tent

are

a 7

(Q-s

ort o

f 4 it

ems,

see A

ppen

dix

4; Q

-sor

t of 7

item

s, se

e App

endi

x 5)

Topi

cs th

at a

re ta

lked

abo

utW

hat m

ay h

appe

n in

the

futu

re b

ecau

se I

had

canc

er9

36.

240.

01Fe

eling

s whe

n ta

lkin

g abo

ut th

ese t

opics

I f

eel u

nder

stoo

d13

105.

280.

02I f

eel r

elie

ved

105

5.32

0.02

I fee

l sca

red

30

3.88

0.05

Tab

le 2

con

tinu

ed


92

5

2. Re-experiencing. The Q-sort (Appendix 1) showed that TCSs within group 1 had significantly more recollections concerning events surrounding diagnosis and treat-ment than did those within group 2. They thought significantly more frequently back to moments in which they had pain, visits to the emergency room, and waiting for results of tumor marker tests (Table 2). Furthermore, group 1 scored significantly higher than group 2 on the five-point Likert scale with which respondents reported the extent to which they were disturbed by such recollections (Z = -3.05, p = 0.002). They were also significantly more likely to report that these memories make their lives more difficult (Z = -2.8, p = 0.01). These findings confirm the prediction that distressed TCSs more often perceive an impact of the cancer experience on their current lives than those without this stress and that their recollections elicit more emotions.

3. Criteria for DSM-IV PTSD diagnosis. The two groups mentioned comparable num-bers of events that had caused feelings of horror or helplessness (Q-sort, Appendix 1). Group 1 mentioned significantly more events that had caused intense fear. The surgi-cal removal of the testicle and waiting for X-ray results caused intense fear among significantly more TCSs within group 1, and illness resulting from chemotherapy caused more feelings of helplessness among this group (Table 2). These findings con-firm the prediction that the events elicited more emotions in TCSs with cancer-re-lated stress.

4. Thoughts of respondents concerning why they had cancer. Almost all of the participants, both in group 1 (n = 13/15) and group 2 (n = 12/15), reported that testicular cancer was something that had just happened to them and that nobody knew why it hap-pened.

5. Medical problems. Results from the Q-sort (Appendix 2) showed that both groups reported similar levels of medical problems resulting from the treatment of testicu-lar cancer. The TCSs within group 1 were significantly more likely to report suffer-ing from concentration problems and headaches (Table 2). As predicted, group 1 perceived more emotions with respect to these problems. They scored significantly higher on the Likert scale that measured the extent to which these problems upset them (mean = 2.9 as compared to 1.3, Z = -3.33, p = 0.001). Furthermore, they re-ported significantly more changes in their daily lives that had resulted from these problems (mean = 2.4 as compared to 1.2, Z = -3.05, p = 0.001), and they were significantly more anxious about future medical problems (mean for group 1 = 2.1 as compared to group 2 = 1.4, Z = -2.05, p = 0.06).

6. Perceptions of personal changes due to cancer and treatment. The Q-sort (Appendix 3) showed that TCSs within group 1 reported significantly more aspects of their daily lives that had changed than did the TCSs within group 2. They were significantly more likely to report making friends differently and that their plans for the future had changed (Table 2). As expected, the distressed TCSs more often perceive an impact


5

93

of the cancer experience on their current lives than those without this stress, but the suggestion that this impact is experienced as harmful can not be investigated with these data.

7. Communication. A comparison of the scores of the two groups on the five-point Likert scale shows that the TCSs within group 1 reported talking about their ex-periences with cancer significantly more frequently than did those within group 2 (mean = 2.7 as compared to 1.6, Z = -2.93, p = 0.003). Most of the TCSs reported talking about their experiences with their partners (group 1 n = 13, group 2 n = 12). The Q-sort (Appendix 4) showed that TCSs from group 1 talked with their friends about their experiences with cancer significantly more frequently than did TCSs from group 2 (n = 13 as compared to n = 5, chi2 = 10.90, p = 0.001), as well as with their fathers (n = 9 as compared to n = 2, chi2 = 7.99, p = 0.005), mothers (n = 8 as compared to n = 2, chi2 = 6.15, p = 0.01), and therapists (n = 5 as compared to n = 0, chi2 = 6.47, p = 0.01). The TCSs within group 1 talked more frequently about the possible future consequences of their treatment for testicular cancer. Talking about what had happened relieved the TCSs within group 1 more frequently and gave them the feeling of being understood by others (Q-sort, Appendix 5). Talking about negative emotions, however, caused anxiety more frequently among TCSs within group 1 (Table 2). These findings indicate that the cancer event still elicits more emotions in TCSs with cancer-related stress symptoms than in those without these symptoms.

Discussion

As hypothesized, a minority of the men who had been cured of testicular cancer (TCSs) (13%) reported clinically elevated levels of cancer-related stress symptoms. In contrast, 50% reported few or no distress symptoms concerning their experience with testicular cancer. This is the first study to examine the prevalence of cancer-related stress symptoms among men with a form of cancer that has a high chance of recovery. Most studies of cancer-related stress symptoms in adults have focused on populations of patients that are not comparable to TCSs with regard to such variables as prognosis, type of treatment, and sex. The results of such studies, however, are actually compa-rable to those of the current study; in general, only a minority of the cancer patients exhibited problems that indicate cancer-related stress, while the majority were able to adapt well [5;6]. It is therefore likely that other factors than estimated chance of sur-vival contribute to the onset and persistence of cancer-related distress. We investigated if cancer-related stress symptoms related to sociodemographics, objective cancer-re-lated variables, concrete events and subjective perceptions.Based on findings from prior research, we predicted that younger age, a lower educa-tional level, being single, being without paid employment, and having a chronic disease would relate to cancer-related distress. This hypothesis was partly confirmed. Singles, those with less education, and those who were not employed for wages reported higher levels of cancer-related stress symptoms than did other categories of TCSs. It is pos-sible that singles, those with less education, and those who are not employed for wages


94

5

have fewer emotional, cognitive, and material resources at their disposal to help them cope with stressful situations [18].We expected that the association between cancer-related stress symptoms and objec-tive cancer-related variables would be weak. This hypothesis was confirmed. Extensive treatment and the experience of a second malignancy appeared to have no association with higher levels of cancer-related stress symptoms. TCSs who were examined more recently after treatment reported significantly more symptoms than did those treated longer ago. This relationship was weak, however, suggesting that there were TCSs with clinically elevated levels of cancer-related stress symptoms who completed treat-ment farther in the past. This finding is consistent with the results of research among the general population [19] as well as among the population of cancer patients [12], which have found that a substantial proportion of people still exhibit ongoing distress years after a traumatic event.

We have used a qualitative approach to investigate the extent to which specific events and subjective perceptions lead some people to experience ongoing cancer-related di-stress, while others do not. In interviews, we asked 15 TCSs with clinically elevated levels of cancer-related stress symptoms and 15 TCSs without these symptoms about recollections of the period of diagnosis and treatment, as well as the emotions that these recollections elicit. We had no hypothesis regarding which events are related to cancer-related stress, but predicted that the TCSs with cancer-related stress symptoms would experience more emotions when remembering an event. Results showed that only 10% (19 of the 186) events presented to the TCSs were recalled more frequent-ly by TCSs with clinically elevated levels of cancer-related stress symptoms than by those who had no such problems. As predicted, recollections did arouse more emo-tions among the TCSs with cancer-related stress. These respondents reported that the recollections had a more negative impact on their lives. The interviews brought about some other noteworthy findings that we want to discuss here more comprehensively. First, the number of medical problems reported by the TCSs with clinically elevated levels of cancer-related stress symptoms was similar to those reported by the TCSs with no cancer-related stress symptoms. This is surprising because long-term effects of treatment have been found to contribute to the persis-tence of distress [6;12]. The problems reported by the TCSs with these high levels of cancer-related distress (e.g., concentration problems and headaches), however, were of a more ambiguous nature. It may be that these problems were caused by mental health problems, such as coping reactions following traumatic events, and not by the treatment for testicular cancer. This interpretation is strengthened by the findings that TCSs with clinically elevated levels of cancer-related stress symptoms reported being more upset by the effects of their illnesses, that the illness had caused more changes in their daily lives, and that they were more anxious about the possibility of additional medical problems in the future. The TCSs with cancer-related distress felt that their lives were still threatened. Anxiety for the future (i.e., future state of health) causes more intrusions than pondering events that have already occurred [20].Second, the TCSs with clinically elevated levels of cancer-related stress symptoms talked more often with others about their experiences with cancer than did those who


5

95

had no cancer-related stress, whereas one of the criteria for a posttraumatic stress disorder is the avoidance of thoughts, feelings, or conversations that are part of the trauma [21]. Yet, another criterion for posttraumatic stress is intrusion of memories. When the mean scores of the TCSs on the two subscales of the Impact of Event Scale are considered (intrusion: 7 items, mean = 6.4; avoidance: 8 items, mean = 5.2), it seems that intrusion contributes more to the experience of cancer-related stress in this population than avoidance. It might be that distressed TCSs talk more about their experience, because the memories about the cancer experience intrude upon them. A second explanation is that the higher level of communication indicates a coping reac-tion. After all, traumas confront individuals with information that is inconsistent with existing assumptions about themselves, the world, and their place in the world [22]. The trauma can be processed only by integrating the dissonant information within the existing cognitive structure [5]. Therefore, talking about the trauma can help by giving meaning to the events and finding a new perspective on life [23]. Finally, it is possible that the TCSs with clinically elevated levels of cancer-related stress symptoms did not experience posttraumatic stress after all. The IES is used as a selection instrument, but it is not a diagnostic instrument for posttraumatic stress disorder.

The present study has a number of shortcomings. First, the response rate of 50% may have biased the results, even if the non-responders did not differ from responders according to age, time since diagnosis, and type of treatment. Second, this was the first time that the interview, which was translated from English and adapted for our research population, has been used, and its reliability and validity remain to be estab-lished through future research. The validity of the original version of the ITSIS has been established for groups of children and adolescents that have had cancer, as well as for their parents [11]. The removal of some and the adaptation of several items from the ITSIS will have influenced the validity of the translated version. Reliability could not be assessed, because only one interviewer conducted all the interviews. Furthermore, the results of the qualitative part of the study should be considered as preliminary, because the small sample size of interviewed TCSs limit the possibility to generalize the results to the overall population of TCSs.

Conclusions

In summary, we can establish that a minority of TCSs experience clinically elevated levels of cancer-related stress symptoms. The greatest proportion showed few or no problems. Single TCSs, those with less education, and those who were not employed for wages reported higher levels of cancer-related stress symptoms than did their counterparts. Objective illness variables appeared to play hardly any role. Subjective perceptions appeared more important as contributors to distress than did actual medi-cal histories. In particular, the impact that TCSs feel on their current lives and anxiety about the future make the difference in the presence or absence of ongoing stress after TC.


96

5

Practical implications

Despite the preliminary character of the results of this study, it would probably be beneficial for TCSs to be asked during their check-ups about their thoughts concer-ning the threat of testicular cancer and its effects for their current and future lives, as these perceptions do not necessarily coincide with actual medical histories or state of health [8]. TCSs who feel that their lives are still threatened should be referred for psychosocial intervention. Among other approaches, cognitive therapy lends itself well for this type of problem, as it tends to focus on replacing unrealistic beliefs with realistic beliefs.


5

97

Reference List



3. Fleer J, Hoekstra HJ, Sleijfer DT, Hoekstra-Weebers JEHM (2004) Quality of life of survivors of testicu-lar germ cell cancer: a review of the literature. Supportive Care in Cancer 12:476-486

4. Jonker-Pool G, van de Wiel HB, Hoekstra HJ, Sleijfer DT, van Driel MF, van Basten JP, Schraffordt Koops HS (2001) Sexual functioning after treatment for testicular cancer--review and meta-analysis of 36 empiri-cal studies between 1975-2000. Archives of Sexual Behavior 30:55-74

5. Cordova MJ, Andrykowski MA (2003) Responses to cancer diagnosis and treatment: Posttraumatic stress and posttraumatic growth. Seminars in Clinical Neuropsychiatry 8:286-296



8. Hobbie WL, Stuber ML, Meeske K, Wissler K, Rourke MT, Ruccione K, Hinkle A, Kazak AE (2000) Symptoms of posttraumatic stress in young adult survivors of childhood cancer. Journal of Clinical Onco-logy 18:4060-4066

9. Kazak AE, Stuber ML, Barakat LP, Meeske K (1996) Assessing posttraumatic stress related to medical illness and treatment: the Impact of Traumatic Stressors Interview Schedule (ITSIS). Families, Systems & Health 14:365-380

10. Stuber ML, Meeske K, Barakat L, Kazak A (1996) Posttraumatic stress responses in families of childhood cancer survivors. Psychosomatics 37:33

11. Kazak AE, Barakat LP, Alderfer M, Rourke MT, Meeske K, Gallagher PR, Cnaan A, Stuber ML (2001) Posttraumatic stress in survivors of childhood cancer and mothers: Development and validation of the Impact of Traumatic Stressors Interview Schedule (ITSIS). Journal of Clinical Psychology in Medical Settings 8:307-323

12. Smith MY, Redd WH, Peyser C, Vogl D (1999) Post-traumatic stress disorder in cancer: a review. Psycho Oncology 8:521-537


14. Tomich PL, Helgeson VS (2002) Five years later: A cross-sectional comparison of breast cancer survivors with healthy women. Psycho-Oncology 11:154-169

15. Horowitz MJ, Wilner N, Alvarez W (1979) Impact of Event Scale: A measure of subjective stress. Psycho-somatic Medicine 41:209-218

16. Brom D, Kleber RJ (1985) De Schok Verwerkings Lijst (The Impact of Event Scale). Nederlands Tijd-schrift voor Psychologie 40:164-168

17. Van der Ploeg E, Mooren TTM, Kleber RJ, van der Velden PG, Brom D (2004) Construct validation of the Dutch version of the impact of event scale. Psychological Assessment 16:16-26

18. Jacobsen PB, Widows MR, Hann DM, Andrykowski MA, Kronish LE, Fields KK (1998) Posttraumatic stress disorder symptoms after bone marrow transplantation for breast cancer. Psychosomatic Medicine 60:366-371


98

5

19. Kessler RC, Sonnega A, Bromet E, Hughes M, Nelson CB (1995) Posttraumatic-Stress-Disorder in the National Comorbidity Survey. Archives of General Psychiatry 52:1048-1060

20. Green BL, Rowland JH, Krupnick JL, Epstein SA, Stockton P, Stern NM, Spertus IL, Steakley C (1998) Prevalence of posttraumatic stress disorder in women with breast cancer [see comments]. Psychosomatics 39:102-111

21. American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders. American Psychiatric Association, Washington DC


23. Cordova MJ, Cunningham LLC, Carlson CR, Andrykowski MA (2001) Social constraints, cognitive processing, and adjustment to breast cancer. Journal of Consulting and Clinical Psychology 69:706-711


5

99

Appendix 1Q-sort used for content area 1 (recollections of cancer and treatment; all 49 items), content area 2 (flashback - the extent to which undesired recollections occur; 46/49 items, without items 36, 37, and 44), and content area 3 (criteria for DSM/IV PTSD diagnosis; 46/49 items, without items 36, 37, and 44)

1. Undergoing testicular ultrasoundc

2. Surgical removal of the testicleb

3. Finding out I had testicular cancerb

4. Having blood drawnc

5. Undergoing X-ray examinationb

6. Waiting for results of tumor marker testsb

7. Waiting for X-ray resultsb

8. Deciding whether to continue treatmentc

9. Deciding whether to freeze spermc

10. Having to produce semen for examination and possible freezingc

11. The first day of radiationc

12. The first day of chemotherapyc

13. Not being allowed out of bed during the administration of chemotherapyc

14. Having a VAPc

15. Having a line or cathetera

16. Getting shots or needlesa

17. Taking pills/tablets/capsulesb

18. Being worried about getting sick from side effects of treatment (such as fatigue, nausea)b

19. Feeling sick from radiationc

20. Feeling sick from chemotherapya

21. Losing hair as a result of chemotherapyb

22. The smell of my skin and urine during and after a chemotherapy sessionc

23. Hurting or having paina

24. Going to the emergency rooma

25. Having a second operation to remove remaining tumor tissueb

26. Going off treatmenta

27. Coming for check-ups at the policlinicb

28. Staying in the hospitala

29. Seeing doctors or nursesa

30. Being confronted with other cancer patients during hospital admissionc

31. Being confronted with other cancer patients during check-up visitsc

32. Meeting other testicular cancer patients in the hospitalc

33. Being friends with people of my age who have cancer at the hospitalb

34. Missing my partner, children, parents, or relatives during hospital admissionb

35. Receiving visitors during hospital admissionc

36. Support I received from family, friends, and colleagues during my treatmentcd

37. That I discovered who my real friends arecd

38. Seeing people in my family sad or scareda

39. Being ashamed that I have testicular cancerc


100

5

40. Feeling less masculinec

41. Worrying about my fertilityc

42. Being scared that I could diea

43. Knowing people of my own age who died of cancerb

44. That my partner and I grew closer during my illnesscd

45. Being afraid of being abandoned by my partnerc

46. Having to interrupt my studies/go on disabilityc

47. Being fired by my employer/having to discontinue my studiesc

48. Not being promoted at workc

49. Returning to work/studiesb

Appendix 2Q-sort used for content area 5 (medical problems - existing cancer - and treatment-related problems)

1. Difficulty returning to work/studiesc

2. Concentration problemsb

3. Heart problemsa

4. High blood pressurec

5. Headachesa

6. Tingling in hands or feet (neuropathy)c

7. Numbness in fingers or toesc

8. Problems concerning scars from operations (numbness, ugly scars)b

9. Stomach/intestinal problemsb

10. Problems with incontinencec

11. Hearing problems (such as rustling in the ears)a

12. Lung problemsb

13. Altered sense of taste or smellc

14. Reflux of semen into the bladder instead of outsidec

15. Troubles with eyes or seeinga

16. Fertility problemsa

17. Tiring easilya

18. Hurting or having paina

19. Thrombosisc

20. Erectile disordersc


5

101

Appendix 3Q-sort used for content area 6 - b (ideas about changes in yourself as a result of testicular can-cer - how respondents were changed by cancer)

1. The way I treat other peoplea

2. The way other people treat mea

3. How I make friendsa

4. How careful I ama

5. How well I perform at school/workb

6. How I behave in school or at worka

7. How my family treats mea

8. My plans for the futurea

9. How I think about my lifea

10. How masculine I feelc

11. My relationship with my partnerc

12. My emotional lifec

13. My career orientationc

14. How I set prioritiesc

Appendix 4Q-sort used for content area 7 - c (communication - topics of discussion concerning diagnosis and treatment)

1. What happened when I had cancera

2. The way in which I am different because I had cancera

3. What may happen in the future because I had cancera

4. My feelings about what happeneda

Appendix 5Q-sort used for content area 7 - d (communication - how respondents feel when they talk about topics concerning diagnosis and treatment)

1. I feel scareda

2. I feel understooda

3. I think of things I do not want to remembera

4. I feel relieveda

5. I feel that it is not OK to talk about ita

6. I feel that I am upsetting peoplea

7. I feel like there is no point in talking about ita

aliteral translation from the original ITSIS of Kazak et al. (1996)bslightly adapted translation from the original ITSIS of Kazak et al. (1996)cdeveloped especially for the Dutch version for men who have been cured of testicular cancerdused in content area 1, but not in 2 and 3


6Prevalence, changes in, and

correlates of fatigue the first year

after diagnosis of testicular cancer

Joke Fleer

Dirk Th Sleijfer

Harald J Hoekstra

Marrit A Tuinman


Anticancer Research (2005) 25: 4647-4654


6

105

Introduction

Testicular cancer affects men between 15 and 45 years, when they are in the prime of their life. Since the introduction of cisplatin-based polychemotherapy in the late 1970s, (dis-seminated) testicular cancer has become one of the most curable malignancies. Almost 90% of the testicular cancer patients can be cured with existing treatment modalities [1;2]. Consequently, these men will have to face possible sequel of disease and treatment. Fa-tigue is one of the most prevalent sequel of cancer and its treatment and it is known for its potential negative effects on the wellbeing of patients [3].Fatigue is considered to be a subjective and multidimensional construct with several types of expression. For one person fatigue may express itself as physical exhaustion (physical fatigue), in another as diminished concentration and attention (cognitive fatigue), or a third as lack of motivation and interest (mental fatigue) [4]. Therefore, it is important to assess fatigue multidimensionally. The prevalence of fatigue in patients under treatment for different types of cancer varies between 25 and 75%. It remains a problem for one-third of the patients for months, or even years, after treatment completion [4]. Few studies have examined the prevalence and course of fatigue in testicular cancer patients. A pilot study measured the course of fatigue before, during and after chemotherapy in 16 testicular cancer patients, but this study did not use a multidimensional measurement instrument. At all time-points, testicular cancer patients reported low levels of fatigue, and these levels of fatigue did not change over time [5]. One cross-sectional study assessed the prevalence of fatigue in testicular cancer survi-vors using a multidimensional questionnaire. Results showed that 16% of the survivors ex-perienced mental and physical fatigue. Testicular cancer survivors younger than 30 years of age experienced more fatigue than the same age cohort of the general population [6]. Not all patients develop cancer-related fatigue, but the etiology of fatigue is not well un-derstood to date. It is assumed that the origin is multifactorial, so, among others, physical and psychological correlates have been studied (for overviews see e.g., [4;7;8]). Sociode-mographics have rarely been found to relate to cancer-associated fatigue [4], but in ge-neral more fatigue is found in older and unmarried people, as well as in those with a lower social status and educational level [9]. Fatigue has been associated with type of treatment (e.g., surgery, radiotherapy, chemotherapy) during treatment for cancer, but cancer-re-lated characteristics were unrelated to off-treatment fatigue. Furthermore, associations with several physical and psychosocial variables were found [4]. Anemia and subnormal testosterone levels may be relevant physical contributors to fatigue in testicular cancer patients.Anemia may be caused by the cancer itself, but it can be aggravated by intensive treatment [10]. Chemotherapy, and to a lesser extent orchidectomy, may lead to a decrease of free and total levels of serum testosterone [11-13], and fatigue is one of the symptoms that has been associated with hypogonadism [14]. However, the link between fatigue and both anemia and subnormal testosterone levels has never been examined in testicular cancer patients. Most research into the psychological correlates of fatigue has focused on the association between depression and fatigue [4]. Depression was reported to predict fatigue in testicu-lar cancer survivors, but anxiety was found to be a more important contributory factor


106

6

[6]. People differ in their proneness toward experiencing anxiety [15], and it might be that this personality trait modifies the sensitivity to fatigue. Previous research has shown that cancer patients with an anxious disposition pay more attention to physical sensa-tions, such as fatigue, than less anxious patients [16;17]. Furthermore, a recent study has confirmed the association between trait anxiety and fatigue in breast cancer patients [18]. Therefore, it would be interesting to further explore the relationship between trait anxiety and cancer-related fatigue in testicular cancer.The aim of the current study is to (a) examine the effects of time and treatment modality on the levels of fatigue reported by testicular cancer patients during the first year after diagnosis; (b) compare levels of fatigue of testicular cancer patients with a reference group of Dutch people without a history of cancer; (c) investigate relationships between fatigue on the one hand and sociodemographics, anemia serum and testosterone levels on the other hand. Because it has been suggested that the degree of fatigue before the start of treatment may be one of the strongest predictors of post-treatment fatigue [19], the rela-tionship between earlier and later levels of fatigue were also investigated; (d) investigate the concurrent and prospective contribution of trait anxiety on fatigue.


Participants

Patients diagnosed with a stage I or with disseminated nonseminomatous testicular tu-mor were eligible for this study. All patients received their diagnosis after their surgical removal of the testicle (orchidectomy). After orchidectomy, the patients were clinically staged, to establish the presence of metastases. Patients without metastases (stage I) were closely monitored in follow-up and were informed about their high cure rate of approxi-mately 99%. Patients with metastases received chemotherapy and, based on their prog-nostic classification, they were informed about their cure rates (varying from 50-90%). After chemotherapy was completed. All chemotherapy-treated patients were restaged. The restaging procedure consisted of measurement of serum tumor markers and CT scanning of the chest and abdomen. Patients without biochemical or radiological abnor-malities were considered to have reached a complete remission, All eligible patients, who consecutively attended the University Medical Center Groningen (UMCG) in The Netherlands for treatment between April 2001 and March 2004, were approached. Exclusion criteria were age younger than 18 years at study entry, insufficient command of the Dutch language, and previous treatment for cancer. Twelve (17%) of the 71 patients approached refused to participate. Of the 59 patients who completed the first assessment, 4 did not complete the second questionnaire. Three of these dropouts refused further participation and 1 was too ill to complete the questionnaire. Of the 55 patients who completed the second assessment, 1 patient died before the third assessment. Two patients dropped out, because they had to receive additional treatment. The final group included 52 patients, of whom 15 (28.8%) were treated with orchidectomy alone, and 37 (71.2%) with orchidectomy and cisplatin-based chemotherapy. The patients treated with chemotherapy received 4 courses of cisplatin, etoposide and bleomycin, every 3 weeks.


6

107

Procedure

The study was introduced to the patients after orchidectomy was performed and written informed consent was obtained. Participants with a stage I non-seminoma received a questionnaire at the following three time points: within 1 month after orchidectomy (T1), 3 months after T1 (T2) and 1 year after T1 (T3). Patients with disseminated non-semino-ma received a questionnaire at the following 3 time-point: after orchidectomy and before the start of chemotherapy (T1), after chemotherapy, which is approximately 3 months after T1 (T2), and 1 year after T1 (T3).The study was approved by the Medical Ethics Committee of the University Medical Center Groningen.

Measurements

SociodemographicsData on the following sociodemographic variables were collected at T1: age, educational level, marital status and employment status. The highest educational level completed was measured on a seven-point scale: primary school [1], and lower vocational [2], lower se-condary [3], middle secondary [4], high secondary [5], higher vocational degrees [6], and university [7]. Employment status consisted of the categories ‘not employed for wages’ (consisting of students, being unemployed, being unable to work and being retired) and ‘employed for wages’.

FatigueFatigue was assessed with the Multiple Fatigue Inventory (MFI-20), which is a self-report questionnaire consisting of 5 scales, each including 4 items: general fatigue, physical fa-tigue, reduced activity, reduced motivation and mental fatigue. The psychometric proper-ties for this questionnaire have been found to be good [20]. In the present study, internal consistency of the subscales varied between α = 0.76 and α = 0.93 for the 3 measurement times of the group of testicular cancer patients. Smets et al. [21] have reported data for the MFI-20 from an unselected group of 139 healthy Dutch individuals (male: 44%, mean age: 46 ± 16 years, married/cohabiting: 64%), and this group was used as reference.

AnemiaThe presence of anemia was determined using the guidelines of the Common Toxicity Criteria, version 3 [22]. An Hb level ≥ 8.7 mmol/L indicates no anemia, 6.2-8.6 mmol/L grade 1 anemia, 4.9-6.1 mmol/L grade 2 anemia, 4.0-4.8 mmol/L grade 3 anemia and <4.0 mmol/L grade 4 anemia. Hemoglobyn levels (Hb) were only determined in pa-tients treated with orchidectomy plus chemotherapy group (n = 37).

Testosterone serumTestosterone levels were considered as subnormal when the total testosterone level was less than 10 nmol/L [11]. Total testosterone was measured using radioimmunoassay (Packard 1500/1600/2700, PerkinElmer, Groningen, The Netherlands). These data were only as-sessed in patients treated with chemotherapy. Three patients were excluded because both


108

6

testicles were removed and consequently they received testosterone replacement therapy. Therefore, data were available for 34 patients.

Trait anxietyTrait anxiety was measured with the Dutch version of the State-Trait Anxiety Inventory trait form (STAI-t) at T1 [15;23]. The STAI-t is a psychometrically sound 20-item self-report measure of anxiety proneness. The items have 4 response categories, ranging from ‘not at all’ (1) to ‘very much so’ (4), resulting in scores from 20-80. In the present study, internal consistency (Cronbach alpha) of this scale was 0.93.


Repeated measures analysis of variance was used to examine changes in levels of fatigue, with type of treatment (orchidectomy alone versus orchidectomy plus chemotherapy) as a between-groups factor. Mann-Whitney U tests were used to compare men treated with orchidectomy alone with men treated with orchidectomy and chemotherapy. Indepen-dent T-tests were used to test differences between testicular cancer patients at T1 and T3 and individuals without a history of cancer. Descriptive statistics, Pearson correlations, T-tests, and Mann-Whitney U tests were used to examine relationships between fatigue and sociodemographics, anemia, testosterone, and earlier levels of fatigue. A Pearson cor-relation coefficient lower than 0.30 indicates a poor relationship, a coefficient between 0.30 and 0.50 a moderately strong relationship, and a coefficient higher than 0.50 a strong relationship [24]. The predictive effect of trait anxiety on fatigue was examined concurrently (within time, T1 trait anxiety and T1 fatigue), and prospectively (T1 trait anxiety and T2 and T3 fatigue, controlling for T1 fatigue). By statistically controlling for levels of the dependent variable (i.e., fatigue) at T1, inferences can be made about the direction of the causal influence.

Results

Descriptives

The mean age of the respondents was 28.7 years, ranging from 18 to 44. The highest educational level completed varied from primary school to university degree, with a mean educational level of 4.0 (SD = 1.3). More than half of the patients were married or coha-biting (59.6%), the rest was single. Most patients were employed for wages (78.8%).

Fatigue during the first year after diagnosis for testicular cancer: effects of time and type of treatment

Repeated measures analyses of variance showed a significant time effect in physical fa-tigue, reduced activity and mental fatigue. Furthermore, significant interactive effects of time and type of treatment were found in all subscales, except for reduced motivation, which signifies that levels of fatigue changed over time in different ways for the groups.


6

109

MFI

-20

subs

cale

aT

1 T

2T

3T

est o

f ove

rall

tren

d

Mea

n (S

D)

Mea

n (S

D)

Mea

n (S

D)

effec

tF

p

Gen

eral

fatig

ueor

chid

ecto

my

11.0

(4.8

)9.

7 (4

.5)

10.3

(4.2

)T

ime

2.30

0.10

chem

othe

rapy

10.2

(4.4

)12

.5 (4

.8)

8.9

(4.8

)G

roup

0.00

30.

87

Inte

ract

ion

5.09

0.00

8

Phys

ical

fatig

ueor

chid

ecto

my

10.5

(4.3

)8.

5 (4

.5)

8.7

(4.4

)T

ime

5.96

0.00

4

chem

othe

rapy

9.4

(4.5

)13

.0 (4

.9)*

*7.9

(4.4

)G

roup

0.54

0.47

Inte

ract

ion

9.45

<0.

001

Red

uced

act

ivity

orch

idec

tom

y11.

1 (5.

1)8.

9 (4

.6)

8.3

(3.5

)T

ime

11.00

<0.

001

chem

othe

rapy

11.2

(4.5

)12

.2 (4

.1)*

7.6 (3

.7)

Gro

up0.

850.

36

Inte

ract

ion

3.93

0.02

Red

uced

mot

ivat

ion

orch

idec

tom

y8.

1 (2.

8)7.0

(2.8

)7.1

(2.1)

Tim

e1.2

90.

28

chem

othe

rapy

8.1 (

4.3)

8.

2 (2

.7)

7.4 (3

.5)

Gro

up0.

600.

44

Inte

ract

ion

0.68

0.41

Men

tal f

atig

ueor

chid

ecto

my

11.5

(5.3

)8.

9 (4

.2)

7.8 (3

.7)

Tim

e6.

720.

002

chem

othe

rapy

10.1

(4.3

)10

.5 (4

.9)

9.5

(4.7

)G

roup

0.27

0.61

Inte

ract

ion

4.64

0.01

Tab

le 1.

Mea

ns o

n th

e M

FI-2

0 su

bsca

les a

nd lo

ngitu

dina

l ana

lyse

s of f

atig

ue

a high

er sc

ore i

ndica

tes m

ore f

atigu

e; M

ann-

Whi

tney

U te

sts o

rchi

decto

my v

ersu

s orc

hide

ctom

y plu

s che

mot

hera

py; *

p≤0.0

5; **

p≤0.0

1


110

6

Inspection of the mean scores in the orchidectomy plus chemotherapy group, showed that levels of fatigue increased from T1 to T2, and decreased from T2 to T3. A continuous decrease from T1 to T3, or a decrease from T1 to T2 followed by stabilization was found for the orchidectomy group (Table 1). No significant effect of type of treatment was found (designated as ‘group’ in Table 1), indicating no significant differences between the means of the 2 groups on the fatigue subscales at the different measurement times.

Because interactive effects were found and groups were small, it was decided to compare the means of the orchidectomy and orchidectomy plus chemotherapy groups at the dif-ferent measurement times. Mann-Whitney U tests showed that the orchidectomy plus chemotherapy group reported significantly more physical fatigue and reduced activity at T2 than the patients who had undergone an orchidectomy only (Table 1).

Comparison of levels of fatigue of testicular cancer patients shortly after orchidec-tomy and one year later with those of individuals without a history of cancer

Because the mean scores of the orchidectomy group and orchidectomy plus chemotherapy group did not differ significantly at T1 and T3, T-tests were performed with the total pa-tient group. Analyses showed that testicular cancer patients reported significantly higher levels of reduced activity (t = 3.34, p = 0.001) and mental fatigue (t = 2.91, p = 0.004) at T1. At T3, levels of fatigue of testicular cancer patients did not differ from those of healthy individuals (Table 2).

T1 T3 Reference group

Mean (SD) Mean (SD) Mean (SD)

General fatigue 10.4 (4.4) 9.3 (4.6) 9.9 (5.2)

Physical fatigue 9.7 (4.5) 8.1 (4.4) 8.8 (4.9)

Reduced activity 11.2 (4.6)** 7.8 (3.6) 8.7 (4.6)

Reduced motivation 8.1 (3.9) 7.3 (3.1) 8.2 (4.0)

Mental fatigue 10.5 (4.6)* 9.0 (4.5) 8.3 (4.8)

Table 2. Means and standard deviations of testicular cancer patients at T1 and T3 (n = 49) and individuals without a history of cancer (n = 139; Smets et al., 1998)

T1 versus reference group: *p≤.01; **p≤.001


6

111

Correlates of fatigue

SociodemographicsAt T1, age correlated significantly with reduced motivation (r = 0.32, p < 0.05). At T2, a significant correlation was found between age and all subscales of the MFI, except for mental fatigue (r = 0.29 to r = 0.36, p < 0.05). At T3, age correlated significantly with general fatigue (r = 0. 35, p < 0.05) and reduced motivation (r = 0.59, p < 0.001). All correlations were positive and moderately strong to strong, indicating that the older the patient the more fatigue he experienced. The only significant relationship found between educational level and fatigue was at T3. More highly educated patients reported more general (r = 0.30, p < 0.05) and physical fatigue (r = 0.29, p < 0.05). Marital status and employment status were not significantly related to fatigue.

AnemiaAt T1, only 1 chemotherapy-treated patient had a grade 3 level of anemia, and at T2, 2 patients had a grade 2 level of anemia. The remaining patients were not or slightly anemic (Grade 1 anemia) at any one of the 3 time-points. The group of patients with anemia was too small to investigate the effect on fatigue. However, the mean scores on the MFI-20 of the 3 patients with a grade 2 or 3 level of anemia fell within the range of mean scores of the other patients.

TestosteroneOnly 1 patient had a testosterone level less than 10 nmol/L at T1. At T2 and T3, none of the patients had testosterone levels less than 10 nmol/L. Therefore, we could not in-vestigate the effect of testosterone on fatigue. Yet, the mean scores on the subscales of the MFI-20 of the patient with an initially low testosterone level fell within the range of scores of the other patients.

Concurrent Prospective

T1-T2 T2-T3 T1-T3STAI-t andT1 fatigue

STAI-t andT2 fatigue

STAI-t andT3 fatigue

r r r r Partial ra Partial ra

General fatigue 0.40** 0.51*** 0.47*** 0.51*** 0.16 0.32*

Physical fatigue 0.30* 0.46*** 0.52*** 0.54*** 0.20 0.33*

Reduced activity 0.27* 0.37** 0.36** 0.54*** 0.34* 0.32*

Reduced motivation 0.52*** 0.52*** 0.56*** 0.46*** 0.17 0.24

Mental fatigue 0.61*** 0.73*** 0.55*** 0.64*** 0.19 0.31*

Table 3. Correlations between earlier and later levels of fatigue and between trait anxiety (STAI-t) and fatigue concurrently and prospectively

acontrolling for fatigue at T1; *p≤0.05; **p≤0.01; ***p≤0.001


112

6

Earlier levels of fatiguePearson’s product moment correlation coefficients were significant between the different measurement times, ranging from 1 poor relationship to 6 moderately strong and 8 strong relationships (Table 3). High correlation coefficients indicate consistency between the measurement times, so individuals were quite stable in their reports of levels of fatigue over time. They were least stable in their reports on the subscale reduced activity.

Concurrent relationship between trait anxiety and fatigueTrait anxiety correlated positively with all fatigue subscales at T1. The lowest, although moderately strong, correlation was found between trait anxiety and reduced motivation. The strongest relationship was found between trait anxiety and mental fatigue (Table 3).

Prospective effect of trait anxiety on fatigueAfter controlling for fatigue at T1, only a significant relationship between trait anxiety and reduced activity at T2 was found. In contrast, trait anxiety was significantly associ-ated with 4/5 subscales at T3. The exception was a non-significant relationship between trait anxiety and reduced motivation (Table 3). These findings indicate little causal effect of trait anxiety on fatigue at T2, whereas at T3 a causal effect of trait anxiety on most subscales of fatigue was found.

Discussion

The current study is the first to examine, with a multidimensional measurement instru-ment, the prevalence, changes in, and correlates of fatigue among testicular cancer patients treated during the first year after orchidectomy. The prevalence of fatigue changed over time and varied between the treatment groups. At the first assessment, shortly after hea-ring the diagnosis and the surgical removal of the affected testicle, the orchidectomy and orchidectomy plus chemotherapy group reported similar levels of fatigue. They reported significantly more reduction in activity and mental fatigue than individuals without a history of cancer. Both the physical recovery from the operation and anesthesia, and the distressing emotions that follow from hearing a cancer diagnosis are likely to have con-tributed to these higher levels of fatigue. Three months later, the orchidectomy group reported lower levels of fatigue, whereas the orchidectomy plus chemotherapy-treated pa-tients experienced increased fatigue. This increase may be explained by the chemotherapy treatment received by these patients. Patients in this group were assessed immediately af-ter completion of chemotherapy and chemotherapy is known to cause fatigue [4;7]. One year after orchidectomy, both the orchidectomy and orchidectomy plus chemotherapy groups reported similar levels of fatigue as individuals without a history of cancer. These findings suggest that fatigue seems not to be an enduring problem for the majority of tes-ticular cancer patients, irrespective of the treatment received. Apparently, treatment has an impact on fatigue only shortly after orchidectomy and completion of chemotherapy. This finding elucidates why previous studies did not find an effect of treatment modality on fatigue in long-term survivors of testicular cancer [6;25].


6

113

The relationships between fatigue and sociodemographics, physical factors, and trait anxiety were studied among testicular cancer patients. Of the sociodemographics, age was related to at least 1 of the fatigue subscales at all measurement times. Similar to our study, increased fatigue with age has been found in both the general population [9;26] and cancer populations [6;27;28]. It should be noted that the testicular cancer patients were relatively young compared to the reference group. So, it is possible that, if we had com-pared the testicular cancer patients with people of their own age, higher levels of fatigue would have been found for the testicular cancer patients. Associations with marital and social status have also been reported in the general popula-tion [9], but we could not identify such relationships in the current study. Furthermore, our finding that a higher educational level was related to physical fatigue 1 year after diag-nosis disagreed with findings in the general population, showing that shorter education was associated with higher fatigue scores [9]. The occurrence of both anemia and hypogonadism was rare in our study population. The majority of patients had hemoglobin levels in the normal or in the mild anemic range and normal testosterone levels at all measurement times. This is consistent with previous reports that only a minority of testicular cancer patients will develop moderate or severe anemia [29] or hypogonadism [11-13]. Both the patients with anemia and the 1 hypogo-nadic patient reported fatigue levels within the range of the whole group of testicular cancer patients. Studies with large sample sizes are needed to gain definitive insight into the impact of anemia and hypogonadism on fatigue levels. Earlier levels of fatigue were predictive of later levels of fatigue. In broad terms, those reporting higher levels of fatigue after orchidectomy also reported higher levels of fatigue 3 months and 1 year later. Similar findings have been reported before [3;30]. This stresses the importance of identifying fatigued individuals early on in the disease trajectory to reduce post-treatment fatigue, because the longer fatigue lasts, the more it will disrupt the daily lives of patients [31;32]. Testicular cancer patients with a more anxious disposition reported more fatigue after orchidectomy. Prospective analyses showed that trait anxiety had a predictive effect only in the subscale reduced activity at the second measurement time, suggesting that at this time-point, other variables are more important for the prediction of fatigue. The greater percentage of our sample (71%) consisted of patients treated with orchidectomy plus chemotherapy, so it is possible that chemotherapy-related side effects outweighed the im-pact of trait anxiety. Unfortunately, our sample size was too small to reliably investigate this hypothesis. Interestingly, 1 year after orchidectomy trait anxiety had a significant causal effect on all fa-tigue subscales. Research has shown that both highly anxious individuals [33] and chroni-cally fatigued individuals [34;35] use maladaptive strategies, such as avoidant coping, relatively often. Avoidance coping strategies lead people into activities or mental states (such as withdrawal and rumination) that keep them from directly addressing problems [36]. So, the highly anxious testicular cancer patient might, for example, cope by avoiding physical exercise, which is known for its physical and psychological benefits (e.g., [37;38]). Further research is needed to clarify the relationships between coping strategies, trait anxiety and fatigue among testicular cancer patients and survivors, especially because co-ping strategies have rarely been studied in relation to cancer-related fatigue [39].


114

6

Conclusions

Based on the results of the current study it can be concluded that treatment for testicu-lar cancer has an impact on levels of fatigue mainly in the short-term. Testicular cancer patients did experience more fatigue immediately after orchidectomy and, for those who received chemotherapy, following completion of chemotherapy. One year after orchidec-tomy, patients treated with orchidectomy alone as well as those treated with orchidectomy and chemotherapy reported levels of fatigue comparable to those of individuals without a history of cancer. Older age was related to increased fatigue. Earlier levels of fatigue were predictive of later levels fatigue. Furthermore, concurrent and causal relationships were found between trait anxiety and fatigue. Older testicular cancer patients, those who already report being fatigued shortly after orchidectomy and those with a more anxious disposition, might benefit from a multidisciplinary intervention that focuses on physical exercise training and the strengthening of coping strategies to effectively reduce fatigue. A multidisciplinary approach of fatigue seems relevant, because when increased levels of fatigue were found in this study, it was found in all domains.


6

115

Reference List



3. Smets EMA, Visser MRM, Garssen B, Frijda NH, Oosterveld P, De Haes JCJM (1998) Understan-ding the level of fatigue in cancer patients undergoing radiotherapy. Journal of Psychosomatic Research 45:277-293

4. Servaes P, Verhagen C, Bleijenberg G (2002) Fatigue in cancer patients during and after treatment: prevalence, correlates and interventions. European Journal of Cancer 38:27-43

5. Trask PC, Paterson AG, Fardig J, Smith DC (2003) Course of distress and quality of life in testicu-lar cancer patients before, during, and after chemotherapy: Results of a pilot study. Psycho Oncology 12:814-820

6. Fossa SD, Dahl AA, Loge JH (2003) Fatigue, anxiety, and depression in long-term survivors of testicu-lar cancer. Journal of Clinical Oncology 21:1249-1254

7. Tavio M, Milan I, Tirelli U (2002) Cancer-related fatigue (Review). International Journal of Oncology 21:1093-1099

8. Stasi R, Abriani L, Beccaglia P, Terzoli E, Amadori S (2003) Cancer-related fatigue - Evolving concepts in evaluation and treatment. Cancer 98:1786-1801

9. Watt T, Groenvold M, Bjorner JB, Noerholm V, Rasmussen NA, Bech P (2000) Fatigue in the Danish general population. Influence of sociodemographic factors and disease. Journal of Epidemiology and Community Health 54:827-833

10. Pujade-Lauraine E, Gascon P (2004) The burden of anaemia in patients with cancer. Oncology 67:1-4

11. Nuver J, Smit AJ, Sleijfer DT, van Gessel AI, van Roon AM, van der Meer J, van den Berg MP, Hoek-stra HJ, Sluiter WJ, Gietema JA (2005) Left ventricular and cardiac autonomic function in survivors of testicular cancer. European Journal of Clinical Investigation 35:99-103

12. Nord C, Bjoro T, Ellingsen D, Mykletun A, Dahl O, Klepp O, Bremnes RM, Wist E, Fossa SD (2003) Gonadal hormones in long-term survivors 10 years after treatment for unilateral testicular cancer. Eu-ropean Urology 44:322-328

13. Brennemann W, StoffelWagner B, Helmers A, Mezger J, Jager N, Klingmuller D (1997) Gonadal func-tion of patients treated with cisplatin based chemotherapy for germ cell cancer. Journal of Urology 158:844-850

14. Shores MM, Sloan KL, Matsumoto AM, Moceri VM, Felker B, Kivlahan DR (2004) Increased inci-dence of diagnosed depressive illness in hypogonadal older men. Archives of General Psychiatry 61:162-167

15. Spielberger D, Gorsuch L, Lushene E (1970) STAI manual. Consulting Psychologists Press, Palo Alto

16. Stark D, Kiely M, Smith A, Morley S, Selby P, House A (2004) Reassurance and the anxious cancer patient. British Journal of Cancer 91:893-899

17. Cameron LD, Leventhal H, Love RR (1998) Trait anxiety, symptom perceptions, and illness-related responses among women with breast cancer in remission during a tamoxifen clinical trial. Health Psy-chology 17:459-469


116

6

18. Sugawara Y, Akechi T, Okuyama T, Matsuoka Y, Nakano T, Inagaki M, Imoto S, Fujimori M, Hosaka T, Uchitomi Y (2005) Occurence of fatigue and associated factors in disease-free breast cancer patients without breast cancer. Supportive Care in Cancer 13:628-636

19. Smets EMA, Visser MRM, Willems-Groot AFMN, Garssen B, Oldenburger F, van Tienhoven G, De Haes JCJM (1998) Fatigue and radiotherapy: (A) experience in patients undergoing treatment. British Journal of Cancer 78:899-906

20. Smets EMA, Garssen B, Bonke B, Dehaes JCJM (1995) The Multidimensional Fatigue Inventory (MFI) Psychometric Qualities of An Instrument to Assess Fatigue. Journal of Psychosomatic Research 39:315-325

21. Smets EMA, Visser MRM, Willems-Groot AFMN, Garssen B, Schuster-Uitterhoeve ALJ, De Haes JCJM (1998) Fatigue and radiotherapy: (B) experience in patients 9 months following treatment. Brit-ish Journal of Cancer 78:907-912

22. Common Toxicity Criteria, Version 3. National Cancer Institute. 7-4-2005.

23. Van der Ploeg HM, Defare PB, Spielberger CD (2000) Handleiding bij de Zelf-Beoordelings Vragen-lijst (ZBV). Swets Test Publishers, Lisse

24. Cohen J (1988) Statistical power analysis for the behavioral sciences. Erlbaum, Hillsdale, NJ


26. Loge JH, Ekeberg O, Kaasa S (1998) Fatigue in the general Norwegian population: Normative data and associations. Journal of Psychosomatic Research 45:53-65

27. Bower JE, Ganz PA, Desmond KA, Rowland JH, Meyerowitz BE, Belin TR (2000) Fatigue in breast cancer survivors: occurrence, correlates, and impact on quality of life. Journal of Clinical Oncology 18:743-753

28. Loge JH, Abrahamsen AF, Ekeberg, Kaasa S (2000) Fatigue and psychiatric morbidity among Hodg-kin’s disease survivors. Journal of Pain and Symptom Management 19:91-99

29. Weissbach L, Bussar-Maatz R, Flechtner H, Pichlmeier U, Hartmann M, Keller L (2000) RPLND or primary chemotherapy in clinical stage IIA/B nonseminomatous germ cell tumors? Results of a prospective multicenter trial including quality of life assessment. European Urology 37:582-594

30. Maliski SL, Kwan L, Orecklin JR, Saigal CS, Litwin MS (2005) Predictors of fatigue after treatment for prostate cancer. Urology 65:101-108

31. Irvine DM, Leslie V, Graydon JERN, Bubela N (1998) Fatigue in women with breast cancer receiving radiation therapy. Cancer Nursing 21:127-135

32. Molassiotis A, Chan CWH (2004) Fatigue patterns in Chinese patients receiving radiotherapy. Euro-pean Journal of Oncology Nursing 8:334-340

33. Deisinger JA, Cassisi JE, Whitaker SL (1996) Relationships between coping style and PAI profiles in a community sample. Journal of Clinical Psychology 52:303-310

34. Nijs J, De Meirleir K, Duquet W (2004) Kinesiophobia in chronic fatigue syndrome: Assessment and associations with disability. Archives of Physical Medicine and Rehabilitation 85:1586-1592

35. Afari N, Schmaling KB, Herrell R, Hartman S, Goldberg J, Buchwald DS (2000) Coping strategics in twins with chronic fatigue and chronic fatigue syndrome. Journal of Psychosomatic Research 48:547-554


6

117

36. Holahan CJ, Moos RH (1987) Personal and Contextual Determinants of Coping Strategies. Journal of Personality and Social Psychology 52:946-955

37. Dimeo F, Bertz H, Finke J, Fetscher S, Mertelsmann R, Keul J (1996) An aerobic exercise program for patients with haematological malignancies after bone marrow transplantation. Bone Marrow Trans-plantation 18:1157-1160

38. Dimeo FC (2001) Effects of exercise on cancer-related fatigue. Cancer 92:1689-1693

39. De Jong N, Courtens AM, Huijer Abu-Saad H, Schouten HC (2002) Fatigue in patients with breast cancer receiving adjuvant chemotherapy: a review of the literature. Cancer Nursing 25:283-297


7Different recruitment methods

identify different individuals

Joke Fleer

Harald J Hoekstra

Roy E Stewart

Raoul E Nap

Dirk Th Sleijfer


Submitted


7

121

Introduction

Since the number of cancer survivors is growing, there is an increasing interest in how cancer patients progress with their lives after treatment, and how they regain a reasonable quality of life (QoL) [1]. Therefore, QoL has become an important outcome variable in medical and psychosocial oncology research [2]. Conclusions about the QoL of cancer patients are always based on data obtained from patients who participate in studies. In order to extrapolate results to the overall population of cancer patients, both adequate sample sizes and samples representative of the total population are needed [3;4]. In QoL studies, different methods are used to recruit patients. Many studies recruit pa-tients through hospital databases. By using postal questionnaires, researchers can easily invite all patients treated at a clinic within a certain time period to participate. When the response rate is high, it is likely that the sample is representative of the total population of patients. Unfortunately, response rates of studies in which postal questionnaires are used are often under 70% [5]. There are also studies that approach support groups, that use the internet or an advertisement in papers or magazines for the recruitment of patients [6]. Yet, it is possible that these methods identify atypical groups of patients, even if a high response rate is achieved. Consequently, the validity of results may be influenced by the source from which patients are recruited. There are indications from the literature that recruitment through different sources in-deed identifies populations with different characteristics. For example, it has been found that members from patient support groups reported less anxiety, depression, stress, and a better QoL than non-members [7;8]. However, members from a patient association were found to experience more distress than those recruited through a hospital database [9]. It is not clear what causes these differences in outcomes, but it may have to do with the fact that patients who become member of a patient association have a greater need to receive support from and offer support to their fellow cancer survivors [10]. A higher likelihood to seek support has been associated with several sociodemographic- and cancer-related characteristics, such as being female, being unmarried, being closer to diagnosis, and a higher educational level [7;8;11;12]. Thus, it could be that members from a patient as-sociation have specific sociodemographic and cancer-related characteristics. Because so-ciodemographic- and cancer-related characteristics have also been found to relate to QoL outcomes in cancer survivors (e.g., [13-15]), it may very well be that the comparability of QoL of patients recruited through a patient association from those recruited through a hospital database is affected by the differences in these characteristics.This study aims to gain insight into (1) the types of individuals that different recruitment sources identify and (2) the effect of recruitment source on QoL outcomes. To do so, we will compare testicular cancer survivors (TCSs) recruited through a patient association with those recruited through a hospital database on sociodemographics, cancer-related data and QoL. Based on the literature, we hypothesize that the two groups will differ on sociodemographic- and cancer-related characteristics. We think that the members from the patient association will be younger, more often unmarried, higher educated and closer to treatment completion. We also hypothesized that the patient association group will be a more specific group compared to those recruited through the hospital database, and thus show less variance in sociodemographic- and cancer-related characteristics. Further-


122

7

more, we assume that the two groups will differ in QoL. However, we do not hypothesize on the direction of this difference, as published literature on this topic is contradictory [7]. If the groups differ in QoL, we will investigate the effect of recruitment source on QoL.


Respondents

Respondents were approached through two recruitment sources:1. The Groningen University Medical Center. 701 men treated for testicular cancer (TC)

between 1977 and 2003 were invited to participate in a QoL study. Fifty percent (n = 354) of the TCSs agreed to participate and completed the questionnaire. Comparison of patients who participated in the study with those who did not showed that the two groups did not differ in age, time since diagnosis, and type of treatment received.

2. The Dutch patient association for TC patients and survivors. The main goal of this association is to provide an opportunity for men treated for TC to get in touch with each other and to provide and receive information about their disease. Eighty-five of the 116 members (73.3%) completed the questionnaire. No information was available on sociodemographic and treatment-related characteristics of the non-responders.

Questionnaires

Sociodemographic and treatment-related characteristicsThe following sociodemographic and treatment-related characteristics were assessed: age at time of study, age at time of diagnosis, time since diagnosis, educational level, marital status, employment status, type of treatment received, and experience of a second cancer event. The highest level of education completed was measured along the following seven-point scale: primary school (1), lower vocational (2), lower secondary (3), middle secon-dary (4), high secondary (5), higher vocational (6), and advanced university (7).

Quality of lifeThe RAND-36 [16] is an internationally used valid and reliable generic self-report ques-tionnaire to assess QoL. The Dutch version of the RAND-36 [17] was used. It contains eight subscales: physical functioning (10 items), social functioning (2 items), role limitations due to physical problems (4 items), role limitations due to emotional problems (3 items), mental health (5 items), vitality (4 items), pain (2 items), general health perception (5 items), and one item on health change. After transformation, scores on the subscales vary from 0 to 100, with a higher score indicating a better health status. In the present study, internal consistency of the subscales for the total sample was good (varying from 0.80 to 0.92).


7

123


Dichotomous variables were created for marital status, employment status and type of treatment. Marital status consisted of the categories “no partner” (consisting of single, divorced and separated patients) and “partner” (consisting of those who were married or cohabiting). Employment status included the categories of “not employed for wages” (consisting of students, unemployed persons, those who were unable to work, and reti-rees) and “employed for wages” (consisting of those who were employed for wages). The following categories were used for type of treatment: “surgical treatment (consisting of the categories orchidectomy and orchidectomy + retroperitoneal lymph node dissection) and “combined treatment” (consisting of the categories orchidectomy + radiotherapy, or-chidectomy + chemotherapy, and orchidectomy + chemotherapy + surgical resection of residual tumor mass). Levene’s test for homogeneity of variance was used to examine the equality of variance between the patient association and the hospital database. Unpaired T-tests and Chi2 tests were performed to study the hypothesis that the two groups differ on sociodemographics, cancer-related characteristics and QoL. Analyses of covariance were used to examine the effect of recruitment source on QoL. Analyses of covariance also enable us to investigate if the differences in QoL between the two groups remain after controlling for sociodemographic and cancer-related variables. The covariates were those sociodemographic- and cancer-related characteristics on which the groups differed significantly.Results were considered statistically significant if the probability of occurrence was 0.05 or less.

Results

Comparison of sociodemographic and cancer-related characteristics

Levene’s test for homogeneity of variance showed that the variances in age, age at diagno-sis, and time since diagnosis were not equal for the two groups. Inspection of the standard deviations and ranges of the continuous variables showed less variance in scores in the patient association group than in the hospital group. T-tests showed that the patient as-sociation group was younger, more recently diagnosed and higher educated. Levene’s test for homogeneity of variance also showed that the variances in marital status, type of treatment, and second cancer event were not equal for the two groups. Inspection of the ratio’s of the dichotomous variables showed that the ratio’s of (1) having a partner to not having a partner, (2) surgical to combined treatment, and (3) having experienced a second cancer event to not having experienced a second cancer event were different for the groups. Members from the patient association had received more extensive treatment, and had more often experienced a second cancer event than the patients from the hospital group (Table 1).


7

124

Hos

pita

l gro

upN

= 3

54*

Pati

ent a

ssoc

iati

on g

roup

N =

85*

Lev

ene’s

test

for

hom

ogen

eity

Com

pari

son

betw

een

grou

ps

MSD

(ran

ge)

MSD

(ran

ge)

Fp

tp

Age

(in

year

s) 43

.712

.1 (1

8-79

)40

.88.

7 (1

9-62

)11.

910.

001

2.58

0.01

Age

at d

iagn

osis

(in y

ears)

32

.910

.4 (1

6-72

)34

.37.8

(17-

55)

7.65

0.00

6-1

.43

0.16

Tim

e sin

ce d

iagn

osis

(in y

ears)

10.7

6.8

(0.3

-25)

6.5

4.4

(0.7

-25)

44.8

4<

0.00

17.1

6<

0.00

1

Educ

atio

nal l

evel

4.2

1.7 (1

-7)

5.0

1.5 (2

-7)

1.90

0.17

-3.9

6<

0.00

1

N%

N%

Fp

Chi

2p

Mar

ital s

tatu

s4.

030.

051.0

90.

30

Part

ner

298

84.7

6880

.0

No

part

ner

5415

.317

20.0

Empl

oym

ent s

tatu

s1.0

00.

320.

270.

61

Em

ploy

ed fo

r wag

es27

577

.964

75.3

Not

em

ploy

ed fo

r wag

es78

22.1

2124

.7

Tre

atm

ent

42.4

7<

0.00

17.2

20.

007

Surg

ical

100

28.6

1214

.3

Com

bine

d25

071

.472

85.7

Seco

nd ca

ncer

even

t 37

.62

<0.

001

10.5

90.

001

Yes

318.

818

21.2

No

322

91.2

6778

.8

Tab

le 1.

Soc

iode

mog

raph

ic a

nd c

ance

r-re

late

d ch

arac

teri

stic

s of t

he tw

o gr

oups

and

com

pari

son

betw

een

the

grou

ps

*N is

sligh

tly v

aria

ble b

ecau

se o

f miss

ing d

ata


7

125

Qua

lity

of li

feH

ospi

tal g

roup

N =

354

*Pa

tien

t ass

ocia

tion

gro

upN

= 8

5*L

even

e’s te

st fo

r ho

mog

enei

tyT

-tes

ts fo

r co

mpa

riso

n be

twee

n gr

oups

MSD

MSD

Fp

tp

Phys

ical

func

tioni

ng88

.219

.287

.617

.61.1

90.

340.

250.

80

Soci

al fu

nctio

ning

85.8

19.9

77.2

22.8

1.30

0.11

3.48

<0.

001

Rol

e lim

itatio

ns (p

hysic

al)

84.0

30.7

71.1

40.4

1.74

<0.

001

2.73

0.00

7

Rol

e lim

itatio

ns (e

mot

iona

l)85

.431

.175

.737

.91.4

90.

022.

190.

03

Men

tal h

ealth

76.9

15.9

72.8

19.1

1.44

0.03

1.83

0.07

Vita

lity

64.3

19.5

39.2

16.1

1.48

0.03

12.3

4<

0.00

1

Pain

88.3

18.7

89.8

18.2

1.05

0.80

-0.6

30.

53

Gen

eral

hea

lth p

erce

ptio

n71

.419

.663

.223

.21.4

00.

043.

030.

003

Hea

lth c

hang

e52

.417

.555

.117

.31.0

30.

89-1

.28

0.20

Tab

le 2

. Des

crip

tives

on

qual

ity

of li

fe su

bsca

les f

or th

e tw

o gr

oups

and

com

pari

son

betw

een

the

grou

ps

*N is

sligh

tly v

aria

ble b

ecau

se o

f miss

ing d

ata


126

7

Comparison of the two groups on QoL

Levene’s test for homogeneity of variance showed that the variances in the subscales role limitations due to physical and emotional problems, mental health, vitality and general health perception were not equal for the two groups. T-tests showed that the two groups differed in means on the subscales social functioning, role limitations due to physical and emotional problems, vitality and general health perception. The patient association group showed more variance and lower means on the subscales (Table 2).

Comparison of the two groups on QoL after controlling for sociodemographic and cancer-related characteristics

Analyses of covariance were used to examine if the differences in QoL remained after controlling for sociodemographic and cancer-related variables. The covariates were the sociodemographic- and cancer-related characteristics on which the groups differed sig-nificantly: age, time since diagnosis, educational level, type of treatment, and second cancer event (see Table 1). Still significantly more variance was found in the scores of the members from the patient association on the subscales role limitations due to physi-cal and emotional problems, mental health and vitality than of the hospital group. Also, lower mean scores on the subscales social functioning, role limitations due to physical and emotional problems, vitality and general health perception were found in the patient association group than in the hospital group (Table 3).

Discussion

One of the prerequisites for a methodologically strong study is adequate recruitment of patients in order to obtain a representative sample of the population of patients. If recruitment is not optimal, it may have an impact on the generalizability of study re-sults [18]. Although this is common scientific knowledge, there is surprisingly little em-pirical research available on the comparability of patient populations included through different types of recruitment. As a consequence, researchers may underestimate the impact of recruitment source on outcomes. Therefore, we have compared testicular cancer survivors (TCSs) recruited through a patient association with those recruited through a hospital database on sociodemographic- and cancer-related characteristics and on QoL. The first aim was to examine if the two recruitment sources represented different types of individuals. Based on the literature [7;8;11;12] we hypothesized that sociodemographic- and cancer-related characteristics of the two groups would differ. Furthermore, we hy-pothesized that the members from the patient association would be a more specific group compared to those recruited through the hospital database, and thus show less variance in characteristics. Our results confirm these hypotheses. Members from the patient associa-tion were significantly younger, more recently treated, higher educated, more often treated with extensive treatment, and had more often experienced a second cancer event than


7

127

Qua

lity

of li

feH

ospi

tal g

roup

N =

354

*Pa

tien

t ass

ocia

tion

gro

upN

= 8

5*L

even

e’s te

st fo

r ho

mog

enei

tyT

-tes

ts fo

r co

mpa

riso

n be

twee

n gr

oups

MSD

MSD

Fp

tp

Phys

ical

func

tioni

ng90

.218

.188

.015

.61.3

50.

101.0

20.

31

Soci

al fu

nctio

ning

85.0

19.9

76.7

22.2

1.24

0.19

3.31

0.00

1

Rol

e lim

itatio

ns (p

hysic

al)

78.2

30.3

66.4

38.2

1.59

0.00

52.

600.

01

Rol

e lim

itatio

ns (e

mot

iona

l)84

.331

.274

.538

.21.5

00.

012.

170.

03

Men

tal h

ealth

73.1

15.8

69.2

19.5

1.53

0.01

1.71

0.09

Vita

lity

59.8

19.3

35.7

15.2

1.61

0.01

12.2

4<

0.00

1

Pain

93.9

18.0

94.2

17.6

1.04

0.84

-0.15

0.88

Gen

eral

hea

lth p

erce

ptio

n61

.118

.754

.721

.41.3

20.

102.

710.

007

Hea

lth c

hang

e67

.117

.167

.517

.31.0

20.

88-0

.170.

86

Tab

le 3

. Des

crip

tives

and

com

pari

son

of th

e tw

o gr

oups

on

qual

ity

of li

fe, a

fter

con

trol

ling

for

soci

odem

ogra

phic

s and

can

cer-

rela

ted

char

acte

rist

ics o

n w

hich

the

two

grou

ps d

iffer

ed si

gnifi

cant

ly

*N is

sligh

tly v

aria

ble b

ecau

se o

f miss

ing d

ata


128

7

those recruited through the hospital group. Also, patients within the patient association were more alike than patients within the hospital database. These findings support the hypothesis that different recruitment sources identify populations with dissimilar patient characteristics. These results also strengthen the notion that recruitment through patient associations leads to the identification of a population with more specific characteristics than recruitment through hospital databases. We also compared the QoL of the two groups. Although we assumed that the two groups would differ in QoL, we did not hypothesize on the direction of this difference, as pu-blished literature on this topic contradicts [7-9]. The results of the present study show that members from the patient association reported a worse functioning on most of the QoL domains. However, this group also had larger variances in scores, which indicates that within the patient association the reports on QoL were more heterogeneous than in the hospital group. An explanation for the greater variety in QoL scores of the members of the patient association may lie in the motives of patients to become and remain a mem-ber of a patient association. People may become a member because they have problems adapting to the cancer experience and feel a need to receive psychosocial support from fellow cancer survivors [10;11]. After they have regained a satisfactory QoL, some remain a member of the patient association to give psychosocial support to new members [10]. This line of reasoning is supported by a study on correlates of participation in a breast cancer support group, that showed that new members reported higher levels of anxiety, depression and stress than established members [7]. The smaller variety in QoL scores in the hospital group may, in part, be the result of the 50% response rate. Considering the high mean scores of this group, it may be that those with a less favorable QoL were less willing to participate in our study. The second aim of the study was to investigate the impact of recruitment source on QoL. No considerable changes in QoL were found after controlling for the sociodemographic- and cancer-related variables on which the two groups differed significantly. The two groups still had unequal means and variances on the same subscales. This indicates that there must be other factors that affect the differences in QoL between the groups. It has been suggested that subjective perceptions, such as anxiety for a recurrence of a tumor, experience of higher impact of the illness on daily lives, and anxiety for the long-term consequences of treatment can influence levels of distress [19-22]. It would be plausible if patients who perceive more anxiety, distress, and a negative impact on daily living are more likely to become a member of a patient association [9]. As a matter of fact, cancer patients have indeed indicated that one of the reasons to become a member of a patient association was the experience of feelings of powerlessness in the situation and for their future lives [10].In conclusion, the results of this study support the notion that different recruitment sources identify different patient populations and that this has quite an impact on out-come variables. However, differences in sociodemographic and cancer-related characteris-tics appear not to determine the differences in QoL. Further empirical research is needed to investigate which other factors may explain the differing QoL scores. The results of this study do allow us to warn researchers to be careful with recruitment, because it may lead to selection bias that has consequences for the generalizability of the results. Findings of the present study are based on a QoL study but may be relevant for all research in which


7

129

patients are recruited. At least, researchers should always present information about the recruitment strategies that they used, because it gives the reader the opportunity to value a study at its true worth [23]. If results are not generalizable it is not possible for the on-cologist to give solid advice to patients on relevant issues [24].


130

7

Reference List

1. Weitzner MA, Meyers CA, Stuebing KK, Saleeba AK (1997) Relationship between quality of life and mood in long-term survivors of breast cancer treated with mastectomy. Supportive Care in Cancer 5:241-248

2. Velikova G, Stark D, Selby P (1999) Quality of life instruments in oncology. European Journal of Cancer 35:1571-1580

3. Bootsma-van der Wiel A, van Exel E, de Craen AJ, Gussekloo J, Lagaay AM, Knook DL, Westendorp RGJ (2002) A high response is not essential to prevent selection bias. Results from the Leiden 85-plus study. Journal of Clinical Epidemiology 55:1119-1125

4. Scholle SH, Peele PB, Kelleher KJ, Frank E, Jansen-McWilliams L, Kupfer D (2000) Effect of different recruitment sources on the composition of a bipolar disorder case registry. Social Psychiatry and Psychia-tric Epidemiology 35:220-227

5. Rothemich SF, Woolf SH, Johnson RE, Marsland DW (2000) Evaluating the administration of an office-based health survey in a primary care practice. Journal of Clinical Epidemiology 53:1002-1012

6. Garrett SKM, Thomas AP, Cicuttini F, Silagy C, Taylor HR, Mcneil JJ (2000) Community-based re-cruitment strategies for a longitudinal interventional study: the VECAT experience. Journal of Clinical Epidemiology 53:541-548

7. Stevens MJ, Duttlinger JE (1998) Correlates of participation in a breast cancer support group. Journal of Psychosomatic Research 45:263-275

8. Katz D, Koppie TM, Wu D, Meng MV, Grossfeld GD, Sadesky N, Lubeck DP, Carroll PR (2002) So-ciodemographic characteristics and health related quality of life in men attending prostate cancer support groups. Journal of Urology 168:2092-2096

9. Hagedoorn M, Buunk BP, Kuijer RG, Wobbes T, Sanderman R (2000) Couples dealing with cancer: role and gender differences regarding psychological distress and quality of life. Psycho Oncology 9:232-242

10. Carlsson C, Baigi A, Killander D, Satterlund Larrson U (2005) Motives for becoming and remaining member of patient associations: a study of 1,810 Swedish individuals with cancer associations. Supportive Care in Cancer 13:1035-1043

11. Berglund G, Bolund C, Gustafsson UL, Sjoden PO (1997) Is the wish to participate in a cancer rehabilita-tion program an indicator of the need? Comparisons of participants and non-participants in a randomized study. Psycho Oncology 6:35-46

12. Taylor SE, Falke RL, Shoptaw SJ, Lichtman RR (1986) Social support, support groups, and the cancer-patient. Journal of Consulting and Clinical Psychology 54:608-615

13. Schultz PN, Beck ML, Stava C, Vassilopoulou-Sellin R (2003) Health profiles in 5836 long-term cancer survivors. International Journal of Cancer 104:488-495


15. Langeveld NE, Stam H, Grootehuis MA, Last BF (2002) Quality of life of young adult survivors of child-hood cancer. Supportive Care in Cancer 10:579-600



18. Goodwin PJ, Leszcz M, Quirt G, Koopmans J, Arnold A, Dohan E, Hundleby M, Chochinov HM, Na-varro M (2000) Lessons learned from enrollment in the BEST study--a multicenter, randomized trial of group psychosocial support in metastatic breast cancer. Journal of Clinical Epidemiology 53:47-55


7

131

19. Hobbie WL, Stuber ML, Meeske K, Wissler K, Rourke MT, Ruccione K, Hinkle A, Kazak AE (2000) Symptoms of posttraumatic stress in young adult survivors of childhood cancer. Journal of Clinical Oncol-ogy 18:4060-4066




23. Froelicher ES, Lorig K (2002) Who cares about recruitment anyway? Patient Education and Counseling 48:97

24. Bertetto O, Bracarda S, Tamburini M, Cortesi E (2001) Quality of life studies and genito-urinary tumors. Annals of Oncology 12 (suppl. 3):S43-S48


8General discussion


8

135

General discussion

In the first part of this final chapter, some overall conclusions about the two primary research aims will be drawn. Then, some general methodological considerations of measurement ap-proaches and recruitment strategies in quality of life studies of cancer survivors will be made. After that, the usefulness of the theoretical model that was used as a framework in this thesis will be discussed and suggestions will be made about directions for future research. The chap-ter will end with a discussion on the implications of the findings for clinical practice.

Overall conclusions

Quality of life of testicular cancer survivorsIn the introduction of this thesis it was hypothesized that the specific combination of clinical features of testicular cancer may make a testicular cancer survivor vulnerable for an impaired quality of life. Our study results did not confirm this hypothesis. Findings of the present study indicate that, on the whole, testicular cancer survivors experience a good quality of life, which includes satisfaction with life, spiritual well-being (e.g., they consider their life as meaningful), and no lasting fatigue. In view of the finding that little or no effect of time since completion of treatment was found, one may conclude that this finding is valid for the short-, median-, as well as long-term survivors.Does this mean that the cancer experience had no effect on the lives of the survivors at all, not even shortly after completion of treatment? The answer must be no. Firstly, these findings are based on answers to generic questionnaires. Such questionnaires often fail to show a reduc-tion in the quality of life of people who have been diagnosed with a life-threatening or chronic disease, but this does not mean that the disease did not have an impact on a survivor’s life [1]. We also included disease-specific questions, which focused on the impact of the cancer experience on the current survivor’s life. The answers to these questions indicated that a small percentage did experience cancer-related distress (e.g., intrusive thoughts, nightmares) and did have a more negative outlook on life since testicular cancer (e.g., more anxiety and worries about the future, feeling more vulnerable and less in control). The majority reported, however, that their outlook on life had become more positive (e.g., more appreciation of life, reprioriti-zing what is important in life) and over half of the survivors stated that negative and positive effects of testicular cancer on their current life co-exist. These apparently opposing outcomes suggest that cancer may lead to distress as well as to positive personal growth. Yet, given the high overall valuation of quality of life, it does not seem that the reported distress interferes with everyday functioning in most of the testicular cancer survivors. A second explanation offered for the finding that the testicular cancer survivors’ quality of life was comparable with that of people not confronted with cancer is that survivors may adjust by changing their internal standards, values and conceptualization of quality of life (i.e., response shift) [1], disengaging from some beliefs and goals and taking up others [2], or by finding positive benefit from the experience (i.e., meaning-making process) [3]. However, the design and methodology of our study did not allow us to capture such processes possibly used by survivors to adapt to their changed life situation.


136

8

Predictors of quality of life of testicular cancer survivorsAlthough most of the testicular cancer survivors appear to function well, a minority did report lower levels of quality of life and clinically elevated levels of cancer-related stress symptoms. This minority might benefit from psychosocial help. To detect these individuals we investi-gated if characteristics of the disease and treatment, personal variables, and life stresses were risk factors for an impaired quality of life or cancer-related distress. The relationship of objective disease- and treatment-related characteristics (i.e., time since completion of treatment, type of treatment, second cancer event, and hemoglobin-, and tes-tosterone levels) with the outcome variables was weak to negligible. This finding contradicts with clinical intuition. For instance, one would suppose that a testicular cancer survivor treated with surgery and chemotherapy would experience more physical and psychosocial late-effects then a patient treated with surgery only. Nevertheless, this study was not the only one to show a lack of effect on quality of life of objective characteristics of disease and treatment. Our fin-dings are in line with what is often reported in the literature: objective disease- and treatment-related characteristics are weak predictors of well-being. Findings of this study, as well as of many earlier studies (e.g., [4-7]), suggest that subjective perceptions may be more important determinants of functioning. We found, for example, that testicular cancer survivors who still perceive a negative impact of testicular cancer on their current lives (e.g., intrusive memories and changes in daily life as a result of late effects of treatment for testicular cancer) and those who report more anxiety about the future (e.g., fear for recurrence or medical problems) show more distress. The personal variables and life stresses that related to an impaired well-being in testicular cancer survivors were: unemployment, not doing what feels meaningful, an anxious disposi-tion, concurrent chronic disease, and the experience of a cumulative number of negative life events. Associations between these variables and (aspects of ) quality of life have also been found in the general population, indicating that the testicular cancer survivors do not deviate from the normal population in risk pattern. Still, knowledge of these risk variables is relevant, because it may assist clinicians in their attempt to identify individuals who will have difficulty adjusting.

In sum, this thesis provides more insight into factors that affect quality of life of testicular can-cer survivors. However, based on this attained knowledge it will still be difficult for health care providers to identify survivors at risk for continuing cancer-related distress. Therefore, more research is needed to identify further factors that relate to their quality of life. The predictor variables under investigation were derived from the research model of Holland and colleagues [8]. In the paragraph “Theoretical reflections and future research” we will discuss whether this model provides a useful framework for future research and suggestions will be made about di-rections for research. First, however, some methodological considerations will be discussed.

Methodological considerations

The studies described in this thesis all include a section in which the strengths and weaknesses of the used designs and methods (e.g., response rate, range in time since treatment) are dis-cussed, and thus, these reflections will not be repeated here. However, there are two additional


8

137

methodological issues that we want to discuss comprehensively here. This section will focus on measurement approaches and recruitment strategies.The design and methodology of our study did not provide the opportunity to capture the cognitive processes that cancer survivors use to come to a judgment of their quality of life. We used traditional questionnaires, in which the weighting of the domains (e.g., physical functioning, social functioning, mental health) was fixed. However, these domains may not have the same relevance or importance for all individuals. Furthermore, the relevance or im-portance of a particular domain does not necessarily remain static with the passage of time or after major life events, such as cancer [9]. For instance, a cancer survivor can experience objec-tive physical limitations and problems as a result of chemotherapy, but still evaluate his quality of life as satisfactory because the experience with cancer has made him realize that it is far more important to spend quality time with his family than to feel physically fit. Individualized measures of quality of life (e.g., SEIQoL) do allow respondents to indicate which areas of life are most important, rate their level of functioning or satisfaction with each, and indicate the relative importance of each to their overall quality of life [10]. A multi-method approach (i.e., both quantitative and qualitative measurement instruments) may be informative as well, as much is yet to learn about the cognitive processes cancer survivors use to come to a judgment of their quality of life and how the cancer experience has influenced the value that survivors attribute to the different domains of quality of life [11]. Since quality of life assessment is increasingly used as a major outcome parameter in health care, both for clinical decisions and policy making, it is vital that the measurement of quality of life accurately reflects how patients arrive at their appraisal and how it should be interpreted. Another impediment of traditional measurement approaches is that they are mostly focused on limitations and problems, without considering positive elements that contribute to quality of life. Linking-up with many other studies (e.g., [12;13]), this study shows that cancer can have a positive as well as a negative impact on quality of life. Measurement of quality of life should therefore include the possibility that it can be evaluated both in positive and negative terms.Recruitment is the second methodological consideration of concern for both researchers and health care providers of cancer patients and survivors. We found that recruitment through dif-ferent sources identifies populations with different characteristics and that this has an impact on quality of life outcomes. This finding stresses the importance of presenting information about recruitment strategies in articles, because that gives the reader the opportunity to value a study at its worth [14]. If results are not generalizable to the population of patients under study, it is not possible for the oncologist to give solid advice to patients on relevant issues.

Usefulness of theoretical model and future research

The model used as a framework for this thesis (see Figure 1) was derived from the research model developed by Holland and colleagues [8]. This research model proved useful for the identification of factors possibly related to outcomes (e.g., quality of life, distress). Therefore, the model can be used to identify further possible predictors (e.g., personality variables and coping styles, illness-related behaviors, doctor-patient communication). However, despite the usefulness of this research model for the categorization of possible risk variables, it may be less useful as a framework for guiding research whose goal is to capture the adaptation process. Other frameworks, such as stress-coping theory (e.g., [15]), response shift theory [1] or the


138

8

theory on self-regulation [2] may provide better guidance in understanding the process of adapting to changes in life situations.

Based on the findings in this thesis, there are some more suggestions for future research. First, the conclusion that the most distressed survivors are those who still experience a threat from their cancer experience on their lives needs further validation, because it was based on the answers of only 30 testicular cancer survivors. We believe that this is a relevant topic to study in this patient population, because a recently published, qualitatively strong study found that a very high proportion of testicular cancer survivors (31%) reports fear for a recurrence at a mean of eleven years after treatment completion [16]. Results from research in other types of cancer also indicate that the degree of anxiety, or worry, may be an important component of overall adjustment (e.g., [17;18]). Moreover, the preoccupation with current threat of the disease has been incorporated into the DSM-IV [19] and the theoretical model of Ehlers and Clark’s [20] cognitive model of persistent posttraumatic stress disorder. This model of Ehlers and Clark may prove a useful theoretical framework for research into cancer-related distress. Second, our finding that having an additional chronic disease related to an impaired well-being needs additional investigation, especially because a recent study also found that self-reported side-effects of treatment are strongly associated with reductions in quality of life in testicular cancer survivors [21]. It might be that self-perceived physical impairments, whether they are induced by the treatment for testicular cancer or not, may make testicu-lar cancer survivors more vulnerable. There is one model that specifically focuses on the underlying relationships between self-reported physical impairments and the experience of an impaired well-being and, thus, may be more effective in unraveling the mechanisms in testicular cancer survivors than the stress-coping paradigm: the illness-intrusiveness theoretical framework of Devins et al. [22;23]. This theory maintains that illness- and treatment-related disruptions may interfere with continued involvements in valued activi-ties and interests and plans for the future, which may lead to lower levels of overall quality of life. Testicular cancer patients are diagnosed in the prime of their life, when career, interpersonal relationships and starting a family are major issues. With the use of this theoretical framework, future research may elicit if the after-effects of illness does indeed intrude on the planned life trajectories of these men and represents an underlying stressor that contributes to distress in testicular cancer survivors. Third, researchers might want to investigate the relationship between meaning and cancer-related distress. We found that having a sense of meaning was related more strongly to quality of life than to cancer-related distress. This suggests that having a sense of meaning is more important to the experience of positive quality of life than to coping with cancer-related post traumatic stress symptoms. Yet, it has also been argued that having a sense of meaning pro-vides a resource that facilitates successful coping with severe life stressors [3;24-26]. Therefore, more research is needed to investigate the underlying mechanisms between meaning and can-cer-related distress. Both the revised model of stress and coping [15] and the model of global and situational meaning [3] may serve as a frame for this research.


8

139

Clinical implications

Most testicular cancer survivors experience a good quality of life and are not in need of a psychosocial intervention, but a minority does experience cancer-related distress and might benefit from such an intervention. Yet, how can clinicians detect these distressed individuals? They cannot base their judgment on the disease- and treatment-related characteristics of the survivor, because the impact of these factors on quality of life and cancer-related distress ap-peared to be minimal. In this thesis we suggested a way to identify distressed testicular cancer survivors, and we found that those who still experience a threat from their cancer experienced the highest level of distress. Therefore, it might be useful to ask testicular cancer survivors during their check-ups about their thoughts concerning the threat of testicular cancer for their current and future lives, as these perceptions do not necessarily coincide with actual state of health [4]. Testicular cancer survivors who feel that their life is still threatened by cancer may benefit from a discussion with their physician about the possible late effects of testicular cancer. Additional cognitive-behavioral therapy, which tends to focus on replacing unrealistic beliefs with realistic beliefs and teaches individuals how to cope with difficult situ-ations, may help them to accept the uncontrollable aspects of their life situations (e.g., chance of recurrence and possible late effects). Furthermore, it may help to redirect their energy to the controllable aspects of their lives (e.g., diet, smoking cessation, exercise), help them to reframe the experience in a positive light (i.e., meaning-making process), and if necessary, guide them in generating new, valued and realistic goals.

Testicular cancer Outcome variables

Figure 1. Schematic representation of the research model in psycho-oncology of Holland and colleagues [8]

Personal variablessociodemographic characteristics meaningtrait anxiety

Life stressesconcurrent chronic diseaserecently experienced life events

---

--

Disease and treatment variablestime since treatmenttype of treatmentsecond cancer eventhemoglobin leveltestosterone level

-----

Quality of life (multidimensional)Cancer-related stress symptoms MeaningSatisfaction with lifeFatigue (multidimensional)

Related variables


8

140

Reference list


2. Carver CS, Scheier MF (2000) Scaling back goals and recalibration of the affect system are processes in normal adaptive self-regulation: understanding ‘response shift’ phenomena. Social Science & Medicine 50:1715-1722


4. Hobbie WL, Stuber ML, Meeske K, Wissler K, Rourke MT, Ruccione K, Hinkle A, Kazak AE (2000) Symptoms of posttraumatic stress in young adult survivors of childhood cancer. Journal of Clinical Oncol-ogy 18:4060-4066




8. Holland JC (2002) History of psycho-oncology: overcoming attitudinal and conceptual barriers. Psycho-somatic Medicine 64:206-221

9. Hickey AM, Bury G, Boyle CA, Bradley F, O’Kelly FD, Shannon W (1996) A new short form individual quality of life measure (SEIQoL-DW): application in a cohort of individuals with HIV/AIDS. BMJ 313:29-33

10. Ring L, Hofer S, Heuston F, Harris D, O’Boyle CA (2005) Response shift masks the treatment impact on patient reported outcomes (PROs): the example of individual quality of life in edentulous patients. Health and Quality of Life Outcomes 3:55

11. Zebrack B (2000) Cancer survivors and quality of life: a critical review of the literature. Oncology Nursing Forum 27:1395-1401

12. Andrykowski MA, Brady MJ, Hunt JW (1993) Positive psychosocial adjustment in potential bone marrow transplant recipients: cancer as a psychosocial transition. Psycho Oncology 2:261-276

13. Lechner SC, Zakowski SG, Antoni MH, Greenhawt M, Block K, Block P (2003) Do sociodemographic and disease-related variables influence benefit-finding in cancer patients? Psycho Oncology 12:491-499

14. Froelicher ES, Lorig K (2002) Who cares about recruitment anyway? Patient Education and Counseling 48:97

15. Folkman S (1997) Positive psychological states and coping with severe stress. Social Science & Medicine 45:1207-21

16. Dahl AA, Haaland CF, Mykletun A, Bremnes R, Dahl O, Klepp O, Wist E, Fossa SD (2005) Study of anxiety disorder and depression in long-term survivors of testicular cancer. Journal of Clinical Oncology 23:2389-2395

17. Langeveld NE, Grootenhuis MA, Voute PA, de Haan RJ (2004) Posttraumatic stress symptoms in adult survivors of childhood cancer. Pediatric Blood & Cancer 42:604-610


19. American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders. American Psychiatric Association, Washington DC


8

141

20. Ehlers A, Clark DM (2000) A cognitive model of posttraumatic stress disorder. Behaviour Research and Therapy 38:319-345

21. Mykletun A, Dahl AA, Haaland CF, Bremnes R, Dahl O, Klepp O, Wist E, Fossa SD (2005) Side effects and cancer-related stress determine quality of life in long-term survivors of testicular cancer. Journal of Clinical Oncology 23:3061-3068

22. Devins GM, Bezjak A, Mah K, Loblaw DA, Gotowiec AP (2005) Context moderates illness-induced life-style disruptions across life domains: a test of the illness intrusiveness theoretical framework in six common cancers. Psycho Oncology July 4 [Epub ahead of print]

23. Devins GM (1994) Illness intrusiveness and the psychosocial impact of lifestyle disruptions in chronic life-threatening disease. Advances in Renal Replacement Therapy 1:251-263

24. Frankl VE (1959) Man’s search for meaning. Hodder & Stoughton, London

25. Janoff-Bulman R (1989) Assumptive worlds and the stress of traumatic events: applications of the schema construct. Social Cognition 7:113-136

26. Tedeschi RG, Calhoun LG (1995) Trauma and transformation - growing in the aftermath of suffering. SAGE Publications, London


9Summary


9

145

Summary

Men who are diagnosed with testicular cancer are generally young. As a consequence of the treatment they receive nowadays, they are likely to cure from the disease, even when they have metastases. This means that they have to live with possible short- and long-term sequel of diagnosis and treatment for many years. Survivors who do not have the resources necessary to cope with such physical and psychosocial sequel, may find it more difficult to adjust to their altered life situation and experience an impaired quality of life. These men might need professional assistance to enhance their skills to competently manage their lives as a cancer survivor. Yet, how do health care providers know which testicular cancer survivor is at risk for distress and might benefit from an intervention? Therefore, it is needed to investigate factors predictive of quality of life, in order to detect and refer at risk individuals to an appropriate intervention. The aim of this thesis was to examine the quality of life of testicular cancer survivors and to identify risk factors for an impaired quality of life.

The introductory chapter (chapter 1) elaborates on the concept of quality of life and the possible effects of cancer on the quality of life of patients and survivors. After that, the working model that was used as a framework for the study is introduced, and an overview of the chapters is presented.

The second chapter of the thesis describes a critical review of the current state of know-ledge on the physical, psychological and social well-being of testicular cancer survivors. A literature search identified 23 studies that met the inclusion criteria. These studies were published between 1980 and 2003 and most of the studies suffered from methodological shortcomings (e.g., small sample sizes, non-validated questionnaires). Therefore, a qua-lity assessment, using methodological and disease- and treatment-related criteria, was performed to distinguish stronger- from poorer-quality studies. Seven studies appeared qualitatively stronger. Results showed that the majority of the survivors in both stronger- and poorer-quality studies reported a good quality of life. Prospective and retrospective studies showed that quality of life increased after completion of treatment and that nega-tive consequences (e.g., fatigue, distress, fear of tumor recurrence) on life decreased as compared to the situation directly after diagnosis. The stronger-quality studies paid little attention to the social dimension of quality of life (marital functioning, social support, and functional life) and to positive effects of the experience with testicular cancer. Poorer-quality studies reported that, besides a decrease in negative consequences of testicular cancer on the patients’ life, the experience of testicular cancer had positive effects also (e.g., emotional growth, stronger relationships with family and friends). Based on this review, one could assume that testicular cancer survivors generally experi-ence a good quality of life. However, it seemed not justified to draw such a conclusion, because of the methodological shortcomings of most studies and because the stronger-quality studies did not investigate all dimensions of quality of life with the same exten-siveness. Therefore, it was suggested that future research should include large groups of both short- and long-term survivors, sociodemographically matched comparison groups, and validated questionnaires. These methodologically stronger studies should focus on


146

9

the identification of risk profiles, to detect distressed individuals possibly in need of pro-fessional psychosocial care.

The study described in chapter 3 aimed to assess the quality of life of testicular cancer survivors and attempted to overcome some of the methodological shortcomings identified in the review. The first goal was to compare the quality of life of testicular cancer survivors (n = 354; range in time since completion of treatment = 3 months to 25 years) to that of a reference group of Dutch men. Secondly, it aimed to identify survivors with an impaired quality of life, based on sociodemographic variables, disease- and treatment-related vari-ables, and recently experienced life events. Results showed that, on the whole, testicular cancer survivors experienced a quality of life similar to that of the reference group of Dutch men. Furthermore, quality of life was comparable among the short-, median-, and long-term survivors, those treated with surgery alone and those treated with more exten-sive treatment, and those who had experienced a second cancer event and those who had not. Survivors at somewhat higher risk for an impaired quality of life were older and had experienced more negative life events. Yet, the combination of unemployment and pre-sence of a chronic disease appeared to be the strongest predictor for an impaired quality of life. Responses to an open question indicated that the unemployed survivors with a chronic disease significantly more often experienced a negative impact of testicular cancer on their current quality of life than their counterparts. In conclusion, it appears that testicular cancer survivors generally experience a good qua-lity of life. The lack of effect of objective disease- and treatment-related variables suggests that survivors quickly adapt to the circumstances created by the experience of cancer and its treatment, even after extensive treatment or a second cancer event. The joint burden of unemployment and chronic disease on testicular cancer survivors’ quality of life has also been found in the general population. This finding makes sense when it is considered that chronic diseases can cause impairments in social activities and work, which may lead to lower levels of quality of life. Nevertheless, the variance explained by the studied variables was low, indicating that more important predictors remain to be identified.

Having a sense of meaning has been identified as an important contributor to a person’s psychosocial well-being. People with a strong sense of meaning view their life from a meaningful perspective, which forms a guide from which goals are directed. People try to fulfill these life goals. Testicular cancer survivors may find it difficult to attain meaning to their cancer experience and, because of the short- and long-term after effects of the diagnosis and treatment, some goals may become unattainable. This may have a negative impact on their quality of life and cause distress. The first aim of the study described in chapter 4 was to assess the sense of meaning of testicular cancer survivors. The second aim was to investigate if the cancer experience had changed the survivors’ outlook on life. The third aim was to assess the relative contribution of meaning in the prediction of psycho-social well-being and cancer-related distress, in addition to sociodemographics, disease- and treatment-related variables, concurrent chronic disease, and recently experienced life events. Results showed that testicular cancer survivors experience life as meaningful. In line with the literature, it was found that single survivors, as well as childless survivors, those without employment, those with a chronic disease and those who had experienced


9

147

more negative life events experienced somewhat less meaning than their counterparts. Disease- and treatment-related characteristics were not related to meaning. Approxi-mately two-third of the survivors experienced a change in outlook on life; a majority (60%) reported a more positive outlook on life since the disease. Meaning was the most important predictor for psychosocial well-being, but it appeared to have only a weak effect on cancer-related distress. In sum, testicular cancer survivors appear to have a strong sense of meaning, and this sense of meaning is more important for their psychosocial well-being than sociodemo-graphics, disease- and treatment-related variables, concurrent chronic disease and recently experienced life events. The effect of meaning on cancer-related distress was less strong, suggesting that it might be more important to the experience of a positive quality of life than to coping with cancer-related distress. Yet, prior research has suggested that meaning may serve as a protective factor against cancer-related distress. Therefore, further research might focus on the underlying mechanisms through which meaning and cancer-related distress relate.

A cancer experience can be so distressing that it produces specific cancer-related psycho-logical symptoms, such as intrusive and emotionally upsetting thoughts and memories of the disease period and the avoidance of unpleasant feelings and recollections. These can-cer-related stress symptoms can persist months or even years after completion of treat-ment and they may have a negative impact on the daily life of the survivors. Cancer-related stress symptoms have never been studied among testicular cancer survivors. Therefore, the primary goal of the study described in chapter 5 was to investigate the prevalence of cancer-related stress symptoms among testicular cancer survivors. The secondary goal was (a) to gain quantitative insight into the extent to which sociodemographics, disease- and treatment-related variables and concurrent chronic disease are associated with levels of cancer-related stress symptoms and (b) to assess qualitatively whether concrete events and subjective perceptions relate to the experience of cancer-related stress symptoms. Responses to the questionnaire showed that 13% of the 354 testicular cancer survivors experienced cancer-related stress symptoms to such an extent that psychosocial care is recommended. Single testicular cancer survivors, those with less education, and those who were not employed for wages reported somewhat higher levels of cancer-related stress symptoms. Time since treatment, type of treatment, the experience of a second cancer event and concurrent chronic disease appeared to play almost no role. Based on the answer patterns to the questionnaire, two groups were selected for an interview: 15 survivors with clinically elevated levels of cancer-related stress symptoms and 15 with-out these symptoms. The interview focused on the relationships between concrete events (e.g., undergoing an ultrasound) and subjective perceptions (e.g., anxiety about future medical problems) and cancer-related stress symptoms. Results showed that the number of cancer-related events recalled by survivors with clinically elevated levels of cancer-re-lated stress symptoms differed little from those without these symptoms. However, those with clinically elevated levels of distress did report being more upset by the recollection of such events. Also, they reported that they were more upset by the effects of their illness, in the sense that it had caused more changes in their daily life and that they were anxious about the possibility of additional medical problems in the future. The testicular cancer


148

9

survivors with clinically elevated levels of cancer-related stress symptoms indicated that their life was still threatened by the cancer experience. In sum, the data indicate that an important subset of testicular cancer survivors expe-riences cancer-related stress symptoms. Sociodemographics and disease- and treatment-related variables appear not to be strong indicators of cancer-related stress, whereas the impact survivors feel on their current life and anxiety about the future does seem to give an indication of the presence of stress symptoms. These findings may offer new directions for communication between health care providers and survivors. Assessment of thoughts concerning threats of cancer and perceived effects on their current and future life may give better insight into distress testicular cancer survivors experience than the actual medical history. Testicular cancer survivors who indicate that they perceive that their life is still threatened might benefit from a psychosocial intervention that focuses on replacing un-realistic beliefs with realistic beliefs.

Fatigue is one of the most prevalent sequel of cancer and its treatment and it is known for its potential negative effects on the well-being of patients and survivors. The study described in chapter 6 is the first to examine the prevalence of and changes in different do-mains of fatigue among testicular cancer patients during the first year after orchidectomy. Furthermore, relationships between fatigue on the one hand and sociodemographics, hemoglobin- and testosterone levels, and trait anxiety on the other hand were studied. Fifteen patients treated with orchidectomy and 37 treated with orchidectomy plus che-motherapy were assessed within one month after orchidectomy, 3 months later (which was immediately after chemotherapy for the chemotherapy group), and 12 months after orchidectomy. Results showed that testicular patients did experience more fatigue imme-diately after orchidectomy and, for those who received chemotherapy, following comple-tion of chemotherapy. One year after orchidectomy, patients treated with orchidectomy alone as well as those treated with orchidectomy and chemotherapy reported levels of fatigue comparable to those of individuals without a history of cancer. Older age related (moderately) strong to increased fatigue. Earlier levels of fatigue were predictive of later levels of fatigue. Furthermore, concurrent and prospective effects were found of trait anxi-ety on fatigue. The numbers of patients suffering from anemia or low testosterone levels appeared small. Therefore, we could not examine the impact of these variables on fatigue levels. In conclusion, testicular cancer survivors seem to experience the highest levels of fatigue shortly after orchidectomy and chemotherapy. Fatigue does not seem to be an enduring problem for the majority of the patients, irrespective of treatment received. Older patients, those who already reported being fatigued shortly after orchidectomy, and those with a more anxious disposition are at risk for cancer-related fatigue. These patients might bene-fit from a multidisciplinary intervention that focuses on physical exercise training and the strengthening of coping strategies to effectively reduce fatigue.

In chapter 7 a methodological issue is addressed. In quality of life studies, several methods are used to recruit patients. Many studies recruit patients through hospital databases, but there are also studies that approach support groups, use the internet or an advertisement in papers or magazines as a recruitment source. Recruitment through different sources


9

149

may identify populations with different characteristics and this may bias the outcomes. Until now, little empirical attention has been paid to the impact of recruitment source on quality of life. Therefore, this study aimed to gain insight into (1) individuals that different recruitment sources identify and (2) the effect of recruitment source on quality of life. To do so, we compared testicular cancer survivors recruited through a patient association (n = 85) with those recruited through the UMCG database (n = 354) on sociodemogra-phics, disease- and treatment-related characteristics and quality of life. Results showed that members from the patient association were significantly younger, more recently treated, higher educated, more often treated with extensive treatment, had more often ex-perienced a second cancer event, and reported a worse functioning on most of the quality of life domains than those recruited through the hospital group. Even after controlling for sociodemographic- and cancer-related variables, a worse functioning on most of the quality of life domains was found in members from the patient association as compared to those from the UMCG database. In conclusion, the results of this study support the notion that different recruitment methods identify individuals with different sociodemographic and disease- and treat-ment-related characteristics and different quality of life scores. However, differences in sociodemographic and disease- and treatment-related characteristics do not seem to determine these quality of life differences. This indicates that there must be other fac-tors that cause the differences in quality of life between groups. These results may warn researchers to be careful with recruitment, because it may lead to selection bias that has consequences for the generalizability of the results. If results cannot be generalized, it is not possible for the clinician to give solid advice to patients on relevant issues.

Finally, in chapter 8 the main findings of the studies described in this thesis are discussed and placed within a broader context. This chapter also reflects on methodological issues. Furthermore, the usefulness of the model that was used as a framework in this thesis is discussed and suggestions are made about directions for future research. The chapter ends with a discussion on the implications of the findings for clinical practice.


10Samenvatting


153

10

Samenvatting

Zaadbalkanker wordt voornamelijk bij jonge mannen gediagnosticeerd. Met de huidige behandeling hebben deze mannen een grote kans op genezing, zelfs in het geval van uit-zaaiingen. Genezen patiënten kunnen echter de rest van hun leven geconfronteerd worden met mogelijke korte- en lange termijn gevolgen van diagnose en behandeling. Sommige van deze patiënten kunnen het moeilijk vinden om zich aan te passen aan deze gevolgen, waardoor zij een lagere kwaliteit van hun leven ervaren. Zij kunnen baat hebben bij pro-fessionele psychosociale begeleiding, waardoor ze manieren aanleren om beter om te gaan met hun veranderd levensperspectief. Het is echter nog onduidelijk welke mannen na behandeling voor zaadbalkanker het risico lopen op psychosociale problemen en gebaat zijn bij een dergelijke begeleiding. Een manier om daarachter te komen, is te onderzoeken welke factoren kwaliteit van leven voorspellen. Het doel van dit proefschrift was om de kwaliteit van leven van mannen die genezen zijn van zaadbalkanker te onderzoeken en om risicofactoren te identificeren die bijdragen aan een slechtere kwaliteit van leven.

Het inleidende hoofdstuk (hoofdstuk 1) gaat in op het concept ”kwaliteit van leven” en op de mogelijke gevolgen van kanker voor de kwaliteit van leven van genezen patiënten. Daarnaast wordt het theoretisch model dat ten grondslag ligt aan het onderzoek geïntro-duceerd en een overzicht gegeven van de hoofdstukken in het proefschrift.

Hoofdstuk 2 geeft een overzicht van het gepubliceerde onderzoek dat is gedaan naar het fysieke, psychologische en sociale welbevinden van mannen die genezen zijn van zaad-balkanker. Via literatuurdatabases werden 23 tussen 1980 en 2003 gepubliceerde studies gevonden, die voldeden aan de inclusiecriteria. De meeste van deze studies vertoonden methodologische tekortkomingen, zoals kleine onderzoeksgroepen en het gebruik van niet-gevalideerde vragenlijsten. Om deze reden werd een kwaliteitsanalyse gedaan waar-bij, op basis van methodologische criteria en ziekte- en behandelingskenmerken, kwali-tatief sterkere studies werden onderscheiden van zwakkere. Zeven studies bleken kwali-tatief sterk te zijn. Uit zowel de kwalitatief sterkere als uit de zwakkere studies bleek dat het grootste deel van de genezen zaadbalkankerpatiënten een goede kwaliteit van leven ervaart. Prospectieve en retrospectieve studies lieten zien dat hun kwaliteit van leven ver-beterde na het beëindigen van de behandeling en dat de negatieve gevolgen van zaadbal-kanker (zoals vermoeidheid, stress, angst voor het terugkomen van de tumor) afnamen in de loop van de tijd. De kwalitatief sterkere studies besteedden weinig aandacht aan het sociaal welbevinden (huwelijkstevredenheid, sociale steun en werksituatie) en de mogelij-ke positieve gevolgen van de ervaring met zaadbalkanker. De kwalitatief zwakkere studies lieten echter zien dat er, naast een afname in negatieve gevolgen van zaadbalkanker op het leven van de patiënten, ook positieve gevolgen werden ervaren (zoals emotionele groei en een betere band met familie en vrienden). Op basis van deze literatuurstudie zou geconcludeerd kunnen worden dat mannen die genezen zijn van zaadbalkanker over het algemeen een goede kwaliteit van leven rap-porteren. Echter, omdat de meeste studies methodologische tekortkomingen vertonen en omdat de kwalitatief sterkere studies niet alle dimensies van kwaliteit van leven hebben onderzocht, kon een dergelijke conclusie niet getrokken worden. Om een beter inzicht


154

10

te krijgen in de kwaliteit van hun leven zal toekomstig onderzoek grotere onderzoeks-groepen met zowel korte- als lange termijn overlevers, sociodemografisch overeenko-mende vergelijkingsgroepen en gevalideerde vragenlijsten moeten omvatten. Deze studies zouden zich dan kunnen richten op het identificeren van risicoprofielen die gebruikt kun-nen worden bij het vinden van de overlevers die kankergerelateerde stress ervaren.

De studie die beschreven wordt in hoofdstuk 3 had als doel om de kwaliteit van leven van patiënten die zaadbalkanker hebben overleefd te onderzoeken. Bij dit onderzoek werd een aantal van de in de literatuurstudie beschreven methodologische tekortkomingen onder-vangen. Als eerste werd de kwaliteit van leven van genezen zaadbalkankerpatiënten (n = 354; tijd sinds behandeling: 3 maanden - 25 jaar) vergeleken met een referentiegroep van Nederlandse mannen. Ten tweede werden overlevers die risico lopen op een verminderde kwaliteit van leven geïdentificeerd op basis van sociodemografische variabelen, ziekte- en behandelingsvariabelen en recent meegemaakte levensgebeurtenissen. De kwaliteit van leven van de mannen die genezen zijn van zaadbalkanker bleek gelijk aan die van de refe-rentiegroep. De verstreken tijd sinds afronding van de behandeling, het type behandeling en het meemaken van een tweede kankerdiagnose bleken geen invloed te hebben op de kwaliteit van leven. Mannen die een minder goed kwaliteit van leven rapporteerden waren wel wat ouder en hadden meer negatieve levensgebeurtenissen in het voorgaande jaar meegemaakt. Echter, het niet hebben van een betaalde baan in combinatie met het heb-ben van een chronische ziekte was de sterkste voorspeller van een minder goede kwaliteit van leven. De antwoorden op een open vraag lieten zien dat, in vergelijking met andere overlevers, mannen zonder betaalde baan en met een chronische ziekte ook vaker een negatieve invloed van zaadbalkanker op hun huidige kwaliteit van leven rapporteerden. Op basis van deze resultaten kan geconcludeerd worden dat mannen die genezen zijn van zaadbalkanker een goede kwaliteit van leven rapporteren. Het feit dat er geen relatie werd gevonden tussen objectieve ziekte- en behandelingsvariabelen en kwaliteit van leven wijst erop dat de overlevers in staat zijn om zich snel aan te passen aan hun door kanker veranderde levensomstandigheden, zelfs na intensieve behandeling of een tweede kanker-diagnose. De negatieve invloed van de combinatie werkloosheid en chronische ziekte op kwaliteit van leven is ook bij de algemene bevolking aangetroffen. Deze bevinding is begrij-pelijk gezien het feit dat chronische ziekten sociale activiteiten en werk kunnen verstoren, wat vervolgens kan leiden tot een lagere kwaliteit van leven. Niettemin, de onderzochte variabelen hadden maar een beperkte invloed op de kwaliteit van leven in deze groep, dus waarschijnlijk zijn andere predictoren meer bepalend hiervoor.

Zinbeleving draagt in sterke mate bij aan het psychosociaal welbevinden van mensen. Mensen die hun leven als zinvol beleven, bezien hun leven vanuit een betekenisvol per-spectief en ontlenen daaraan een aantal levensdoelen. Bovendien zijn zij bezig met het vervullen van deze levensdoelen. Mannen die zaadbalkanker hebben gehad kunnen het moeilijk vinden om hun ziekteperiode als zinvol te ervaren en de ziekteperiode kan ervoor zorgen dat bepaalde levensdoelen onhaalbaar worden. Dit kan een negatieve invloed heb-ben op psychosociaal welbevinden en stress veroorzaken. Het eerste doel van de studie die beschreven wordt in hoofdstuk 4, was te onderzoeken of mannen die genezen zijn van zaadbalkanker het leven als zinvol ervaren. Als tweede is onderzocht of de kankerervaring


155

10

ervoor heeft gezorgd dat ze een andere kijk op het leven hebben gekregen. Ten derde is onderzocht in hoeverre zinbeleving, sociodemografische variabelen, ziekte- en behande-lingsvariabelen, het hebben van een chronische ziekte en recent meegemaakte gebeurte-nissen psychosociaal welbevinden en kankergerelateerde stress kunnen voorspellen. De resultaten lieten zien dat mannen die genezen zijn van zaadbalkanker hun leven als zinvol ervaren. In overeenstemming met onderzoek naar deze aspecten in de algemene popula-tie, bleek dat de overlevers zonder partner, of zonder kinderen, of zonder betaalde baan, of met een chronische ziekte, of degenen die meer negatieve levensgebeurtenissen hadden meegemaakt het bestaan als minder zinvol ervaren. Ziekte- en behandelingsvariabelen waren niet gerelateerd aan zinbeleving. Ongeveer tweederde van de mannen gaf aan dat ze door de ervaring met zaadbalkanker anders tegen het leven aankijken; de meerder-heid (60%) gaf aan dat ze nu een positievere kijk op het leven hebben. Zinbeleving was de belangrijkste voorspeller van psychosociaal welzijn, maar voorspelde in mindere mate kankergerelateerde stress.Samenvattend kan gesteld worden dat mannen die genezen zijn van zaadbalkanker het leven als zinvol ervaren en dat dit gevoel de belangrijkste voorspeller was van psychoso-ciaal functioneren. Zinbeleving had minder invloed op kankergerelateerde stress, wat erop wijst dat het wellicht belangrijker is voor het ervaren van een positieve kwaliteit van leven dan voor het omgaan met kankergerelateerde stress. Eerder onderzoek suggereert echter dat zinbeleving kan beschermen tegen kankergerelateerde stress. Toekomstig onderzoek kan de onderliggende mechanismen waardoor zinbeleving en kankergerelateerde stress met elkaar samenhangen wellicht blootleggen.

De ervaring met kanker kan zo traumatisch zijn dat het specifieke psychologische symp-tomen kan veroorzaken, zoals het ongewenst herbeleven van emotionele gebeurtenissen uit de ziekteperiode en het vermijden van situaties die onplezierige gevoelens en herin-neringen oproepen. Deze kankergerelateerde stress symptomen kunnen maanden of zelfs jaren na het beëindigen van de behandeling blijven bestaan. Kankergerelateerde stress symptomen zijn niet eerder onderzocht bij mannen die genezen zijn van zaadbalkanker. Het eerste doel van de in hoofdstuk 5 beschreven studie was het onderzoeken van de prevalentie van kankergerelateerde stress symptomen onder mannen die genezen zijn van zaadbalkanker. Daarnaast is (a) op kwantitatieve wijze onderzocht hoe sociode-mografische variabelen, ziekte- en behandelingsvariabelen en het hebben van een chro-nische ziekte relateren aan kankergerelateerde stress symptomen en (b) op kwalitatieve wijze onderzocht of concrete gebeurtenissen en subjectieve percepties relateren aan kan-kergerelateerde stress symptomen. Dertien procent van 354 mannen die genezen zijn van zaadbalkanker had last van klinisch verhoogde kankergerelateerde stress. Alleenstaande, lager opgeleide en werkeloze mannen rapporteerden meer stress symptomen. Tijd sinds behandeling, type behandeling, een tweede kankerdiagnose en het hebben van een chro-nische ziekte waren slechts zwak gerelateerd aan de mate van stress symptomen. Op basis van de antwoordpatronen op de vragenlijst werden twee groepen mannen geselecteerd voor een interview: 15 mannen met een klinisch verhoogd niveau van kankergerelateerde stress symptomen en 15 mannen zonder deze symptomen. De interviews richtten zich op de relatie tussen concrete ziektegerelateerde gebeurtenissen (zoals het ondergaan van een echografie) en subjectieve percepties (angst voor toekomstige medische problemen)


156

10

en kankergerelateerde stress symptomen. Het aantal objectief meegemaakte gebeurtenis-sen verschilde niet tussen mannen met en zonder kankergerelateerde stress symptomen. Echter, de mannen met een klinisch verhoogd niveau van kankergerelateerde stress gaven wel aan dat ze meer van slag raakten door de herinneringen aan deze gebeurtenissen. Ook gaven ze aan dat ze meer van slag zijn door de gevolgen van zaadbalkanker, omdat die heb-ben geleid tot meer veranderingen in het dagelijkse leven. Ook gaven ze aan dat ze banger zijn voor medische problemen in de toekomst. De mannen met klinisch verhoogde stress symptomen hadden het gevoel dat hun leven nog steeds bedreigd wordt door kanker.Samenvattend kan gesteld worden dat een belangrijke minderheid van de mannen die genezen is van zaadbalkanker nog kankergerelateerde stress symptomen ervaart. Socio-demografische en ziekte- en behandelinggerelateerde variabelen lijken geen sterke indica-toren te zijn voor kankergerelateerde stress, terwijl het ervaren van een negatieve invloed van de ziekte op het huidige leven en angst voor de toekomst er wel aan gerelateerd lij-ken te zijn. Als clinici vragen naar gedachten over de dreiging van kanker en de ervaren effecten op het huidige en toekomstige leven, krijgen ze wellicht beter inzicht in de mate van kankergerelateerde stress symptomen dan als ze zich uitsluiten richten op de objec-tieve medische geschiedenis. Mannen die het gevoel hebben dat hun leven nog steeds bedreigd wordt zouden baat kunnen hebben bij een psychosociale interventie gericht op het vervangen van onrealistische door realistische cognities.

Vermoeidheid is een van de meest voorkomende gevolgen van de behandeling van kanker en kan een negatieve invloed hebben op het welbevinden van patiënten en overlevers. In hoofdstuk 6 wordt de eerste studie beschreven naar de prevalentie van en de verand-eringen in vermoeidheid bij zaadbalkankerpatiënten gedurende het eerste jaar na orchi-dectomie. Daarnaast is in deze studie onderzocht of vermoeidheid samenhangt met sociodemografische variabelen, hemoglobine- en testosteronniveau en angstdispositie. Vijftien patiënten die alleen een orchidectomie hadden ondergaan en 37 patiënten die ook chemotherapie kregen, hebben na orchidectomie, drie maanden daarna (dit was direct na chemotherapie voor de patiënten die ook chemotherapie kregen) en twaalf maanden na orchidectomie een vragenlijst ingevuld. Zaadbalkankerpatiënten rapporteerden meer ver-moeidheid vlak na orchidectomie en, in het geval van chemotherapie, na het afronden van chemotherapie. Een jaar na orchidectomie lag het vermoeidheidsniveau van beide groepen patiënten op een hoogte vergelijkbaar met dat van mensen die geen kanker hadden gehad. Een hogere leeftijd was matig tot sterk gerelateerd aan meer vermoeidheid. Verder bleek dat de patiënten die op het eerste meetmoment een hoger niveau van vermoeidheid rap-porteerden, dit ook deden op het tweede en derde meetmoment. Angstdispositie bleek een voorspeller van zowel huidig als toekomstig niveau van vermoeidheid. Het aantal patiënten met anemie of een laag testosteronniveau was te klein, zodat het niet mogelijk was om de invloed ervan op vermoeidheid te onderzoeken. Op basis van deze bevindingen kan gesteld worden dat zaadbalkankerpatiënten het meest vermoeid zijn kort na orchidectomie en chemotherapie. Vermoeidheid lijkt niet een blij-vend probleem te zijn voor de meerderheid van de patiënten, ongeacht de behandeling die ze hebben ondergaan. Oudere patiënten, patiënten die kort na orchidectomie al vermoeid zijn en degenen met een angstige aard lijken het risico te lopen op langdurige vermoeid-heid. Deze patiënten zouden baat kunnen hebben bij een multidisciplinaire interventie


157

10

gericht op lichaamsbeweging en het versterken van copingstrategieën die effectief zijn bij het verminderen van vermoeidheid.

In hoofdstuk 7 wordt een methodologisch onderwerp behandeld. In kwaliteit van leven studies worden verschillende methoden gebruikt om patiënten te werven. Veel studies werven patiënten via ziekenhuisdatabases, maar er zijn ook studies waarbij patiënten worden gevraagd via patiëntenverenigingen, het internet, of advertenties in kranten en tijdschriften. Het werven van patiënten via verschillende bronnen kan leiden tot popula-ties met heel verschillende karakteristieken en dat kan weer van invloed zijn op de gene-raliseerbaarheid van de resultaten. Tot nu toe is er nog maar weinig empirisch onder-zoek gedaan naar de relatie tussen wervingsbronnen en kwaliteit van leven. De studie die beschreven wordt in dit hoofdstuk had als doel om inzicht te krijgen in (1) de ka-rakteristieken van de patiënten die geworven worden via verschillende bronnen en (2) de invloed van de wervingsbron op kwaliteit van leven. Daartoe werden leden van de patiëntenvereniging voor mannen die zaadbalkanker hebben gehad (n = 85) en overlevers van zaadbalkanker uit de UMCG database (n = 354) vergeleken op sociodemografische variabelen, ziekte- en behandelingsvariabelen en kwaliteit van leven. De leden van de patiëntenvereniging bleken significant jonger, korter geleden behandeld, hoger opgeleid, vaker een gecombineerde behandelding te hebben gehad en vaker een tweede keer kanker te hebben gehad. Verder hadden ze significant lagere scores op de meeste schalen van de kwaliteit van leven vragenlijst. Ook als werd gecorrigeerd voor de sociodemografische en ziekte- en behandelingsvariabelen waarop de twee groepen verschilden, rapporteerden de leden van de patiëntenvereniging nog steeds een lagere kwaliteit van leven dan mannen die benaderd waren via het UMCG. Er werd geconcludeerd dat het werven van patiënten via verschillende bronnen inderdaad leidt tot populaties met verschillende sociodemografische en ziekte- en behandelingsgere-lateerde karakteristieken en verschillende kwaliteit van leven scores. Echter, verschillen in sociodemografische en ziekte- en behandelingsvariabelen lijken dit verschil in kwaliteit van leven niet te veroorzaken. Dit betekent dat andere factoren het verschil in kwaliteit van leven tussen de groepen bepalen. Deze resultaten laten zien dat onderzoekers voor-zichtig moeten zijn met het benaderen van patiënten voor hun onderzoek, want de bron waaruit ze hun onderzoekspopulatie werven kan van invloed zijn op de resultaten en dus op de generaliseerbaarheid daarvan.

In hoofdstuk 8, ten slotte, worden de belangrijkste bevindingen van het onderzoek be-sproken en in een breder kader geplaatst. Daarnaast wordt er een aantal methodologische kwesties in beschouwing genomen en worden er suggesties gedaan voor verder onder-zoek. Het hoofdstuk sluit af met een bespreking van de implicaties van dit onderzoek voor de klinische praktijk.


Dankwoord


161

Dankwoord

Allereerst wil ik de deelnemers aan het onderzoek bedanken voor het invullen van de vragenlijsten. Dit proefschrift was er niet geweest zonder hun medewerking.

Verder wil ik een aantal mensen bedanken voor hun bijdrage bij het opzetten van het onderzoek, het analyseren van de verkregen resultaten en het schrijven van de artikelen. Dit proefschrift had ik nooit kunnen voltooien zonder hun hulp.

Als eerste, Dr JEHM Hoekstra-Weebers, mijn copromotor en dagelijkse begeleider. Josette, wat hebben wij de afgelopen jaren veel samen gedaan! Denk aan de vele uren die we om tafel hebben gezeten om de eerste, tweede of weet-ik-hoeveelste versie van mijn artikelen te bespreken. Maar denk ook aan onze reis naar de Verenigde Staten, de con-gresbezoeken die we samen hebben afgelegd en de gesprekken en discussies die we heb-ben gevoerd. Ik heb veel van je geleerd over goed onderzoek doen, maar ik heb het ook erg gezellig met je gehad. Bedankt daarvoor!

Prof dr HJ Hoekstra, promotor. Harald, jij hebt me zo’n vijf jaar geleden enthousiast ge-maakt voor dit onderzoek, je hebt me gemotiveerd tijdens het promotietraject en je hebt altijd vertrouwen in me gehad. Je weet niet half hoe belangrijk dat voor me is geweest. Dank je wel!

Prof dr DTh Sleijfer, promotor. Dirk, jij hebt al mijn artikelen met een kritische blik gelezen. Je wist de overbodige zinnen en inconsistenties feilloos te vinden en wilde altijd weten wat de relevantie voor de praktijk was. Jouw begeleiding heeft absoluut bijgedragen aan de kwaliteit en helderheid van de artikelen. Bedankt!

Prof dr EC Klip, promotor. Ed, jou wil ik bedanken voor je bereidheid om voortdurend mee te denken over het onderzoek. Elke keer dat ik met je om tafel zat leerde ik weer wat bij over de psychologie van ziekte en gezondheid.

Imi Veldman, onderzoeksassistent van onschatbare waarde en paranimf. Imi, elke chaoot zou iemand als jou naast zich moeten hebben. Zonder jou was de dataverzameling nooit zo soepel en georganiseerd verlopen en was de onderzoeksperiode nooit zo gezellig ge-weest. Bedankt voor alles wat je voor me hebt gedaan, zowel tijdens als na het onderzoek. Bedankt ook voor de vele kopjes koffie en de fijne gesprekken. Laten we elkaar niet uit het oog verliezen.

Martijn Lutke Holzik, ook jou wil ik bedanken. Het lijkt al weer zo lang geleden dat we samen aan het project zijn begonnen, jij als arts, ik als psycholoog. Ik heb twee erg leuke jaren met je gehad en ben blij dat we elkaar nog steeds af en toe zien en spreken.

Marrit Tuinman, collega, paranimf en vriendin. Marrit, na Martijn kwam jij op het project en het zat gelijk goed tussen ons. Ik heb veel aan je gehad bij het schrijven van de artike-len, want je structureerde me en wist mijn gedachten soms beter te formuleren dan ikzelf


162

kon. Ook buiten het werk heb ik je leren kennen en ik kan je zeggen, het is goed praten en lachen met jou. Zullen we gewoon doorgaan met elke dag bellen en elkaar vaak zien?

Raoul Nap en Roy Stewart hebben me geholpen met de statistiek van het zevende hoofd-stuk. Raoul en Roy, jullie hebben me allebei met buitengewoon veel geduld keer op keer op keer op keer op keer op keer op keer op keer uitleg gegeven over de statistische tech-nieken die jullie gebruikten. Gelukkig was er ook altijd tijd voor gesprekken over films, huisdieren, mensen en het leven. Op een of andere manier duurden de overleggen altijd langer dan gepland. Raoul en Roy, bedankt voor jullie hulp en tijd.

Prof dr JW Groothoff, Prof dr JM Nijman en Prof dr MAG Sprangers: bedankt voor het beoordelen van het proefschrift.

Ten slotte wil ik nog een aantal collega’s noemen die niet direct betrokken waren bij het project, maar die er wel voor hebben gezorgd dat ik enorm veel heb gelachen tijdens mijn promotietraject: Antonius Breeuwsma, Brunhild Eijsenga, Suzan ten Heuvel, Miranda Hollander, Gea Huizinga, Rudy Komdeur, Niels Koopmans, Thomas Muller, Tom Pereboom, Annemieke Visser, Wendel Vonkeman, Mattijs de Vries, Ellen van Weert en Gonda Wiersma. Wat leuk dat ik jullie heb leren kennen!


Northern Centre for Healthcare Research (NCH)

This thesis is published within the research program Disorder, Disability and Quality of Life of the Northern Centre for Healthcare Research. More information regarding the institute and its research can be obtained from our internet site: www.med.rug.nl/nch.

Previous dissertations from the program Disorder, Disability and Quality of Life

Arnold R (2004) Quality of life in chronic obstructive pulmonary disease and chronic heart failure.PROMOTORES: prof dr R Sanderman, prof dr GH KoëterCO-PROMOTORES: dr AV Ranchor, dr MJL De Jongste

Keers JC (2004) Diabetes rehabilitation; effects and utilisation of a multidisciplinary education programme.PROMOTORES: prof dr R Sanderman, prof dr ROB GansCO-PROMOTORES: dr J Bouma, dr TP Links

Pool G (2003) Surviving testicular cancer; sexuality and other existential issues.PROMOTORES: prof dr HBM van de Wiel, prof dr DTh Sleijfer, prof dr HJ HoekstraCO-PROMOTOR: dr MF van Driel

Stiegelis HE (2003) A life less ordinary; cognitive adaptation and psychological functioning among cancer patients treated with radiotherapy.PROMOTORES: prof dr R Sanderman, prof dr AP BuunkCO-PROMOTOR: dr M Hagedoorn

Heuvel ETP van den (2002) Supporting caregivers of stroke patients; an intervention study.PROMOTORES : prof dr B Meyboom-de Jong, prof dr R SandermanCO-PROMOTOR: dr LP de WitteREFERENT: dr LM Schure

Schroevers MJ (2002) Short- and long-term adaptation to cancer; a comparison of patients with the general population.PROMOTOR: prof dr R SandermanREFERENT: dr AV Ranchor

Hoekstra-Weebers JEHM (2000) Parental adaptation to pediatric cancer. PROMOTORES: prof dr EC Klip, prof dr WA KampsREFERENT: dr JPC Jaspers

Doeglas DM (2000) Functional ability, social support and quality of life: a longitudinal study in patients with early rheumatoid arthritis. PROMOTORES: prof dr WJA van den Heuvel, prof dr R SandermanCO-PROMOTOR: dr ThPBM Suurmeijer

Nijboer C (2000) Caregiving to patients with colorectal cancer: a longitudinal study on caregiving by partners.PROMOTORES: prof dr GAM van den Bos, prof dr R SandermanCO-PROMOTOR: dr AHM Triemstra

Tiesinga LJ (1999) Fatigue and Exertion Fatigue: from description through validation to application of the Dutch Fatigue Scale (DUFS) and the Dutch Exertion Fatigue Scale (DEFS). PROMOTORES: prof dr WJA van den Heuvel, prof dr ThWN DassenCO-PROMOTOR: dr RJG Halfens

Jong GM de (1999) Stress, stress management and issues regarding implementation. PROMOTORES: prof dr PMG Emmelkamp, prof dr JL PescharREFERENT: dr R Sanderman

Alberts JF (1998) The professionalized patient: sociocultural determinants of health servicesutilization.PROMOTOR: prof dr WJA van den Heuvelreferent: dr R Sanderman


Objective and subjective predictors of cancer-related stress symptoms in testicular cancer survivors

Documents

Transcript of Objective and subjective predictors of cancer-related stress symptoms in testicular cancer survivors