International Psychogeriatrics: page 1 of 11 C© International Psychogeriatric Association 2014doi:10.1017/S1041610214000581

“Being all alone makes me sad”: loneliness in older adultswith depressive symptoms

...........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................

Ilse M. J. van Beljouw,1 Eric van Exel,1 Jenny de Jong Gierveld,2 Hannie C. Comijs,1

Marjolijn Heerings,1 Max L. Stek1 and Harm W. J. van Marwijk3

1Department of Psychiatry, VU University Medical Center / GGZ inGeest and EMGO+ Institute for Health and Care Research, A.J. Ernststraat 1187,1081 HL Amsterdam, the Netherlands

2Netherlands Interdisciplinary Demographic Institute (NIDI), P.O. Box 11650, 2502 AR, The Hague, the Netherlands3Department of General Practice, VU University Medical Center and EMGO+ Institute for Health and Care Research, Van der Boechorststraat 7,1081 BT Amsterdam, the Netherlands

ABSTRACT

Background: The consequences of co-occurring persistent loneliness and late life depression are yet unknown.The aim of this study was to get a deeper insight into the mental health consequences of loneliness in olderpersons with depressive symptoms and their perspectives of emotional distress by using a mixed-methodsstudy design.

Methods: Two hundred and forty nine community-dwelling older persons with depressive symptomsaccording to the Patient Health Questionnaire-9 (�6) were included. A validated cut-off score on theLoneliness Scale was used to distinguish lonely elders from elders who were not lonely. Quantitative andqualitative data were used to examine differences in mental health and perspectives on emotional distressbetween lonely and not lonely older persons with depressive symptoms.

Results: Loneliness was highly prevalent among older persons with depressive symptoms (87.8%). Lonelypeople suffered from worse mental ill-health (e.g., more severe depressive symptoms, more often a depressivedisorder and a lower quality of life) compared to not lonely individuals. Depressive symptoms were regardedas a logical consequence of loneliness. Lonely people perceived little command over their situation: causes ofloneliness were attributed externally to perceived deficits in their social networks and they mainly expressedthe need to be listened to.

Conclusion: Our findings underline the importance of paying considerable attention to (severe) lonelinessin older adults with depressive symptoms given its high prevalence and serious mental health consequences.Future studies should look into whether addressing loneliness when discussing depressive symptoms in clinicalpractice may provide an opportunity to better adjust to older persons’ depression perceptions and mighttherefore improve care utilization.

Key words: aged, depression, loneliness, health knowledge, attitudes, practice, qualitative research

Introduction

Persistent loneliness may pose a serious concernin older adults since it is related to poorerphysical health (Luanaigh and Lawlor, 2008),impaired cognitive functioning (Wilson et al.,2007), increased mortality (Luanaigh and Lawlor,2008; Holwerda et al., 2012), and particularlydepression (Alpass and Neville, 2003; Adams et al.,

Correspondence should be addressed to: E. van Exel, Department of Psychiatry,VU University Medical Center/GGZ inGeest and EMGO+ Institute forHealth and Care Research, A.J. Ernststraat 1187, 1081 HL Amsterdam, theNetherlands. Phone: +0031 20 788 45 35. Email: e.vanexel@ggzingeest.nl.Received 18 Nov 2013; revision requested 10 Jan 2014; revised version received4 Mar 2014; accepted 5 Mar 2014.

2004; Cacioppo et al., 2006b; Cohen-Mansfield andParpura-Gill, 2007). Prevalence rates of lonelinessin older adults are high with estimates around 25%(Adams et al., 2004; Stek et al., 2005).

Feeling lonely is a subjective experience thatis distinct from objective social isolation. Socialisolation relates to social integration indices suchas no (longer) being married, living alone, orthe number of contacts with friends and familymembers. Feelings of loneliness however refer toa discrepancy between the individual’s desired andactual social relations, either in quantity or quality(Peplau and Perlman, 1982). Therefore, lonelinessis not equivalent to social isolation: socially isolatedpersons do not always feel lonely and lonely

2 I. M. J. van Beljouw et al.

people are not necessarily socially isolated (de JongGierveld and Havens, 2004).

Subjective loneliness has been consistently andstrongly related to the presence of depressivesymptoms in older adults. Loneliness is two tothree times more prevalent in older adults witha depressive disorder than in nondepressed elders(Adams et al., 2004; Stek et al., 2005). Several cross-sectional studies of older persons have shown thatloneliness is one of the most important determinantsof depressive symptoms (Alpass and Neville, 2003;Adams et al., 2004; Cacioppo et al., 2006b; Cohen-Mansfield and Parpura-Gill, 2007).

This strong association between depression andloneliness raises questions about the distinctivenessof both constructs. Factor analytic studies provideevidence for their conceptual difference by showingthat items from depression and loneliness ratingscales load on separate factors (Cacioppo et al.,2006a). Accordingly, from a clinical perspective, itbecomes clear that both phenomena appear to berelated but separate entities: although depression iscommon in lonely older adults, almost half of thedepressed elders do not report feeling lonely (Steket al., 2005).

Yet the exact nature of the relationship is poorlyunderstood. The above-mentioned studies wereall performed in pooled samples of older adultswith and without depressive symptoms. To ourknowledge, no single study has yet been conductedin a sample consisting exclusively of older adultswith depressive symptoms. The consequences ofco-occurring loneliness and late life depression aretherefore unknown.

The perspectives of lonely older persons withdepressive symptoms on this matter have notbeen addressed either. The only qualitative studyperformed on this subject in a pooled samplesuggests that deeper insight into older persons’illness perceptions may have a considerablepotential to improve depression management. Thisstudy implied that older persons view lonelinessas a precursor to depression and that it might beless stigmatizing to express feelings of lonelinessthan to disclose depressive symptoms (Barg et al.,2006). We consider the absence of any mentionof loneliness in depression guidelines a significantomission.

Current research does not yet provide a compre-hensive understanding of the consequences of co-occurring loneliness and late life depression or offerrecommendations for the prevention and treatmentof depression in lonely older persons. Therefore,more insight into the complex association betweendepression and loneliness, and in particularthe perspectives of older persons themselves, isneeded. The present study uses both quantitative

and qualitative methods to answer the followingresearch questions: (1) are feelings of loneliness inolder persons with depressive symptoms associatedwith more serious mental health consequences?And (2) how would older persons with depressivesymptoms and loneliness experience thesefeelings?

Methods

DesignThis study was conducted as part of the “Lust forLife”-trial, aimed at preventing and treating de-pression in older adults by implementing a steppedcare intervention program. This program wasimplemented in 18 general practices and one homecare facility in the Netherlands. These practices andthe home care facility were randomized into fourclusters following a stepped-wedge randomized-cluster design. Although starting moments werephased, the intervention program was implementedin all clusters by the end of the study. Enlisted olderpersons with depressive symptoms were recruitedby screening. They then completed a baselineand entered a waiting period for the interventionprogram to be implemented in their cluster.Subsequently, they were invited for an intakesession to discuss participation in the interventionswhen depressive symptoms persisted. All subjectswere followed up from baseline.

For the current study, a mixed-methods ap-proach was used to provide a more comprehensiveview on the close connection between depressivesymptoms and loneliness than either quantitativeor qualitative methods can provide on their own.Quantitative methods are appropriate to relatesubjects’ objective mental health to the presenceof loneliness. Qualitative methods are suitable togain an understanding of older adults’ perceptionsof emotional distress and perceived needs. Inthe present study, both methods were integratedto verify and corroborate findings from differentapproaches about this single phenomenon (trian-gulation) and to elaborate on or clarify the resultsof one method by means of the other (complement-arity). The VU University Medical Centre EthicalReview Board approved of the study (No. 2010/084).

RecruitmentSubjects were recruited from 18 participatinggeneral practices and a home care facility in theNetherlands (Figure 1). All enlisted persons of65 years and older (N = 9,661) were informedabout the intervention program and invited to fill in

Loneliness in elders with depressive symptoms 3

Figure 1. Recruitment of subjects. ∗ 29 subjects were excluded for reasons of the insufficient mastery of the Dutch language (n = 9),

deceased/could not be reached (n = 18), or moving outside the research area (n = 2).

a screening questionnaire for depressive symptoms,the self-report Patient Health Questionnaire-9(PHQ-9; Spitzer et al., 1999). The questionnairewas returned by 5,492 persons (56.8%), of whom758 (13.8%) scored above the predefined cut-offscore of 6. Four hundred and sixty three personsrefused participation and 32 persons were excluded(9 had insufficient mastery of the Dutch language,3 moved outside the research area, and 18 could notbe reached). Data on loneliness were missing for 14subjects, leading to a final sample of 249 personsfor the current study.

Data collection

QU A N T I T A T I VE DATA C OLLECTION

After screening assessment and obtaining writteninformed consent, subjects were invited to completea baseline structured telephone interview andfill in a written questionnaire. Written follow-up

measurements took place every three months afterbaseline for two years.

QUALITATIVE DATA COLLECTION

Qualitative data reported on in this paper wereobtained from semi-structured interviews to gaina deeper insight into subjects’ experiences ofemotional distress and reasons for accepting ordeclining the “Lust for Life”-intervention program,reported on elsewhere (van Beljouw et al.,submitted). Therefore, interviews were performedafter the intake session had taken place and subjectshad decided on accepting or declining the program.A convenience sample of all cluster three subjectseligible for intake (PHQ � 6, n = 38) were invited bytelephone to participate in the interviews after thegoal of the research had been explained to them.Twenty six subjects agreed to being interviewed,six persons refused participation, five could not be

4 I. M. J. van Beljouw et al.

interviewed within the time frame of this study, andone subject could not be reached. Two respondentsdid not provide recent data on loneliness and weretherefore excluded from the current analyses. Of the24 remaining interviewed subjects, 5 refused and 19accepted the intervention program. Interviews wereconducted at the subjects’ homes by two researchers(MH or IvB) within 5 weeks after the intake sessiontook place, and lasted between 31 and 92 minutes.

Measurements

QU A N T I T A T I VE M EASUREMENTS

Feelings of loneliness were measured by thevalidated De Jong Gierveld Loneliness Scale (DeJong-Gierveld and Kamphuis, 1985), citing 11statements such as “I miss having people around,”that can be rated on a 3-point rating scale. Acutting score of �3 was used to distinguish notlonely from lonely people (van Tilburg and de JongGierveld, 1999). A second cutting score was usedto distinguish severe loneliness from no or mildfeelings of loneliness by splitting sum scores at themedian. This cutting score of �9 was in accordancewith the cut-off reported by Van Tilburg and DeJong Gierveld (1999) to distinguish severely lonelypersons from not/mildly lonely individuals.

Background characteristics were gathered bymeans of written questionnaires. For mental (ill-)health, diagnoses of past and current (in the past2 weeks) major depressive disorder and dysthymiaas well as the age of onset and number of previousdepressive episodes were assessed by the MiniInternational Neuropsychiatric Interview (MINI;(Sheehan et al., 1998), a short structured diagnosticinterview conducted by telephone by trainedresearch staff. Current dysthymia was only exploredwhen a current depressive disorder was absent. Forsymptom severity, scores on the Patient HealthQuestionnaire-9 (PHQ-9; Spitzer et al., 1999)during screening were reported. Feelings of anxietywere derived from two items from the MedicalOutcomes Study Short Form 36 (SF-36; Ware andSherbourne, 1992). Since higher scores were relatedto less feelings of anxiety, scores were invertedto facilitate interpretation. Self-rated quality oflife was measured with Cantril’s Ladder (Cantril,1966).

QU A LI T A T I VE MEASUREMENTS

Qualitative semi-structured interviews startedwith an open question about the most importantreason(s) for accepting or declining the interventionprogram. To fully capture older persons’perspectives and beliefs, we consequently askedsubjects about their self-perceived emotionalproblems in a semi-structured way. Subjects’ own

terminology to describe their emotional distress(e.g., feeling “gloomy,” “lonely,” or “anxious”),was adopted by the interviewer in subsequentquestions on self-perceived causes of emotionalproblems and need for care.

Data analysis

QUANTITATIVE DATA A NALYSIS

Simple and multiple logistic regression techniqueswere used to determine which mental health factorswere associated with severe loneliness, adjusting forage, sex, and educational level. Determinants witha nonlinear relationship (PHQ-scores, number ofdepressive episodes) were divided into quartiles.SPSS version 20 was used for all analyses.

QUALITATIVE DATA ANALYSIS

Semi-structured interviews were digitally recordedand transcribed verbatim. Transcripts were readand reread and a coding tree was built around keythemes such as “perception of current emotionaldistress” using Atlas.ti 5.2. Within the key themes,different levels of codes were constructed (e.g.,“type of emotional distress” or “cause of emotionaldistress”), each consisting of several sub-codes (e.g.,“mood-related problems,” “feelings of loneliness,”and “feelings of anxiety or worrying”). Interviewswere independently coded by two researchers (MHand IvB), and discussed until consensus wasachieved. Data saturation was reached since no newthemes emerged from the interviews.

Coded interviews were scrutinized for underlyingthemes and associations between themes by apply-ing grounded theory approach (Strauss and Corbin,1990). Emerging themes were compared betweennot/mildly lonely subjects and severely lonelypersons. This distinction was based on recent scoreson the De Jong Gierveld Loneliness Scale (De Jong-Gierveld and Kamphuis, 1985), collected during aregular follow-up measurement within 6–15 weeksaround the semi-structured interview. To minimizesubject burden in validation procedures (Barbour,2001), subjects checked and – if applicable –added to an oral summary of the discussedthemes provided by the interviewer at the end ofthe interview. Coauthors attended regular teammeetings in which original data, (sub-)codes, andresults were presented and discussed (Tong et al.,2007).

Results

SampleResults from quantitative analyses were basedon data from all included subjects (n =

Loneliness in elders with depressive symptoms 5

Table 1. Demographic and clinical characteristics of included (left) and interviewed (right) persons (by meansof semi-structured interviews)

I N C L U D E D I N T E R V I E W E D

SUBJECTS N = 249 SUBJECTS N = 24% (N)/M E A N ± SD % (N)/M E A N ± SD

.........................................................................................................................................................................................................................................................................................................................

Demographic characteristicsAge in years 75.2 ±6.7 76.1 ±7.3Sex, female 71.1 (177) 75.0 (18)Region, urban 76.7 (191) 91.7 (22)Household composition, alone 67.1 (167) 75.0 (18)Marital status

Married/cohabiting 32.9 (82) 25.0 (6)Widowed/divorced 56.6 (141) 75.0 (18)Never married 10.4 (26) 0 (0)

Clinical characteristicsDepressive symptom severity1

6–7 19.7 (49) 16.7 (4)8–10 29.7 (74) 29.2 (7)11–14 28.9 (72) 37.5 (9)15–27 21.7 (54) 16.7 (4)

Major depression (MDD)No diagnosis 57.0 (142) 54.2 (13)Current MDD 20.1 (50) 4.2 (1)

Past MDD 22.9 (57) 41.7 (10)Dysthymia§

No diagnosis 82.3 (205) 95.8 (23)Current dysthymia 9.6 (24) 0 (0)Past dysthymia 7.2 (18) 4.2 (1)

Loneliness2

Lonely (�3) 87.8 217 83.3 (20)Severely lonely (�9) 48.6 (121) 41.7 (11)

1Measured with the PHQ-9, range 0–27.2Measured with the Jong Gierveld Loneliness Scale; loneliness at baseline shown for all included respondents, and for interviewed subjectswithin 6–15 weeks around the interview.§Numbers do not add up to 100% due to missing values (n = 2).

249); qualitative results were derived from semi-structured interviews with a subsample of 13not/mildly lonely and 11 severely lonely individuals.Demographic and clinical characteristics of bothsamples are shown in Table 1. Feelings ofloneliness were highly prevalent among older adultswith depressive symptoms: 87.8% (n = 217)perceived loneliness. Almost half of all subjectsfelt severely lonely (48.6%, n = 121). Table 2shows differences in background characteristicsbetween not/mildly lonely and severely lonelypersons.

Mental health consequences of co-occurringloneliness and depressionTable 3 shows the results of logistic regressionanalyses comparing subjects with depressivesymptoms and severe feelings of loneliness tonot/mildly lonely individuals on several personal

and clinical factors. Qualitative information wasused to elaborate on or add to these findings.

Older people with depressive symptoms whowere severely lonely suffered from worse mentalill-health than not/mildly lonely older adults.Severe loneliness was associated with more severedepressive symptoms, a current major depressivedisorder, a current or past dysthymic episode, andmore feelings of anxiety. Also a lower self-ratedquality of life was related to perceiving severefeelings of loneliness.

The association between loneliness and worsemental ill-health also became apparent from thequalitative data. When asked whether subjectsexperienced emotional problems, a wide range ofsymptoms and complaints was described. Althoughall subjects scored above the cut-off score of six onthe Patient Health Questionnaire, indicative of thepresence of depressive symptoms, several not/mildlylonely persons mentioned other problems than a

6I.M

.J.vanB

eljouw

etal.

Table 2. Differences in background characteristics between subjects who feel not/mildly lonely versus those who feel severely lonely (N = 249); resultsfrom simple and multiple logistic regression analyses.

UNADJUSTED ADJUSTED 1

NOT/M I L D L Y L O N E L Y S E V E R E L Y L O N E L Y

N = 128 N = 121% (N)/M E A N ± SD % (N)/M E A N ± SD OR 95%CI OR 95%CI

...........................................................................................................................................................................................................................................................................................................................................................................................................................................................

Background characteristicsGender, female 68.8 (88) 73.6 (89) 1.26 0.73–2.19 1.22 0.70–2.13Age, 65–69 years 25.8 (33) 21.5 (26) Ref Ref

70–74 25.0 (32) 31.4 (38) 1.51 0.75–3.03 1.41 0.70–2.8575–80 20.3 (26) 25.6 (31) 1.51 0.73–3.15 1.37 0.65–2.92� 81 28.9 (37) 21.5 (26) 0.89 0.44–1.83 0.82 0.39–1.72

Education5–9 years 34.4 (44) 34.2 (41) ref Ref

10 years 44.5 (57) 45.8 (55) 1.04 0.59–1.82 0.99 0.56–1.76� 11 years 21.1 (27) 20.0 (24) 0.95 0.48–1.91 0.94 0.46–1.94

Region, urban 75.0 (96) 78.5 (95) 1.22 0.68–2.20 1.23 0.67–2.25Marital status

Married/cohabiting 45.3 (58) 19.8 (24) ref RefWidowed/divorced 46.1 (59) 67.8 (82) 3.36∗∗∗ 1.88–6.01 3.59∗∗∗ 1.94–6.66Never married 8.6 (11) 12.4 (15) 3.30∗ 1.32–8.20 3.39∗ 1.33–8.65

Household composition, alone 55.5 (71) 79.3 (96) 3.08∗∗∗ 1.76–5.41 3.24∗∗∗ 1.79–5.84Income (net per month, €)

600–900 26.5 (26) 40.4 (38) ref ref901–1299 7.1 (7) 14.9 (14) 1.37 0.49–3.85 1.46 0.51–4.151300–2099 40.8 (40) 26.6 (25) 0.43∗ 0.21–0.87 0.47∗ 0.23–0.98� 2100 25.5 (25) 18.1 (17) 0.47 0.21–1.03 0.52 0.22–1.24

†p < 0.10, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.1Adjusted for age, sex, and educational level (if applicable).

Lonelinessin

eldersw

ithdepressive

symptom

s7

Table 3. Differences in mental health characteristics and perceived need for care between subjects who feel not/mildly lonely versus those who feel severelylonely (N = 249); results from simple and multiple logistic regression analyses

UNADJUSTED ADJUSTED 1

NOT/M I L D L Y L O N E L Y S E V E R E L Y L O N E L Y

N = 128 N = 121% (N)/M E A N ± SD % (N)/M E A N ± SD OR 95%CI OR 95%CI

...........................................................................................................................................................................................................................................................................................................................................................................................................................................................

Depressive symptom severity2

6–7 25.0 (32) 14.0 (17) ref ref8–10 34.4 (44) 24.8 (30) 1.28 0.61–2.72 1.33 0.63–2.8311–14 24.2 (31) 33.9 (41) 2.49∗ 1.18–5.27 2.43∗ 1.14–5.1815–27 16.4 (21) 27.3 (33) 2.96∗∗ 1.33–6.61 2.93∗∗ 1.31–6.57

Major depression (MDD)No diagnosis 62.5 (80) 51.2 (62) ref refCurrent MDD 13.3 (17) 27.3 (33) 2.51∗∗ 1.28–4.91 2.45∗ 1.23–4.85Past MDD 24.2 (31) 21.5 (26) 1.08 0.58–2.01 1.10 0.58–2.11

Age of onset depression 44.1 ± 21.9 45.7 ± 21.8 1.00 0.99–1.02 1.01 0.99–1.03Nr. of depressive episodes

1 27.3 (12) 16.4 (9) ref ref2–3 22.7 (10) 27.3 (15) 2.00 0.62–6.49 2.05 0.62–6.774–10 25.0 (11) 36.4 (20) 2.42 0.78–7.54 2.51 0.79–7.99� 11 25.0 (11) 20.0 (11) 1.33 0.40–4.44 1.38 0.40–4.76

Dysthymia§

No diagnosis 92.1 (117) 73.3 (88) ref refCurrent dysthymia 5.5 (7) 14.2 (17) 3.23∗ 1.28–8.12 2.99∗ 1.17–7.64Past dysthymia 2.4 (3) 12.5 (15) 6.65∗∗ 1.87–23.67 6.83∗∗ 1.89–24.69

Age of onset dysthymia§ 52.6 ±21.3 49.0 ±22.5 0.99 0.96–1.03 0.99 0.96–1.03Feelings of anxiety3 7.5 ±1.8 8.3 ±1.6 1.31∗∗∗ 1.13–1.53 1.30∗∗ 1.12–1.52Quality of life4 6.5 ±1.4 6.0 ±1.3 0.75∗∗ 0.61–0.92 0.75∗∗ 0.61–0.91

†p < 0.10, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.1Adjusted for age, sex, and educational level.2Measured with the PHQ-9, range 0–27.3Measured with the SF-36 anxiety items, range 3–12 (scores were inverted, so that more severe feelings of anxiety correspond to higher scores).4Measured with the Cantril’s Ladder, range 0–10.§Numbers do not add up to 100% due to missing values (n = 2).

8 I. M. J. van Beljouw et al.

depressed mood such as having trouble sleepingor feelings of anxiety. Yet mood-related problemswere acknowledged by all subjects who felt severelylonely.

Further, about three-quarters of the severelylonely persons also perceived their emotionaldistress as (highly) burdensome. For instance, a 74-year old male (resp#14) stated: “At this age, I reallydon’t like having to struggle through all those years tocome by myself.” In contrast, only a minority of oldernot/mildly lonely persons described their emotionalproblems (if present in their opinion) as hindering.For example, an 81-year old female (resp#11) said:“Sure, I have my ups and downs. But that is part oflife, isn’t it? You don’t need to be upset by that.”

Experience of loneliness and depressivesymptomsThree major themes emerged from the in-depthinterviews about older persons’ experiences ofdepressive symptoms and severe loneliness: (1)the self-perceived relationship between depressivesymptoms and loneliness; (2) self-perceived causesof severe loneliness; and (3) self-perceived needs toalleviate emotional distress.

THE R ELA T I ONSHIP BETWEEN DEPRESSIVE

S Y M P T O M S A N D L O N E L I N E S S

Most severely lonely persons attributed theirdepressive symptoms to feeling lonely or tolimitations of their social network. A 74-year oldfemale (resp#21) explained: “Well yes, I am alone. Ihave lost my husband, I have lost my parents. I don’thave anyone. That also makes me gloomy”. Or, asa 79-year old female (resp#7) stated: “I feel sadpretty often, because I think: darn, I’m alone, but they[my children] won’t call.” Although feeling lonelywas also perceived as the cause of their depressivesymptoms by a minority of not/mildly lonely olderadults, most of them mentioned other reasons.They referred to a wide variety of aging-relatedcircumstances such as physical constraints, griefover the loss of loved ones, losing their homes bymoving to retirement housing, or no longer beingable to fulfill (volunteer) work.

SE L F-P E R C E I V E D C A U S E S OF SE V E R E

L O N E L I N E S S

Severely lonely persons with depressive symptomsattributed their feelings of loneliness to externalrather than internal factors. Changes in or perceivedshortcomings of their social networks were regardedas logical causes of their loneliness. Severely lonelysubjects referred to the loss of others due to illness,death, divorce, moving away, or to infrequentcontact with children. For example, a 74-year-

old male (resp#14) commented: “I am actuallydisappointed in people. [ . . . ] Also in my own children.I hardly ever see them. [ . . . ] It makes me sad, asthe world around you has become so small.” Or, as a79-year-old female (resp#7) said: “I used to havenice friends, but unfortunately I have lost them. Ialways have to do everything by myself.” Some othersmentioned external factors that maintained theirfeelings of loneliness, e.g., limited mobility or havinglittle money to spend on activities. An 86-year-oldfemale (resp#17) explained: “I used to have a lotmore social contacts. [ . . . ] But I couldn’t visit them,and they can not come to me, they are all becoming old.We are all in the same boat.”

Two older severely lonely adults pointed topersonal characteristics that limited them inengaging in or seeking new social contacts, suchas constantly bringing up their own diseases inconversations with others or having a low self-esteem: “[I do] not have such a strong self esteemanymore. And also having very little social contact. Sothese things reinforce each other a little.” (a 66-year-oldmale, resp#24).

SE L F-P E R C E I V E D NE E D S TO A L L E V I A T E

E M O T I O N A L DI S T R E S S

The majority of not/mildly lonely older adultsperceived an important role for themselves inalleviating their emotional distress, by expressinga need to learn new skills to handle emotionalproblems or to exchange experiences with otherelders with the same problems. For example, a 75-year old female said: “Well, some kind of tool to pickmyself up a little. [ . . . ] Like: think about that, or tryto do that . . . ” (resp#3). The majority of severelylonely persons however disclosed more passiveneeds for care. They mainly expressed the needto be listened to. This need could be fulfilled bysomeone they confided in, who was not necessarilya professional. For instance, a 73-year old woman(resp#12) stated: “Because you don’t have anyone totalk to. Or to tell your story to. That I regret. That issomething that I would really like.” Only one personexpressed a need to talk to a professional care giver:“ . . . to be able to talk to someone about that. [ . . . ]The therapeutical conversation. To see how someone elselooks at that. From a more professional perspective” (a66-year-old male, resp#24). The need to meet other(like-minded) elders to share common interests oractivities with was only mentioned by two severelylonely persons.

Discussion

This is the first study examining the consequencesof co-occurring loneliness and depression in a

Loneliness in elders with depressive symptoms 9

sample exclusively consisting of older adults withdepressive symptoms in detail. Our unique mixed-method findings showed that severe loneliness isa serious concern in older people with depressivesymptoms, given its high prevalence and worsemental ill-health. From older persons’ perspectives,it became clear that depressive symptoms wereseen as a logical consequence of severe loneliness.Severely lonely persons furthermore perceived littlecommand in the experience and possible solutionsto their feelings of loneliness: the causes ofloneliness were attributed externally to perceiveddeficits in their social networks and they mainlyexpressed the need to be listened to.

The mental health consequences of co-occurringloneliness and late life depression had not beenstudied before in a sample consisting exclusivelyof older persons with depressive symptoms. Wefound that a worse mental ill-health was morefrequently seen in severely lonely older adults withdepressive symptoms compared to not/mildly lonelydepressed persons. Our findings are underlined bytrends observed in studies using pooled samples ofolder adults with and without depression, showinga linear relation between loneliness and depressionseverity (Alpass and Neville, 2003; Adams et al.,2004; Cacioppo et al., 2006b; Cohen-Mansfieldand Parpura-Gill, 2007). The exact nature of theassociation between loneliness and depression isyet poorly understood, although it does seem clearfrom factor-analytic studies that they are relatedbut distinct phenomena (Cacioppo et al., 2006a).Although older persons themselves perceived(severe) loneliness as a gateway to depression, as hasbeen reported previously (Barg et al., 2006), resultsfrom existing longitudinal studies on the directionof this association remain diverging (Heikkinenand Kauppinen, 2004; Cacioppo et al., 2006b;Cacioppo et al., 2010). Our cross-sectional data arenot suitable to provide answers to this question.

Our quantitative results have shown that severelylonely persons with depressive symptoms perceivedlittle command in the experience and possiblesolutions to their emotional distress. Feelings ofloneliness were attributed externally to perceiveddeficiencies in their social networks due to (aging-related) losses of loved ones and limited contactwith family members, especially with children.This finding that important fractured relationshipsare perceived as precipitating loneliness as peopleage is in line with previous research (McInnisand White, 2001; Smith, 2012). Growing older isinevitably accompanied by losses of important otherpeople, but the way this is dealt with underlines animportant difference between lonely and not lonelyolder persons. Kirkevold et al. (2013) note that theability to cope with losses by moving on and staying

connected to others is more frequently seen in notlonely older persons, while lonely individuals appearto be more often paralyzed by the loss of importantrelationships. Also, not lonely older persons tryto deal with infrequent contact with children bynot depending too heavily on them but to focuson social relationships outside of their families. Incontrast, lonely elders feel disappointed in theirchildren and their limited contact is perceived as animportant cause of loneliness (McInnis and White,2001; Kirkevold et al., 2013), as also confirmed byour findings. Lonely persons rather wait for othersto connect with them, instead of taking matters intotheir own hands.

The most important clinical implication derivedfrom our findings is that it could be of great valueto enquire about (severe) feelings of lonelinesswhen discussing depressive symptoms with olderadults in the doctor’s office. Terminology could beadapted more to older persons’ illness perceptionsby acknowledging the important role of lonelinessin the onset and existence of current depressivesymptoms. This might in turn increase olderpersons’ willingness to accept depression treatmentsince most severely lonely elders with depressivesymptoms view their loneliness as the mostsignificant problem and their treatment might,therefore, require a tailored approach. Since ourresults suggest that severely lonely persons are lessinclined to seek help, it might especially temptthose with mild feelings of loneliness into acceptingtreatment. It should be noted however that thisrecommendation applies to older persons with mildto moderate depressive symptoms; it seems lessapplicable to the treatment of persons with severedepression.

The consequences of our findings for currentdepression guidelines however remain unclear.

It is unknown whether lonely older persons withdepressive symptoms are willing to discuss theirsevere loneliness in the doctor’s office, whetherthey perceive a need to alleviate these feelings andwhether the recommendation to address feelings ofloneliness actually improves depression treatmentuptake and outcomes. Furthermore, addressingloneliness in older persons with depressivesymptoms alone might not be sufficient to improvethe initiation and management of depressiontreatment. It is unclear whether interventions aimedat alleviating loneliness might have a considerablepotential to treat or prevent depressive symptoms inolder persons. Future research is needed to providean answer to these questions.

This is the first study of the mentalhealth consequences of co-occurring lonelinessand late life depression in older adults withdepressive symptoms. The combined approach

10 I. M. J. van Beljouw et al.

of qualitative and quantitative methods allowedfor a thorough examination of this subject fromdifferent perspectives. However, our study also hasits limitations. As can be expected, results areonly based on the perspectives of subjects whowilling to be screened and gave informed consentto participate in the study. Also, measurementbias cannot be ruled out in the assessmentof depressive symptoms by a validated self-report questionnaire such as the PHQ-9. Further,due to time constraints, purposive sampling forthe semi-structured interviews was less feasible,leading for instance to a limited number of malesubjects who were interviewed since depressionoccurs less frequently in males. Also, althoughperceived loneliness is associated with personalitycharacteristics, especially neuroticism (Cacioppoet al., 2006a; Luanaigh and Lawlor, 2008), we wereunable to relate respondents’ personality traits totheir perceived feelings of loneliness (or control forthem) since this information was not collected inthe original study.

In conclusion, our findings underline theimportance of paying considerable attention to(severe) loneliness in older adults with depressivesymptoms, given its high prevalence and seriousconsequences. Addressing loneliness in the doctor’soffice may provide an opportunity to better adjustto older persons’ depression perceptions and mighttherefore improve care utilization.

Conflict of interest

None.

Description of authors’ roles

I. van Beljouw carried out the study, carried out andcoded the semi-structured interviews, performedthe qualitative and quantitative analyses, and wrotethe paper. M. Heerings carried out and coded thesemi-structured interviews, performed qualitativeanalyses and cowrote the paper. E. van Exel, H.van Marwijk, and M. Stek had the original ideafor the study, designed the study, obtained thegrants, helped in the analyses and interpretation ofthe data, supervised the project, and cowrote thepaper. H. Comijs supervised the project, helpedin the analyses and interpretation of the data andcowrote the paper. J. de Jong Gierveld helped in theanalyses and interpretation of the data and cowrotethe paper. All authors approved the final version ofthe manuscript.

Acknowledgments

This work was supported by The NetherlandsOrganisation for Health Research and Development(grant number 60.61900.98.353).

References

Adams, K. B., Sanders, S. and Auth, E. A. (2004).Loneliness and depression in independent living retirementcommunities: risk and resilience factors. Aging & MentalHealth, 8, 475–485.

Alpass, F. M. and Neville, S. (2003). Loneliness, health anddepression in older males. Aging & Mental Health, 7,212–216.

Barbour, R. S. (2001). Checklists for improving rigour inqualitative research: a case of the tail wagging the dog?British Medical Journal, 322, 1115–1117.

Barg, F. K., Huss-Ashmore, R., Wittink, M. N., Murray,G. F., Bogner, H. R. and Gallo, J. J. (2006). Amixed-methods approach to understanding loneliness anddepression in older adults. Journals of Gerontology Series B:Psychological Sciences and Social Sciences, 61, S329–S339.

Cacioppo, J. T. et al. (2006a). Loneliness within anomological net: an evolutionary perspective. Journal ofResearch in Personality, 40, 1054–1085.

Cacioppo, J. T., Hughes, M. E., Waite, L. J., Hawkley,L. C. and Thisted, R. A. (2006b). Loneliness as a specificrisk factor for depressive symptoms: cross-sectional andlongitudinal analyses. Psychological and Aging, 21, 140–151.

Cacioppo, J. T., Hawkley, L. C. and Thisted, R. A.(2010). Perceived social isolation makes me sad: 5-yearcross-lagged analyses of loneliness and depressivesymptomatology in the Chicago Health, Aging, and SocialRelations Study. Psychological Aging, 25, 453–463.

Cantril, H. (1966). The Pattern of Human Concerns. NewBrunswick: Rutgers University Press.

Cohen-Mansfield, J. and Parpura-Gill, A. (2007).Loneliness in older persons: a theoretical model andempirical findings. International Psychogeriatriatrics, 19,279–294.

De Jong-Gierveld, J. and Kamphuis, F. (1985). Thedevelopment of a Rasch-type loneliness scale. AppliedPsychological Measurement, 9, 289–299.

de Jong Gierveld, J. and Havens, B. (2004). Cross-nationalcomparisons of social isolation and loneliness: introductionand overview. Canadian Journal of .Aging, 23, 109–113.

Heikkinen, R. L. and Kauppinen, M. (2004). Depressivesymptoms in late life: a 10-year follow-up. Archives ofGerontology and Geriatrics, 38, 239–250.

Holwerda, T. J. et al. (2012). Increased risk of mortalityassociated with social isolation in older men: only whenfeeling lonely? Results from the Amsterdam Study of theElderly (AMSTEL). Psychological Medicine, 42,843–853.

Kirkevold, M., Moyle, W., Wilkinson, C., Meyer, J. andHauge, S. (2013). Facing the challenge of adapting to alife ‘alone’ in old age: the influence of losses. Journal ofAdvanced Nursing, 69, 394–403.

Loneliness in elders with depressive symptoms 11

Luanaigh, C. O. and Lawlor, B. A. (2008). Loneliness andthe health of older people. Intermational .Journal of GeriatricPsychiatry, 23, 1213–1221.

McInnis, G. J. and White, J. H. (2001). Aphenomenological exploration of loneliness in the olderadult. Archives of Psychiatric Nursing, 15, 128–139.

Peplau, L. A. and Perlman, D. (eds.) (1982). Loneliness, aSourcebook of Current Theory, Research and Therapy. NewYork: Wiley.

Sheehan, D. V. et al. (1998). The Mini-InternationalNeuropsychiatric Interview (MINI): the development andvalidation of a structured diagnostic psychiatric interviewfor DSM-IV and ICD-10. Journal of Clinical Psychiatry, 59(Suppl 20), 22–33.

Smith, J. M. (2012). Toward a better understanding ofloneliness in community-dwelling older adults. Journal ofPsychology, 146, 293–311.

Spitzer, R. L., Kroenke, K. and Williams, J. B. (1999).Validation and utility of a self-report version ofPRIME-MD: the PHQ primary care study. Primary CareEvaluation of Mental Disorders. Patient HealthQuestionnaire. Journal of the American Medical Association,282, 1737–1744.

Stek, M. L., Vinkers, D. J., Gussekloo, J., Beekman,A. T. F., van der Mast, R. C. and Westendorp,

R. G. J. (2005). Is depression in old age fatal only whenpeople feel lonely? American Journal of Psychiatry, 162,178–180.

Strauss, A. and Corbin, J. (1990). Basics of QualitativeResearch. Grounded Theory Procedures and Techniques.London: Sage.

Tong, A., Sainsbury, P. and Craig, J. (2007). Consolidatedcriteria for reporting qualitative research (COREQ): a32-item checklist for interviews and focus groups.International Journal of Qualilty of Health Care, 19,349–357.

van Beljouw, I. M. J. et al. (submitted). Pulling out all thestops: What motivates 65+ year-olds with depressivesymptoms to participate in an outreaching personalisedintervention program? Ageing and Mental Health.

van Tilburg, T. G. and de Jong Gierveld, J. (1999).Cesuurbepaling van de eenzaamheidsschaal (Cutting scoreson the De Jong Gierveld Loneliness Scale). Tijdschrift voorGerontologie en Geriatrie, 30, 158–163.

Ware, J. E. J. and Sherbourne, C. D. (1992). The MOS36-item short-form health survey (SF-36). I. Conceptualframework and item selection. Medical Care, 30,473–483.

Wilson, R. S. et al. (2007). Loneliness and risk of Alzheimerdisease. Archives of General Psychiatry, 64, 234–240.