BRITISH JOURNAL OF PSYCHIATRY (2006), 188, 173^179

Influence of carer expressed and participating patients, and were conducted before randomisation into the trial. on relapse in non-affective psychosis The trial was based in four National Health Service (NHS) trusts in London and East Anglia in the UK. Within each of E. KUIPERS, P. BEBBINGTON, G. DUNN, D. FOWLER, D. FREEMAN, these trusts, recruitmentwas from speci- P. WATSON, A. HARDY and P. GARETYGARE T Y fied in-patient and out-patientout-patient teams which agreed that all patients who met the eligibility criteria would be asked to participate in the trial. These services were canvassed at least fortnightly for patients with psychosis who were relapsing. Patients Background High expressed emotion High expressed emotion in carers predicts who fulfilled the eligibility criteria were in carers predicts relapse in psychosis, but an increased relapse rate in asked to give their informed consent. (Bebbington & Kuipers, 1994; Butzlaff & Patients with carers who were in contact it is not known why this is so.In our Hooley, 1998). This finding led to the de- (including telephone contact) with them cognitive model of psychosis, we velopment of family interventions, which for at least 10 h a week were also asked postulated thatthe effectis mediated were recently endorsed by the National to give their consent to be contacted. These through affective changes. Institute for Clinical Excellence (NICE) carers were then asked for their consent to guidelines on schizophrenia (National enter the trial. Patients were recruited at Aims Toinvestigate the relationships Institute for Clinical Excellence, 2003). the time of a re-emergence of positive between carer expressed emotion, However, it is not known how critical, symptoms, either from a previously patients’symptoms and carer hostile or overinvolved family relationships recovered state or from a state of persistent lead to the re-emergence of psychosis. Our symptoms. For patients with persistent characteristics during a recent relapse of cognitive model of psychosis (Garety et aletal,, symptoms, a significant exacerbation in psychosis.psychosis. 2001) posits that family environments positive symptoms, typically leading to achieve this via affect, and that patients hospital admission, was required. The MethodMethod Atotal of 86 patients and living with carers who show high expressed inclusion criteria were as follows: current carers were investigatedinvestigatedin in a cross- emotion would have higher levels of diagnosis of non-affective psychosis sectional design. and and lower self- (schizophrenia, schizoaffective psychosis, esteem, but not more symptoms of delusional disorder; ICD–10 F20; World ResultsResults Patients whose carers showed psychosis.psychosis.PreviousPrevious studies have shown that Health Organization, 1992); age 18–65 high expressed emotion had significantly high expressed emotion is ‘not an artefact years; a second or subsequent episode higherlevels of anxiety and depression, of patient morbidity’ (Leff & Vaughn, starting not more than 3 months before 1985: p. 105). As in earlier studies the patient consented to enter the trial; but not more psychotic symptoms or (Barrowclough & Hooley, 2003; Raune etet and a rating of at least 4 (moderate lower self-esteem.Linear regression alal, 2004), we also predicted that carers severity) on at least one positive psychotic showed thatcarers’critical comments who show high expressed emotion would symptom of the Positive and Negative predicted anxietyin patients.Critical have higher levels of ‘burden’, stress, Syndrome Scale (PANSS; Kay, 1991) at depression and avoidant coping style, and the first time of meeting. comments were related to low carer self- lower self-esteem. A total of 86 patients and their carers esteem and avoidantcoping strategies. who had consented to take part in the Low carer self-esteem was also related to PRP trial were included in this study. carer depression, stress and carer METHOD ‘burden’, and to low patient self-esteem. DesignDesign Participants This was a cross-sectional study. The data The study sample consisted of participants Conclusions Our hypothesis was were obtained by trained assessors during recruited for the Psychological Prevention partiallysupported.Carercriticismwas the baseline phase of the randomised of Relapse in Psychosis (PRP) Trial associated with patient anxiety,low carer controlled trial, before allocation. (ISRCTN83557988) and their immediate self-esteem and poor carer coping carers. The PRP is a UK multicentre strategies.Familyinterventions should randomised controlled trial of cognitive– Carer measures focus ononimproving improving these after a relapse of behavioural therapy and family inter- Camberwell Family Interview (CFI; Vaughn & vention for psychosis, designed to test Leff, 1976)1976)Leff, . This is a semi-structured ques- symptoms of psychosis. hypotheses about outcome and about the tionnaire that assesses how well carers get Declaration of None.None. psychological processes associated with on with the person who has had a recent psychosis for both carers and participants. episode of psychosis. It covers family re- Funding detailed in Acknowledgements. Studies of psychological processes linked lationships, arguments, time spent together, with psychosis were incorporated into symptoms and role functioning. With the the baseline assessment of carers and carer’s consent this interview is recorded

173173

Downloaded from https://www.cambridge.org/core. 01 Oct 2021 at 03:44:05, subject to the Cambridge Core terms of use. KUIPERS ETE T AL

on audiotape, and it is subsequently rated inventory that assesses different coping variables (anxiety, depression, self-esteem for expressed emotion. Ratings are based styles (on a scale where 1¼never, 2never,2¼rarely,rarely, and overall symptoms of psychosis). not only on content of speech but also on arely, 3arely,3¼sometimes and 4¼a lot). The total We next related the components of prosodic variables such as pitch, speed score for each scale is obtained by adding expressed emotion (emotional over- and tone. Five scales are rated: critical the items together. The present study used involvement, , critical comments, comments (frequency count); hostility two questions per scale from the short form warmth and positive remarks) to patients’ (score of 0, 1 or 2); warmth (0–5); emo- of the measure (Carver & Scheier, 1994). symptoms using Pearson’s correlations, tional overinvolvement (0–5); and positive As in a previous study by Raune et aletal followed by multiple linear regression to remarks (frequency count). More than six (2004), we used an avoidant coping sub- control for potential confounding effects. critical comments, any hostility, or a rating scale consisting of 8 items (2 items Finally, we looked at the correlations of 3 or higher for emotional overinvolve- from each of the following: behavioural between carer expressed emotion variables ment categorise a carer as showing high disengagement, mental disengagement, and carer characteristics. expressed emotion. Taped interviews were alcohol/drug use and denial). rated by assessors previously trained by Dr Christine Vaughn to give reliable ratings Patient measures RESULTSRESULTS of expressed emotion. High correlations or phi coefficients were obtained for all ex- PANSS (Kay, 1991). This is a 30-item instru- Patient and carer characteristics pressed emotion scales: 440.76 for critical ment (rated on a scale from 1 to 7) for Around half of the carers were the parents comments, hostility, emotional over- the assessment of phenomena associated of patients, and three-quarters of the involvement, warmth, positive remarks with schizophrenia. Symptoms during the remainder were spouses or partners and overall expressed emotion category. past 72 h are rated, and higher scores indi- (Tables 1 and 2). The mean age of carers cate more severe symptoms. The positive was consequently higher than that of symptoms sub-scale and negative symp- Experience of Caregiving Inventory (ECI; patients (carer mean age 52.9 years, range toms sub-scale each consist of 7 items, Szmukler et al, 1996). This is a 66-item 26–86 years; participant mean age 36.3 and there is also a general pathology sub- instrument that assesses the subjective years, range 18–46 years). Patients were scale consisting of 16 items. experience of caregiving in eight areas defined by not being in their initial episode, (difficult behaviour, negative symptoms, and the mean duration of illness was 11.2 stigma, problems with services, effects on Self-Esteem Scale (Rosenberg, 1965). (see.(see years (range 551 year to 44 years). the family, need to provide back-up, depen- above).above). In total, 72% of the patients were male, dency and loss), together with two areas of and 84% were White. The relatively small positive experiences of caring (positive BeckDepressionInventory^II(BDI^II;Becketal,BeckDepression Inventory^II(BDI^II;Becketal, proportion of patients from a Black ethnic personal experiences and positive aspects 1996)1996). This established instrument con- background (7%) may be accounted for of the relationship). The questionnaire sists of 21 items, each of which is by the low overall proportion of patients measures how often carers have thought measured on a scale ranging from 0 to 3. with carers in the inner-city areas of the about each issue during the past month The total BDI–II score thus ranges from 0 study locations. Nearly 80% of the patients before interview (on a rating scale where to 63, with a high score representing a were unemployed, which is consistent with 00¼never, 1never,1¼rarely, 2rarely,2¼sometimes, 3¼oftenoften higher level of symptoms. Depression is the poor general employment prospects of and 4and4¼nearly always). measured for the previous 2 weeks. Birch- people with psychosis (Marwaha & woodwood et aletal (2000) have reported a high Johnson, 2004). Less than a quarter of the patients were living with partners, and Self-Esteem Scale (Rosenberg, 1965). This.This correlation (rr¼0.91) between the BDI and nearly two-thirds were single. For carers, measure consists of 10 items, each the interview-based Calgary Depression the gender ratio was (as usual) reversed, measured on a 4-point scale ranging from Scale for Schizophrenia (Addington et aletal,, with only 30% being male. ‘strongly agree’ to ‘strongly disagree’. After 1993), which confirms that the BDI can In total, 36% of carers were given a reverse scoring, the items were summed and be used to assess depression in psychosis. high overall rating of expressed emotion, divided by 10 to obtain a mean self-esteem and 30% had a high rating for expressed score. A high score represents low self- Beck Anxiety Inventory (BAI; Beck et al, emotion on the basis of critical comments, esteem.esteem. 1988)1988). This measure consists of 21 items, each of which measures common anxiety 24% on the basis of emotional over- involvement and 13% on the basis of symptoms. The total anxiety score ranges General Health Questionnaire (GHQ; Goldberg some hostility. The mean number of from 0 to 63, with a higher score represent- &Williams,1988)& Williams, 1988). The 28-item version of critical comments was 3.5 (range 0–32). ing a higher level of anxiety. Anxiety is this instrument was used, with scoring of The mean rating for emotional over- measured for the previous week. 0–4. It has a total score and four subscales involvement was 1.8 (range 0–5), whereas (somatic symptoms, stress, social func- the mean hostility was 0.26 (range 0–3). tioning and depression). In this study we Data analysis The mean rating for warmth was 2.3 (range focused on the stress and depression sub- All analyses were performed using SPSS 0–4), with 43% of carers having a rating of scales.scales. for Windows (version 12.01). Using more than 2. The mean rating for positive independent-sample tt-tests, we first investi- remarks was 1.9 (range 0–9). These ratings COPE Inventory (Carver et al, 1989; Carver & gated the relationships between high and were lower than those reported in the lit- Scheier, 1994). This is a multidimensional low expressed emotion carers and patient erature. For instance, 48% of carers made

174

Downloaded from https://www.cambridge.org/core. 01 Oct 2021 at 03:44:05, subject to the Cambridge Core terms of use. CARER EXPRESSED EMOTION IN NON- AFFECTIVE PSYCHOSIS

Ta b l e 11Tab Attributes of carers and patients expressed emotion had a significantly higher mean score of 25.8 (s.d.¼15.9,15.9, Attribute Low expressed High expressed Total PP¼0.046; Table 3). Similarly patients whose carers had low expressed emotion emotionemotion emotionemotion n/N (%)(%) had a mean BDI–II score of 22.6 n/N (%) n/N (%)(%) (s.d.(s.d.¼13.7), whereas those whose carers Gender of patient had high levels of expressed emotion had Male 40/55 (72.7)(72.7)40/55 22/31 (71.0)(71.0)22/31 62/86 (72.1) a significantly higher BDI–II score of 28.7 (s.d.(s.d.¼12.1,12.1, PP¼0.045). Contrary to our pre- FemaleFemale 15/55 (27.3) 9/31 (29.0) 24/86 (27.9)(27.9)24/86 diction, there were no significant differ- Gender of carer ences in patients’ self-esteem scores on the Male 17/55(30.9)17/55 (30.9) 8/298/29(27.6) (27.6)25/84 (29.8) Rosenberg scale. However, as we had pre- FemaleFemale 38/55 (69.1)(69.1)38/55 21/29 (72.4)(72.4)21/29 59/84 (70.2) dicted, there were also no significant differ- Patient employment status ences in patients’ scores on the PANSS Employed 8/55 (14.5)(14.5)8/55 3/31 (9.7)(9.7)3/31 11/86 (12.8) negative, PANSS positive or PANSS general Unemployed 42/55 (76.4)(76.4)42/55 26/31 (83.9)(83.9)26/31 68/86 (79.1)(79.1)68/86 sub-scales (although the latter approached Economically inactive 5/55 (9.1) 2/31 (7.4)(7.4)2/31 7/86 (8.1)(8.1)7/86 significance, as it partly comprises anxiety Carer employment status and depression scores). Employed 21/53 (39.6) 10/26 (38.5)(38.5)10/26 31/79 (39.2)(39.2)31/79 We next related the components of Unemployed 19/53 (35.8) 10/26 (38.5)(38.5)10/26 29/79 (36.7)(36.7)29/79 expressed emotion (emotional over- Economically inactive 13/53 (24.5) 5/26 (19.2)5/26(19.2) 18/79 (22.8)(22.8)18/79 involvement, hostility, critical comments, warmth and positive remarks) to patients’ Patient marital status scores on the BAI, BDI–II, Rosenberg Single 32/54 (59.3) 22/31 (71.0)(71.0)22/31 54/85 (63.5)(63.5)54/85 self-esteem and PANSS scales (Table 4). Married 18/54 (33.3)(33.3)18/54 4/31 (12.9)(12.9)4/31 22/85 (25.9)(25.9)22/85 There were significant correlations Divorced, widowed 4/54 (7.4) 5/31 (16.1)9/85 (10.6) between carer critical comments, emotional or separated overinvolvement and patients’ BAI scores, Patient cohabiting but not depression as measured on the Yes 13/54 (24.1)(24.1)13/54 7/30 (23.3)(23.3)7/30 20/84 (23.8)(23.8)20/84 BDI–II. Patients’ self-esteem scores were No 41/54 (35.9) 23/30 (76.7)(76.7)23/30 64/84 (76.2)(76.2)64/84 significantly correlated with patients’ Patient ethnicity BDI–II scores, as would be expected, but White 48/55 (87.3)(87.3)48/55 24/31 (77.4)(77.4)24/31 72/86 (83.7) not with any components of carer Black 2/55 (3.6) 4/31 (12.9)6/86 (7.0) expressed emotion. Patients’ PANSS scores were also not correlated with carer ex- OtherOther 5/55 (9.1)(9.1)5/55 3/31 (9.7)8/86 (9.3) pressed emotion, with the exception of positive remarks and the PANSS general Ta b l e 2 Carer age, face-to-face contact and expressed emotion sub-scale. It was unclear why the latter should be associated (it is not a general AttributeAttribute nn MeanMedian s.d. Range finding). Carer hostility was correlated with critical comments, and also with emotional Carer age (years) 7152.9 55.0 13.1 26^8826^88 overinvolvement; the negative aspects of Duration of face-to-face contact (h)71 39.2 35.0 23.9 7^84 11 expressed emotion were interrelated. Emotional overinvolvement86 1.8 2.0 1.2 0^5 Warmth was associated with positive Hostility86 0.260.260.0 0.75 0^3 remarks, and was negatively correlated with critical comments. Critical comments86 3.5 2.0 4.6 0^32 Given the correlations that were found Positive remarks 861.9 1.01.0 1.81.80^9 between patients’ BAI score, carers’ critical Warmth 86 2.02.02.0 1.2 0^4 comments and emotional overinvolve- 1. The carers who had 7 hours of contact a week also had more than 3 hours of telephone contact a week. ment, patient anxiety was chosen as the dependent variable in a linear regression either one or no critical comments. There Expressed emotion in carers using all of the component expressed was no relationship between the level of ex- and patients’ symptoms emotion ratings. Only the frequency of pressed emotion and any of the demo- critical comments made by carers predicted graphic variables analysed (gender of Carers with high and low expressed anxiety in patients (PP¼0.01). The rating of participant, gender of carer, participant or emotion were compared with regard to hostility only contributed to the model at a carer employment status, ethnicity and differences inpatients’ symptoms using trend level (PP¼0.092).0.092). age), although high levels of expressed independent-samplesampleindependent- tt-tests. Patients whose When the same analysis was repeated emotion were relatively unusual when the carers had low expressed emotion had a using patients’ BDI–II scores as the carer was married to the participant mean BAI score of 18.7 (s.d.¼14.5),14.5), dependent variable, none of the (13%).(13%). whereas those whose carers had high expressed emotion components were

175175

Downloaded from https://www.cambridge.org/core. 01 Oct 2021 at 03:44:05, subject to the Cambridge Core terms of use. KUIPERS ETE T AL

Ta b l e 3 Patient symptom ratings according to expressed emotion (EE) of carers associated with carer coping, but only indirectly via expressed emotion. Patient symptom ratings nn Means.d. tt d.f. PP MeanMean There were also significant correlations between carers’ and patients’ symptoms. (two-tailed)(two-tailed) difference Low carer self-esteem was associated with BAI score low patient self-esteem (rr¼0.29,0.29, PP¼0.017)0.017) Low EEEELow 5118.7 14.5 772.032.03 74740.046 777.17.1 and with carer depression (rr¼0.58,0.58, High EE 2925.8 15.9 PP550.0005) and stress (rr¼0.53,0.53, PP550.0005)0.0005) on the GHQ. Low carer self-esteem was BDI^II also related to patients’ BDI–II score Low EEEELow 54 22.622.613.7 772.03 78780.045 776.1 ((rr¼0.25,0.25, PP¼0.029), and was non- High EE 3028.7 12.1 significantly correlated with patients’ BAI PANSS positive sub-scale score (score(rr¼0.23,0.23, PP¼0.052), but was signifi- Low EEEELow 5516.5 5.2 771.581.58 84840.12 771.9 cantly correlated with negative aspects of High EE 3118.4 5.6 caregiving on the ECI (the so-called PANSS negative sub-scale11 burden of care; rr¼0.33,0.33, PP¼0.013). The Low EEEELow 5514.9 5.0 770.94 45.545.50.35 771.41.4 ECI negative score was in turn associated High EE 3116.3 7.4 with high carer stress scores on the GHQ PANSS general sub-scale ((rr¼0.48,0.48, PP550.0005) and with patient BDI Low EEEELow 5533.1 7.8 771.581.58 84840.06 773.23.2 scores (scores(rr¼0.28,0.28, PP¼0.036).0.036). The significant relationships are High EE 3136.4 6.9 illustrated in Fig. 1. BAI, Beck Anxiety Inventory; BDI^II, Beck Depression Inventory ^ II; PANSS, Positive and Negative Syndrome Scale. 1. Equality of variance not assumed.

found to contribute significantly (see relationships between carer and patient DISCUSSION Table 5).Table5). attributes.attributes. Carers’ critical comments were signifi- Carer criticism and patient anxiety Expressed emotion, other cantly correlated with low carer self-esteem As we had predicted, carer expressed emo- characteristics of carers, ((rr¼0.30,0.30, PP¼0.0008) and with the avoid- tion was related to one aspect of patient and patients’ symptoms ance coping strategy score on the COPE symptoms, namely affect. High levels of As in earlier studies, we had predicted that inventory (rr¼0.26,0.26, PP¼0.026), and were carer expressed emotion, particularly criti- components of carer expressed emotion negatively associated with reinterpretation cism, predicted high levels of patient would be related to measures of burden, on the COPE inventory (rr¼770.29,0.29, anxiety. Positive expressed emotion ratings, stress, low self-esteem and unhelpful coping PP¼0.009). Carer hostility towards the such as warmth, were not directly involved, strategies (e.g. avoidance) in carers. Pear- patient was also correlated with low carer although they were lower in highly critical son’s correlations were used to examine self-esteem (rr¼0.24,0.24, PP¼0.033). However, relationships. As would be expected, the these relationships as well as the overall low carer self-esteem was not directly negative aspects of relationships appeared

Ta b l e 4 Correlations between carer expressed emotion and patients’symptoms11

EmotionalEmotional HostilityCritical WarmthPositive PatientPatient PatientPatient PatientPatient PANSSPANSS PANSSPANSS PANSS over-over- comments remarks BAI BDI^II self-self- positive negative general involvement scorescore scorescore esteemesteem sub-scalesub-scale sub-scalesub-scale sub-scalesub-scale scorescore scorescore score

Emotional overinvolvement1.00 0.34** 0.38** 0.120.12 770.070.23* 0.08 770.060.11 0.20 0.10 Hostility 1.000.60** 770.18 770.110.03 0.090.09 770.050.05 770.03 770.10 770.14 Critical comments 1.00 770.29** 770.12 0.26*0.26*0.15 0.02 0.08 770.150.02 Warmth 1.000.32** 0.07 770.130.13 770.060.06 770.06 0.200.200.13 Positive remarks 1.000.12 770.080.08 770.010.10 0.12 0.23* Patient BAI score 1.000.52** 0.22 0.28*0.28*0.12 0.46** Patient BDI^II score 1.000.54** 0.18 0.34**0.34** 0.39**0.39** Patient self-esteem 1.000.05 0.18 0.13 PANSS positive sub-scale score 1.000.06 0.51** PANSS negative sub-scale score 1.000.48** PANSS general sub-scale score 1.00

BAI, Beck Anxiety Inventory; BDI^II, Beck Depression Inventory^II; PANSS, Positive and Negative Syndrome Scale. **PP550.05, **PP550.01.0.01. 1.1.Pearson’s Pearson’s correlations.

176

Downloaded from https://www.cambridge.org/core. 01 Oct 2021 at 03:44:05, subject to the Cambridge Core terms of use. CARER EXPRESSED EMOTION IN NON- AFFECTIVE PSYCHOSIS

Ta b l e 55Tab Linear regression of components of expressed emotion (Raune(Raune et aletal, 2004), but were not, in our study, related to negative aspects of care- Standardised coefficients 95% CI giving. Instead we found that negative evaluations of caregiving (high carer BetaBeta tPt P ‘burden’) were directly related to carer stress and patient depression, but were not Anxiety (BAI) directly related to expressed emotion. Emotional overinvolvement 0.151.24 0.2200.220 771.16 to 4.97 Hostility 770.24 771.71 0.0920.092 7710.08 to 0.77 Theoretical issues Critical comments 0.392.65 0.010 0.31 to 2.19 Positive comments 0.0070.06 0.951 771.80 to 1.921.921.80 Our model (Garety et aletal, 2001) was partly supported. It is plausible, as we had pro- Warmth 0.121.00 0.322 771.63 to 4.90 posed, that expressed emotion impacts on Depression (BDI^II) patients via affect, with critical comments Emotional overinvolvement 0.040.35 0.73 772.27 to 3.243.242.27 in particular being correlated with high Hostility 770.020.02 770.150.150.88 775.33 to 4.54 levels of patient anxiety. Carers inevitably Critical comments 0.120.80 0.440.44 770.52 to 1.19 find their role stressful and depressing at Positive comments 770.030.03 770.220.220.82 771.90 to 1.52 times, and this is sometimes demonstrated Warmth 771.00 770.780.780.44 774.10 to 1.791.794.10 by their own low self-esteem and by high levels of hostility towards the patient. Carer BAI, Beck Anxiety Inventory; BDI^II, Beck Depression Inventory^II. ‘burden’ is related to carer stress and patient depression, but not directly in to be the most upsetting to patients who treated with caution and requires this sample to poor relationships (only had recently had a relapse of symptoms of replication prospectively. indirectly via low carer self-esteem). psychosis. We had predicted a more general effect of negative relationships on patients, with high levels of patient Carer characteristics Study limitations depression and low patient self-esteem also and expressed emotion Our failure to replicate the finding that involved. However, our subsequent Relationships between carer characteristics carer criticism is related to low patient analyses suggested instead that anxiety and expressed emotion were less pro- self-esteem may have been due to our use was the main emotion found in patients nounced in this sample. Carer criticism of the Rosenberg measure, which has been with carers who showed high levels of and hostility towards patients were related criticised for its lack of specificity (Barrow- expressed emotion. Furthermore, we found to low carer self-esteem, and to carer stress cloughclough et aletal, 2003). However, our finding that carer criticism was implicated as the and depression. Carers’ critical comments that, in terms of relationships to expressed mechanism although, given that this was a were also directly related to an avoidant emotion, anxiety is a key feature in patients cross-sectional study, this result must be coping strategy, as we found previously suggests that low self-esteem may be only

Fig. 11Fig. Significant inter-correlations between attributes of carers and patients.

177

Downloaded from https://www.cambridge.org/core. 01 Oct 2021 at 03:44:05, subject to the Cambridge Core terms of use. KUIPERS ETE T AL

one consequence of difficult relationships. The role of anxiety identified in this study CLINICAL IMPLICATIONS revives the idea, first investigated in the 1970s, that high levels of are && Carer criticism seems to affect patients primarily by making them anxious. related to relapse in people with schizo- phrenia who live with relatives who show && Patients whose carers had high levels of expressed emotion did not have more high levels of expressed emotion (e.g. symptoms of psychosis than those whose carers had low expressed emotion soon TarrierTarrier et aletal, 1979; Sturgeon et aletal, 1981).,1981). after a relapse. It also links in with epidemiological evi- && dence that anxiety is a central feature of Family interventions should focus on reducing carer criticism and thereby the schizophrenia syndrome (Turnbull & decreasing patient anxiety, and improving carer self-esteem and carers’ ability to Bebbington, 2001). cope.cope. Another limitation of our study was that all of the patients had experienced a LIMITATIONS relapse of their symptoms of psychosis && This was a cross-sectional study,whichstudy, which limits our ability to make causal inferences within the past 3 months. This may have and investigate outcomes over time. meant that we had ‘ceiling’ effects in our symptom measures, and if so, they were && The carers had lower mean levels of expressed emotion than in previous studies all likely to have been high. This would soon after a relapse, and this may have reduced the power available for our analyses. have reduced the variance, and it may offer a partial explanation for the low && Because all of the patients had recently relapsed, there may have been less correlations. symptom variance to be explained. This study was also limited by the fact that, even soon after a relapse, the levels of expressed emotion were lower than those reported in some other studies, yielding E. KUIPERS, PhD, FBPsS, Department of Psychology,Institute of Psychiatry,King’s College London, London; relatively few carers with high ratings. P.P.BEBBINGTON, BEBBINGTON, PhD, FRCP,FRCPsych, Department of Mental Health Sciences, Royal Free and University Low levels of criticism (with a mean of College London Medical School, London; G. DUNN, PhD,Biostatistics Group, Division of Epidemiology and around three critical comments) may have Health Sciences,University of Manchester,Manchester; D.FOWLER, MSc, School of Medicine, Health Policy reduced the power available for our ana- and Practice,University of East Anglia, Norwich; D.FREEMAN, PhD,P.WATSON, MSc, A. HARDY,BSc, lyses. There was no evidence that expressed P.GARETY,P.GARETY,PhD, PhD, FBPsS, Department of Psychology,Psychology,Institute Institute of Psychiatry,King’s College London, London,London,UK UK emotion was not being rated appropriately. CorrespCorrespondence:ondence: Professor E.Kuipers,DepartmentE.Kuipers, Department of Psychology,PO Box 77,Institute of PsychPsychiatry,iatry, The low levels of expressed emotion may King’s College London,London SE5 8AF,UK.E-mail: e.kuipers@@iop.kcl.ac.uk have been because our carer sample was largely located in Essex and Norfolk, rather (First received 3 December 2004; final revision 9 March 2005; accepted 22 March 2005) than in inner-city locations in London. Fewer carers were identified in the inner- city areas than elsewhere, and an appreci- able number of these refused to participate found the demands of caring particularly problem-solving (Falloon et aletal, 1984;,1984; in the trial. Recruitment was pursued vigor- difficult.difficult. AndersonAnderson et aletal, 1986; Barrowclough & ously, but this did not overcome the Tarrier, 1992; Kuipers et aletal, 2002). The problem of some carers being unwilling to results of this study suggest that this route identify themselves as such, and others Implications for family is particularly important. being unwilling to participate in a treat- interventions The interrelationships between carer ment trial, in line with the findings of some High levels of criticism by carers appear to attributes confirm that carer stress, burden other recent studies (e.g. Szmukler et aletal,, be the main feature of high expressed and poor coping strategies are related to 2003).2003). emotion that affects patients, and they the carer’s own . This suggests that The low levels of expressed emotion possibly exert this effect via anxiety. We family intervention may need to improve may on the other hand be a good sign, in already know that high levels of anxiety carer understanding of difficulties and that carers now have more resources in patients are associated with symptoms optimise their coping strategies, moving available to them than they did when of psychosis that may precede relapse the latter away from avoidance and these studies began in the 1970s. The (Freeman & Garety, 2003), and that they towards reinterpretation or cognitive re- greater amount of information available, are an epidemiological feature of the appraisal. Improving these aspects might and the insistence in current UK clinical schizophrenia syndrome (Turnbull & then reduce negative relationships (critical guidelines on schizophrenia that carers Bebbington, 2001). In this case, anxiety comments and hostility) and subsequently should themselves be the focus of clinical may provide a more specific pathway for improve carer self-esteem, depression and support (e.g. Department of Health, 1999; interventions, particularly family interven- care ‘burden’. This might be the route to National Institute for Clinical Excellence, tions. A key feature of family interventions reducing the stress and burden of caring roles, 2003), may have altered the behaviour has always been to reduce tension and which have been notably resistant to more and attitudes of some individuals who improve negotiation, communication and general interventions (e.g. Barrowclough

178

Downloaded from https://www.cambridge.org/core. 01 Oct 2021 at 03:44:05, subject to the Cambridge Core terms of use. CARER EXPRESSED EMOTION IN NON- AFFECTIVE PSYCHOSIS

et aletal, 1999; Szmukler et aletal, 2003). Specific between self-evaluation, family attitudes, and Leff, J. & Vaughn, C. E. (1985) Expressed Emotion in difficulties with carer self-esteem and symptomatology. Journal of Abnormal Psychology,, 112112 ,, FamiliesFamilies. London: Guilford Press. 92^99.92^99. depression, leading to negative evaluations Marriott, A., Donaldson, C.,Tarrier,C., Tarrier, N., et al (2000)(2000) of caregiving, might also be improved Bebbington, P. & Kuipers, L. (1994) The predictive Effectiveness of cognitive^behaviouralcognitive ^ behavioural family utility of EE in schizophrenia: an aggregate analysis. intervention in reducing the burden of care in carers of by interventions based on cognitive– Psychological Medicine,, 2424, 707^718. patients with Alzheimer’s disease. British Journal of behavioural therapy (e.g. Marriott et aletal,, PsychiatryPsychiatry,, 176,557^562.,557^562. Beck, A.A.T., T., Epstein, N., Brown, G., et aletal (19 8 8) AnAn 2000). More targeted approaches in family inventory for measuring clinical anxiety: psychometric Marwaha, S. & Johnson, S. (2004) Schizophrenia and intervention might enable us to improve properties. Journal of Consulting and Clinical Psychology,, employment. A review. Social Psychiatry and Psychiatric both carer stress and patient outcomes. 5656, 893^897. Epidemiology,, 3939, 337^349. Beck, A.T., Steer, R. A. & Brown, G. K. (1996) BDI^IIBDI^II National Collaborating Centre for Mental Health ManualManual. San Antonio:The Psychological Corporation. ACKNOWLEDGEMENTS (2003)(2003) Schizophrenia. Full National Clinical Guideline on Birchwood, M., Iqbal, Z., Chadwick, P., et aletal (2000) Core Interventions in Primary and Secondary Care. London.London Cognitive approach to depression and suicidal thinking in & Leicester: Royal College of Psychiatrists & British WearegratefultoallofthepatientsandtheircarersWe are grateful to all of the patients and their carers psychosis. I.Ontogeny of post-psychotic depression. Psychological Society. who consented to take part in the study, and to the British Journal of Psychiatry,, 177177,,516^528. 516^528. Raune, D., Kuipers, E. & Bebbington, P. E. (2004) clinical teams in the South London and Maudsley Butzlaff,R.L.&Hooley,J.M.(1998)ExpressedExpressed Expressed emotion at first episode psychosis: NHS , Camden and Islington Mental Health emotion and psychiatric relapse: a meta-analysis. investigating a carer appraisal model. British Journal of and Social Care Trust, North East London Mental Archives of General Psychiatry,, 55,547^552. PsychiatryPsychiatry,, 184184, 321^326.,321^326. Health Trust and Norwich Mental Health Trust, who provided access to their clients.Weclients. We also thank Carver,C. S. & Scheier, M. F. (1994) Situational coping Rosenberg, M. (1965) Society and the Adolescent Self- Image. Princeton, NJ: Princeton University Press. Louise Isham and Katherine Ruffell, who completed dispositions in a stress transaction. Journal of Personality and Social Psychology,, 6666,184^195. some of the assessments and ratings of expressed Sturgeon, D., Kuipers, L., Berkowitz, R., et aletal (19 81) emotion. This study was funded by a Wellcome Trust Carver,C. S., Scheier, M. F. & Weintraub, J. K. (1989) Psychophysiological responses of schizophrenic patients Programme grant (no. 06452). Assessing coping strategies: a theoretically based to high and low expressed emotion relatives. BritishBritish approach. Journal of Personality and Social Psychology,, 5656,, Journal of Psychiatry,, 13813 8, 40^45.,40^45. 267^283. Szmukler,G. I., Burgess, P., Herrman, H., et aletal (19 9 6)6)(19 REFERENCES Department of Health (1999) National Service Caring for relatives with serious mental illness: the Framework ^ Mental Health. London: Department of development of experience of caregiving inventory. Addington, D., Addington, J. & Maticka-Tyndale, E. Health. Social Psychiatry and Psychiatric Epidemiology,, 31,, (19 93) Assessing depression in schizophrenia:Theschizophrenia: The 137137^148. ^ 14 8. Calgary Depression Scale. British Journal of Psychiatry,, Falloon, I. R. H., Boyd, J. L. & McGill, C.W. (1984) 163 (suppl. 22), 39^44. Family Care of Schizophrenia. New York: Guilford Press. Szmukler,G.,Szmukler, G., Kuipers, E., Joyce, J., et aletal (2003)(2003) An exploratory randomised controlled trial of a support Anderson, C. M., Reiss, D. J. & Hogarty, G. E. (1986) Freeman, D. & Garety, P. A. (2003) ConnectingConnecting neurosis and psychosis: the direct influence of emotion programme for carers of patients with a psychosis. Social Schizophrenia and the Family.NewYork:Guilford Psychiatry and Psychiatric Epidemiology,, 3838, 411^418.,411^418. Publications. on delusions and hallucinations. Behaviour Research TherapyTherapy,, 4141, 923^947. Tarrier, N.,Vaughn, C. E., Laden, M. H., et aletal (19 79) Barrowclough, C. & Tarrier, N. (1992) Families of Bodily reactions to people and events in schizophrenia. Schizophrenic Patients: Cognitive Behavioural Intervention.. Garety, P., Kuipers, E., Fowler, D., et aletal (2001)(2001) AA Archives of General Psychiatry,, 3636, 311^315.,311^315. London: Chapman & Hall. cognitive model of the positive symptoms of psychosis. Psychological Medicine,, 3131,189^195.,189^195. Barrowclough, C. & Hooley, J. M. (2003) AttributionsAttributions Turnbull,Turnbull,G.&Bebbington,P.(2001) G. & Bebbington, P. (2001) Anxiety and the and expressed emotion: a review. Clinical Psychology Goldberg,D.P.&Williams,P.(1988)A User’s Guide to schizophrenic process: clinical and epidemiological Review,, 2323, 849^880.,849^880. the General Health Questionnaire.Windsor: NFER evidence. Social Psychiatry and Psychiatric Epidemiology,, Nelson. 3636, 235^243. Barrowclough, C.,C.,Tarrier, Tarrier, N., Lewis, S., et aletal (19(1999) 9 9) Randomised controlled effectiveness trial of a needs- Kay, S. R. (1991) Positive and Negative Syndromes in Vaughn, C. E. & Leff, J. P. (1976) The influence of family based psychosocial intervention service for carers of Schizophrenia: Assessment and Research.NewYork: and social factors on the course of psychiatric illness. people with schizophrenia. British Journal of Psychiatry,, Brunner Mazel. British Journal of Psychiatry,, 129129,125^137.,125^137. 174,505^511. Kuipers, E., Leff, J. & Lam, D. (2002) Family Work for WorldHealthOrganization(1992)WorldHealthOrganization(19 92) The ICD ^10 Barrowclough, C.,C.,Tarrier, Tarrier, N., Humphreys, L., et aletal Schizophrenia: a Practical Guide (2nd edn). London: Classification of Mental and Behavioural Disorders: Clinical (2003)(2003) Self-esteem in schizophrenia: relationships Gaskell Press. Descriptions and Diagnostic Guidelines..Geneva:WHO. Geneva: WHO.

179179

Downloaded from https://www.cambridge.org/core. 01 Oct 2021 at 03:44:05, subject to the Cambridge Core terms of use.