1997 - Profiles Of Subjective Quality Of Life In Schizophrenic In- And Out-patients Samples

PSYCHIATRY RESEARCII ELSEVIER

Psychiarry Rcseurch60(l q'r?)15l-166

Profilesof subjectivequaiity of life in schizophrenicin- and out-patientsamples Wolfgarg Kaisera*,StefanPriebe', Wally Barr', Karin Hoffmannb, peter Huxley. MargareteIsermanfla,Ute-Uhike Roder-Wannerb, 'Krafie ta4 Spondau,Oenli.h.r Beteich Cnesikg.Bttul)!, Cn{insdn.27 33,D.1358s Bertin, cmn! zbteillaEltu Sozialpstchiah., Feie UniEE&t Benin, I'lakieMllee 19,D.145A Alnn, cenonr ' Schootof Psrchitt4 .ni BehaLiowatlhiences, The UniN6 ) oJ Man(h4kt, rNod RM4 MtNh6@ M13 opL, UK Rerned I r De*mber loqr: rc\scd 2r luh laco.a(epjrd , O.roner t9oo

For the use of subjectilequalityof life as an evaluaxoDcrnerion,n lhould bc known i{ nmures re retiable.ro wbal extenl they are infloencedby other variablesand rbether diffelence!and similalitiescan be oerecreoacros lreatmentsitxatiors. Quality of life profiles(Berhne! Lrhensgudlitllspro /Lancaqhire ou.tir] of Life profite) of 440 schizopbrenicpatients(from Bedin, Germany and from Wales, UK) w€re eMdined. Reliabihrresdiffered betwecnlife domainsand Sroups.Tle innuenceof orher variableswas moderateand vaned betseen rhe srouDs. Seve.alsisnifcant differencesbetweensubsamplcs could be shovn berweenjn,patiemswith a shoner prese;t si{} and oul-patients.In addition,inrerestingsimilaritiesi. profije pauemsbetweenin-patjentswith a longersray(> 2 yea.sland ouGpatientscanbe fouDd.Beingadrnitkd to a psychiatrichospirals€enslo havean influenceon the tevel and structureof subjectivequaiity of life for sone rirne.Being in a psichiarrjchospiralfor a toDgerrime see6s10 coincidewith a stabilizalionof leveland srrudure of subjeclivequaliryof tife: Fururereportson subjectivequatiryoi life shouldiDciudediagnostically homogeneous sanpling and conr.ol the @ffelationwith psychopatholos/. Further researchis neededto chnfy the impad of olher va;ables(e.e.lengrhofstay, cognitivevarinbles,ireatmenrfealures) on patients'satisfactionin differentsettings. O 1997ElsevierScienceIreland Lrd. AU righrsresened (qrrolds

Communilymental healdrsenices;Evaluationsrudies;I])ng-tlm care;ScbizopbreDia

1 lntroduction Sincethe early 1980stie conceptof subjecrive qu:,ti[ of life ha5 as,umedmore :'nd more rig. "Cor6ponding author. Tel.i +49 (O) 30 37014551,tax: + 4 9 ( 0 1 3 03 t 0 1 3 5 0 5

nificancein both sociolog/and socialpsycholos/ (Stracket al., 1991;Diener, 1984)dnd in research in communirypsychiarryrohver (l Jd.,Iilgb.. The applicationof the conceptin the field of mental health was origirated by Lebman et al. (1982)in studieswhichexploredthe qualiryof life among chronicallymenially disabledpeople in communiryfacilitiesin l-os Angele. and in a

01651731/9tl$r7.00 O 199t ElsevierScienceIreland Lrd. All righrsreseFed P . 1 .Sf 0 1 65 - 1 ? 8 I ( 9 6 ) 0 2 9 58 - 7

154

W leuta

ol / Py.hut4 R6drh66

comparativestudywith chronicmenralpatientsin a statehospitaland conmunitytreatmentsettings in Baltimore (Irhman et al., 1986).In his second study l€hman reported that significantmain effecls.benreenfour samples o[ communiDresi. denlsand state hospitalpatientswith mixeddiagnoses(affectjvedisorder,6-257,; schizophrema, 54-86%; mentallyfetarded,0 197.),on ihe subjectnerariable.of his Oualiryof Life lnteFJie$. could no longer be found after adjushents to rake accounrof rllnessand demogrrphic!ariables.Other studies(Simpsonet al., 1989;Warner and Huxley, 1993) showed few differences betweencommuniryseltings,but lowerscoresm a fe\r domains for long-rermin-parient..There is someevidencethat differencesin communitvsettjngs could be a result of different interventions, like different componentsof case management (Hudey and warner, 1992)or educationgroups (Atkinson et al., 1996). For use as an evaluationc terion, though, it should be known whether subjectivemeasures difier betweendiagno.ric-all! consrstenlgroupsin difierent settingsand to what extent subjective quality of life is influenced by other variables. The presentstud-\7 set out to examinethe following queslions in six diagnosticaliyhomogeneous samplesof schiophrenjc in' and out-pa, lients1. Can subjectivemeasuresof quaiity of life be assumedro be reliable? 2. Is there an impact of deoographicor illness relatedvariabieson subjectivequality of life? 3. Can differencesor similaritiesbe found across different treatmentsituations? 2. M€thods 2.1. Descnption of different settings We comparedfive groupsof schizophrericrall from Berlin, Germany- plus a samplefrom Wales, UK. The German patientshad received ICD l0 dragnoses aF20)duDng lor schuophrenL their hospiial stay fuom psychiatristswho were trainedin the applicalionof ICD 10 criteria.The Welsb palients had a clinical diagnosis of

(t@7) ItJ 1o6

schizophrenia bascdon ICD 10 criteria-Mentally retardedpatientsor patientswith additionalneurologicaldisorden or substanceabusewere excluded. SampleA are patients(n : 68) during the first 3 weeks of their first admissionto 7 differen! psychiatrichospitalsor departmentsof psychiatry at gcneralhospitalsin Berlin dnd Potsdam,Germany.They were consecutively admittedbetween May and December1994.SampleB - patients (n = 76) with a prcsenthospitalstayof between6 patients monthsand 2 years- and sampleC (n = 99)with a presenthospitalstayof more than 2 years were examjnedin a studyof deinstitutionaljsation,which includedall patientsof th.ee dislrict.ol (tormerWest)Berlin$ ho $ere hospitalizedduring the fint half of 1994and had a stay in hospitalof more than 6 months(Priebeet al., 1996).SampleD were patients (n:64) of the oulpatientclinic of NeFenklinik Spandau'(and former in-patientsof that hospital).Patientsare referred ro out-patient-clinics(Institursambulan, zen)becauseof the lengthofpreviousstaysin the hospital andlor needsfor additionalsocial support in addition to pharmacological/psychiatric treatment.SampleE are out-patients(r = 79) of the Departmenlof SocialPsychiatryat the Freie Uni!e,siril Berlin. which is a comprehensive community care system serving the district of Charlottenburg(in the centre of Berlin). It includes three partial hospitalizationprograms, community-based seryiccsaDdvadousout-patient facilities.Continuityof careis guaranteedby case managers. SampleF consistsof out-patients(, = 51) treatedbythe Rlyl CommunityMentalHealth Team, which offers social srpport and pharmacologicalrreahenl for ahout 350 patienrsin a semi-ruralarea on lhe Nonh Walescoait in the UK. The team cariry for all of these patients contaitu'ucial sorkersand nurses.and paLients are refered for the samereasonsas sampleD.

'During $c rine oflhc data collcclio. NepenHinik Spandauwasa psychiatrichospnal.Since01/01/95 il wasadnDs batively cha.ged inlo a depanoenr oi a ecneral hspnal: kankenhaus Spandau,OertlrcherBereichGriesingeslraB..

w K'iJu a al. / Pqrhiar,, R6@rch 66 4997) 15j- 166

For the comparisonof the similairy of the profile pattems (seeSection3 and Table 6) 6ndings ftom a self-administered applicationof the Berliner Irbensqualit.itsprofil to medical srudents ai the Freie UniversitetBerlin (n = 175; mean age+S.D.: 26.013.6;women:45%) are also sho*n. 2.2. Sampb charucteri.stics: demography ln-patientswith a presentstayover 2 years(C) and out-patient-clinicparients(D) are the oldest (nlean age+ S.D.: 53.4:t 14.0and 49.0i 11.0* C > A,B,E,F; D > A.E). Fint adrnissions (A) are significantly younger (meant S.D.: 30.2a 10.3)than all other gruups; the orhers (B,E,F) are in between- range of means:42.4-43.4(differencesin age:dI = 5, F = 31.2, P < 0.001.- Bonferroni ad.iusted multiple meanj comparisons, at least P < 0.05). SampleA hasa predominance ofwomen(71%), and sampleF of men (68%), the other samples range ftom 417, ro 47V. women (differencesm gender:df:5, x' :21.8, P < 0.001). A number of oufpatients of the psychiatric hospital (D) live in group homes or supervhed aparrnents (22/o). A tlird of rhe Welsh sample (F) lives in hostelsbut, on the other hand, 16% own a hous€.For the in-patientsamples(excluding first adrnissions) and the Welsb samplemore peopleare single (B.C.F:bl o6a, rs. A,D,E: 43-4'lVo)(df:5, x' : 14.6,P < 0.05). 2.3. Samplecharacbi:tics: i rrcssrclated uariables Table I (meansexcepl neurolepric rrearmetrl. standarddeviationsare not sholm for the sakeof clarityofpresentationof the table)showsrhar the durationof illnessis longestfor rhe in-patientsof sampleC and out-patjentsof the psychiatrichospitai (D) Gonf€roni adjusted mulripie me.ln comparisons,at leasl P < 0.05; letters together with meansindicatesignificantlydifferinggroups). Tle cumularivenumber ol ho5pirahzJlions is greater for tbe in-patient samplesB and C and rhe cumulati\e durarion oi hospitalzalionis greatestfor those with the longer present stay (C). A5 well. pslchoparhological disrurbrnceis

155

significandy greater for in-patients (A-C) and particulariylower jD oulpatient sampleF. Most of the patientsin all the samplesare oD neuroleptics.Information about doseswai available only for the ceman samples (A E). They are significantly higher for in-patients wjth longer present hospital stays (B,C). Relatively few patierts (<5%) receive additional continuous medication with antiparkinsonian, anti-depressanl,anti-adiety d gs or lithium or carbamazepine. 2.4. lnxtrumenE and data collcction Quality of life was examined using Oliver's (Oiiver, 1991,1992;Oliver et al., in press)l-ancashire Quality of Life Profile Gample F) and its German ve.sion, the Berliner Lebensqualitatsprofil(Priebe et al., 195) (samplesA througbE). Both versionscontain identicalitems and domainsfor direct comparison. Tle l-ancashrreOuality ol LiIe Profile is an adaptationi.r the United Kingdom of Lehrnan's original work and €onsistsof demographicand objectivedata dnd subjectivesatisfactionscalesas well. For the assessment of subjectivequality of life the inlewjews use seven-pointsatisfaction scales(from 'terdble' to 'deiighted'in the original version)and resultscan be calculatedfor different areasof quality of life. Psychopathology was rated using the BPRS (Overall and Gorbam, 1962).Information about previoushospitalizationand diagnoses were takeD from medical records. Neuroleptic doses were transformedinto chlorpromazineequivalentsacco.ding lo Jahn and Mussgay(1989). The quality of life assessments were admm$" tered togetherwith the psychoparhological raring (for some patients in two sessions).Ttey were don€ by psychiatrists(samplesA and E), clinical (samplesB, C and D) and a trained psychologists socialworker (sampleF). The in-patientsdid not know the interviewers,who were not involvedin lheir presenttreatment,from personalcontacts. Tbe raten of all of the out-patientsampleswere members of the treatment teams - althougb again not directly involvedin the presenttreaC ment of mos! cases.

W Koistet ol. /Psr.hiatry Revorch 66 (197) 153 166

e

5

S :

;

EE E;;

f . i

:l

!-

6

E :

.:'E sr

^

d>F3 l h 1 : l ! ] E A ;,i

-

|

lA:3s

;

*l)i

* a..cE o:d5

e EY 1 e 9 .

a ,

4 n =

-

c ; ; i ! .

: E o

! ! E -e- E !

3

!cj:;s;:;

w tui:{ etal / Pt thiatq R6ea4h 6611997)t53 166

2.5. Statisticalanalyses Relinbilig estimateswere calculatedusing a slightly transformed formula giver by Horst (1954).Additionaily Cronbach'salpha was calculaled for a profile set consisting of the items/scalesfor: general well-being,leisure, 6nances.living situation.satet!.socialrehtions. mentalhealth (work and family excludedfor reasonsmentionedlater). A principalcomponentsfactor analysiswasundcrtaken to test the homogencityof the profile and the varianceof the differen! sub-areas. Multiple regressions were calculatedto deteci joint vadance between quality of life variables andlor other significanily differing variables betweenthe groups. ANOVds (first factor: 'setting',secondfactor: 'psychopathology') were carried out to show differencesbetweenthe groups.Resultswere Bonas were all other statislics felroni adjusted Coeflicientswere also cal€ulatedfor the reliabiliry of the resulringprofile pattcms for each group and for the similaritybetweenthe profiles usinga derivativeftom Lienert (1969)of Cattell's /r lcartell. lqaq) as a coefficrcntror srmil"riry betweenprofile patterns. 3. Results 3.1. Interuiews ln sampleA (first admissioN)only one of the patientswho were able to be inte iewedrefused. For the other in-patientssamples(B and C) 19% of the original217interviewscouldnot be carried o8 (12% of the palients were psychopathologi cally too distu.bedand 7% refused1()be inlerviewed).Fof the oulpatienlsamples the Dumber of patients\'/ho could not be inteNiewedbecause of severepsychopatholos/was relativelysmauer (D, 3; E, 8; F, 0). Resultsof the satisfactionscoresfor the areas of 'work' ard 'family' could be obtainedonly for ponroosoi the srmples.becausemany palicnls '70%; D, were uremployed:(A, 41Vc;B, 83o/c:C, 860/o;E, 64Vo;F, 827,) or had no contact wilh

(A, 6%; B,28%, C,35Ea;D,11%; their families: E, 20%; F,6%).As a result,for somemultivariate 'work' and 'family' had to be statisticalanai)ses excludedor calculatedseparately. 3.2. Reliabilities Table 2 shows coefficicnts for an estimate of 'instrumental reliability' (Hors! 1954; Lienert, 1969).For the whole samplethey were sufiicienl for group comparisonsalthough valucs in some singlescalesfor some subsamplesare somewhat toolow(<0.7). Tbe valuesfor Cronbach'salpha,which implies the precondilionlhai quality of life is a consistent the consistencyand conceptand which a.csesses homogfnciqof rhe qudlin ot lile profile.dgain are sufficienlfor the whole samplebut somewhat problematicfor samplesA and B. This correspondslo the results of a principal componentsfactorialanalysis(seeTable 3) which supportsfor all patients acrossall settingsthe assumptionofhomogeneityof the quality of life profile. A single factor solutioD accouDtedfor 301 uf lhe varianceot this'generul5JlisfJcrion factor'. A first factor (F1) accountingfor even ior more variance(42-,{67,)car b€ deinonstrated samplesC-F,whereassecondfaclors(F2) in these samples(% of variance:1517) are characterized by diffcrent items.Again, samplesA and B show less homogeneous3-faclor solutionswith a reduced varianceof thc {irst factor ot 30Eo and 3l%, respectiveb'. 3.3. Relationrbetueenqualiry of lile and demogrophicor ilhess related Lvnables In all the samples.there was no signjnc.ant rr.ociarionberqeen.e\ and'ubjecliv( 5ali\laction in rhc difterent domains of qualiry of life (pointbiscrial correlations).Nor *as there any significantassociationwiih levels of neuroleptic nedication(chlorpromazineequivalenis). fable 4 repo'l' rhe re.ulL\oi sepJlrt( re$e\_ sion analysesincludingvariableswith sig ficant correlations(giving only the amountsof shared This varianceabove l\q. for tbe subsamples). !arihehveen olher sho\r. thar the i6$ci:,lion

V Kaijt et ot. / Ptchbtrtt ReteaEh 66 (199?)153-166

a 5

t'E

i 6Ss

qif

t: €6 l i !

|

l af :

cqr:.9\r-9v).:.:

!a E 1 i E ;

o i o o o o o o i

!ir-.909\iiqq.:\

U E

3 d

c!9eoc.:.:occ99

n

F l1

a.

4E844€6$q ' :

:

.

6 :

" - 1 € : :

-E is =€

iEEia:;i€

6iEisrtid

W Kaue. et al. / Pslctuah! Researh 66 497) 153 166

o

o o o c i i 6

n 6

o o o

i o

. i €

|: v.i

9

ci

i-i o

E l r E i*r

";E€F!, o

:9

; -=a lr !{ sl

;

' 3 l 5 , E Eai e E :

l .

i : i r i

r:

vt 9.-

o

i

e

a E ii€5 3

.i a :E= -

- - s " p ;i 3 i ! !

F E

s E i E : 93;i!F

lv Ktit..a ol / Psrchiot| Rdearh 66 (1997)153 166

$ss*ssEs* 2

i 6

.9

E!

|

E

.

Afs

I

6a! 5 g5 e

;

E 9.F

.9 E ^ .9 E

<€

E l

!

+z A

.

-

:;i.ifF:-:; 3i;i;;:;E

W Kain et aL./ Ptchia'.r R.vorch 6{ (1997) 153- I66

16r

3.a. Qusliry of tifeprofles - dilferent:esand ables(age,durationof iliness,cumulativenumber and duration of hospitalizatjonand BPRS-totalscorerand rhe qudliryof lite profiledomainsi! moderat€.For the whole sampleit does not ex'rnentalhealth'.The BPRS The results of ANOVAS of the salisfaction ceed 107. exceptfor ot generdlwell-beinglnd eighl domains scores total scoreis alwaysnegativelycorrelatedwith lhe variables,the setting as the first dependent as subjcctivequality of life in differert domains.For (BPRS total score, and psychopathologr' factor the whole sampleadditionallyage and the nummedian:40) as the secondfactor did not showany ber of hospitaliations arc positivelycorrelated signilicantinteractionbetwe€nboth facton in any ('safely' and 'work') and the duratioDof iliness domain. 'safety'. showsa negativecorrelationwilh Significantdifierencesbelweenthe groups/setFor thc subsamplesthe pattern of association 'generalwell-being'and in tjngs can be found in 'mental 'living varies markedly by setting and quality of life situation', 'safety', the areas of domain.The BPRS total scoreaccounlsfor most h(rlrh .rnd work ,seeTJhle5 givingBonferroni of thc negativeconclations with qualjty of lifc adjustedP valuesand Bonferroniadjustcdmultrvariables.The cumulativenumber and duration at leasl P < 0.05;lette^ ple mean comparisoDs, of hospitalization,duration oI illness and age togethcr with meansindicatesignjficantlydiffershow severalsignificantcorrelationswhich differ mg groupsrjn their directionsand arensbetweenthe groups. Compaisons betwecn the six gloups show As it might be confusingto presentthe values within thc ouapatientsno significantdifferences tor atl 35 significantp valueswe refer to those and within the in patients three significantdjffor equivalert multiple regressionanalyseswith ferencesbut a number of signiicanl differences 'generalsalisfactionfacthe factor scoresof the (total: i4) between in'patient and our-patient tor' (acrossall subjects)as dependentvariables. groups:for the firsi admissions(A): 8, for tn-paFor the total samplethe resultswere:n':0.09, tientswith a presenlstaybetween6 monthsand 2 F: 10.5,P < 0.m1;BPRStotalscore:b = -0 25, years (B): 5; and for those with a stay of more P < 0.001;age: b :0.12, P < 0.05; total number than 2 years(C): L of hospitalizations:b: -0.12, P < 0.05 (b= The mean.coreslor dre o groupsin domain( standardpartial regressioncoefficients). of the quality of life intcwiewsare givenas proIn separatc analysesfor the subsamplesfor fil€ patiernsof scoresin Fig. l. Cattell (1949)had (P data sampleA no vaiable reachedthe limirs < 0-05) defineda patlem or profil€ of psychological lo enter the analysis.For sampleB the exPlained as '... a systemof measurableparts in a whole. variarceis 17% (R' :0.17, F=11.9, P<0.01) The persislent relations which tie parts into a rnd rhe cumulJrivedurarionof hospiraliTarions whole and distinguishil from oiher wholes.botb cho\..r i,gnrficanr uni!ariare currclalioo neg:,lrve in fact and in human perception,range flom (b- -0.42, P<0-01). The valucsfor tlle other simple quantitativeand spatiai relations to ihe mosl complexfunctional, causai,evideniial,and relations'(p. 280.). psychological (For the representationof psychologicaltest. sanpleC: R1 : 0.16,F - 5.68,P < 0.01;BPRS it hasbecomeconventio_ or questionnaire'resuits tolui score:b = -0.11, P 0.01:(umularire nal 1()join lhe values by a 1ine, although this time in the hospital:, - 0.26,P < 0.05; tends to give a misleadingimpressionof move_ sampleD: R? = 0.09,lr = 6.36,P < 0.05;BPRS ment betweenthe points.) total score:b: -0.31, P < 0.05; The reiiabilitiesof the profileswere:A,0 79; B, ramplcL: R' 0.0o.F = 5.3-.P. 0.05:BPRS 0.80;C, 0.64;D, 0.54iE, 0.60;F, 0.50 I' order to tolai score:b: -0.29, P < 0.05; be inte.pretableihe reliability should be >05 sampleF: R: = 0.09,F= 4.11,P< 0.05;BPRS (Lienert, 1969)which is reachedby all profiles tolal score:b = -0.30, P <0.05.

162

,f. Kaiw et ot. / Psychiaiy tuseath 66 4997) 153 166

v z z v v z v v z +o.l€r.i€ci-

n o?eqevt

9v? i

i F

.E;a lnP i

i dF :

l 5

63i

rivivi6qiqi+uivi

a Ei

vtq9r:..t.9v]c99

2 l F i

a E

=

E i f

qq-q?n-'1c-q

q

"

-

r

;

::

s::iS;eiF

w Ktis* d "1./ PrdutD

R5ut.h Ck tlo97) ti3-to6

163

l ffi4Esl

: ; 1

Fig. l. Proile parlcms.

Sinilanlies bet*een proile pattehs (Cauelit ,p, sinilariies B Fint addisions l.-palicntj <2ycaft

B l.-patients<2years

c

In-paiienrs>2yean D Out-parienl
> 0.85 ae printe.t in botd g?c)

c D tn-patienll Out-patient- Depannenr of Scjal Cjmnunirv Menlat pEvchiarry >2yeE clinic BealrhTe@

0.93 0.85

0,92

0.65

0.67

0.i7

0.80

0.9f,

0.92

l).s9

o.92

0.96

0.83

0,93

0.63

0.66

0.79

0.84

0.8?

0.59

W.Koiss et ul. / Pslchbtry Rdeorch 66 (1997)153-166

Table 6 showsthe overail similaritiesof the prolilc-pattens accordingto our .e-classification (values > 0.85 prinled in bold twe). In-patients wirh a longer stal fC) and the pijtrent'irom Wales (F) share most simil:rity with all other profile patterns.The out-patientclinic patients (D) are the most differert ftom all oiher prr-rfrl<-paurrns. Frr!r admtsions{Ai and in-patientswith a shortersiay (B) are rather similar to the other in-patient sanples and to the Welsh pulienls nol lo the cerman uul-pxlienl5amples. Table 6 also showsfor comparativepurposes rhe sirnilaritiesto a profile pattem of a group of medicalstudents(rneans:generalwelfbeing, 5.3; leisure, 5.7; finances,4.5; living situation, 4.8; safcry, 5-2; social relations,5.5; mental health, 5.5;work, 4.8;famil,v,5.2).The studentsapproach greater similarity with the German ouGpatients and againwith the in-patienrsof sampleC (> 2 years'presentstayin the hospjtal). 4. Discussion

toadingsfo. each group and with a third of the first factor percentof variance.In samplesA and B. lhe percenlol vsflrnceof the 6rrl faclol is lacremarkublyreduceddnd t$o idioryncraric tors' can be found. Even if a one-factorsolution tesults it may be pr€mature to make a decisionfor or againsta homogeneous,single fxclor/dimension solution for the constructionofmeasuresfor qualityof Iife profilesfor the mentallyill. There is an addilional need to cla{ry the concepttheoreticallyand to examineother innuenceson it. Being admitted for the first time or afler a relapse to a pslrchiatri€ hospitalseemsro be of significantimportancefor the structureof subjectivequality of life for some time. This cannotbe explainedonly by the sever_ it_vof the illnessas samplesA and B do not differ from the lory-tem in-patients(C) with respectto the BPRS total scores.Other possibleinfluences re.C.p(r\ondliDldcto15. cugnitire!lrirl'ler. arti tudes,life events)havenot been d€tecledfor the mentallyill so far. 4.2. Relatian of tubje.tiDe qualitt of llfe to atfut

1.1. Methodalogicalimpliations of fndings The meaDrelirbiliq csrimates in differ(ntdc mdrns ol the &'l,ncr Leb<nsquJllSr.profil/ kncashire Qu:llity of Life Profilewere srfficient for comparisons beiween groups (although they remain somewhatproblematjcin the areas of leisure,socialrclalions,menlal heallh aDdsafety) and differ betwcenthe subsamples. For all of thesevariableswith lower reliabiliiics. exceptmental health, the score is a mean from two questions, each of which is of a rather parienls, high cognitivecomplexity.Schizophrenic variety problems who suffcr from a of cognitive might ha!e stecilicditljculliesrn diFIe'entiaring ihejr answers,when cognitivelycomplexconceplual demandsare made. The valuesfor Cronbach'salphaas well as lhc results of the factor analysesindicate that lhe honogeneiryfor the wholeprofilediffen between the sampleswith a shorrertirnc in the hospital(A aDdB) dnd lhr orher'.Foldmnle. C-F rherei' a dominalionof a 'generalsatisfactionfactor' and a second'idiosyncraticfacror' wilh diffeing item

has lhe greule\rcorrelalron P.)choparholog) with all life domainsand witb the generalqualityoHife faclor. The associalionbetweenother devariablesand qualit-v mographicorillness-related of life variesbetweengroups.So psychopathology is lhe only robust predictor of subjectivequalily of life acrossall settings. We should stress,though, that these findings are limited to thc ptrtientswho were interviewed. If those not intefliewed had been taken inio accountlhe 'reaf impact of psychopatholog/on quality of life in schizophrenics might have been higher1.3. Dwrences ond similanics Cornparedio Lehmanh (Lehman et al., 1986) earlier findingsour studyshowsthat the opportuniry ro demonstratereal differencesin qualilv of life bctweencornmuniryresidentsand hospitalized psychiatdcpaticnts seems lo be betler if diagnosticallymore homogeneoussamplcs are siudied.

w rrisdn at./ Pr.hi^ry Resorchbb 1191)7) 153-j66

Most of the differences exist between in, and out-patientsand no significanrdifferencescan be demonstratedwithin the out-patienrs.Within the in-patientsthere seemsto be an impact of their lengthofpresentstayon satisfaction ratingswhich could be explained b), udaptation level theoiy, which was one of the 6^t tleories to explain (Srrucket al.. I9alr. lmp'c\sneevisutisfacLion dencefor that theorycamefrom a studyby Brickmann ct al. (1978) with peoptc, who haye had 'eriouc itcciJenc anJ becomc paraplegic or quadriplcgic:after a time they becamenearly as satisfiedas the geDeralpopulalion,though never quite as satjs{ied. With regardto the sirnilarityofprofile patrerns, the results show tlat rhe 'baselinesimiladty' is rather high comparedto orher psychological data and thereforewe had to raise the criterion for a higher degreeof similarity > 0.85.The resulrsof lhe anallciiof profrleparernsshowrhrr rhereis a slatisticalmethodavailablcwhich producesunderstandableresults. We believe that the techniquecouldbe cxtendedro clinicalpractjcewhere proGlesof groupsand of singlecases(reiatedto group mcansor as pre- and post test results)can be comparedin a way wbich avoidsrelianceupon intuition or subjectiveinrerpretation of differThe profiie-pa$ernanalysisshowsas wetj rhar in-patientswith a longer sray in rhe hospitat(C, > 2 years) seem reasonablycontenl with rheir overalliife situationand they are more similar to oulpatient-groupsand a non-psychiatricsample ihan the other in,patientswho havea shorterstay (A, B) eventhoughthe psychopathological disturbanceis on tlle samelevelin all in-patientgruups. For the W€lshsampleit appearsto be rhe case that the greater heterogcneiryof the sample (especiallyin respectof their objectiveliving situatron)scemsto be a problem.The combinationof home ownershipand mdny patientsiiving in hostels suggeststhat the srmple is more heterogeneous,and lhis may have influencedthe resulls for iNtance, by includinga grealerproportion oi asymptomaticpatients.It is also possiblethar diff€rencesberweenrhe urban and rural way of life had an influenceon subjecrivequaiiry of life, which has been found to be the case in other

165

studies of the risk of rehospitalization (Angermeyerand Kiihn, 1986). 5. Conclusions

1. Beingadmittedto a psychiarrichospitdlseems to have an influenceon the level and struciure of subjective quality of life for some time. 2. Beinga longerlime in a psychiaLric ho,piral seemsto coinciJeqilh a srabilLarion ol letel and structure of subjecrivequaliry of life (\e\cnheless.rhe speci6ccondirioDs of hopitaliation in different hospitals andlor countdesmight lead 10 differing resutts.) 3. Even though reliabilities are suficienr for mosrsubsamples andvariJbles. turtheicmprrical and theoreticalwork has ro be done to clarifythecunceptanddetec!rhe influence oi other variables.Apart from theoreticalquestions (personalityfactors,etc.)a focusshould be on rhc illenlification ot tearures ol rhefirsr $ee[s or monlhsLn the hospiralthat cirnbe ma pulatedfor improvingqualitvof life. Thc ramet true for characrerislics ot out pi,lrenr 4. Psychopatholos/ is lhe only robust predicior tor subl. i\e qualjryot Iie. Bur o\emll lhe impacrseemslo be moderuleand doc, no. atf(ct lhe $hol( conceplof \ubieclive quaLry of life ttnd the patients'satisfactionratings. 5. A irali.licaltechnique wasapphedwhichenableddifferenrprofileprIe'ns lo be comprredrnd o!erall\imilrririe,bErqe(ngroup5 to be revealcd. 6. Future reports on subjectivequality of Ufe should include diagnosticauyhonogeneous samprng. R€ferences Atk,nson,t.A., Coia, D.A., Harpcr Cihour. w' and H.Aer, J.P.(1996)The inpact ol educalionCroupstor peoplewnh schizophredaon socialtun.tioningand qualiivoi life. 3r.r Ps!.hbbr 168, 199-2A1. Angerneler, M.C. and Kiihd, L. {1986) Rehospnal isierungsrisiko pr_vchish l('anker: Stadt versus tjnd. Ner-

W Ka6t 4 ot. / Psrclthtry R6.4tch 66 (tD?) ls] Brickhann, P., Coales, D. and Janof-Bulrnan, R. (1978) tn! t€ry winncs and aeident vicliN: is hapPi.N reladve?./ PeR SocPrtchol 36,911-927. Caltell. R.B. 0949) rp and other oel6cienls ol pallem sots 16nly. Pslchonei! 1,4,279-298. DEner, E. (1984) subjeaive wel-bein8. Pslchol Bul 235, 542 575. Horst, A.P. (1954) Tlre estidation of imnediarc .ctest reliabilir!. Edrc Psr.hol Meos !4. J02-7OA tturey, P. and Wader, R. (1992)Cde m!.a8enent, quality of life, and sarisf6ction vitb senics oi long-tem Psych'a6]Jj.pari. ts. Hosp connuni.r Pgrhitt4 43,199 802. Jahn, T and Musgay, L (1989)Die statisthcheKontrouc nijelich€r Meditabenrcneinnn$o in expeimenralpsychologischen Schtopnrenicsludten. Z IAin PsJchol Pstchoth$ la,257-291. L € h m a n ,A F . , w a ' d . \ . C a n d L n n . L . L r l o 8 2 ) C b r u n ' . oental patienrs: the qualiry ol life issue. ,4n ,/ Prttudry 139,t27t-1216. bhnan, A.F., Possiden&,S and {awker, F- (1986) Tl,e qlality of lile oI chronjcpatientsin a stale hospitaland i. Ps/chiaiy 37, @mnunfty residences. Hap Comujry 901-90?Li.neft, A.A.. (1969J T.nauftau und Testanory. Bela vert^e, Oliver, J.?J. (1991 1992)The socialcare dne.livc: develoP henl of a quality of life pronb for se i. connunity scFicesfor the nentarryill. S& wo* st sci Reu 3,5-45.

166

Oliver, J.!.J.. guney, t.J., Bridges,K a.d Mohmad, H. (.1996) Qualitr ol L,le ond Matut Hcatth snic6. Ro!tOlivcr, J.P.J.,Hutley, P.J.,Pnebe,S. a Xaisr, W. (in Prc$) Measunng the quality o! lile of saerely menlally ill people Ning lhe rr6hn€ Qu.lity of Life Prcfile. S@ Pl} /,idrl Ov€ral, LE. and Gorhan, D.R. (1962)Thc Bnel Psychiairjc Raing Scale. Pq{id n.p 10, 799-812. ftiebe, S., Gruttes, T., Hcitze, M., Hotrnmn, C. and Jikel, A. (1995)SubjektiveElaluatio.skriletien i. der ps]thiatrischenVesorgu.g Erhebungsderhod.nirfFor$hu.g nnd Pruis Prchian PM 22, L40-144. Pnebe,S,, Hofimann,K, lsemann, M. and Kaher W. (1996) Kiinische Merkmale la.Ezeithospiralisiener laio.ten 'aeil I der Berliner Enthospilalisien.Bstudi€. Pqdri4t S i n F U t r . CJ , H v d e . . . F . a n d! a r a g h e rB, ' l o 8 a ) T h e c t u u n i cally denlally il! in comunity facilities. A study of qnal'ly ol lilE. Bt I Pslthittt! \54,71-82. strack F., A.gxle,M. and schwaE,N. (Eds.)(199r)&rrecri'E Wel'geiq. A^ I^kditcipltury P, pectu. Perganon Pts, Warner, R. and Hurley, P. (1993) Psfchopadology a.d quar'ty of life .mong nentally ill paric.ts i. th€ onmunnt. ,rJ Plr.niatry L63, 505-5u9.