Psychological Medicine, 1990, 20, 857-865 Printed in Great Britain
Families coping with schizophrenia: coping styles, their origins and correlates MAX BIRCHWOOD1 AND RAYMOND COCHRANE From the Department of Clinical Psychology, All Saints Hospital, Birmingham; and School of Psychology, University of Birmingham
An analysis of the coping styles adopted by relatives of schizophrenic patients has been identified by many reviewers as essential to an understanding of the complex interactions between patient and caregiver and to the origins of relatives' expressed emotion (EE). This study reports a taxonomy of coping behaviour derived from interviews with relatives of schizophrenic patients. It was found that relatives adopted broad styles of coping across all areas of patients' behaviour change. Relationships were uncovered between the styles and (a) relatives perceived control, burden and stress, (b) patients' social functioning, severity of behavioural disturbance and progress of the illness. It is suggested that advising relatives of changes in their coping styles in the course of family intervention must be tempered by an understanding of their origins in patients' behaviour. Further research is recommended to identify the coping styles associated with the high EE/low EE research classification.
SYNOPSIS
INTRODUCTION Recent reviews of the role of family life in the course of schizophrenia seem to have reached broadly similar conclusions (Koenigsberg & Handley, 1986; Falloon, 1988; Kuipers & Bebbington, 1988; Leff, 1989; Tarrier, 1989). While each acknowledge the robustness of the predictive efficacy of expressed emotion (EE), within and between cultures, together with the promising results of the family intervention studies, the origins and correlates of the index are less clear: 'it is evident that understanding of the origins of expressed emotion remain crude but it is likely to reside in complex interactions between patient and caregiver' (Falloon, 1988, p. 270). Three issues in particular point to the need to examine directly the strategies employed by families in interacting with a relative with schizophrenia. The first centres on whether relatives' EE is a 'trait' or a 'state' (Brown, 1985). The observation that criticism is associated with long-standing social impairment (Vaughn, 1986) and ongoing behaviour dis1 Address for correspondence: Dr Max Birchwood, Department of Clinical Psychology, All Saints Hospital, Lodge Road, Winson Green, Birmingham B18 5SD.
turbance (Brown et al. 1972) and that certain patterns of behaviour are more difficult for relatives to cope with, including unpredictable behaviour (Greenley, 1986) and' negative' symptoms (Kuipers & Bebbington, 1985) suggest an interaction between patient behaviour and relatives coping responses in the genesis of EE (Falloon, 1988). The burden this creates for families is now well documented (e.g. Gibbons et al. 1984) but whether and how it interacts with relatives' coping and emotional responses is unclear (Kuipers & Bebbington, 1988). The second issue centres on the temporal stability of EE: a substantial minority of high EE relatives become less critical (e.g. Hogarty et al. 1986) and a small number of low EE relatives more critical (Tarrier et al. 1988) as time progresses from the acute admission. Kuipers & Bebbington (1988) have proposed that EE level and relatives' coping efficacy interact to determine the stability or otherwise of EE. Birchwood & Smith (1987) go further and propose that EE includes a state component which is a ' thermometer' of developing interactions ('transactions') between patient and relative in which family members are attempting to cope and adjust to the emergence of schizophrenia. Third, the lack of understanding of EE's interactional underpinnings
857
858
M. Birchwood and R. Cochrane
has led to some uncertainty with regard to the design of appropriate interventions, with the result that the intervention approaches, although similar in content, have not been directly informed by such an understanding; if anything the reverse has been the case (Smith & Birchwood, 1990). This study therefore attempts to identify strategies and styles of coping employed by relatives in response to behavioural changes witnessed in the family member with schizophrenia and to identify any relationship they may have to family burden and to the patient's social impairment, behavioural disturbance and progress of his illness. The derivation of a taxonomy of relatives' coping behaviour will enable future study of the coping styles of relatives in the high/low EE research classification. A homogenous sample in terms of age and time since first episode was studied in order to: (a) control for families' length of exposure to schizophrenia and (b) as a means of measuring the progress of the illness independent of illness duration. Since no method of assessing coping behaviour was available, new methodologies were developed and are described here.
METHOD Assessments and measures 1. Coping behaviour A scheduled interview procedure was developed to elicit coping strategies adopted by relatives in relation to six areas of behavioural change most commonly observed (these include: withdrawal, positive symptoms, loss of independence, aggression, refusing medication and restlessness (Creer & Wing, 1973; Gibbons et al. 1984)). The interview centred on relatives' response to behaviours identified as occurring in their household. These behaviours were identified from a set of 17 most commonly noted by relatives in the study by Creer & Wing (1973) and a pilot study (Birchwood, 1983) and cover the six areas listed above (see Table 1). In order to facilitate disclosure and to ensure that the behaviour under discussion was present, the relative viewed a video in which the 17 behaviours were role-played in distinct 'scenes'. Thus, the video scenes acted a s ' category models' to cue relatives into identifying whether such
Table 1. Patient behaviour changes covered in the coping interview Behaviour category Withdrawal
Symptoms
Loss of independence
Aggression Overactivity Medication compliance
Constituent behaviours Declining social contact Emotional detachment from relatives Remaining in bed during the daytime Indifference to past social activities Talking and listening to "voices Expressing passivity ideas Persecutory delusions involving parents Persecutory delusions extra familial Expressing delusions of reference Retardation impairs self-help activity Patient requests relative to undertake routine domestic tasks Personal appearance neglected Indifference to employment Threats to family Provoking family discord Restlessness Refusing medication
behaviour occurred in their own home. This realistic video was professionally prepared at the University of Birmingham TV and Film Unit using professional actors. Once the presence of each behaviour was established in accordance with predetermined criteria, coping probes enquired about its day-to-day management using connotatively neutral language (eg. 'When (s)he behaves in this way how do you usually react?'). This interview was administered by a research psychologist. The interviews were tape recorded and transcribed for subsequent analysis. For each behaviour endorsed as present, the relative rated: the frequency of the behaviour on the scale 'daily', 'at least once a week', 'at least once a month', 'less than once a month', assigned scores 4 through to 1; a four-point rating of perceived control over the behaviour; fear for safety of self or family and a rating of tolerance were also made but not reported upon here. 2. Stress in relatives The Symptom Rating Test (SRT, Kellner & Sheffield, 1973; modified by Cochrane, 1980) is a well validated self-report measure of stress and
Family coping behaviour
includes anxiety, depression, inadequacy and somatic components. 3. Burden Relatives were interviewed on a further occasion by another psychologist (M.B.). As part of this interview, relatives were questioned about the impact which the illness had had on three specific areas: restriction of social and recreational activities; social and family consequences (stigma, social alienation, unemployment, marriage, effect on siblings) and feelings of burden and strain (Birchwood, 1983). Relatives rated each area of burden on a simple four-point scale. Analysis of the data, based on the sample reported here, showed that the scales were highly intercorrelated yielding a high reliability coefficient (Cronbach's alpha = 08). These ratings were therefore summed. This scale correlated r = 0-51 (P < 001) with the SRT. 4. Psychopathology The Present State Examination (PSE, Wing et al. 1974) was administered to patients participating in this study by the first author (M.B.), who received training in its use. 5. Symptom-related (SBDS) The 22-item scale is patients' behavioural to residual psychotic Birchwood, 1987).
859
between 15 and 24 months prior to the point of follow-up in the present study (it was anticipated that the data collection would take up to 9 months, so for some patients the 24th month coincided with the end of the study). This criterion was later relaxed to 18 months in order to increase the sample size (the final sample had a mean illness duration of 2-6 years). The case notes of all those with a psychiatric illness were screened using the ' Syndrome Checklist' (Wing et al. 1974) and those corresponding to the CATEGO ' S + ' class, with no evidence of organic brain syndromes, learning difficulty or extensive drug abuse were included (N = 98). Of these, 20 were not living with a relative; 8 had committed suicide; 6 were not stabilized on a neuroleptic drug and 11 refused to take part, leaving a final sample of 53 (Table 2). Comparison of those consenting to take part in the study with those who were eliminated did not reveal any significant differences on available data (age, sex, age at onset, living circumstances). Table 2. Demographic characteristics of the sample Sex by living group
behavioural disturbance a sensitive measure of disturbance attributable symptoms (see Smith &
Parental Marital Total Age at follow-up (years)
Male
Female
Total
31 2 33
18 2 20
49 4 53
Relatives' social class* (%)
6. Social Functioning Scale (SFS) (51-4) Mean 25-7 1 (14-3) 4-3 II This self-report scale of social functioning S.D. (11-4) III includes 7 subscales (social engagement, in(22-9) Unemployed terpersonal functioning, prosocial activities, recreation, independence (performance and • R e g i s t r a r - G e n e r a l C l a s s e s I , 2 , 3 = I ; 4 = I I ; 5 , 6 = III. competence) and employment/occupation) and a full-scale score. It has extensive normative data and psychometric validation (Birchwood et al. 1990) and has been used in family inter- RESULTS vention studies (Barrowclough & Tarrier, 1990). 1. Strategies and styles of coping The classification of coping behaviour was Subjects undertaken on the basis of the topographical A sample of schizophrenic patients was sought similarity of coping behaviours and not through with a 2-year history of illness who were living any theoretical considerations. Following conwith their families. The records of two large siderable development, eight categories were urban psychiatric hospitals were screened for delineated. Coercion includes strategies where patients with a first admission for schizophrenia the relative adopts a punitive approach including over a given 9-month time interval which was criticism of the behaviour or of the individual
860
M. Birchwood and R. Cochrane
himself, verbal or physical aggression, threats, attempts to shame or embarrass or any idiosyncratic reactions which are intended to provoke confrontation. Avoidance manoeuvres are strategies which minimize relatives' exposure to the individual's behaviour either by withdrawing from the situation or through the adoption of short-term expedients to curtail the behaviour (e.g. where relatives complete tasks for the individual in which he seems slow). The Ignore/ Accept category includes 'indifferent' reactions where relatives perceive the behaviour as unproblematical, where there is a benign acceptance of the behaviour (e.g. as part of the individual's personality) or as part of a deliberate strategy of non-response. Collusion includes those strategies where relatives actively condone or support the behaviour. The Constructive category represents' special action' taken by the relative in order, as they see it, to ameliorate the behaviour (excluding Coercive tactics). No distinction was made between the varieties of such tactics owing to insufficient endorsement of this category. Resignation is similar to the Ignore/ Accept category but in this instance represents the outcome of a developmental change where initial efforts to control the behaviour fail and relatives express a sense of powerlessness (' I've tried it all before, nothing changes so what's the point?'). Reassurance was a strategy reported only in relation to symptoms, here relatives present a stable and calm exterior to the individual emphasizing the security of the home and their relationship. Disorganized reactions are those where relatives express feelings of desperation and helplessness and engage in many strategies without consistency and without any clear dominant strategy emerging. (See appendix for examples of these strategies.) In those families reporting the presence of aggression, quite different strategies emerged which were categorized as follows: Submission responses are those which involve an acquiescence to the demands or challenges of the individual; Conflict includes aggressive, retaliatory responses of a verbal nature where the relative 'stands up' to aggressive demands or provocations; Avoidance strategies were described by some relatives where they felt they had learned to avoid confrontation either by minimizing certain types of demands made on the individual or escaping from the situation. In
view of the uniqueness of these coping strategies and the infrequency of reported aggression (5 % of the sample), this behaviour category is excluded from the following analyses. A score of 4 was given if a strategy was used exclusively, 3 if used predominantly, 2 if adopted with equal frequency to another, 1 if a minor or rare strategy and 0 if never used. The coping strategies were carefully defined including 'anchor' examples and then ten transcripts were independently classified and scored by one of the authors (M.B.) and a research psychologist. The reliability coefficients for the coping categories using the above scoring system averaged r = 0-79 (range 0-66 to 0-91), with no differences between raters in their mean scores (Table 3). Table 3. Reliability data for the coping styles: inter-rater correlations and mean scores
r Rater A(X) Rater B(X)
r Rater A(X) Rater B(X)
Constructive
Collusion
Ignore/ Accept
0-72 0-21 015
0-66 0-24 0-39
0-89 1-47 1-25
0 91 0-30 0-32
Avoidance
Coercion
Disorganization
Reassur a nee
0-75 0-24 0-39
0-90 0-69 0-73
0-77 0-23 Oil
0-72 006 008
Resignali
The first question addressed was whether relatives adopted an overall style of coping independent of the kind of behaviour encountered. This required two steps of data reduction. First, the scores for each kind of coping strategy was averaged across all observed behaviours within each of thefiveareas identified in Table 1 (i.e. withdrawal, symptoms etc.) excluding aggression. Thus, at this stage each family has a score on each of the coping strategies for each of the areas of observed behavioral change ('strategy scores'). This represents the family's profile of coping behaviour for each of the areas of behavioural change. Second, these scores were averaged across all areas of behavioural change. Thus, each family also has a score for each coping category across all areas of behavioural change ('style scores'). Since reassurance was used only in relation to symptoms,
Family coping behaviour
861
Table 4. Corrected correlations between 'Strategies' and overall 'Styles' of coping Coping category Behaviour Withdrawal Symptoms Dependency Restless Refusing medication Mean
Constructive
Collusion
0-80* • 0-43" 0-85" 0-68" 0-23 0-60
0-87" 0-92" 0-70"
t t
0-83
Ignore/ Accept
Resignation
Avoidance
Coercion
0-90* 0-64" 075" 0-69" 0-85" 0-77
0-40* 0-60" 0-70" 0-81"
0-72" 0-93" 0-82" 0-30
0-63
069
0-81" 014 0-80" 0-80" 0-62" 0-64
t
t
Disorgan ized 0-97" 0-60" 0-73" 008 0-60
t Correlation could not be computed.
a 'style' score could not be computed and is excluded from the following analysis. Taking each coping category in turn, corrected correlations between each strategy score and the overall style scores are computed. All such correlations were significant at the 1 % level. These are summarized in Table 4. This suggests that relatives show a degree of consistency in coping strategies between the areas of behavioural change they witness and supports the notion that relatives employ broad 'styles' of coping. There were two notable exceptions to this: relatives' adopting a Coercive style did so with considerable consistency except in relation to positive symptoms; and relatives' response to 'refusing medication' did not usually reflect their overall coping styles. The percentage of relatives using each coping style as a 'predominant' or 'exclusive' style (mean strategy score ^ 3) was as follows: Disorganized, 5-5%, Coercion 20-5%; Avoidance 24-3%; Resignation 24-3%; Ignore/Accept 23-2% Constructive 8-2%; Collusion 17-8%. 2. The relationship between coping styles and social impairment Patients were trichotomized into three social adjustment groups (high, average, low) at the 33rd and 66th percentile points of the sample used in the standardization of the Social Functioning Scale (SFS) (Birchwood et al. 1990) and their relatives' coping styles contrasted using ANOVAs testing for a linear trend. The results showed that Coercion is more commonly adopted by relatives with patients of low social functioning (F = 3-60, P < 005). This was especially so in respect of patients scoring high on social withdrawal (P < 001) and those
without employment or daytime occupation (P < 002). Conversely, the Ignore/Accept style was adopted more by relatives of patients with higher social functioning (F = 2-32, P = 008). All remaining coping styles were statistically non-significant. 3 Coping styles, behavioural disturbance and residual symptoms The relationship between coping styles and the nature and severity of the behavioural disturbance is shown in Table 5. Two measures of behavioural disturbance were used. The first was a summation of relatives' ratings of frequency of observed behaviour elicited during the interview; the second, Symptom Behavioural Disturbance Scale (SBDS), an undifferentiated measure of behavioural disturbance. The results show that the adoption of many of the styles of coping are, by and large, independent of the level of prevailing behavioural disturbance, with two important exceptions. First, some relatives are more likely to become Disorganized the greater the level of overall disturbance, particularly in relation to positive symptoms and restless agitation; with higher levels of disturbance arising from residual symptoms, relatives were more likely to use Reassurance. Second, Coercion is a more probable strategy where the individual shows greater withdrawal and inactivity (cf. previous section). Positive/negative symptoms were measured based on the PSE results: patients were dichotomized into presence v. absence of each kind of symptom. Point-biserial correlation coefficients between these and the coping styles, revealed weak, non-significant relationships. Both scales of behavioural disturbance cor-
862
M. Birchwood and R. Cochrane Table 5. Correlations between coping strategies and behavioural disturbance Behavioural disturbance
Coping strategies Constructive Collusion Ignore/accept Resignation Avoidance Coercion Disorganized Reassurance
Withdrawal
Symptoms
-014 -003 -019 006 014 0-43" 0-27 005
003 000 -018 -009 -000 -012 0-62" 0-81"
Dependence
Restlessness
-0-24 005 -013 0-30 -004 017 -004 -005
009 -009 -008 0-23 005 -015 0S8" 018
•/>
Families coping with schizophrenia: coping styles, their origins and correlates MAX BIRCHWOOD1 AND RAYMOND COCHRANE From the Department of Clinical Psychology, All Saints Hospital, Birmingham; and School of Psychology, University of Birmingham
An analysis of the coping styles adopted by relatives of schizophrenic patients has been identified by many reviewers as essential to an understanding of the complex interactions between patient and caregiver and to the origins of relatives' expressed emotion (EE). This study reports a taxonomy of coping behaviour derived from interviews with relatives of schizophrenic patients. It was found that relatives adopted broad styles of coping across all areas of patients' behaviour change. Relationships were uncovered between the styles and (a) relatives perceived control, burden and stress, (b) patients' social functioning, severity of behavioural disturbance and progress of the illness. It is suggested that advising relatives of changes in their coping styles in the course of family intervention must be tempered by an understanding of their origins in patients' behaviour. Further research is recommended to identify the coping styles associated with the high EE/low EE research classification.
SYNOPSIS
INTRODUCTION Recent reviews of the role of family life in the course of schizophrenia seem to have reached broadly similar conclusions (Koenigsberg & Handley, 1986; Falloon, 1988; Kuipers & Bebbington, 1988; Leff, 1989; Tarrier, 1989). While each acknowledge the robustness of the predictive efficacy of expressed emotion (EE), within and between cultures, together with the promising results of the family intervention studies, the origins and correlates of the index are less clear: 'it is evident that understanding of the origins of expressed emotion remain crude but it is likely to reside in complex interactions between patient and caregiver' (Falloon, 1988, p. 270). Three issues in particular point to the need to examine directly the strategies employed by families in interacting with a relative with schizophrenia. The first centres on whether relatives' EE is a 'trait' or a 'state' (Brown, 1985). The observation that criticism is associated with long-standing social impairment (Vaughn, 1986) and ongoing behaviour dis1 Address for correspondence: Dr Max Birchwood, Department of Clinical Psychology, All Saints Hospital, Lodge Road, Winson Green, Birmingham B18 5SD.
turbance (Brown et al. 1972) and that certain patterns of behaviour are more difficult for relatives to cope with, including unpredictable behaviour (Greenley, 1986) and' negative' symptoms (Kuipers & Bebbington, 1985) suggest an interaction between patient behaviour and relatives coping responses in the genesis of EE (Falloon, 1988). The burden this creates for families is now well documented (e.g. Gibbons et al. 1984) but whether and how it interacts with relatives' coping and emotional responses is unclear (Kuipers & Bebbington, 1988). The second issue centres on the temporal stability of EE: a substantial minority of high EE relatives become less critical (e.g. Hogarty et al. 1986) and a small number of low EE relatives more critical (Tarrier et al. 1988) as time progresses from the acute admission. Kuipers & Bebbington (1988) have proposed that EE level and relatives' coping efficacy interact to determine the stability or otherwise of EE. Birchwood & Smith (1987) go further and propose that EE includes a state component which is a ' thermometer' of developing interactions ('transactions') between patient and relative in which family members are attempting to cope and adjust to the emergence of schizophrenia. Third, the lack of understanding of EE's interactional underpinnings
857
858
M. Birchwood and R. Cochrane
has led to some uncertainty with regard to the design of appropriate interventions, with the result that the intervention approaches, although similar in content, have not been directly informed by such an understanding; if anything the reverse has been the case (Smith & Birchwood, 1990). This study therefore attempts to identify strategies and styles of coping employed by relatives in response to behavioural changes witnessed in the family member with schizophrenia and to identify any relationship they may have to family burden and to the patient's social impairment, behavioural disturbance and progress of his illness. The derivation of a taxonomy of relatives' coping behaviour will enable future study of the coping styles of relatives in the high/low EE research classification. A homogenous sample in terms of age and time since first episode was studied in order to: (a) control for families' length of exposure to schizophrenia and (b) as a means of measuring the progress of the illness independent of illness duration. Since no method of assessing coping behaviour was available, new methodologies were developed and are described here.
METHOD Assessments and measures 1. Coping behaviour A scheduled interview procedure was developed to elicit coping strategies adopted by relatives in relation to six areas of behavioural change most commonly observed (these include: withdrawal, positive symptoms, loss of independence, aggression, refusing medication and restlessness (Creer & Wing, 1973; Gibbons et al. 1984)). The interview centred on relatives' response to behaviours identified as occurring in their household. These behaviours were identified from a set of 17 most commonly noted by relatives in the study by Creer & Wing (1973) and a pilot study (Birchwood, 1983) and cover the six areas listed above (see Table 1). In order to facilitate disclosure and to ensure that the behaviour under discussion was present, the relative viewed a video in which the 17 behaviours were role-played in distinct 'scenes'. Thus, the video scenes acted a s ' category models' to cue relatives into identifying whether such
Table 1. Patient behaviour changes covered in the coping interview Behaviour category Withdrawal
Symptoms
Loss of independence
Aggression Overactivity Medication compliance
Constituent behaviours Declining social contact Emotional detachment from relatives Remaining in bed during the daytime Indifference to past social activities Talking and listening to "voices Expressing passivity ideas Persecutory delusions involving parents Persecutory delusions extra familial Expressing delusions of reference Retardation impairs self-help activity Patient requests relative to undertake routine domestic tasks Personal appearance neglected Indifference to employment Threats to family Provoking family discord Restlessness Refusing medication
behaviour occurred in their own home. This realistic video was professionally prepared at the University of Birmingham TV and Film Unit using professional actors. Once the presence of each behaviour was established in accordance with predetermined criteria, coping probes enquired about its day-to-day management using connotatively neutral language (eg. 'When (s)he behaves in this way how do you usually react?'). This interview was administered by a research psychologist. The interviews were tape recorded and transcribed for subsequent analysis. For each behaviour endorsed as present, the relative rated: the frequency of the behaviour on the scale 'daily', 'at least once a week', 'at least once a month', 'less than once a month', assigned scores 4 through to 1; a four-point rating of perceived control over the behaviour; fear for safety of self or family and a rating of tolerance were also made but not reported upon here. 2. Stress in relatives The Symptom Rating Test (SRT, Kellner & Sheffield, 1973; modified by Cochrane, 1980) is a well validated self-report measure of stress and
Family coping behaviour
includes anxiety, depression, inadequacy and somatic components. 3. Burden Relatives were interviewed on a further occasion by another psychologist (M.B.). As part of this interview, relatives were questioned about the impact which the illness had had on three specific areas: restriction of social and recreational activities; social and family consequences (stigma, social alienation, unemployment, marriage, effect on siblings) and feelings of burden and strain (Birchwood, 1983). Relatives rated each area of burden on a simple four-point scale. Analysis of the data, based on the sample reported here, showed that the scales were highly intercorrelated yielding a high reliability coefficient (Cronbach's alpha = 08). These ratings were therefore summed. This scale correlated r = 0-51 (P < 001) with the SRT. 4. Psychopathology The Present State Examination (PSE, Wing et al. 1974) was administered to patients participating in this study by the first author (M.B.), who received training in its use. 5. Symptom-related (SBDS) The 22-item scale is patients' behavioural to residual psychotic Birchwood, 1987).
859
between 15 and 24 months prior to the point of follow-up in the present study (it was anticipated that the data collection would take up to 9 months, so for some patients the 24th month coincided with the end of the study). This criterion was later relaxed to 18 months in order to increase the sample size (the final sample had a mean illness duration of 2-6 years). The case notes of all those with a psychiatric illness were screened using the ' Syndrome Checklist' (Wing et al. 1974) and those corresponding to the CATEGO ' S + ' class, with no evidence of organic brain syndromes, learning difficulty or extensive drug abuse were included (N = 98). Of these, 20 were not living with a relative; 8 had committed suicide; 6 were not stabilized on a neuroleptic drug and 11 refused to take part, leaving a final sample of 53 (Table 2). Comparison of those consenting to take part in the study with those who were eliminated did not reveal any significant differences on available data (age, sex, age at onset, living circumstances). Table 2. Demographic characteristics of the sample Sex by living group
behavioural disturbance a sensitive measure of disturbance attributable symptoms (see Smith &
Parental Marital Total Age at follow-up (years)
Male
Female
Total
31 2 33
18 2 20
49 4 53
Relatives' social class* (%)
6. Social Functioning Scale (SFS) (51-4) Mean 25-7 1 (14-3) 4-3 II This self-report scale of social functioning S.D. (11-4) III includes 7 subscales (social engagement, in(22-9) Unemployed terpersonal functioning, prosocial activities, recreation, independence (performance and • R e g i s t r a r - G e n e r a l C l a s s e s I , 2 , 3 = I ; 4 = I I ; 5 , 6 = III. competence) and employment/occupation) and a full-scale score. It has extensive normative data and psychometric validation (Birchwood et al. 1990) and has been used in family inter- RESULTS vention studies (Barrowclough & Tarrier, 1990). 1. Strategies and styles of coping The classification of coping behaviour was Subjects undertaken on the basis of the topographical A sample of schizophrenic patients was sought similarity of coping behaviours and not through with a 2-year history of illness who were living any theoretical considerations. Following conwith their families. The records of two large siderable development, eight categories were urban psychiatric hospitals were screened for delineated. Coercion includes strategies where patients with a first admission for schizophrenia the relative adopts a punitive approach including over a given 9-month time interval which was criticism of the behaviour or of the individual
860
M. Birchwood and R. Cochrane
himself, verbal or physical aggression, threats, attempts to shame or embarrass or any idiosyncratic reactions which are intended to provoke confrontation. Avoidance manoeuvres are strategies which minimize relatives' exposure to the individual's behaviour either by withdrawing from the situation or through the adoption of short-term expedients to curtail the behaviour (e.g. where relatives complete tasks for the individual in which he seems slow). The Ignore/ Accept category includes 'indifferent' reactions where relatives perceive the behaviour as unproblematical, where there is a benign acceptance of the behaviour (e.g. as part of the individual's personality) or as part of a deliberate strategy of non-response. Collusion includes those strategies where relatives actively condone or support the behaviour. The Constructive category represents' special action' taken by the relative in order, as they see it, to ameliorate the behaviour (excluding Coercive tactics). No distinction was made between the varieties of such tactics owing to insufficient endorsement of this category. Resignation is similar to the Ignore/ Accept category but in this instance represents the outcome of a developmental change where initial efforts to control the behaviour fail and relatives express a sense of powerlessness (' I've tried it all before, nothing changes so what's the point?'). Reassurance was a strategy reported only in relation to symptoms, here relatives present a stable and calm exterior to the individual emphasizing the security of the home and their relationship. Disorganized reactions are those where relatives express feelings of desperation and helplessness and engage in many strategies without consistency and without any clear dominant strategy emerging. (See appendix for examples of these strategies.) In those families reporting the presence of aggression, quite different strategies emerged which were categorized as follows: Submission responses are those which involve an acquiescence to the demands or challenges of the individual; Conflict includes aggressive, retaliatory responses of a verbal nature where the relative 'stands up' to aggressive demands or provocations; Avoidance strategies were described by some relatives where they felt they had learned to avoid confrontation either by minimizing certain types of demands made on the individual or escaping from the situation. In
view of the uniqueness of these coping strategies and the infrequency of reported aggression (5 % of the sample), this behaviour category is excluded from the following analyses. A score of 4 was given if a strategy was used exclusively, 3 if used predominantly, 2 if adopted with equal frequency to another, 1 if a minor or rare strategy and 0 if never used. The coping strategies were carefully defined including 'anchor' examples and then ten transcripts were independently classified and scored by one of the authors (M.B.) and a research psychologist. The reliability coefficients for the coping categories using the above scoring system averaged r = 0-79 (range 0-66 to 0-91), with no differences between raters in their mean scores (Table 3). Table 3. Reliability data for the coping styles: inter-rater correlations and mean scores
r Rater A(X) Rater B(X)
r Rater A(X) Rater B(X)
Constructive
Collusion
Ignore/ Accept
0-72 0-21 015
0-66 0-24 0-39
0-89 1-47 1-25
0 91 0-30 0-32
Avoidance
Coercion
Disorganization
Reassur a nee
0-75 0-24 0-39
0-90 0-69 0-73
0-77 0-23 Oil
0-72 006 008
Resignali
The first question addressed was whether relatives adopted an overall style of coping independent of the kind of behaviour encountered. This required two steps of data reduction. First, the scores for each kind of coping strategy was averaged across all observed behaviours within each of thefiveareas identified in Table 1 (i.e. withdrawal, symptoms etc.) excluding aggression. Thus, at this stage each family has a score on each of the coping strategies for each of the areas of observed behavioral change ('strategy scores'). This represents the family's profile of coping behaviour for each of the areas of behavioural change. Second, these scores were averaged across all areas of behavioural change. Thus, each family also has a score for each coping category across all areas of behavioural change ('style scores'). Since reassurance was used only in relation to symptoms,
Family coping behaviour
861
Table 4. Corrected correlations between 'Strategies' and overall 'Styles' of coping Coping category Behaviour Withdrawal Symptoms Dependency Restless Refusing medication Mean
Constructive
Collusion
0-80* • 0-43" 0-85" 0-68" 0-23 0-60
0-87" 0-92" 0-70"
t t
0-83
Ignore/ Accept
Resignation
Avoidance
Coercion
0-90* 0-64" 075" 0-69" 0-85" 0-77
0-40* 0-60" 0-70" 0-81"
0-72" 0-93" 0-82" 0-30
0-63
069
0-81" 014 0-80" 0-80" 0-62" 0-64
t
t
Disorgan ized 0-97" 0-60" 0-73" 008 0-60
t Correlation could not be computed.
a 'style' score could not be computed and is excluded from the following analysis. Taking each coping category in turn, corrected correlations between each strategy score and the overall style scores are computed. All such correlations were significant at the 1 % level. These are summarized in Table 4. This suggests that relatives show a degree of consistency in coping strategies between the areas of behavioural change they witness and supports the notion that relatives employ broad 'styles' of coping. There were two notable exceptions to this: relatives' adopting a Coercive style did so with considerable consistency except in relation to positive symptoms; and relatives' response to 'refusing medication' did not usually reflect their overall coping styles. The percentage of relatives using each coping style as a 'predominant' or 'exclusive' style (mean strategy score ^ 3) was as follows: Disorganized, 5-5%, Coercion 20-5%; Avoidance 24-3%; Resignation 24-3%; Ignore/Accept 23-2% Constructive 8-2%; Collusion 17-8%. 2. The relationship between coping styles and social impairment Patients were trichotomized into three social adjustment groups (high, average, low) at the 33rd and 66th percentile points of the sample used in the standardization of the Social Functioning Scale (SFS) (Birchwood et al. 1990) and their relatives' coping styles contrasted using ANOVAs testing for a linear trend. The results showed that Coercion is more commonly adopted by relatives with patients of low social functioning (F = 3-60, P < 005). This was especially so in respect of patients scoring high on social withdrawal (P < 001) and those
without employment or daytime occupation (P < 002). Conversely, the Ignore/Accept style was adopted more by relatives of patients with higher social functioning (F = 2-32, P = 008). All remaining coping styles were statistically non-significant. 3 Coping styles, behavioural disturbance and residual symptoms The relationship between coping styles and the nature and severity of the behavioural disturbance is shown in Table 5. Two measures of behavioural disturbance were used. The first was a summation of relatives' ratings of frequency of observed behaviour elicited during the interview; the second, Symptom Behavioural Disturbance Scale (SBDS), an undifferentiated measure of behavioural disturbance. The results show that the adoption of many of the styles of coping are, by and large, independent of the level of prevailing behavioural disturbance, with two important exceptions. First, some relatives are more likely to become Disorganized the greater the level of overall disturbance, particularly in relation to positive symptoms and restless agitation; with higher levels of disturbance arising from residual symptoms, relatives were more likely to use Reassurance. Second, Coercion is a more probable strategy where the individual shows greater withdrawal and inactivity (cf. previous section). Positive/negative symptoms were measured based on the PSE results: patients were dichotomized into presence v. absence of each kind of symptom. Point-biserial correlation coefficients between these and the coping styles, revealed weak, non-significant relationships. Both scales of behavioural disturbance cor-
862
M. Birchwood and R. Cochrane Table 5. Correlations between coping strategies and behavioural disturbance Behavioural disturbance
Coping strategies Constructive Collusion Ignore/accept Resignation Avoidance Coercion Disorganized Reassurance
Withdrawal
Symptoms
-014 -003 -019 006 014 0-43" 0-27 005
003 000 -018 -009 -000 -012 0-62" 0-81"
Dependence
Restlessness
-0-24 005 -013 0-30 -004 017 -004 -005
009 -009 -008 0-23 005 -015 0S8" 018
•/>
