Soc Psychiatry Psychiatr Epidemiol (1999) 34: 254±258
Ó Steinkop Verlag 1999
ORIGINAL PAPER
G. Haddock á N. Tarrier á A.P. Morrison á R. Hopkins R. Drake á S. Lewis
A pilot study evaluating the effectiveness of individual inpatient cognitive-behavioural therapy in early psychosis
Accepted: 11 December 1998
Abstract Background: Recent research indicates that cognitive-behaviour therapy (CBT) can be eective in ameliorating persistent positive symptoms in chronic psychotic patients. The eectiveness of CBT in acute and recent-onset psychosis has been little explored, although a recent pilot study indicated that CBT could signi®cantly improve recovery in acutely psychotic inpatients. Method: Short-term individual CBT was compared to supportive counselling/psychoeducation (SC) as an adjunct to standard inpatient hospital care and medication in 21 inpatients experiencing a recent-onset acute schizophrenic episode. Results: Both groups showed signi®cant reductions in Brief Psychiatric Rating Scale (BPRS) scores following treatment, although there were no group dierences. Time to discharge did not dier signi®cantly between the groups, although there was a greater variance for the SC patients. Two-year follow-up showed no signi®cant dierences between the groups, although the number of patients who relapsed, the number of relapses and the time to recurrence of psychotic symptoms was lower in the CBT group than the SC group. Interestingly, the time to readmission was shorter in the CBT group. Conclusions: CBT and SC are
G. Haddock N. Tarrier Department of Clinical Psychology, University of Manchester, Manchester, UK A.P. Morrison Department of Psychology, Prestwich Hospital, Manchester, UK R. Hopkins R. Drake S. Lewis Department of Psychiatry, University of Manchester, Manchester, UK G. Haddock (&) Academic Clinical Psychology, Tameside General Hospital, Fountain Street, Ashton-u-Lyne OL6 9RW, UK
acceptable treatments for recent-onset acutely psychotic inpatients. A larger randomised controlled trial over multiple hospital sites is warranted.
Introduction The ®rst line treatment for patients experiencing a ®rst episode of psychosis is antipsychotic medication. It is uncommon for patients to be oered any type of speci®c psychological treatment to aid in reducing the severity of the psychosis during the initial stages of a psychotic illness, although recently some small controlled studies have demonstrated that cognitive-behavioural interventions can signi®cantly bene®t chronically psychotic patients (Kuipers et al. 1998; Tarrier et al. 1998a). Tarrier et al. compared CBT to supportive counselling and treatment as usual in a randomised controlled trial with independent and blind assessment. CBT was superior to SC and treatment as usual on a number of outcomes. Kuipers et al. (1997) compared CBT to case management and treatment as usual in a randomised, controlled trial and found CBT to be superior on total Brief Psychiatric Rating Scale (BPRS; Overall and Gorham 1962) scores. Eighteen-month follow-up of patients in this study showed greater bene®ts for the CBT group than the controls (Kuipers et al. 1998). Despite the bene®ts demonstrated in these studies, treatment has generally been focused on patients with chronic schizophrenia and medication-resistant psychotic symptoms, and there has been little research on patients experiencing recent-onset or acute symptoms. One exception of this is a study by Drury et al. (1996a, b) carried out with acutely psychotic inpatients, although the subjects varied in the duration of their illness. Patients admitted to an acute psychiatric ward were randomly allocated to either a cognitive-behavioural treatment consisting of individual sessions, group work, family sessions and activity scheduling, or a control treatment of recreational activities matched for therapist time. The combined cognitive-behavioural treatment
255
signi®cantly speeded the time to recovery of symptoms, by between 25 and 50%, and reduced actual time spent in hospital by approximately 50% compared to the control group. The results also showed that these bene®ts were maintained in terms of lower relapse rates over 9 months for the cognitive-behavioural group. However, this study is limited, in that: 1. Evaluation of outcome was not carried out by a blind and independent assessor. 2. The CBT treatment consisted of a number of elements (individual, group and family sessions), making it dicult to determine which were the essential ingredients for change, and 3. The control treatment did not control for non-speci®c aspects of therapy (such as an understanding and empathic one-to-one relationship etc). As a result, these encouraging results need to be interpreted cautiously, but suggest that applying CBT in the acute stage of a psychotic illness is feasible and may produce long-lasting and signi®cant bene®ts. It could be further hypothesised that these bene®ts may be even greater if the CBT is applied to the very early stages of illness, when the secondary disabilities associated with schizophrenia have had less time to develop. Evidence from early intervention studies on neuroleptic medication suggests that the earlier in the course of illness that a patient receives treatment, the better the long-term prognosis. Taken together, this suggests that further research into the eectiveness of CBT on recent-onset acutely psychotic patients is a priority. This pilot study was carried out for two reasons: ®rst, as a pilot study to test the feasibility and eectiveness of short-term CBT applied to recently diagnosed acutely psychotic inpatients, and second, to ascertain whether the results reported by Drury et al. (1996a, b) could be reproduced with a small sample treated only with individual short-term CBT.
Subjects and methods Subjects The study inclusion criteria were: 1. DSM-IV (APA 1994) diagnosis of schizophrenia or schizoaective disorder con®rmed by an independent psychiatrist attached to the project 2. First treatment for psychosis less than 5 years ago 3. Currently admitted to an acute general psychiatric ward for onset or relapse of psychotic symptoms
more on the appropriate BPRS sub-scales). Patients giving consent were assessed using the BPRS and Psychotic Symptom Rating Scales (PSYRATS ± structured interviews designed to elicit the severity of dimensions of auditory hallucinations and delusions; Haddock et al. 1999). Patients were then randomly allocated to either cognitive-behaviour therapy (CBT) or supportive counselling/psychoeducation treatment (SC). There was no treatment as usual alone condition. Ward sta and responsible medical ocers were informed that patients were taking part in the study, but were not informed of the treatment allocation. All patients received standard ward management, which included treatment with neuroleptics. Treatment was delivered by two clinical psychologists with expertise in the CBT of psychosis (G.H. and A.M.) and was designed to take place during an inpatient hospital stay over a therapy envelope of 5 weeks or until the patient was discharged if this period was shorter. Booster sessions were oered at 1, 2, 3 and 4 months following discharge. The treatments used during the pilot study are described below. Independent assessment using the BPRS and PSYRATS was carried out by the project psychiatrist, who was blind to treatment allocation, and took place on entry to the study, at end of treatment and following the ®nal booster session (4 months following discharge). Days in hospital for the initial episode and any subsequent episodes, number of relapses, time to relapse of psychotic symptoms and time to ®rst readmission were recorded for each patient from entry to the study and for 2 years following entry to the study by a blind independent assessor (R.D.) using patient case notes. Cognitive-behavioural treatment The cognitive-behavioural treatment followed a detailed treatment manual designed by the authors and used in earlier CBT for psychosis research studies. It was primarily focused on the treatment of auditory hallucinations and delusions, associated symptoms and problems (for example, anxiety, depression, self-esteem) and relapse prevention. For some patients, psychotic symptoms rapidly remitted during the ®rst few days of their admission. In these cases the cognitive-behavioural intervention was focused on relapse prevention, treatment of associated problems and enhancing medication adherence. Treatment was designed to follow a cognitive-behavioural approach shown to be eective with chronic psychotic patients, and involved the following elements: 1. Assessment and engagement 2. Formulation of key problems 3. Intervention directed at reducing the severity or occurrence of key problems 4. Relapse prevention/keeping well The following speci®c CBT strategies were used: formulation, guided discovery, symptom monitoring, exposure/focusing strategies for managing voices, hypothesis/reality testing, re-framing attributions, rational responding, coping strategy enhancement, distraction techniques, anxiety management, depression/self-eseem work, medication compliance/motivational interviewing, schema work, relapse prevention and keeping well strategies. The booster sessions involved review and consolidation of strategies used during inpatient sessions, with particular emphasis on keeping well strategies. Supportive counselling and psychoeducation
Method Consecutive patients admitted to the four wards targeted were screened using the above criteria. Patients meeting initial screening criteria were then independently assessed within 10 working days of admission by the project psychiatrist (R.H.) to obtain consent to enter the study and to con®rm entry criteria (those listed above plus hallucinations and/or delusions that reached a criterion of 4 or
This treatment was also carried out according to a treatment manual designed by the authors and was matched in terms of therapy time and therapist to the CBT treatment. It aimed to control for nonspeci®c aspects of therapy by delivering basic assessment, psychoeducation and counselling in a supportive, warm, genuine, empathic and unstructured style. Speci®c techniques used involved: re¯ection, paraphrasing, summarising, feedback, avoidance of criticism, information about schizophrenia, medication and side eects.
256 Regular supervision of sessions (approximately 1 h per week) was used in order to maintain treatment ®delity and quality of both treatments. Supervision focused on adherence to the treatment manual, which was monitored through tape recordings of individual sessions with patients and through intensive case discussion.
Results Thirty-six patients were identi®ed as being eligible to take part in the study over approximately 1 year (27 males and 9 females). Of these, 21 entered the trial and 20 completed treatment. Twelve of the 36 patients refused to consent to enter the trial (7 males and 5 females) and 3 were excluded following initial identi®cation due to the following reasons; 1 failed to meet entry criteria at assessment, 1 was discharged prior to assessment and 1 was not assessed within the 10-day recruitment cut-o period. One patient in the CBT treatment withdrew from the study after three sessions because he did not feel it was helpful. Nine patients completed the CBT treatment and 11 completed the supportive counselling/psychoeducation (SC) treatment. All subjects in CBT were male. There were two females and nine males in SC. The patient who dropped out was male. The average age was 28.1 years (SD = 7.24 years) for the CBT group and 30.0 years (SD = 7.90) for the SC group. There were no signi®cant dierences between the groups. All patients who received treatment ®tted the Oce of Population Censuses and Surveys' (OPCS) category of `white', except for one patient in SC, who ®tted OPCS category of `black other' (Aspinall 1995). Ten patients were detained under the Mental Health Act (®ve in SC and ®ve in the CBT group). Patients received a mean of 10.2 (SD = 5.1) sessions of CBT and 9.1 (SD = 4.36) sessions of SC during the inpatient phase and 1.67 and 0.91 booster sessions respectively. The total number of minutes of therapy delivered diered between the two groups, with the CBT group receiving slightly more therapy time overall than patients in SC (480 vs 311 min). The number of inpatient sessions delivered varied considerably. Three patients received 3 sessions, one received 4 sessions, three received 6 sessions, two received 8 sessions, one received 9 sessions and the rest received between 11 and 18 sessions. The number of sessions received was equally spread between the two groups, and the dierences between groups on number of sessions and duration of therapy were non-signi®cant. Attendance at booster sessions was poor, with only two patients (in the CBT group) attending all four booster sessions. Three CBT patients and seven SC patients did not attend any booster sessions. Outcome on days in hospital and relapse There was an apparent dierence in the number of days until discharge from hospital between the two groups, with the CBT group spending a mean of 37.89 days
(SD = 17.12 days) in hospital and the SC group spending a mean of 55.09 days (SD = 47.00 days) in hospital. However, although these data were normally distributed, there was a signi®cant dierence between the variances of the groups, indicating that a nonparametric analysis was necessary. This revealed no signi®cant dierences between the groups on the number of days in hospital. This was supported when the medians were examined with the CBT and SC groups having a median of 37 days (range 11±65 days) and 34 days (range 20±175 days) respectively. This illustrates the wider variance in outcome in the SC group. Two-year follow-up data were collected from case notes. The percentage number of relapses, the mean number of relapses, the median time to relapse of psychosis (de®ned as a documented change in clinical management directly resulting from an increase in psychotic symptomatology), the median time to readmission and total number of days in hospital since study entry are shown in Table 1. When the changes in symptom outcomes were examined, there were signi®cant eects over time, but no speci®c group dierences or interactions. Analysis of variance of total BPRS scores at study entry and at end of inpatient treatment revealed a highly signi®cant main eect over time (F[1,16] = 11.80, P = 0.003), but no signi®cant main eect for group or interaction. When individual sub-scales from the BPRS were examined, the majority showed a reduction in symptom severity following treatment, and this was signi®cant at greater than the 1% level for unusual thought content (F[1,16] = 14.34, P = 0.002), grandiosity (F[1,16] = 10.0,P = 0.006) and conceptual disorganisation (F[1,16] = 9.29, P = 0.008). Multiple testing was carried out on the BPRS sub-scales; therefore, a 1% level of signi®cance was taken to reduce the likelihood of overestimating chance ®ndings. Signi®cant changes were not seen for hallucinations, although this was probably related to the small number of patients who were hallucinating (®ve in the CBT group, four in the SC group). Data on only eight patients from the CBT treatment and only ten patients from the SC treatment could be analysed on the BPRS variables, due to missing data. Due to poor attendance at booster sessions, few data on symptoms were collected on outcome at the ®nal booster
Table 1 Relapse and days in hospital over 2 years (CB cognitivebehaviour therapy, SC supportive counselling/psychoeducation) CBT Median no. of days in hospital Mean no. of days to 1st readmission % relapse Mean no. of relapses Median no. of days to relapse of psychosis
SC
59 (range 0±196)
88 (range 21±280)
316 (range 5±730)
639 (range 43±730)
44 0.78 730 (range 58±730)
73 1.2 527 (range 43±730)
257 Table 2 Mean Brief Psychiatric Rating Scale (BPRS) scores before and after inpatient treatment
CBT SC
Mean pre-treatment BPRS scores
Mean post inpatient treatment BPRS scores
53.0 (SD = 7.0) 53.2 (SD = 8.2)
46.8 (SD = 8.75) 38.3 (SD = 17.4)
assessment point, therefore these data were not analysed. BPRS data are shown in Table 2. Analysis of the PSYRATS revealed a similar picture. There was a signi®cant main eect over time for the total score on the delusions scale (F[1,17] = 7.34, P = 0.015), but no main eect for group or an interaction. All of the delusion sub-scales showed a decrease over time for both groups. Only the sub-scale measuring disruption from delusions showed a signi®cant decrease over time overall, at a minimum of the 1% level of signi®cance (F[1,17] = 15.16, P = 0.001), although the sub-scales of preoccupation and duration of delusions approached signi®cance (F[1,17] = 4.55, P = 0.048; F[1,17] = 5.63, P = 0.03). Analysis of the auditory hallucinations sub-scale showed a similar picture, with decreases on total auditory hallucination scores and all sub-scale scores for both groups. No signi®cant main eects or interaction were revealed on any of these items. One patient from the SC group was excluded from the PSYRATS analysis due to missing data.
Discussion The results indicate that applying CBT with recent-onset acute psychotic patients is feasible and is acceptable to signi®cant numbers of patients (only one patient withdrew). The refusal rate to enter the study (30% of eligible patients) was high, but consistent with that observed in studies with chronic patients, suggesting that this group of psychotic patients are not less likely to engage in `talking' treatments, despite their mental state requiring an acute inpatient hospitalisation. In addition, the patients who took part in the study engaged in therapies that were delivered up to three times per week. This was acceptable to the patients, and did not result in an exacerbation of symptomatology in any instance. This small pilot study failed to ®nd a signi®cant difference between the groups in terms of symptoms and length of hospital stay during the initial episode, and no signi®cant dierences in a number of relapse variables over a 2-year follow-up period. This may suggest that either both treatments are equally eective or that neither are eective. The lack of a no-treatment control group in this pilot limits the conclusions that can be made. However, the apparent lack of dierence in outcomes between the treatment groups in the present study may have been due to the small sample size, although an estimated power analysis indicated that the numbers should have been adequate from the eect size demon-
strated in the Drury et al. (1996a, b) study. It is possible that the eect size shown in that study was not robust enough to withstand the greater methodological rigour applied in the present study, or that elements of treatment carried out in the Drury study had a greater in¯uence on their treatment outcomes than individual CBT alone (e.g. the family sessions or group sessions). However, some of the non-signi®cant ®ndings from the study warrant discussion. The SC had a higher variance in hospital stay than the CBT group. If it is assumed that this dierence in variance was not present on entry to the trial, it is possible that the two treatments are having dierent eects. Speculatively, it may be that many acutely psychotic patients will get better as a result of hospital treatment, but that there may be some patients who need additional treatment to aid their recovery. The signi®cant dierence between the CBT and SC groups in the variance of the outcome measures may indicate a dierence in response to these two treatments, with patients receiving SC showing a much greater variability in their improvement than those who received CBT. It is also possible that the `dose' of treatment was not sucient for some patients. Treatment only continued until discharge and, although post-hospital booster sessions were oered, the take-up was poor. This meant that some patients only received a small number of sessions of either treatment, which may have resulted in them receiving an inadequate therapeutic dose. Further research should ensure that adequate `doses' of therapy are received. For example, in the recent study by Tarrier et al. (1998a) with chronic patients, 24 sessions were delivered in total, and the acute patients in the Drury et al. (1996a, b) study received approximately 8 h of therapy per week. Finally, there was some indication that there may be dierences in the longer-term outcome between the groups. It is possible with larger patient numbers and an increase in therapeutic dose that a clearer picture of the relative eectiveness of CBT compared to other psychotherapeutic treatments could be gained. The study highlighted a number of diculties in carrying out work with acutely psychotic inpatients in the early stages of illness. The majority of patients were young and male; therefore, there may have been particular developmental issues relating to these factors that should be addressed in therapy. In addition, the patients targeted are those who are most likely to drop out of treatment. Tarrier et al. (1998b) reported that patients who drop out of CBT treatment are likely to be male, single, unemployed and unskilled, with a low-level educational attainment and low pre-morbid IQ. In addition, drug and alcohol use as a precursor to the inpatient episode was common, suggesting that strategies to promote a reduction in alcohol and drug use and relapse prevention strategies to help patients reduce their likelihood of harmful drug or alcohol use should be incorporated into treatment interventions. Finally, although the majority of patients were in the early stages of illness (less than 5 years since onset), some symptoms often
258
considered to be the secondary disabilities associated with illness were apparent for some patients, e.g. negative symptoms, withdrawal, social and ®nancial diculties. It is possible that CBT delivered early, at a ®rst or second episode, would be more likely to avoid these complications and might be more eective than with a more chronic sample. Further research addressing the limitations highlighted above is needed to assess the eectiveness of CBT and to examine whether it is superior to other types of talking treatments. This is being examined in an ongoing, randomised, controlled, multi-site trial supported by the Medical Research Council (Lewis et al. 1995). The study is aimed at 354 patients experiencing their ®rst or second episode of psychosis. A number of issues that were highlighted in the pilot study are being addressed. Sessions are delivered within a therapy envelope of 5 weeks followed by four booster sessions. The sessions are carried out while the patient is an inpatient (or acute day patient), and continue even if they are discharged during the therapy envelope period. Therapy is delivered in patients' homes or on other sites to increase the likelihood of attendance. This means that patients receive a therapy `dose' of between 15 and 20 h delivered during the 5-week treatment envelope, plus the booster sessions. All patients receive medical and nursing care as usual, and consecutive referrals to acute inand day-patient services who meet inclusion criteria are randomly allocated to one of three therapy conditions: CBT, supportive counselling only or to treatment as usual only. The SC treatment diers from that delivered in the pilot in that only the non-speci®c aspects of therapy are delivered (empathy, warmth and genuineness). Psychoeducation is not included, as this may also be a speci®c component of CBT; therefore, excluding it from SC ensures that there is no overlap between the CBT and comparison treatment. A treatment as usual control has also been included to control for contact time. All therapy sessions are carried out according to a protocol, are supervised and tape-recorded. Tape samples are assessed using a modi®ed version of the Young and Beck Cognitive Therapy Scale (unpublished, Philadelphia Cognitive Therapy Center, 1980), to ensure that ®delity to treatment is maintained. It is hoped that these methodological improvements will address the diculties in carrying out work with acutely psychotic
inpatients in the early stages of illness highlighted above.
References Aspinall PJ (1995) Department of Health's requirement for mandatory collection of data on ethnic group of inpatients. BMJ 311: 1006±1009 Chadwick P, Birchwood M (1994) The omnipotence of voices. A cognitive approach to auditory hallucinations. Br J Psychiatry 164: 190±201 Drury V, Birchwood M, Cochrane R, Macmillan F (1996a) Cognitive therapy and recovery from acute psychosis. 1. Impact on psychotic symptoms. Br J Psychiatry 169: 593±601 Drury V, Birchwood M, Cochrane R, Macmillan F (1996b) Cognitive therapy and recovery from acute psychosis. 2. Impact on recovery time. Br J Psychiatry 169: 602±607 Garety P, Fowler D, Kuipers L, et al. (1998) Cognitive-behavioural therapy for drug resistant psychosis. Br J Psychiatry 67: 259± 271 Haddock G, McCarron J, Tarrier N, Faragher EB (1999) Scales to measure dimensions of hallucinations and delusions: the psychotic symptom rating scales (PSYRATS). Psychological Medicine (in press) Haddock G, Barrowclough C, Tarrier N, Lewis S (1996) Evaluation of family and individual cognitive-behaviour therapy for recent onset schizophrenia suerers with associated substance misuse problems. Grant protocol, funded by North West Research and Development Liaison Group fund Kingdon D, Turkington D (1991) The use of cognitive-behaviour therapy with a normalising rationale in schizophrenia. J Nerv Ment Dis 179: 207±211 Kuipers E, Garety P, Fowler D, Freeman D, Dunn G, Bebbington P, Hadley C, Jones S (1997) The London-East Anglia randomised controlled trial of cognitive behaviour therapy for psychosis: eects of the treatment phase. Br J Psychiatry 171: 319±327 Kuipers E, Fowler D, Garety P, Chisholm D, Freeman D, Dunn G, Bebbington P, Hadley C (1998) London-East Anglia randomised controlled trial of cogntive-behavioural therapy for psychosis. III. Follow-up and economic evaluation at 18 months. Br J Psychiatry 173: 61±68 Lewis S, Tarrier N, Haddock G, Bentall RP, Kinderman P (1995) Cognitive-behaviour therapy in early schizophrenia. Medical Research Council project grant application. Overall J, Gorham D (1962) The Brief Psychiatric Rating Scale. Psychol Rep 10: 799±812 Tarrier N, Yusupo L, Kinney C, McCarthy E, Gledhill A, Haddock G, Morris J (1998a) A randomised controlled trial of intensive cognitive-behaviour therapy for chronic schizophrenia. BMJ 317: 303±307 Tarrier N, Yusupo L, McCarthy E, Kinney C, Wittkowski A (1998b) Some reasons why patients suering from chronic schizophrenia fail to continue in psychological treatment. Behav Cogn Psychother 26: 177±181