British Journal of Psychiatry (1990), 157, 853 859The Social Functioning ScaleThe Development and Validation of a New Scale of Social Adjustmentfor use in Family Intervention Programmes with Schizophrenic PatientsMAX BIRCHWOOD, JO SMITH, RAY COCHRANE, SHEILA WETTON and SONJA COPESTAKESocialfunctioningasanoutcomevariableinfamily interventionswith schizophrenicpatientshas been a relativelyneglectedarea. The requirementsof a scale of social functioningtomeasurethe efficacy of family interventionsinclude:the measurementof skill/behaviourrelevantto the impairmentsandthe demographyof thisgroup;the abilityto yieldconsiderableinformationwith an economyof clinicaltime; and the establishmentof ¿ comparative’n edthroughcomparisonbetweensubscalesandwith appropriatereferencegroups.Resultsfromthreesamplesshowthat the SocialFunctioningScaleisreliable,valid,sensitiveandresponsiveto change.Within the last ten years, significant advances in themanagement of schizophrenia have come fromstudies of psychosocial intervention in the familyenvironment in which the patient resides. Thesestudies have largely focused on relapse as the majoroutcome variable; where social functioning has beenmeasured, the results are equally favourable (Falloonet a!, 1984).This focus on relapse is surprising since impairmentof social functioning is widespread in schizophreniaand may reflect a primary impairment as well as asecondary disability (Bellacketa!, 1990).Deteriorationin interpersonal relationships forms part of thedefining characteristics of the syndrome and withdrawal and impairment in life-role functioning(social/recreation activity, independence/daily livingskills, etc.) are listed as residual symptoms in DSMIII-R (American Psychiatric Association, 1987). Thefamily interventions of Tarrier et a! (1988) andBirchwood & Smith (1987) have directly targeted theraising of social functioning. These interventions require a comprehensive assessment of social functioningin which a direct comparison is made between itsdifferent components, and relative to establishednorms, in order to identify an individual’s strengthsand weaknesses.The methods of assessing social adjustment havebeen reviewed by Weissman (1975, 1981). Sheindicates (1975, p. 1251) that many of the availablescales have limitations for use with chronic disorderssuch as schizophrenia since this group may not befunctioning in the roles that are assessed (e.g. currentwork, marital and parental roles). She advises thatassessment of social functioning in a schizophrenicpopulationshouldthereforeassessmore ¿ fundamental’characteristics, such as level of independence(competence and performance), social engagement!
withdrawal, friendships/interpersonal functioning,and daily activities.One particular limitation of available scales forfamily intervention is that they require a normativejudgement by raters; for example, the SocialBehaviour Schedule of Sturt & Wykes (1986) ratesbehaviour in terms of severity of problem. As Plattet a! (1980) have pointed out, norms vary withcharacteristics such as age, sex, employment status,presence of disability, etc., and may lack externalvalidity. In the case of schizophrenic patients whocontinue to live with their families, an unusual anddifficult normative judgement about social adjustment will be required since this will refer to a ¿ disabled’group that will be largely unemployed,male and single. Furthermore, the judgementsrequired in some scales refer to the presence ofproblems rather than strengths.One further problem of many of the availablescales for family interventions is the use of what canbe a lengthy interview requiring trained raters (e.g.the SBAS; Platt et a!, 1980). Those cliniciansinterested in implementing family interventions in aservice will regard the investment of time requiredas excessive in the context of what is certain tobe a treatment demanding of resources (Smith &Birchwood, 1990).The ideal characteristics of a social adjustmentscale for use with family interventions may thereforebe summarised as follows:(a) comprehensive and permitting comparisonsbetween subscales(b) sensitive to the likely level of impairment andrelevantto the community tenureof this group(c) independent of the normative judgement of ¿ independent’raters853

854 BIRCHWOOD ET AL(d) norms available for a population with a comparable age, sex and (un)employment structure(e) able to yield considerable information with aneconomy of clinical time(f) available in self-report or informant versions(g) conforming to accepted criteria of reliability,validity and sensitivity.The Social Functioning Scale (SFS) was constructedspecifically to tap those areas of functioning that arecrucial to the community maintenance of individualswith schizophrenia. In this respect, the content ofthe SFS was informed by the areas focused on in thesuccessful psychosocial intervention programmes ofHogarty et a! (1979), Paul & Lentz (1977), Stein &Test (1980), and the impairments and disabilitiesassessed by the Disability Assessment Schedule(World Health Organization, 1980). The seven areasselected are shown in Table IV and include:(a) socialengagement/withdrawal (time spent alone,initiation of conversations, social avoidance)(b) interpersonal behaviour (number of friends!heterosexual contact, quality of communication)(c) pro-social activities (engagement in a range ofcommon social activities, e.g. sport)(d) recreation (engagement in a range of commonhobbies, interests, pastimes etc.)(e) independence-competence (ability to performskills necessary for independent living(f) independence-performance (performance ofskills necessary for independent living)(g) employment/occupation (engagement in productive employment or structured programmeof daily activity).The SFS was designed with two requirements inmind. The first was to provide a detailed assessmentof strengths and weaknesses, both to guide anintervention and to provide the clinician with somepossible specific goals, subject to negotiation withthe individual and relatives. Thus it was intended thatthe SFS would establish ¿ comparative’need (Bradshaw,1972)based on comparison with reference groups asdistinct from ¿ felt’or ¿ expressed’ need (whichrequires negotiation/discussion with the individual)or action-based needs assessment (Brewin et a!,1987). The second requirement was the ability tosynthesise such detailed coverage into coherent,reliable continua.Assessing personal and social functioning is notstraightforward. Some measures have assessed rolefunctioning and require judgement about the extentto which an individual fulfils a social role (e.g.worker, parent). As indicated above, these assessments require a normative judgement, which may
prove unreliable. The SFS uses a different approachby enumerating basic skills, social behaviour, etc.which informants record as present or absent,thereby avoiding ¿ evaluative’decisions. In thisrespect, the SFS has some similarities with the ¿ MRCNeeds for Care Assessment’ which was developedfor the long-term mentally ill in residential settings(Brewin eta!, 1987). The SFS also distinguishes lackof competence from lack of performance: lack ofcompetence refers to the absence or loss of a skill;lack of performance refers to non-use or disuse ofan available skill. This distinction was measuredsolely in relation to skills necessary for independentliving, as it was felt that informants would havedifficulty in achieving this distinction in other areas(e.g. social skills v. social performance). The SFS wasdeveloped by Birchwood (1983) and underwentextensive development through psychometric analysisbefore the final version was established.The present study examined the reliability, validity,sensitivityand utility for familyinterventionsof the SFS.MethodSeveraldistinctgroupsof subjectswererecruitedat differentstages of this study.A sample of 334 schizophrenicout-patients (Table I) allwith a clinical diagnosis of schizophrenia and conformingto the broad CATEGO ¿ S’class including S +, S?, P + andO (Wing et a!, 1974) was a compendium taken frompreviousand ongoing researchstudies (Birchwood,1983;Birchwood et al, 1989; Smith & Birchwood, 1987, 1990).Each of the samples of which this is a compendium weredefined on the same clinical criteria, had been living (orwere in daily contact) with their relative from the first orsecond episode, and were drawn from the same catchmentarea. All patients were in contact with the mental healthservices by virtue of their attendance at out-patient clinics,or referralto a community psychiatricnurseor the clinicalpsychology services. It should be noted that this sample wastaken during a period of investigation into families’ serviceneeds (Smith & Birchwood, 1990), when the large majorityof patients livingwith their familiesknown to the servicewere approached. In the total sample 88´ !. were taking oralor depot neuroleptics, and 24´ !. were in productive paidemployment. In keeping with other studies (e.g. Tarrieretal, 1988), the sample contained an excess of males. Twosubgroups of this sample took part in reliability analyses:30setsof parents completedthe SFSindependentlyabouttheir schizophrenic offspring; in a further 25 cases, relatives’data were compared with the SFS completed by 25symptom-free patients.A sample of 100 normal subjects were recruited via theirrelatives. Relatives were approached by interviewers in ¿ keysites’ throughout the catchment area from which the patientsample was drawn (e.g. shopping and job centres), andrequested to complete the SFS about an offspring or relativewith whom they were in close contact. Interviewers wereinstructed to suggest a male relative or offspring in two out