Satisfaction With Mental Health ServicesAmong People With Schizophrenia inFive European Sites: Results From the

EPSILON Study

by Mirella Ruggeri, Antonio Lasalvia, Qiulia Bisoffi, Qraham Thomicroft,Jose Luis Vazquez'Barquerot Thomas Becker, Martin Knapp,

Helle Charlotte Knudsen, Aart Schene, Michele Tansella,and the EPSILON Study Qroup

Abstract

Patient satisfaction with services is an important out-come variable that is increasingly used in mentalhealth service evaluation. This study includes 404 peo-ple with schizophrenia in five European sites andaddresses five questions focused on site, service, andpatient characteristics as variables that might explainservice satisfaction, using the Verona ServiceSatisfaction Scale. Patient satisfaction differed signifi-cantly across sites (highest in Copenhagen, lowest inLondon). In all sites, patients were least satisfied withinvolvement of relatives in care and information aboutillness. A multiple regression model showed that lowerlevels of total service satisfaction were associated withliving in London or Santander, being retired/unem-ployed, having more hospital admissions, having moresevere psychopathology, having more unmet needs, orhaving lower satisfaction with life. This modelexplained 31 percent of variance in service satisfac-tion. Our data show that service satisfaction can beseen as a result of (1) the ability of the service to pro-vide a standard of care above a certain quality thresh-old, and (2) the perception of each patient that the carereceived has been tailored to the patient's own prob-lems.

Keywords: Schizophrenia, patient satisfaction,community mental health services.

Schizophrenia Bulletin, 29(2):229-245,2003.

Satisfaction with services has been given increasing atten-tion in mental health services research, as it represents akey component of the patients' perspective in outcomeassessment. In this context, patient satisfaction can beviewed as a measure of outcome per se and/or as a factor

in the process of care influencing other outcomes (Ruggeri1994). Assessment of service satisfaction in severely men-tally ill patients has been relatively neglected because of aview, by clinicians and researchers, that a lack of insightmay compromise the validity of self-reported outcomes.While these difficulties should not be discounted, over thepast few years a growing body of evidence has shown thatthe self-reports of people with psychotic disorders are reli-able and convey valid and useful information (Naber et al.1994; Awad et al. 1995; Voruganti et al. 1998). This sug-gests that the severity of illness in itself does not necessar-ily undermine the ability of patients to report their viewsand experiences (Awad and Voruganti 2000).

As an outcome variable, satisfaction has been hypoth-esized to be the consequence of various factors, includingexpectations about services, attitudes toward life, self-esteem, illness behavior, previous experience with ser-vices, and service characteristics (Svensson and Hansson1994; Barker et al. 1996). Nevertheless, the publishedwork on the predictors of service satisfaction is inconsis-tent and generally of poor scientific quality. Conflictingresults have been obtained on the relationship betweenpatients' satisfaction and their sociodemographic charac-teristics, type of diagnosis, and severity of illness (Larsenet al. 1979; Hansson 1989; Kelstrup et al. 1993; Perreaultet al. 1996; Leavey et al. 1997; Greenwood et al. 1999).On balance, the results of previous studies show thatusers' characteristics have a weak to moderate associationwith service satisfaction. The only variable shown to beclearly and consistently associated with service satisfac-tion is self-perceived quality of life (Ruggeri et al. 1998,

Send reprint requests to Professor M. Ruggeri, Dipartimento di Medicinae SanitA Pubblica, Sezione di Psichiatria, Universita di Verona, OspedalePoliclinico, 37134 Verona, Italy; e-mail: mirella.ruggeri@univr.it

229

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Schizophrenia Bulletin, Vol. 29, No. 2, 2003 M. Rugged et al.

2001, 2002; Rohland et al. 2000; Berghofer et al. 2001;Druss et al. 2001), but the nature and the cross-culturalstability of this relationship remain to be clarified.

In contrast, service characteristics do seem to play amajor role in service satisfaction. For example, a numberof previous studies have consistently reported that patientstreated in community-based mental health services expresshigher levels of satisfaction than those receiving hospital-based mental health care (Hoult and Reynolds 1984;Elbeck and Fecteau 1990; Merson et al. 1992; Dean et al.1993; Audini et al. 1994; Marks et al. 1994; Leese et al.1998; Boardman et al. 1999; Henderson et al. 1999).Moreover, two studies conducted recently in England onpatients with psychotic disorders reported that highernumbers of unmet needs for care tend to be associatedwith lower service satisfaction (Leese et al. 1998; Board-man et al. 1999). Unfortunately, most studies on servicesatisfaction have been conducted using unstandardizedmeasures of overall satisfaction (Rugged 1994) and byassessing the performance of experimental rather than rou-tine psychiatric services. Moreover, no studies have beendesigned to compare service satisfaction in people withschizophrenia who live in different countries, using stan-dardized instruments.

This study was conducted as a comparison betweenfive European sites and was specifically designed toaddress five key questions. Three were focused on the roleof service characteristics, two were focused on the role ofpatients' characteristics in relation to service satisfaction,and all five were established a priori (Becker et al. 1999):(1) Does the overall level of service satisfaction differacross the five European sites? (2) Are the weaknesses andthe strengths of mental health services similar across thefive European sites? (3) Is higher service satisfaction asso-ciated with a lower number of patients' unmet needs forcare? (4) Is service satisfaction related to patients'sociodemographic characteristics and illness severity? (5)Is service satisfaction associated with patients' satisfactionwith life?

Methods

Research Setting. This research was conducted as a partof the EPSILON (European Psychiatric Services: InputsLinked to Outcomes and Needs) Study, a comparative,cross-national, cross-sectional study of the characteristics,needs for care, quality of life, caregiver burden, patternsof care, associated costs, and satisfaction levels of peoplewith schizophrenia in five European sites. Study centerswere located in Amsterdam (The Netherlands),Copenhagen (Denmark), London (U.K.), Santander(Spain), and Verona (Italy). The criteria used to identifystudy centers, and the key characteristics of each study

site, are given in Becker et al. (1999). Briefly, the criteriaused to identify study centers were similar to thoseemployed in other European research consortia (Dowricket al. 1998): (1) experience in health services research,mental health epidemiology, and development and cross-cultural adaptation of research instruments; (2) access tomental health services providing care for local catchmentareas; (3) a national health service providing communitymental health care; and (4) geographical and culturalspread across the European Union.

Subjects. The EPSILON study was conducted with atotal sample of 404 subjects with an ICD-10 (WHO1992a) research diagnosis of schizophrenia (F20 code,corresponding to 295 DSM-IV code). The number ofpatients for each site varied from 52 (Copenhagen) to 107(Verona). Cases included were adults aged 18-65 yearsinclusive (Becker et al. 1999). In the first stage of thestudy, administrative prevalence samples of people with adiagnosis of schizophrenia or other psychotic disorders(ICD-10, F20-F25) were initially identified either frompsychiatric case registers (in Copenhagen and Verona) orfrom the caseloads of local specialist mental health ser-vices (inpatients and outpatients). All patients in contactwith mental health services during the 3-month periodpreceding the start of the study were selected. Cases iden-tified were diagnosed using the Item Group Checklist ofthe Schedule for Clinical Assessment in Neuropsychiatry(WHO 1992/?). Only patients with an ICD-10 researchdiagnosis of schizophrenia were included in the study.The exclusion criteria were current residence in prison,secure residential services, or hostels for long-termpatients; coexisting learning disability (mental retarda-tion), primary dementia, or other severe organic disorder;or extended inpatient treatment episodes longer than 1year. Fuller details on sample selection have been pub-lished (Becker et al. 1999).

Instruments. Satisfaction with mental health serviceswas assessed using the European version of the VeronaService Satisfaction Scale (VSSS-EU), an instrumentdeveloped by careful translation, back-translation, andcultural adaptation from the original VSSS (Knudsen etal. 2000). It is designed for use in comparative cross-national research projects as well as in routine clinicalpractice in mental health services across Europe and hasbeen shown to have good levels of internal consistencyand test-retest reliability (Ruggeri et al. 2000). VSSS-EUconsists of 54 items (see column 1 in the Appendix),which conceptually cover seven dimensions: overall satis-faction, professionals' skills and behavior, information,access, efficacy, type of intervention, and relatives'involvement.

230

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

The instrument is designed for self-administration andcan be completed in 20-30 minutes, without prior training.Subjects are asked to give an overall rating of their experi-ence of the mental health services they have attended inthe previous year. For items 1-40, satisfaction ratings areon a 5-point Likert scale (1 = terrible, 2 = mostly unsatis-factory, 3 = mixed, 4 = mostly satisfactory, 5 = excellent).The items are presented with alternate directionality toreduce stereotypic responses. Items 41-54 consist of threequestions each: first the subject is asked if he or she hasreceived the specific intervention (Question A: "Did youreceive the intervention x in the last year?")- If the answeris "yes," the subject is asked about his or her satisfactionon a 5-point Likert scale, as above (Question B). If theanswer is "no," the subject is asked Question C: "Do youthink you would have liked to receive intervention JC?" (6= no, 7 = don't know, 8 = yes). These questions allowmeasurement of the subjects' satisfaction both with theinterventions provided and with the professionals' deci-sion not to provide an intervention. The latter may be con-sidered a measure of underprovision of care from thepatient's point of view.

Other measures included in the study are the BriefPsychiatric Rating Scale (BPRS; Ventura et al. 1993), theGlobal Assessment of Functioning (GAF; APA 1994), theEuropean version of the Camberwell Assessment of Need(CAN-EU; McCrone et al. 2000), and the European ver-sion of the Lancashire Quality of Life Profile (LQL-EU;Gaite et al. 2000).

Statistical Analyses. Patients' sociodemographic, serviceutilization, and clinical characteristics were comparedusing chi-square, one-way analysis of variance (ANOVA),and Kruskal-Wallis tests, where appropriate. In the case ofchi-square tests, standardized residuals were generated toidentify the source of any significant difference. If theadjusted residual had a value of -3 or below, then that sitewas considered to have a significantly lower rating.Similarly, the site was considered to have a significantlyhigher rating if the value was +3 or above.

The VSSS-EU total mean score and VSSS-EUdimension scores were compared across sites. Unadjustedand adjusted mean values and their confidence intervalswere calculated. For the unadjusted means, the VSSS-EUtotal score and the VSSS-EU dimension scores were com-puted, using as independent variables the study centersonly (Amsterdam was used as the reference category). Toproduce scores adjusted to the mean levels of all thecovariates, the study centers and patients' backgroundcharacteristics (age, gender, ethnicity, marital status, lan-guage, employment status, lifetime admissions, yearssince first contact, level of functioning, and psychopathol-ogy) were entered, in turn, into linear regression models as

independent variables. When necessary, transformationswere made to normalize the variable distributions (accord-ing to Box-Cox method results). The analyses were per-formed using the adjust command in STATA Release 6.0(StataCorp 1999).

Intersite differences in service satisfaction were testedby means of ANOVA. Simes modified Bonferroni test wasused to ascertain pairwise significant differences betweencenters for both unadjusted and adjusted mean scores(Simes 1986). Differences between the scores obtained inthe VSSS domains within each center were also assessedby Simes modified Bonferroni test.

The percentage of patients dissatisfied in each VSSS-EU dimension, and the total percentage of dissatisfiedpatients, were compared across the five sites. Patientswere considered dissatisfied when their mean scores werebelow 3.5. Unadjusted and adjusted proportions of dissat-isfied patients were calculated across the five EPSILONsites using the same procedure as for the mean values,applying logistic regression models.

To explore which variables best explain patient satis-faction, a series of linear block regression analyses (SPSSInc. 2000) were performed, using in turn, as dependentvariables, the VSSS-EU total mean score and the meanscores of the VSSS-EU dimensions. Study sites (Amster-dam as reference category), sociodemographic character-istics, service utilization, psychopathology, care needs,and quality of life were used as putative explanatory vari-ables. The sociodemographic characteristics includedwere sex, age, marital status (single, married, other), livingsituation (alone, with family, with others), years of educa-tion, employment status (employed, sheltered work,unemployed, housewife/retired/student), ethnicity (white,other), and language (national, other).

The service utilization variables included in the mod-els were lifetime admissions (log transformed) and yearssince first contact with mental health services. The othervariables used were BPRS (mean scores minus reciprocaltransformed), GAF (total mean score), CAN-EU (totalmet and unmet needs), and LQL-EU (total mean score,and mean subjective life satisfaction scores in ninedomains). Block 1 included the study site variable only,Block 2 the sociodemographic and service utilization char-acteristics, and Block 3 the clinical variables (GAF andBPRS). Subsequently (Block 4), met and unmet needswere included, where the latter were recoded as a categor-ical variable (no unmet needs; 1-2 unmet needs; 3-4unmet needs; more than 5 unmet needs). Block 5 thenincluded the LQL-EU total mean score. Finally, anotherregression model was built, with Block 5 constituted bythe mean score of each LQL-EU dimension.

Block 1-3 variables have been constrained to enter inthe model, even when they were nonsignificant. For

231

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Schizophrenia Bulletin, Vol. 29, No. 2, 2003 M. Rugged et al.

blocks 4 and 5, a backward stepwise selection was used toretain in the final model only significant associations (p toenter: 0.05; p to remove: 0.1).

ResultsPatients' Characteristics Across the EPSILON StudySites. Patients' sociodemographic characteristics in thefive EPSILON sites and in the whole sample are reportedin table 1.

Patients' mean age differed significantly across sites,with the oldest in London and the youngest in Copenhagen.Ethnicity differed significantly, because of large ethnicminority populations in London and Amsterdam. Patients inAmsterdam, Copenhagen, and London were more likely tolive alone than those in Santander and Verona, who tended tolive with their families. This is also reflected in the signifi-cant difference in accommodation type. Most patients wereeither unemployed or were students, and this was especiallyso in London. Years of education also differed: patients inAmsterdam had received the most education, while those inVerona had received the least.

Service utilization and clinical characteristics of thepatients in the EPSILON sites and in the whole sample arereported in table 2.

Time since first contact with psychiatric services didnot differ across sites, while there were significant differ-ences in lifetime admissions (with the highest number inLondon and the lowest in Santander). Global functioningand level of psychopathology were roughly similar acrossthe sites. The total number of needs, the number of unmetneeds, and the subjective quality of life differed signifi-cantly between sites, with the highest number of totalneeds found in Amsterdam and the lowest in Santander,while the highest number of unmet needs was in Amster-dam and the lowest in Copenhagen. The highest satisfac-tion with life was reported by patients in Copenhagen andthe lowest was found in London.

Specific Service Interventions Provided andUnderprovision of Care Across the EPSILON Sites.Table 3 shows the percentage of subjects who reportedhaving received specific service interventions in the previ-ous year and the percentage of subjects wishing to havespecific service interventions that they had not received.

The vast majority of patients received psychopharma-cological treatments in all the EPSILON sites. In London,psychotherapy and rehabilitation were provided to thefewest patients. In Santander, the range of social interven-tions listed in the VSSS was provided to the fewest sub-jects and the highest underprovision of care was detected.Amsterdam, Copenhagen, and Verona tended to providethe full spectrum of interventions available.

Service Satisfaction Across the EPSILON Sites. Table4 shows the unadjusted and adjusted means in the variousVSSS-EU dimensions and in the VSSS-EU total scoreacross the five EPSILON sites.

Satisfaction along the various VSSS-EU dimensionssignificantly differed across sites. When background char-acteristics were adjusted for, the difference remained sig-nificant, with the exception of the domain assessing globalsatisfaction. Ranking of sites did not substantially change.In terms of overall satisfaction, professionals' skills andbehavior, self-perceived efficacy, and type of intervention,the highest scores were found in Copenhagen and the low-est in London. Satisfaction with information was highestin Copenhagen and lowest in Santander. Satisfaction withaccess to services was highest in Copenhagen and lowestin Amsterdam and Verona. In terms of relatives' involve-ment, Verona had the highest score and London the lowest.Copenhagen had the highest level of total satisfaction andLondon the lowest. Within each center, most VSSS dimen-sions' scores were significantly different from each other(Simes modified Bonferroni test, p < 0.05), with relatives'involvement and information being the domains with thelowest score in most centers and the total sample.

To better identify the areas with higher dissatisfaction,unadjusted and adjusted percentages of dissatisfiedpatients (mean score below 3.5), as measured by the vari-ous VSSS-EU dimensions and by the total score acrossthe five EPSILON sites, are shown in table 5.

Subjects were fundamentally least satisfied with ser-vices in London and most satisfied in Copenhagen. Again,satisfaction with relatives' involvement and informationprovided were the domains with the highest number ofdissatisfied subjects. In the domain assessing self-per-ceived efficacy, a noticeable number of dissatisfiedpatients were found too. After adjustment for patients'background characteristics, the ranking of sites did notsubstantially change. Adjusted percentages of dissatisfiedpatients in the individual VSSS-EU items are shown in theappendix; they indicate wide variability of scores.

Variables Associated With Service Satisfaction. Asshown in table 6 (column 1), living in London andSantander, being retired/unemployed, having a high num-ber of hospital admissions, having high levels of psy-chopathology, having a high number of unmet needs, andhaving a poor quality of life in social relations and healthwere all associated with low total service satisfaction, andthe final model accounted for 31 percent of the variance.

With regard to the various satisfaction dimensions, amajor role is played by site (Block 1): living in London isassociated with low satisfaction in all dimensions (theonly exception being information/access), and living inSantander is associated with low satisfaction in informa-

232

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

Table 1. Comparison of patients' key sociodemographic characteristics across the EPSILON sites(chi-square and one-way analysis of variance)1

GenderMaleFemale

Marital statusSingleMarriedOther

EthnicityWhite EuropeanOther

LanguageNationalOther

Living situationAloneWith relativesWith others

AccommodationDomesticShelterHospitalOther

OccupationEmployed/studentSheltered work

Amsterdam(n = 61)

%

67.232.8

72.19.9

18.0

5A145.9

78.721.3

49.127.923.0

70.527.9

1.60.0

18.08.2

Unemployed/pensioner 57.4Housewife

Age of respondent, yrsGeneral education, yrs

16.4

Mean (SD)

39.9 (9.9)11.8(2.2)

Copenhagen(n = 52)

%

59.640.4

59.611.628.8

92.37.7

88.511.5

65.41^419.2

80.815.4

3.8

15.43.8

80.80.0

Mean (SD)

39.4 (9.6)9.9 (2.3)

London(n = 84)

%

58.341.7

64.315.520.2

65.534.5

92.97.1

41.736.921.4

79.713.14.82.4

431.2

91.72.3

Mean (SD)

43.8(12.3)11.3(1.5)

Santander(n = 100)

%

59.041.0

71.016.013.0

100.0QJJ

98.02.0

L289.04.Q

100.0QJ20.00.0

20.02.0

65.013.0

Mean (SD)

39.9 (9.3)10.7(2.3)

Verona(n = 107)

%

48.651.4

57.924.317.8

100.0o.o

100.0M

15.074.810.2

91.67.50.00.9

23.44.7

57.914.0

Mean (SD)

43.0(12.4)8.7 (3.2)

All sites(n = 404)

%

57.442.6

64.816.618.6

84.915.1

93.36.7

30.255.714.1

86.710.9

1.21.2

16.83.7

69.69.9

Mean (SD)

41.4(11.1)10.4(2.7)

p value2

x2

0.197

0.102

<0.001

<0.001

<0.001

<0.001

<0.001

ANOVA

0.024<0.001

Note.—SD = standard deviation, ANOVA = analysis of variance.1 Values that are in bold have adjusted standardized residuals > 3.0. Values that are underlined have adjusted standardized residuals lessthan -3.0.

tion/access, efficacy, and type of intervention. The vari-ance explained by site alone ranged between 5.6 percent(information/access) and 22.3 percent (type of interven-tion).

Among the sociodemographic variables (Block 2),younger age was associated with lower satisfaction in thetype of intervention. Being employed had an effect indecreasing satisfaction in specific domains, namely overallsatisfaction, type of intervention, and relatives' involve-ment. Living alone was associated with lower satisfactionin type of intervention, and a high number of admissionswas associated with lower satisfaction in type of interven-tion. However, when added to the site variables, sociode-mographic variables made a minimal contribution to thetotal variance explained.

High level of psychopathology (Block 3) was associ-ated with low satisfaction in information/access efficacyand relatives' involvement, and having a higher number ofunmet needs (Block 4) was associated with low satisfac-tion in information/access and relatives' involvement (butthe contribution of this latter variable to the varianceexplained was very low).

Quality of life (Block 5) was consistently associatedwith service satisfaction in each VSSS-EU dimension:the inclusion of the LQL-EU score gave the greatestcontribution in efficacy, information/access, and type ofintervention. The total variance explained by the finalmodels in the various dimensions ranged from 13.9 per-cent (overall satisfaction) to 34.8 percent (type of inter-vention).

233

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Schizophrenia Bulletin, Vol. 29, No. 2, 2003 M. Ruggeri et al.

CD

o.2a

CA

2CO

Z

oiin

i

25o8c73"<oa>

I•acoo

Ic\i0)

0)

To

CO

sJ2T,

cco

I5 n

co(A

2

S

Q

12.cCO

if9 c S"_i -=•</)

0) o

C II n

1 61

a II ^E ^^ to

coo

pc\i

O CO h; IO CO T - Oo r̂ ^ o r̂ o oO CM O p CM p o

V d ci d ci d V

O CD O)

p ^ •<* p p p tof) ID r II) ri r̂ *_ «^ irt 00 T— CM CO COGO , _ ^ O CM O> h*;

St o

ID r S Nco i n r-. CDi 6T^ iri c6 i-^

p p p cqi n ••- i n

m

mto ^

-^ o

COCMCO

O

CM

CMO) 00

oo) co r^ co co coco (Q ^ O) co ^|co in f- ^ co i -

O O CD CT>

d P d ^ P P P 1 ^ :i- oi e ^ * ri r *^ c\T

^ - c\i — - d c M i - T ^ dCo '^^ f^ ^ ^ ^ ^ ^"' >—' ^ ^in T - C D O T - C O ' - C Ocj • * o o < D . ° P ' - : u ? a ?»- C M i n i - T f c o ^ ' *

O O CM CO

^: P o J - ^ p p p c M7- ^ in i- m n w ^

CM" P

5 c ? ^ ? ^ § S S~ in — d CM i - CM dm ^ - ^ ^ ^ ^ ^ ^ - . — • ~-

co c o i ^ ' - i n ^ t t

i n i n m c o c M ' *

P CM ^co in i -

(Dp p p en•* co d t'

CM^^ co ^^ >—^ ^^ ^^ ^^o _,: o> en to co oN J * n « oi sis. — s si ^- i 2.co m r^ o> TT , - co»- p^ p i - oq co pin in W in co T^ 1-

5 « . . .CM in i - co co CM

co ^ ^ S c? § ?2 - CM CM i - O

co r̂ in r» o oN; ^ j tO CO r-; Tt

ri m r (D ri cicoCO

Eil§

c?

CD

1 <coDC0 .

3

o m o

IIQW

2

O

l.1t33

IooII

LL<

O

8CO

to

(Z aa t3

As shown in table 6 (Block 5), high quality of life inhealth and social relations is associated with higher satis-faction in all the VSSS-EU dimensions (the only excep-tion being relatives' involvement for satisfaction withsocial relations). When the LQL^-EU domain scores weresubstituted for the total score, the most clear-cut improve-ments in the variance explained were found in the follow-ing VSSS-EU dependent variables: satisfaction with thetype of intervention, and professional's skills and behav-ior.

Discussion

This is the first study to compare satisfaction with mentalhealth services in a large sample of schizophrenia patientsacross multiple European sites. The study used a multidi-mensional, setting-specific instrument, the VSSS-EU,whose psychometric properties and cross-cultural applica-bility were previously assessed in all participating sites(Ruggeri et al. 2000). The study has some limitations.First, subjects recruited in the EPSILON sites were repre-sentative of all treated (administrative) prevalence cases ofschizophrenia on the caseloads of mental health servicesin those particular catchment areas. They are likely to rep-resent the large majority of all true cases (Andrews et al.2001), but the extent to which this happens has not beenassessed. Second, the results found in each European sitecannot necessarily be generalized to other mental healthservices in the corresponding countries (Becker et al.1999). Third, for the larger metropolitan areas included inthis study (London and Amsterdam), because the generalpopulation and the nature of mental health services arehighly heterogeneous, our findings may not be representa-tive for the cities as a whole. Nevertheless, London datawere collected in Croydon, a semisuburban borough inSouth London that has exactly the mean score for sociode-mographic characteristic in England and is thereforebroadly representative of the entire national population(Johnson et al. 1997). Moreover, broader patterns of cross-cultural differences, such as the proportion of patients wholive with family members, are reasonably fairly reflectedby the five sites included in this study.

In spite of these limitations, this study represents thefirst multicenter study on schizophrenia patients' satisfac-tion with mental health services and its correlation withother key variables. The only multicenter study of patientsatisfaction published previously (Vicente et al. 1993) didnot target schizophrenia patients and did not use rigorousmethodology and comprehensive standardized assess-ments. The results of this study will now be presented inrelation to each of the five key questions, which were set apriori.

234

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

Table 3. Subjects receiving and subjects not receiving but wishing for specific service interventionsin the various EPSILON sites1

Medication prescription

Individual sessions

Group sessions

Family sessions

Individual rehabilitation

Practical help by the service at home

Informal admission to hospital

Compulsory treatment

Sheltered accommodation

Recreational activities in the service

Recreational activities outside the service

Help in finding open employment

Shelter work

Welfare benefits

ReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWishedReceivedWished

Amsterdam(n = 58), %

98.20.075.516.39.323.321.411.960.015.635.68.929.88.516.710.435.44.260.59.324.412.211.618.625.022.734.119.5

Copenhagen(n = 52),%

1000.050.030.027.714.930.218.651.312.834.813.038.012.09.39.319.18.543.58.712.219.55.017.521.49.562.215.6

London(n = 84),%

94.06.020.5NA8.5NA7.3NA21.7NA9.8NA15.7NA12.0NA21.7NA27.7NA18.3NA10.0NA7.2NA45.1NA

Santander(n = 100),%

1000.085.910.113.326.554.810.748.127.31.231.04.19.210.31.00.024.73.433.014.030.19.653.09.348.810.252.5

Verona(n = 107),%

95.30.096.21.917.29.745.014.059.615.927.813.429.34.010.80.913.43.134.711.926.39.114.419.613.318.420.824.7

Note.—NA = not available (number of valid responses to question was low); VSSS-EU = Verona Service Satisfaction Scale, Europeanversion.1 The percentages reported were computed on the basis of the patients' answers to questions 41-54 of the VSSS-EU (type of interven-tion dimension).

1. Does the Overall Level of Service Satisfaction DifferAcross the Five European Sites? Satisfaction with ser-vices varied substantially across the five European sites,both before and after adjusting for patients' backgroundcharacteristics. Overall, the highest level of satisfactionwas found in Copenhagen and the lowest in London, withthe remaining sites placed in an intermediate position(Santander showed lower satisfaction than Amsterdamand Verona). The explanation for this variability may berelated to the different characteristics of mental health ser-vices and of the social environment across the sites.Differences between study sites with regard to patients'sociodemographic, service utilization, and clinical charac-teristics have been accounted for in the calculation of theadjusted satisfaction scores; therefore, any remaining dif-ference may be interpreted as arising from other site-spe-cific factors. In fact, the five sites differed widely withrespect to cultural and economic factors, national healthcare systems, mental health service organization, and ser-

vice provision. While these aspects do not appear togreatly influence patients' clinical characteristics as mea-sured by the BPRS and the GAF scores, they do seem tobe associated with the overall level of service satisfaction.At the same time, the high degree of variability of thescores in the individual VSSS-EU items across the fivesites, and the absence of any specific pattern of differ-ences between southern and northern European countries,suggest that if cross-cultural differences do exist, theycannot be simplified into a north-south Europeandichotomy alone.

2. Are the Weaknesses and the Strengths of MentalHealth Services Similar Across the Five EuropeanSites? This study has been able to identify the strengthsand the weaknesses, from the perspective of patients, ofeach mental health service assessed and has demonstratedthat they differ in many respects across these Europeansites. This indicates the high degree of service specificity

235

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Table 4. Unadjusted and adjusted mean scores (95% Cl) in the VSSS-EU dimensions across EPSILON sites (1 = terrible; 5 =excellent)1

to

VSSS-EU dimension

Overall satisfactionUnadjustedAdjusted

Professionals' skillsand behavior

UnadjustedAdjusted

InformationUnadjustedAdjusted

AccessUnadjustedAdjusted

EfficacyUnadjustedAdjusted

Type of interventionUnadjustedAdjusted

Relative's involvementUnadjustedAdjusted

Total mean scoreUnadjustedAdjusted

Amsterdam(n = 58),

mean (95% Cl)2

3.90(3.70-4.10)3.97 (3.74-4.20)

3.97(3.83-4.10)4.00(3.85-4.15)

3.66 (3.43-3.89)3.71 (3.44-3.98)

3.63(3.44-3.81)3.60(3.37-3.81)

3.69 (3.51-3.87)3.78 (3.58-3.99)

3.65 (3.55-3.76)3.72 (3.58-3.86)

3.57 (3.31-3.83)3.62 (3.32-3.92)

3.79 (3.67-3.92)3.84 (3.70-3.98)

Copenhagen(n = 51),

mean (95% Cl)3

4.04 (3.83-4.25)4.19 (3.95-4.44)

4.13(3.99-4.28)4.19(4.02-4.35)

3.69 (3.45-3.93)3.82(3.53-4.11)

4.19(3.99-4.38)4.16(3.93-4.40)

3.79 (3.60-3.99)3.87(3.67^.11)

3.72 (3.61-3.84)3.81 (3.66-3.96)

3.32 (3.05-3.59)3.39 (3.07-3.70)

3.89 (3.76-4.03)3.96(3.82-^.12)

London(n = 83),

mean (95% Cl)4

3.45(3.28-3.61)3.42(3.23-3.61)

3.46 (3.34-3.57)3.47 (3.34-3.59)

3.26 (3.07-3.45)3.27 (3.04-3.49)

3.97(3.81-4.12)3.95 (3.77-4.13)

3.06(2.90-3.21)3.00(2.83-3.18)

3.06(2.95-3.16)3.05(2.94-3.17)

2.91 (2.71-3.12)2.88(2.64-3.11)

3.45 (3.35-3.55)3.31 (3.19-3.42)

Santander(n = 100),

mean (95% Cl)5

3.79 (3.64-3.94)3.78 (3.61-3.95)

3.94 (3.83-4.04)3.93 (3.82-4.05)

2.93(2.76-3.11)2.91 (2.71-3.12)

3.84 (3.70-3.98)3.86 (3.70-4.03)

3.41 (3.27-3.55)3.42 (3.26-3.58)

3.42 (3.33-3.52)3.39 (3.29-3.50)

3.39 (3.20-3.57)3.36(3.14-3.58)

3.59 (3.50-3.69)3.58 (3.48-3.69)

Verona(/i = 107),

mean (95% Cl)6

4.01 (3.87-4.16)3.93 (3.74-4.13)

4.00(3.90-4.11)3.93 (3.79-4.06)

3.64 (3.47-3.80)3.44(3.21-3.68)

3.64 (3.51-3.78)3.59 (3.40-3.78)

3.81 (3.68-3.94)3.65 (3.47-3.84)

3.75 (3.66-3.84)3.61 (3.49-3.74)

3.75 (3.56-3.94)3.55(3.29-3.81)

3.86 (3.76-3.95)3.74 (3.62-3.86)

ANOVA testp value

0.0460.283

<0.001<0.001

<0.0010.021

<0.0010.007

<0.0010.013

0.0010.006

0.0040.001

<0.0010.006

Simes modifiedBonferroni

test7

a,b, dNS

a, b, c, da, b, c, d, f, k

e,f,ha, b, e, f, g, h

j , k, m, ni, j , k, I, m

a, b, c, d, f, ha, b, c, d, e, f, h

a, b, c, d, f, ha, b, c, d, e, f, h

a, c, da, b, c, d

a, b, c, d, f, ha, b, c, d, e, f

ophri

**.&

1i fa'<:;—'

\o

2

to

I

Note.—ANOVA = analysis of variance; Cl = confidence interval; NS = nonsignificant; VSSS-EU = Verona Service Satisfaction Scale, European version.1 p values of the ANOVA test and pairwise significant differences between centers and, for each center, between VSSS domains are reported (Simes modified Bonferroni test, p <0.05).2 Amsterdam, high vs. low: Overall satisfaction vs. information/efficacy/type of intervention/relatives' involvement; professionals' skills and behavior vs. information/access/efficacy/typeof intervention/relatives' involvement.3 Copenhagen, high vs. low: Overall satisfaction vs. information/efficacy/type of intervention/relatives' involvement; professionals' skills and behavior vs. information/efficacy/type ofintervention/relatives' involvement; access vs. information/efficacy/type of intervention/relatives' involvement; efficacy vs. relatives' involvement; type of intervention vs. relatives'involvement.4 London, high vs. low: Overall satisfaction vs. information/efficacy/type of intervention/relatives' involvement; professionals' skills and behavior vs. information/efficacy/type of inter-vention/relatives' involvement; information vs. type of intervention/relatives' involvement; access vs. overall satisfaction/professionals' skills and behavior/information/efficacy/type ofintervention/relatives' involvement; efficacy vs. type of intervention/relatives' involvement; type of intervention vs. relatives' involvement.5 Santander, high vs. low: Overall satisfaction vs. information/efficacy/type of intervention/relatives' involvement; professionals' skills and behavior vs. overallsatisfaction/information/efficacy/type of intervention/relatives' involvement; access vs. information/efficacy/type of intervention/relatives' involvement; efficacy vs. information; type ofintervention vs. information; relatives' involvement vs. information.6 Verona, high vs. low: Overall satisfaction vs. information/access/efficacy/type of intervention/relatives' involvement; professionals' skills and behavior vs. information/access/efficacy/type ofintervention/relatives' involvement; efficacy vs. information.7 (a,) Amsterdam high, London low; (b) Copenhagen high, London low; (c) Santander high, London low; (d) Verona high, London low; (e) Amsterdam high, Santander low; (f) Copen-hagen high, Santander low; (g) London high, Santander low; (h) Verona high, Santander low; (i) Amsterdam high, Verona low; (j) Copenhagen high, Amsterdam low; (k) Copenhagenhigh, Verona low; (I) London high, Amsterdam low; (m) London high, Verona low; (n) Verona high, Amsterdam low.

jo00

1

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

Table 5. Unadjusted and adjusted percentages of dissatisfied patients in the VSSS-EU dimensionsacross EPSILON sites (score < 3.5)

VSSS-EU dimension

Overall satisfaction

UnadjustedAdjusted

Professionals' skills and behavior

UnadjustedAdjusted

Information

UnadjustedAdjusted

Access

UnadjustedAdjusted

Efficacy

UnadjustedAdjusted

Type of intervention

UnadjustedAdjusted

Relatives' involvement

Unadjusted

Adjusted

Total mean score

UnadjustedAdjusted

Amsterdam(n = 58),%

29.828.2

22.417.4

36.832.1

27.630.0

39.733.7

33.327.1

38.8

31.4

29.320.5

Copenhagen(n = 51), %

26.016.2

15.711.2

39.230.2

7.86.6

35.328.9

27.417.9

48.9

44.9

19.611.2

London(/? = 83), %

42.243.2

47.040.7

61.463.8

16.917.4

72.377.4

72.376.8

75.0

78.7

60.260.7

Santander(n = 100), %

30.029.4

15.016.2

61.064.4

24.021.6

50.050.2

53.053.9

41.7

44.3

39.042.9

Verona(n = 107),%

20.620.5

18.721.0

36.445.8

33.636.3

28.033.0

22.431.0

34.8

46.5

21.525.9

in levels of service satisfaction. The data obtained in thisstudy confirm that service satisfaction, measured usingstandardized and comprehensive scales, can be a powerfuland specific tool to help clinicians and planners makeinformed decisions about how to improve local serviceprovision when taking direct account of the viewsexpressed by schizophrenia patients.

Moreover, this study has been able to identify whatpatients consider to be unsatisfactory mental health serviceperformance that might occur commonly in psychiatricservices across Europe. Specifically, involvement of rela-tives in the process of care and information about illnessare the satisfaction domains where the mental health ser-vices in most sites show the worst performance. Moreattention may need to be paid to these issues when plan-ning and providing services for people with schizophrenia.

With regard to relatives' involvement, a recent studyreported that patients consider relatives' involvement inthe process of care as one of the most important dimen-sions when evaluating service satisfaction and that themajority of patients are in favor of sharing information

between relatives and professionals (Perreault et al. 1999).With the emphasis on community care, relatives are takingmore responsibility for caring for schizophrenia patients.While this process demands a closer collaborationbetween mental health professionals and family membersor other informal carers, our results suggest that patientsperceive that relatives are not involved enough in theprocess of care. A considerable body of evidence stressesthe importance of family attitudes on the course of, andrecovery from, schizophrenia and also emphasizes theneed for professional, carer, and patient collaboration toattain the best outcome (Kuipers and Bebbington 1990;Pharoah et al. 2000). For this reason, involving familymembers in the process of care and being prepared to takeaccount of their needs are essential to successful commu-nity care provision.

Low satisfaction scores in the information domainhave been consistently reported in a number of previousstudies (Leavey et al. 1997; Leese et al. 1998; Boardmanet al. 1999; Henderson et al. 1999). The lack of informa-tion given to patients by mental health services is an

237

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Table 6. Block regression analysis.1 Factors associated with total satisfaction

BLOCK 1

London vs. AmsterdamSantander vs. Amsterdam

F^adj

BLOCK 2

AgeEducational level (yrs)Employment status

Sheltered (vs. employed)Retired/unemployed (vs. employed)Housewife (vs. employed)

Living situationOther (vs. alone)With family (vs. alone)

^ Ethnicity& White (vs. other)

Admissions (log n)

F^adj

BLOCK 3

BPRS scoreF?adj

BLOCK 4

CAN-EU unmet1-2 vs. 03-4 vs. 025 vs. 0

F^adj

BLOCK 5

LQL-EU score

Total VSSS-EU

-0.54(<0.001)-0.27 (0.006)0.153

-0.02(0.081)

-0.15(0.048)

-0.22(0.019)0.159 (AO.6%)

-0.38 (0.05)0.211 (A5.2%)

-0.11 (0.086)-0.20 (0.022)-0.22 (0.033)0.213 (AO.2%)

0.22(<0.001)0.28 A6.7%)

Overallsatisfaction

-0.55(<0.001)

0.065

0.53 (0.029)0.24(0.051)

-0.26 (0.089)0.076 (A1.1%)

0.108 (A3.2%)

0.108 (A0.0%)

0.23(<0.001)0.139 (A3.1%)

Professionals'skills andbehavior

-0.54(<0.001)

0.149

-0.024 (0.063)

-0.16(0.077)

0.175 (A2.6%)

0.202 (A2.7%)

0.202(A0.0%)

0.20(<0.001)0.255 (A5.3%)

with services

Information/access

-0.34 (0.018)0.056

-0.031 (0.067)

0.063 (AO.7%)

-0.47(0.091)0.080 (A1.7%)

-0.21 (0.030)-0.26 (0.040)-0.35(0.019)0.088 (AO.8%)

0.28(<0.001)0.148 (A6.0%)

and satisfaction in VSSS-EU

Efficacy

-0.79(<0.001)-0.37(0.012)0.133

0.20 (0.076)

0.124 (A-0.9%)

-0.59 (0.043)0.168 (A4.4%)

-0.22 (0.090)-0.29 (0.064)0.168 (A0.0%)

0.24(<0.001)0.241 (A7.3%)

Type ofintervention

-0.67(<0.001)-0.33(<0.001)0.223

0.01 (0.004)

0.37(0.015)

0.13(0.080)0.17(0.039)

-0.21 (0.027)0.265 (A4.2%)

0.288 (A2.3%)

0.288 (A0.0%)

0.21 (<0.001)0.348 (A6.0%)

dimensions

Involvement ofrelatives

-0.81 (<0.001)

0.085

0.50 (0.087)0.33 (0.035)

0.051 (A-3.4%)

-1.13(0.005)0.099 (A4.8%)

-0.46(0.010)-0.63 (0.004)0.125 (A2.6%)

0.33(<0.001)0.171 (A4.6%)

1

ma i

ca

'.tin.

£

Q

N;

u

c

TO

1

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

$5ri

no

s

isfg£ (S

c e oO (0 ^« « 10

_ o

O •:

HI

si!(A

S3ps.o>od

to s

do

O) CO O ) CMq q q qdo d din asq qd d

(O N O)q q too d d

0 0 CO <N

S.S-OO> i - COq i - c«jo d d

CD COi- sf̂ i t^i ^3 »̂

2. o o w

8 8 2 2d odd

T- O 00o q ooP o oo V *-'

o o o

?O - * ifj°- °- <,O CM CO

o d d

oo

sCO CO O> T -q q q eod o d d

g- c o ,»

it I§ a «

• n OL <0 CD

_L i , _L _L_L i333334

COco

I

8

111

IHI

I

IQ.<D

2(0

25"

CO

+C\J

• S PQ.

I15c

co co

o rCD °

. ! = u)J = CDCO 3

- 1 TO

=> 5sLU i 2 CO

^ 28g.

o o.£2™ <n

o oE o

oC CO

~ COCO i -

8 |2 ^

CD

ilss3 c 5 >111 M ^ - - ~

< s I "5"V 111 CD CD

CO OiJ C

important and complex issue, because the level of infor-mation provided and the kind of communication betweenstaff and patients may have a significant impact on thewhole process of care (Thompson 1993). In a study on car-ing behaviors, it was found that psychiatric staff rated"explanations" as the least important caring behavior (VonEssen and Sjoden 1995). Moreover, it is common in clini-cal practice for professionals not to give all the informa-tion that patients need on their diagnosis and prognosis or,because not all patients explicitly seek advice and infor-mation, to wrongly conclude that no information isrequired at all. Finally, when information is given, it maybe provided in a partial or overly simplistic way.

3. Is Higher Service Satisfaction Associated With aLower Number of Patients' Unmet Needs for Care? Aprevious study conducted in England on an epidemiologi-cally representative sample of patients with psychosisreported that having more unmet needs for care is corre-lated with lower service satisfaction (Leese et al. 1998).Another study has shown that availability of communitybeds results in significant reduction of unmet needs andbetter satisfaction with services (Boardman et al. 1999).In the present study, a significant association has beenfound between unmet needs and satisfaction, with lowersatisfaction when more unmet needs are detected. Thisassociation maintained its significance also after havingtaken into account cross-site variability in the separatedomains of satisfaction. The biggest effect was found inthe domain of patients' satisfaction with the involvementof relatives in care. This might suggest that one of themajor problems that services have to face when caring forschizophrenia patients who are more disabled is findingan appropriate strategy to support their relatives. Our datathrow some light on this complex issue; however, the pre-dictive value of unmet needs on patient satisfactionrequires further investigation.

4. Is Service Satisfaction Related to Patients'Sociodemographic Characteristics and IllnessSeverity? The early literature concerning the relation-ships between patient characteristics and satisfaction wasinconclusive (Ruggeri 1994). The present study confirmsthat demographic characteristics of schizophrenia patientshave a very low impact not only on total satisfaction butalso on satisfaction in the various domains and that thisfinding is likely to have cross-cultural stability. Previousstudies have found only a weak relationship betweenpatients' clinical characteristics, such as psychopathologyor global functioning, and service satisfaction (Ruggeri etal. 1998, 1999; Becker et al. 1999). This study has foundthat higher illness severity, as measured by BPRS score, iscross-culturally associated with lower service satisfaction,especially in the domains assessing involvement of rela-

239

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Schizophrenia Bulletin, Vol. 29, No. 2, 2003 M. Ruggeri et al.

tives and self-perceived efficacy, but the amount of vari-ance explained is relatively low.

5. Is Service Satisfaction Associated With Patients'Satisfaction With Life? A previous 6-month follow-upstudy of the outcome of care for a group of patients with afull spectrum of psychiatric disorders in the Verona com-munity-based psychiatric service used graphical chainmodels and reported that the best predictor of patient'ssatisfaction is subjective quality of life (Ruggeri et al.1998). A more recent Verona study on a larger sample ofpsychiatric patients, including individuals with schizo-phrenic and nonschizophrenic disorders, showed that ser-vice satisfaction is closely related to subjective quality oflife (Ruggeri et al. 2001). This finding has been confirmedin other studies (Rohland et al. 2000; Berghofer et al.2001; Druss et al. 2001) In the present study, among allthe explanatory variables considered, high satisfactionwith life has, after study site, the strongest and most posi-tive association with service satisfaction. Interestingly,domains that are very likely to be affected by severe men-tal illness, such as quality of social relations and qualityof health, are those with the highest impact on satisfactionwith services. All these consistent findings tend to confirmthe statement made by Locker and Dunt (1978), who sug-gested that, particularly in long-term care, "quality of carecan become synonymous with quality of life and satisfac-tion with care an important component of life satisfac-tion" (p. 283). The cross-cultural stability of this associa-tion and its predominance over the other associationstested offer the appealing perspective that improvementsto patients' subjective quality of life can be achieved byproviding adequate and individualized care. However, forthe time being, caution should be taken in interpretingthese findings, as conflicting results have been obtainedon possible confounders. On the one hand, in a previouspaper, we have demonstrated that subjective and objectivequality of life have different latent constructs (Ruggeri etal. 2001). Another study has found that self-rated symp-toms, subjective quality of life, self-rated needs, andpatient's assessment of treatments are all substantiallycorrelated (Priebe et al. 1998). On the other hand, it hasbeen found that mood interferes with ratings of quality oflife but has a small influence on satisfaction with care

(Atkinson and Caldwell 1997). Finally, there is little con-sensus on the factors that influence subjective quality oflife in the general population, and results from nationalsurveys do not provide reliable information on cross-cul-tural differences (Veenhoven 1993). Achieving knowledgeabout these background aspects and implementing follow-up studies on changes occurring over time in both qualityof life and satisfaction with care might greatly contributeto a better understanding of this issue.

Conclusions

This study helps to clarify our understanding of servicesatisfaction and may have important practical implica-tions. We have identified areas of unsatisfactory perform-ance in the care of schizophrenia patients in five Europeansites, and this information may be of use to clinicians andplanners in improving psychiatric services. The cross-sec-tional nature of this study allowed analysis of associationsto identify service satisfaction's explanatory variables.Psychiatric services located in deprived metropolitanareas, such as South London, or in areas where few typesof interventions/facilities are available, such as Santander,are more likely to be considered unsatisfactory by patients.Demographic and clinical characteristics are weakly asso-ciated with service satisfaction; subjective quality of life,and especially satisfaction with health and social relations,has a clear-cut association with all service satisfactiondomains. These associations, once established, can betested in future work, including longitudinal studies, toinvestigate the role they might have in predicting satisfac-tion over time.

Furthermore, the fact that only a medium to low per-centage of variance has been explained by our models sug-gests that there is wide individual variability in the factorsthat determine patient satisfaction. Our data therefore sup-port the view that service satisfaction can be seen as thecombined result of (1) the ability of the service to providea standard of care above a certain quality threshold (e.g.,in professional competence, or the availability of specificinterventions, or the physical characteristics of the treat-ment setting), and (2) the perception of the patient that thecare received has been tailored to his or her own problems.

240

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

6.75.00

1.67.13.0

16.716.411.1

6.424.5

0

010.33.3

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

Appendix. Adjusted percentages of dissatisfied patients in the VSSS-EU items across the 5 EPSILONsites

Amsterdam Copenhagen London Santander VeronaVSSS-EU item (n = 58), % (n = 52),% (n = 85), % (n = 100), % (n = 107),%

OVERALL SATISFACTION

11. Amount of help20. Kind of services offered21. Service general sense

PROFESSIONALS' SKILLS AND BEHAVIOR

3. Professionalism of psychiatrists/psychologists

16. Thoroughness of psychiatrists/psychologists

6. Personal manner of psychiatrists/psychologists

5. Ability of psychiatrists/psychologiststo listen

22. Professional competence of nurses/social workers

35. Thoroughness of nurses/social workers25. Personal manner of nurses/

social workers37. Ability of nurses/social workers to listen28. Nurses' knowledge of patient's medical

history2. Behavior and manners of reception staff

33. Instruction on what to do between visits18. Cooperation between service providers17. Referring to GP or other specialists40. Continuity of care10. Confidentiality and respect for patient's

rights7. Punctuality of the professionals

INFORMATION

12. Explanation procedures andapproaches used

29. Information on diagnosis and prognosis19. Publicity on mental health services

offered 10.0 3.4 18.8 33.6 25.5

ACCESS

4. Appearance, comfort level, andphysical layout

8. Costs of the service

EFFICACY

9. Attaining well-being and preventingrelapses

1. Helping patient deal with problems24. Helping patient improve knowledge

of his problems13. Helping to relieve symptoms

1.0

0.6

0

0

00

2.30

15.21.36.87.203.2

1.617.7

1.4

7.6

6.8

3.5

5.10

1.96.8

3.52.82.73.85.72.6

1.92.8

5.1

1.7

0

7.1

6.01.7

00

8.35.48.95.94.3

28.7

3.825.0

7.5

0

1.9

8.1

2.00

0.50.7

7.02.27.96.7

16.624.0

5.51.9

5.8

6.7

0

6.5

3.00

0.83.5

7.82.3

13.83.6

16.77.3

3.30.7

10.114.7

10.115.0

17.320.9

37.643.7

12.329.7

14.914.9

6.13.4

6.78.3

1.513.8

5.80.9

6.86.1

09.6

8.05.3

20.312.7

32.70

21.610.2

37.312.9

12.319.9

11.515.7

9.414.4

241

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Schizophrenia Bulletin, Vol. 29, No. 2, 2003 M. Ruggeri et al.

7.8

4.1

13.817.1

14.0

5.9

11.718.2

24.4

10.7

34.637.5

22.0

7.8

15.832.7

20.5

6.1

21.927.0

11.4 6.3

Appendix. Adjusted percentages of dissatisfied patients in the VSSS-EU items across the 5 EPSILONsites

Amsterdam Copenhagen London Santander VeronaVSSS-EU item (n = 58), % (n = 52), % (n = 85), % (n = 100), % (n = 107), %

26. Improving relationship betweenpatient and relative

34. Helping to improve capacity to lookafter himself

31. Helping to establish good relationshipsoutside family

38. Helping patient improve abilities to work

TYPE OF INTERVENTION

14. Response to crisis during office hours15. Response to emergencies during nights,

weekends41. Medication prescription39. Help for discomfort of side effects from

medications43. Individual sessions48. Group sessions45. Family sessions42. Individual rehabilitation51. Practical help by the service at home50. Informal admission to hospital44. Compulsory treatment in hospital46. Living in sheltered accommodation47. Recreational activities in the service54. Recreational activities outside the service53. Help to find open employment49. Shelter work52. Helping in obtaining welfare benefits

RELATIVES' INVOLVEMENT

30. Ability of psychiatrists/psychologists tolisten to relative

23. Recommendations about how relativecould help

32. Information to relative about diagnosis andprognosis

27. Helping relative to deal better with patient'sproblems

36. Helping relative improve understanding ofpatient's problems

Note.—GP = general practitioner; VSSS-EU = Verona Service Satisfaction Scale, European version.

29.39.5

7.115.212.717.619.68.54.99.112.414.212.511.227.117.1

5.41.3

9.130.811.024.110.814.26.56.04.87.18.413.29.024.9

60.922.9

27.912.611.717.58.812.221.052.97.222.521.913.78.730.2

3.415.3

34.219.440.614.024.231.19.25.418.432.027.664.748.853.5

12.922.7

18.28.811.49.021.119.58.94.64.013.19.024.622.922.7

4.7

14.1

11.9

14.1

10.1

13.6

21.5

23.8

18.7

25.4

15.3

64.3

25.1

24.7

25.6

14.2

21.8

41.0

12.9

21.7

15.6

13.2

21.3

10.9

24.0

242

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

References

American Psychiatric Association. DSM-IV: Diagnosticand Statistical Manual of Mental Disorders. 4th ed.Washington, DC: APA, 1994.

Andrews, G.; Issakidis, C ; and Carter, G. Shortfall inmental health service utilization. British Journal ofPsychiatry, 179:417-425,2001.

Atkinson, M.J., and Caldwell, L. The differential effectsof mood on patients' ratings of life quality and satisfactionwith their care. Journal of Affective Disorders,44:169-175, 1997.

Audini, B.; Marks, I.M.; Lawrence, R.E.; Connoly, J.; andWatts, V. Home-based versus out-patients/inpatients carefor people with serious mental illness. British Journal ofPsychiatry, 165:204-210, 1994.

Awad, A.G.; Hogan, T.P.; Voruganti, L.N.P.; andHeslegrave, R.J. Patients' subjective experiences on antipsy-chotic medications: Implications for outcome and quality oflife. International Clinical Psychopharmacology, 10(Suppl3):123-133,1995.

Awad, A.G., and Voruganti, L.N.P. Intervention researchin psychosis: Issues related to the assessment of quality oflife. Schizophrenia Bulletin, 26(3):557-564, 2000.

Barker, D.A.; Shergill, S.S.; Higginson, I.; and Orrell,M.W. Patients' views toward care received from psychia-trists. British Journal of Psychiatry, 168:641-646, 1996.

Becker, T.; Knapp, M.; Knudsen, H.C.; Schene, A.H.;Tansella, M.; Thornicroft, G.; Vazquez-Barquero, J.L.;and the EPSILON Study Group. The EPSILON study ofschizophrenia in five European countries: Design andmethodology for standardising outcome measures andcomparing patterns of care and service costs. BritishJournal of Psychiatry, 175:514-521, 1999.

Berghofer, G.; Lang, A.; Henkel, H.; Schmidl, R; Rudas,S.; and Schmidtz, M. Satisfaction of inpatients and outpa-tients with staff, environment, and other patients.Psychiatric Services, 52:104-106, 2001.

Boardman, A.P.; Hodgson, R.E.; Lewis, M.; and Allen, K.North Staffordshire Community Beds Study: Longitudinalevaluation of in-patient units attached to community men-tal health centres: I. Methods, outcome and patient satis-faction. British Journal of Psychiatry, 175:70-78, 1999.

Dean, C ; Phillips, J.; Gadd, E.M.; Joseph, M.; andEngland, S. Comparison of community based service withhospital based service for people with acute, severe psy-chiatric illness. British Medical Journal, 307:473-476,1993.

Dowrick, C ; Casey, P.; Dalgard, O.; Hosman, C ;Lehtinen, V.; Vasquez-Barquero, J.L.; and Wilkinson, G.Outcomes of Depression International Network (ODIN).

Background, methods and field trials. British Journal ofPsychiatry, 172:359-363, 1998.

Druss, B.G.; Schlesinger, M.; and Hallen, H.M.Depressive symptoms, satisfaction with health care, and2-year work outcomes in an employed population.American Journal of Psychiatry, 5:731-734, 2001.

Elbeck, M., and Fecteau, G. Improving the validity ofmeasures of patient satisfaction with psychiatric care andtreatment. Hospital and Community Psychiatry,41:998-1001, 1990.

Gaite, L.; Vazquez-Barquero, J.L.; Arriaga Arrizabalaga,A.; Schene, A.; Welcher, B.; Thornicroft, G.; Rugged,M.; Vazquez-Bourgon, E.; Perez Retuerto, M.; Leese,M.; and the EPSILON Study Group. Quality of life inschizophrenia: Development, reliability and internal con-sistency of the Lancashire Quality of Life Profile—European version. British Journal of Psychiatry,177(Suppl 39):49-57, 2000.

Greenwood, N.; Key, A.; Burns, T.; Bristow, M.; andSedgwick, P. Satisfaction with in-patient psychiatric ser-vices. Relationship to patient and treatment factors.British Journal of Psychiatry, 174:159-163, 1999.

Hansson, L. Patient satisfaction with in-hospital psychi-atric care. A study of 1-year population of patients hospi-talised in a sectorised care organisation. EuropeanArchives of Psychiatry and Neurological Sciences,239:93-100, 1989.

Henderson, C ; Phelan, M.; Loftus, L.; Dall'Agnola, R.;and Ruggeri, M. Comparison of service satisfactionamong psychotic patients in South London and SouthVerona. Acta Psychiatrica Scandinavica, 97:1-8, 1999.

Hoult, J., and Reynolds, I. Schizophrenia. A comparativetrial of community orientated and hospital orientated psy-chiatric care. Acta Psychiatrica Scandinavica,69:359-372, 1984.

Johnson, S.; Ramsay, R.; Thornicroft, G.; Brooks, E.;Lelliot, P.; Peck, E.; Smith, H.; Chisholm, D.; Audini, B.;Knapp, M.; and Goldberg, D. London's Mental Health.London, U.K.: King's Fund, 1997.

Kelstrup, A.; Lund, K.; Lauridsen, B.; and Bech, P.Satisfaction with care reported by psychiatric inpatients.Acta Psychiatrica Scandinavica, 87:374-379, 1993.

Knudsen, H.; Vazquez-Barquero, J.; Welcher, B.; Gaite,L.; Becker, T; Chisholm, D.; Ruggeri, M.; Schene, A.;Thornicroft, G.; and the Epsilon Study Group. Translationand cross-cultural adaptation of outcome measurementsfor schizophrenia: EPSILON Study 2. British Journal ofPsychiatry, 176(Suppl 39):S8-S14, 2000.

Kuipers, L., and Bebbington, P. Working in Partnership:Clinicians and Carers in the Management of Long-stand-ing Mental Illness. Oxford, U.K.: Heineman, 1990.

243

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Schizophrenia Bulletin, Vol. 29, No. 2, 2003 M. Rugged et al.

Larsen, D.L.; Attkisson, C.C.; Hargreaves, W.A.; andNguyen, T.D. Assessment of client/patient satisfaction:Development of a general scale. Evaluation and ProgramPlanning, 2:197-207, 1979.

Leavey, G.; King, M.; Cole, E.; Hoar, A.; and Johnson-Sabine, E. First-onset psychotic illness: Patients' and rela-tives' satisfaction with services. British Journal ofPsychiatry, 170:53-57, 1997.

Leese, M.; Johnson, S.; Slade, M.; Parkman, S.; Kelly, F.;Phelan, M.; and Thornicroft, G. User perspective on needsand satisfaction with mental health services. BritishJournal of Psychiatry, 173:409-415, 1998.

Locker, D., and Dunt, D. Theoretical and methodologicalissues in sociological studies of consumer satisfactionwith medical care. Social Science and Medicine,12:283-292, 1978.

Marks, I.M.; Connolly, J.; Muijen, M.; Audini, B.;McNamee, G.; and Lawrence, R.E. Home-based versushospital-based care for people with serious mental illness.British Journal of Psychiatry, 165:179-194, 1994.

McCrone, P.; Leese, M.; Thornicroft, G.; Schene, A.;Knudsen, H.C.K; Vazquez Barquero, J.L.; Lasalvia, A.;Padfield, S.; White, I.R.; Griffiths, G.; and the EPSILONStudy Group. The reliability of the CamberwellAssessment of Need—European version. British Journalof Psychiatry, 177(Suppl 39):34-40, 2000.

Merson, S.; Tyrer, P.; Onyett, S.; Lack, S.; Birkett, P.;Lynch, S.; and Johnson, T. Early intervention in psychi-atric emergencies: A controlled clinical trial. Lancet,339:1311-1314, 1992.

Naber, D.; Walther, A.; Kirshee, T.; Hayek, D.; andHolzback, R. Subjective effects of neuroleptics predictcompliance. In: Gabel, W., and Awad, A.G., eds.Prediction of Neuroleptics ' Treatment Outcome inSchizophrenia: Concepts and Methods. Vienna, Austria:Springer-Verlag, 1994. pp. 469-473.

Perreault, M.; Paquin, G.; Kennedy, S.; Desmarais, J.; andTardif, H. Patients' perspective on their relatives' involve-ment in the treatment during a short-term psychiatric hos-pitalization. Social Psychiatry and PsychiatricEpidemiology, 34:157-165, 1999.

Perreault, M.; Rogers, W.L.; Leichner, P.; and Sabourin,S. Patients' requests and satisfaction with services in anoutpatient psychiatric setting. Psychiatric Services,47:287-292, 1996.

Pharoah, F.M.; Mari, J.J.; and Streiner, D. Family inter-vention for schizophrenia. Cochrane Data BaseSystematic Reviews 2: CDOOOO88, 2000.

Priebe, S.; Kaise, W.; Huxley, P.J.; Roder-Wanner, U.U.;and Rudolf, H. Do different subjective evaluation criteria

reflect distinct constructs? Journal of Nervous and MentalDisease, 186:385-392, 1998.

Rohland, B.M.; Langbehn, D.R.; and Rohrer, J.E.Relationship between service effectiveness and satisfac-tion among persons receiving Medicaid mental health ser-vices. Psychiatric Services, 51:248-250, 2000.

Ruggeri, M. Patients' and relatives' satisfaction with psy-chiatric services: The state of the art of its measurement.Social Psychiatry and Psychiatric Epidemiology,29:212-227, 1994.

Ruggeri, M.; Biggeri, A.; Rucci, P.; and Tansella, M.Multivariate analysis of outcome of mental health careusing graphical chain models. The South-Verona OutcomeProject 1. Psychological Medicine, 28:1421-1431, 1998.

Ruggeri, M.; Gater, R.; Bisoffi, G.; Barbui, C ; andTansella M. Determinants of subjective quality of life inpatients attending community-based mental health ser-vices. The South-Verona Outcome Project 5. ActaPsychiatrica Scandinavica, 105:131-140, 2002.

Ruggeri, M.; Lasalvia, A.; Dall 'Agnola, R.; vanWijngaarden, B.; Knudsen, H.C.; Leese, M.; Gaite, L.;Tansella, M.; and the EPSILON Study Group.Development, internal consistency and reliability of theVerona Service Satisfaction Scale—European Version.British Journal of Psychiatry, 177(Suppl 39):41^8, 2000.

Ruggeri, M.; Santolini, N.; Stegagno, M.; Imperadore, G.;and Dall'Agnola, R.B. La qualita di vita dei pazienti psichi-atrici. Monograph Supplement 4, Epidemiologia e PsichiatriaSociale. Rome, Italy: D Pensiero Scientifico Editore, 1999.

Ruggeri, M.; Warner, R.; Bisoffi, G.; and Fontecedro, L.Subjective and objective dimensions of quality of life inpsychiatric patients: A factor analytical approach. TheSouth-Verona Outcome Project 4. British Journal ofPsychiatry, 178:268-275, 2001.

Simes, R.J. An improved Bonferroni procedure for multi-ple test of significance. Biometrika, 73:751-754, 1986.

SPSS Inc. Statistical Package for Social Sciences:Release 10.0.7. Chicago, IL: SPSS Inc., 2000.

StataCorp. Stata Statistical Software: Release 6.0. CollegeStation, TX: Stata Corporation, 1999.

Svensson, B., and Hansson, L. Patient satisfaction withinpatient psychiatric care. The influence of personalitytraits, diagnosis and perceived coercion. ActaPsychiatrica Scandinavica, 90:379-384, 1994.

Thompson, A. Inpatients' opinion of the quality of acutehospital care: Discrimination as the key to measurementvalidity. In: Fitzpatrick, R., and Hopkins, A. eds.Measurement of Patients' Satisfaction With Their Care.London, U.K.: Royal College of Physicians, 1993. pp.19-72.

244

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

Results From the EPSILON Study Schizophrenia Bulletin, Vol. 29, No. 2, 2003

Veenhoven, R. Happiness in Nations: SubjectiveAppreciation of Life in 56 Nations, 1946-1992. RISBO,Studies in Social and Cultural Transformation no. 2.Rotterdam, The Netherlands: Erasmus University, 1993.

Ventura, J.; Green, M.F.; Shaner, A.; and Liberman, R.P.Training and quality assurance with the Brief PsychiatricRating Scale: "The drift busters." International Journal ofMethods in Psychiatric Research, 3:221-244, 1993.

Vicente, B.; Vielma, M.; Jenner, A.F.; Mezzina, R.; andLliapas, R. Users' satisfaction with mental health services.International Journal of Social Psychiatry, 39:121-130,1993.

Von Essel, L., and Sjoden, P.O. Perceived occurrence andimportance of caring behaviours among patients and staffin psychiatric, medical and surgical care. Journal ofAdvanced Nursing, 21:266-276, 1995.

Voruganti, L.; Heslegrave, R.; Awad, A.G.; and Seeman,V. Quality of life measurement in schizophrenia:Reconciling the quest for subjectivity with the question ofreliability. Psychological Medicine, 28:165-172, 1998.

World Health Organization. The 1CD-10 Classification ofMental and Behavioural Disorders: Clinical Descriptionsand Diagnostic Guidelines. Geneva, Switzerland: WHO,1992a.

World Health Organization. SCAN: Schedules for ClinicalAssessment in Neuropsychiatry. Geneva, Switzerland:WHO, 19926.

Acknowledgments

Many colleagues contributed to the EPSILON Study.Amsterdam: Dr. Maarten Koeter, Karin Meijer, Dr. MarcelMonden, Professor Aart Schene, Madelon Sijsenaar, andBob van Wijngaarden. Copenhagen: Dr. Helle CharlotteKnudsen, Dr. Anni Larsen, Dr. Klaus Martiny, Dr. CarstenSchou, and Dr. Birgitte Welcher. London: ProfessorThomas Becker, Dr. Jennifer Beecham, Liz Brooks,Daniel Chisholm, Gwen Griffiths, Julie Grove, ProfessorMartin Knapp, Dr. Morven Leese, Paul McCrone, SarahPadfield, Professor Graham Thornicroft, and Ian R. White.Santander: Andres Arriaga Arrizabalaga, Sara HerreraCastanedo, Dr. Luis Gaite, Andres Herran, Modesto PerezRetuerto, Professor Jose Luis Vazquez-Barquero, andElena Vazquez Bourgon. Verona: Dr. Francesco Amaddeo,Dr. Giulia Bisoffi, Dr. Doriana Cristofalo, Dr. RosaDall'Agnola, Dr. Antonio Lasalvia, Professor MirellaRuggeri, and Professor Michele Tansella.

This study was supported by the European Commis-sion BIOMED 2 Programme (Contract BMH4-CT95-1151).

We are grateful to Dr. Chiara Bonetto and Dr. Mor-ven Leese for their advice on statistical analyses. Wewould also like to acknowledge the sustained and valu-able assistance of the users, the carers, and the clinicalstaff of the services in the five study sites. In Amster-dam, the EPSILON study was partly supported by agrant from the Nationaal Fonds Geestelijke Volksge-zondheid and a grant from the Netherlands Organizationfor Scientific Research (940-32-007). In Santander, theEpsilon Study was partially supported by the SpanishInstitute of Health FIS (FIS Exp. No. 97/1240). InVerona, additional funding for studying patterns of careand costs of a cohort of schizophrenia patients was pro-vided by the Regione del Veneto, Giunta Regionale,Ricerca Sanitaria Finalizzata, Venezia, Italia (Grant No.723/01/96 to Professor M. Tansella).

The Authors

Mirella Ruggeri, M.D., Ph.D., is Associate Professor ofPsychiatry; Antonio Lasalvia, M.D., Ph.D., is RegistrarPsychiatry; Giulia Bisoffi, D. Stat., is Coordinator ofCore Group Biostatistics; and Michele Tansella, M.D.,is Professor of Psychiatry and Director, Department ofMedicine and Public Health, Section of Psychiatry,University of Verona, Italy. Graham Thornicroft, M.D.,FRCPsych, is Professor of Community Psychiatry andHead of the Health Services Research Department,Institute of Psychiatry, King's College, London, U.K.Jose Luis Vazquez-Barquero, M.D., is Professor ofPsychiatry, Clinical and Social Psychiatric ResearchUnit, Department of Psychiatry, Universi ty ofCantabria, Santander, Spain. Thomas Becker, M.D., isProfessor of Psychiatry, Department of Psychiatry,University of Leipzig, Leipzig, Germany. MartinKnapp, B.A., M.Sc, Ph.D., is Professor of Health Eco-nomics and Director of the Centre for the Economics ofMental Health, Health Services Research Department,Institute of Psychiatry, King's College. Helle CharlotteKnudsen, M.D., is Consultant Psychiatrist, Institute ofPreventive Medicine, Copenhagen University Hospital,Denmark. Aart Schene, M.D., Ph.D., is Professor ofCommunity Mental Health Care, Department of Psy-chiatry, Academic Medical Centre, University of Ams-terdam, The Netherlands.

245

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from

by guest on July 22, 2011schizophreniabulletin.oxfordjournals.org

Dow

nloaded from