Community Mental Health Journal, Vol. 36, No. 5, October 2000

Integrating Cognitive-BehavioralPsychotherapy for Persons withSchizophrenia into a Psychiatric

Rehabilitation Program:Results of a Three Year Trial

William Bradshaw, Ph.D., LICSW

ABSTRACT: Emerging models of cognitive-behavioral treatment (CBT) offer promisingnew intervention strategies in the psychotherapy of schizophrenia. These models, how-ever, have not been integrated into community support programs and evaluated incomparison to standard community treatments. This study examined differences inoutcomes of clients who received long-term day treatment program services (DTP)compared to clients who received individual CBT that was included as part of theirDTP treatment. Twenty-four clients were randomly assigned to DTP treatment or CBT/DTP treatment. Data on standardized measures of psychosocial functioning, symptom-atology and rehospitalizations were collected over the course of three years of treatment.Analysis of variance with repeated measures was conducted to evaluate the effects oftype of treatment (CBT/DTP and DTP) and time (dependent variable scores taken atthe end of treatment years one, two and three) on the three outcome variables. Resultsindicate significant improvement for the CBT/DTP group compared to the DTP groupin the patterns of change over time for psychosocial functioning and symptomology. Inaddition to this significant group/time interaction the time factor was also significantfor both variables. For rehospitalizations the time factor was significant and the group/time factor was nonsignificant. Implications for service delivery to persons with schizo-phrenia and suggestions for future research are discussed.

William Bradshaw is affiliated with the School of Social Work, University of Minnesota, Minneap-olis.

Address correspondence to William Bradshaw, Ph.D., LICSW, School of Social Work, 105 PetersHall, University of Minnesota, 1404 Gortner Avenue, St. Paul, MN 55108; e-mail: bbradshaw@che1.che.umn.edu.

491 2000 Human Sciences Press, Inc.

COMH 36(5) #4344 08-08-00 08:13:27

Community Mental Health Journal492

Recently, there has been renewed interest in psychotherapy withpersons with schizophrenia based on current understandings of psycho-pathology and therapeutic approaches that have tailored interventionsto the unique needs of clients (Wasylenski, 1992). Of particular interestis the use of cognitive behavioral interventions that reduce relapse,enhance coping and improve community adjustment. These emergingapplications of cognitive behavioral methods offer promising new inter-vention strategies in the treatment of schizophrenia (Alford & Correia,1994; Kingdom & Turkington, 1994).

Previous cognitive behavioral treatment (CBT) of schizophrenia hasfocused primarily on modifying delusions and hallucinations in briefoutpatient treatment. Studies report reduction in the frequency andthe strength of belief in hallucinations and delusions through the useof cognitive restructuring techniques (Lowe & Chadwick, 1990; Chad-wick & Lowe, 1990, 1994). Others have found improved coping skills,psychosocial functioning and reduction of symptoms in clients who re-ceived long-term outpatient CBT (Tarrier et al., 1993a,b; Bradshaw,1997).

Results from a series of experimental trials of CBT in the UnitedKingdom have provided additional support for the effectiveness of CBTin reducing psychotic symptomology. While gains were maintained atnine months follow-up, there was no significant improvement in anyother outcome measures (Kuipers et al., 1997a,b; Drury, Birchwood,Cochrane & MacMillan, 1996a,b). These studies targeted psychoticsymptoms and the samples included mood disorders, schizoaffective anddelusional disorders thus making it difficult to determine the impactof CBT on schizophrenia specifically.

CBT has been used as an adjunct to standard treatment of patientsin psychiatric inpatient units and residential programs with reportedreductions in symptomatology, hospitalizations and improved social ad-justment (Kingdom & Turkington (1991, 1994; Perris, 1992). Thesestudies, however, lack designs that control for the multiple treatmentsclients received in these settings. Therefore, what cognitive-behavioralpsychotherapy actually contributed to outcomes in these programs re-mains a question.

This is an important question for several reasons. First, while researchhas supported the effectiveness of outpatient CBT with persons withschizophrenia, it is not the treatment of choice for the vast majority ofpersons with schizophrenia. Second, given the multiple service needsof clients a single treatment modality for schizophrenia is insufficient.Quality care must focus on the integration of multiple treatment per-

COMH 36(5) #4344 08-08-00 08:13:27

William Bradshaw, Ph.D., LICSW 493

spectives (Fenton & McGlashen, 1997). Third, most treatment is pro-vided in community support programs that emphasize milieu treatmentand frequently either do not include individual psychotherapy or offerindividual contact that is not guided by empirically based practice.Fourth, individual psychotherapy may be vital for the long term recoveryof persons with schizophrenia (A Recovery Patient, 1986; Fenton &McGlashen, 1997). Study of the effects of integrating CBT into commu-nity based milieu treatment have important implications for servicedelivery that may enhance the quality and effectiveness of care to per-sons with schizophrenia.

This study examined differences in outcomes of clients who receivedlong-term day treatment services compared to clients who received indi-vidual CBT that was included as part of their DTP treatment. Clientswere randomly assigned to either the treatment control group, DTPtreatment, or the experimental treatment group, CBT/DTP treatment.Research hypotheses included the following: Ho1: CBT/DTP clients willimprove significantly more than DTP clients on measures of psychosocialfunctioning. Ho2: CBT/DTP clients will improve significantly more thanDTP clients on measures of psychiatric symptomatology. Ho3: CBT/DTP clients will have significantly fewer hospitalizations than DTPclients. Analysis of variance with repeated measures was conducted totest these hypotheses.

METHODS

Clients

This study included twenty-four consecutively referred clients who were admitted intothe study after discharge from inpatient psychiatric hospitalization and referral bytheir psychiatrists or case managers for aftercare in a district day treatment program.Criteria for entry into the study included (a) valid diagnosis of schizophrenia based onmeeting DSMIV criteria (b) age between 18–60 (c) persons with mental retardation,organic brain syndrome, or a primary diagnosis of alcoholism or drug abuse wereexcluded. Diagnoses were made by the DTP consulting psychiatrist and by the clinicalsocial workers providing the outpatient psychotherapy. There was 100% concurrenceon diagnosis.

Nine clients dropped out of the study, four from CBT/DTP and five from DTP. Dropoutswere statistically compared to the remaining group members of their assigned groupon the variables of age, education, type of residence, gender, length of illness andbaseline scores on measures of psychosocial functioning and symptomology. They didnot significantly differ from the remaining clients on any demographic variables.

The fifteen clients who completed the study ranged in age from 24–42 (x = 32, sd =7). All were high school graduates and three had up to one year of college study. Ninewere female, six were male, all were Caucasian. The average length of illness was

COMH 36(5) #4344 08-08-00 08:13:27

Community Mental Health Journal494

eleven years (sd = 6). All subjects were unemployed and on disability assistance (SSI).All subjects were on some type of phenothiazine medication, primarily thorazine andhaldol; they were medication compliant. None of the subjects in either group were onnewer neuroleptic compounds such as resperidole, olanzapine or serlect.

Procedures

Control Group: Day Treatment Program. The DTP was a district program in a largemetropolitan area that was designed to provide long term treatment for persons withsevere and persistent mental illness. The program was based primarily on the psychiat-ric rehabilitation model (Anthony, 1980). Clients in the study participated in the pro-gram three days a week from 9:00am–3:00pm. The program consisted of social skillstraining, independent living skills groups, goal groups, occupational and recreationaltherapy, prevocational employment training and medication management. Staff con-sisted of three masters level social workers, three psychiatric nurses, an occupationaltherapist and consulting psychiatrist. Clients were also seen monthly for medicationmonitoring by their psychiatrists and monthly by their county case managers to monitortreatment needs and progress.

Experimental Group: CBT/DTP. Clients in this group participated in the regularDTP activities and were also seen for weekly CBT in the DTP for the duration of thetreatment period. Two of the DTP social workers provided the CBT. They each hadbeen trained in the CBT model, had an average of 10 years experience in mental healthand received weekly supervision of their work by the director of the program.

The CBT model used in this study has been described in detail elsewhere (Bradshaw,1998). The interventions described in each phase were used with all clients. However,the length of each phase varied among clients and there was some overlap betweenphases with the clients.

Phase I: Engagement and Education Months 1–5. The goals of this phase were todevelop the therapeutic alliance regarding roles, goals and rationale of treatment; andto facilitate the client’s understanding of their illness and the process of cognitivebehavioral treatment. Education about schizophrenia included information about etiol-ogy, symptoms, issues in recovery and proper use of medication (Wallace, 1992). Therationale of treatment utilized the normalizing procedure described by Kingdom thatemphasizes vulnerability, stress and coping (Kingdom & Turkington, 1991, 1994). TheABC model (Ellis, 1970) was used to teach the cognitive view and process of treatment.Issues from the client’s daily life were used to highlight the cognitive components offeeling and behavior.

Phase II: Behavioral Treatment Months 6–20. The tasks of the beginning phase oftreatment were to (1) identify reoccurring stresses, signs of stress and prodromal symp-toms of relapse (2) teach behavioral skills to manage stress and physiological arousal.These skills included progressive relaxation, meditation, exercise and managing timeby scheduling activities.

Phase III: Cognitive Treatment Months 20–36. This phase of treatment emphasizedidentification of habitual stressful situations and utilization of cognitive strategies tounderstand and cope with them. Three cognitive interventions were used: thoughtstopping (Wolpe, 1973), cognitive restructuring to promote adaptive coping statementsthat replace negative thoughts and beliefs (Beck, 1984) and positive self-appraisal

COMH 36(5) #4344 08-08-00 08:13:27

William Bradshaw, Ph.D., LICSW 495

training to enhance self-esteem (Bradshaw, 1996). Social skills training was integratedwith cognitive interventions to help clients improve coping with reoccurring stressfulinterpersonal situations.

Outcome Measures

Carpenter and Strauss (1991) have highlighted the importance of using multiple out-come measures in schizophrenia research as a way of tapping the multidimensionalityof psychosocial functioning in schizophrenia. Three outcome variables were used in thepresent study: symptomatology, psychosocial functioning and rehospitalizations.

Symptomatology was measured by the Global Pathology Index of the Hopkins Psychi-atric Rating Scale (Derogatis, 1974). The GPI is an 8 point behaviorally anchored scalethat describes severity of symptoms. Psychosocial functioning was measured by theRole Functioning Scale (RFS) (Goodman, et al., 1993). The RFS is made up of foursubscales: work, social, family and independent living subscales. Each subscale is a 7point behaviorally anchored scale. Rehospitalization was measured by the total numberof psychiatric hospitalizations clients had during the treatment period.

The GPI, RFS and hospitalization data were obtained every six months throughoutthe three year treatment period. Rehospitalization data was obtained from county casemanagement and hospital records. Independent ratings for RFS and GPI were obtainedby the DTP consulting psychiatrist and the clients’ case managers. The psychiatristand case managers received five hours of training in use of the RFS and the GPI. Thecase managers were all masters level social workers and had at least three yearsexperience in mental health. Raters were blind to the subjects’ treatment condition.Interobservor reliability for the GPI was .87 (ICC) and for the RFS .93 (ICC). Power foreach measure was computed assuming an alpha of .05 and a two tailed test (Kraemer &Theiman, 1987). The power for the GPI was .80, for the RFS .70 and for hospitaladmissions .99.

A critical issue in comparisons of treatment groups is fidelity to the treatment model.In order to adequately control for the multiple treatments received by clients a numberof steps were taken to insure that the CBT was the only difference between groups inservices they actually received. Fidelity to the CBT psychotherapy model was monitoredby weekly supervision, review of case materials and periodic review of audiotapes ofsessions. The CBT therapists were also DTP staff. In order to lessen the chance theymight inadvertently add CBT interventions into other DTP services, each of the CBTtherapists was paired with one of the psychiatric nurses as co-leaders of social skillstraining, independent living skills, medication management, and prevocational traininggroups. Each of these groups utilized a treatment manual and fidelity to the protocolof each group was monitored by periodic observation of the groups by use of a one waymirror in the group rooms.

RESULTS

Comparison of clients in the CBT/DTP experimental treatment groupand DTP treatment control group by t-tests and chi-square tests re-vealed no significant differences on demographic variables of sex, mari-tal status, educational level or type of residence. Nor did they differ onthe pretest RFS GPI clinical outcome variables scores.

Analysis of variance with repeated measures was conducted to evalu-

COMH 36(5) #4344 08-08-00 08:13:27

Community Mental Health Journal496

ate the effects of type of treatment and time on three outcome variables:psychosocial functioning, symptomology and rehospitalizations. Thewithin-subjects factors were type of treatment (DTP and CBT/DTP) andtime with three levels (dependent variables scores taken at the end oftreatment years one, two and three). The time effect and treatment xtime interaction effects were tested using the Greenhouse-Geisser test.

Both groups showed marked improvement in psychosocial functioning(RFS) over the three year treatment period. This was highly significantfor the main effect of time (F = 70.05, p < .001). As predicted the improve-ment in functioning was greater for the CBT/DTP group (time × groupinteraction effect F = 6.57, p < .005). Both groups also showed significantreduction in symptomology during the treatment period. This was highlysignificant for the main effect of time (F = 71.01, p < .001). As predictedthe reduction in symptomology was greater for the CBT/DTP group (timeUl group interaction effect (F = 5.20, p < .01). Both groups had markedreductions in the number of rehospitalizations. This was significantfor the main effect of time (F = 7.34, p < .005) and nonsignificant fortime × group interaction (F = 1.73, p < .20)

Results of this study supported hypotheses #1 and #2. Clients inthe CBT/CSP group showed a significant improvement over time inpsychosocial functioning and symptomology when compared to clientsin the DTP group. Data do not support hypothesis #3. There was asignificant pattern of change over time on rehospitalizations, but nosignificant group × time effects between groups for psychiatric rehospi-

TABLE 1

Mean Dependent Variable Scores TreatmentYears 1, 2, 3

CBT/DTP DPT

RFS GPI Readmits RFS GPI ReadmitsTime Mean (sd) Mean (sd)

Year 1 9.38 6.13 5.0 10.85 7.00 2.57(2.33) (1.46) (7.87) (2.73) (.81) (3.82)

Year 2 14.38 4.88 .000 14.00 6.14 2.71(3.02) (1.73) (000) (3.51) (.69) (4.86)

Year 3 20.13 3.50 .000 15.28 5.14 2.29(2.41) (1.19) (.000) (2.43) (.89) (4.27)

COMH 36(5) #4344 08-08-00 08:13:27

William Bradshaw, Ph.D., LICSW 497

TABLE 2

Analysis of Variance GPI, RFS, Readmissions

Sum of MeanSource Squares df Square F Sig

GPITime 57.55 1.86 28.77 71.01 .000Time + Group 4.22 1.86 2.27 5.20 .01Error 10.54 24.13 .44

RFSTime 700.39 1.98 352.54 70.05 .000Time + Group 65.64 1.98 33.03 6.57 .005Error 129.96 25.83 5.03

ReadmissionsTime 2.29 1.89 1.20 7.74 .003Time + Group .51 1.89 .27 1.73 .200Error 1.79 13.00 .138

talizations during the treatment period. The pattern of change betweengroups on all variables was most pronounced between treatment yearstwo and three.

DISCUSSION

This is the first trial of long-term outpatient CBT with persons sufferingfrom schizophrenia. Results of this study add to the growing body ofliterature supporting the use of CBT with this population and providesempirical support specifically for the use of individual CBT with personswith schizophrenia who participate in DTP.

Previous long term studies of supportive and dynamic psychotherapywith schizophrenia showed no significant positive results (Frank &Gunderson, 1990). The CBT used in this study is a departure from pastpsychotherapeutic approaches. The focus of this treatment was on (1)education about illness and illness management (2) the identificationof current stresses and development of coping skills (3) restructuringhabitual negative thoughts and dealing with the impact of illness on

COMH 36(5) #4344 08-08-00 08:13:27

Community Mental Health Journal498

the experience of self. Results of this study indicate that clients withsevere impairments and long histories of illness can benefit from thispsychotherapeutic approach.

Persons receiving CBT/DTP showed significant improvements in psy-chosocial functioning and reduction of symptomatology compared toclients in DTP. The DTP provided a high standard of care using thepsychiatric rehabilitation model and protocols. Particularly noteworthyis that in the CBT/DTP group two clients moved to independent living,three returned to school and three were employed part-time in main-stream jobs; no clients in DTP made these changes. Regarding symptom-atology 50% of the CBT/DTP group had a mild level of symptoms onthe GPI (a score of 3 or less) at the end of the study period while 0%of the DTP group achieved that level of symptomatology. Although therewas no statistically significant time × group effect regarding hospitaliza-tions, the small sample size may have impeded statistical detection ofan effect.

What can we make of the dramatic differences between groups in yeartwo to year three? While stabilization of symptoms can occur relativelyquickly (Drury et el., 1996) recovery is a longer process. Treatment inyears two and three focused on helping clients integrate the experienceof illness, expand self-concepts and strengthen self-esteem through theuse of cognitive restructuring. This suggests the importance of helpingpersons explore the impact of illness on their sense of self and supportthe development of new meanings, interactions and views of one’s selfas a person with illness (Davidson & Strauss, 1992; Bradshaw, 1997).It also highlights the important role psychotherapy may play in thelong-term recovery of persons with schizophrenia.

While the results of this pilot study are promising a number of limita-tions should be noted. First, the study needs to be replicated with largersamples. All clients in the study were white, high school graduates andhad high rates of attendance in treatment. These findings may notgeneralize to other ethnic groups, less educated, compliant or substanceabusing groups. Second, there is the potential of multiple treatmentinterference in that all clients had ongoing contact with a psychiatristand received medication. This contact was limited to brief monthly visitsthat were medication checks and it is unlikely these visits confoundedthe design to any significant degree. All clients were on neurolepticmedication over the course of the study and were reported as medicationcompliant. However, the issue of medications was complicated in a threeyear study. While medications were monitored, not all clients were ableto be kept on the same regime and some results may reflect medication

COMH 36(5) #4344 08-08-00 08:13:27

William Bradshaw, Ph.D., LICSW 499

effects. No clients were switched to newer neuroleptics and other re-search has suggested that within medication compliant samples, medi-cation was not a predictor of outcome (Garety et al., 1997). Futureresearch could use dismantling approaches to identify the critical thera-peutic interventions.

The last two decades have witnessed increased awareness of the multi-ple treatment needs of persons with schizophrenia. The question of whatcombinations of treatment enhance outcomes for which clients and inwhat phase of illness recovery is a critical emerging issue for researchersand practitioners (Brekke, Long, Nesbit & Sobel, 1997).

This question is particularly important regarding the use of CBT withpersons with schizophrenia. Continued exploration of ways to integrateCBT with other psychosocial interventions in a variety of communitycare contexts is needed. Protocols of care that utilize CBT need to beincorporated into treatment guidelines for schizophrenia. Similarly stafftraining modules are needed to facilitate integration of CBT into commu-nity care settings. Identifying the types of clients that are responsive toCBT, the timing of interventions and ways of integrating and evaluatingCBT in different contexts of care have the potential to enhance thequality and effectiveness of services for persons with schizophrenia.

REFERENCES

A Recovery Patient (1986). Can we talk: The schizophrenic patient in psychotherapy. AmericanJournal of Psychiatry, 86, 1468–70.

Alford, B. & Correia, C. (1994). Cognitive therapy of schizophrenia: Theory and empirical status.Behavior Therapy, 25, 17–33.

Anthony, W. (1980). Principles of Psychiatric Rehabilitation. Baltimore: University Park Press.Beck, A. (1984). Cognitive approaches to stress. In R. Woolfolk & P. Lehren Eds.), Principles and

Practice of Stress Management (pp. 125–155). New York: Guilford.Bradshaw, W. (1998). Cognitive-behavioral treatment of schizophrenia: A case study. International

Journal of Cognitive Psychotherapy, 12, 1, 13–26.Bradshaw, W. (1997). Evaluating cognitive-behavioral treatment of schizophrenia: Four single-

case studies. Research on Social Work Practice, 7, 419–445.Brekke, J., Long, J., Nesbit, N. & Sobel, E. (1997). The impact of service characteristics on functional

outcomes from community support programs for persons with schizophrenia: A growth curveanalysis. Journal of Consulting and Clinical Psychology, 65, 464–475.

Chadwick, P. & Lowe, C. (1990). Measurement and modification of delusional beliefs. Journal ofConsulting and Clinical Psychology, 58, 225–232.

Chadwick, P. & Lowe, C. (1994). A cognitive approach to measuring and modifying delusions.Behavior Research and Therapy, 32, 355–367.

Davidson, L. & Strauss, J. (1992). Sense of self in recovery from severe mental illness. BritishJournal of Medical Psychology, 65, 131–145.

Derogatis, L. (1974). The Hopkins symptom inventory. Behavioral Science, 9, 1–15.Drury, V., Birchwood, M., Cochrane, R. & MacMillan, F. (1996a). Cognitive therapy and recovery

from acute psychosis: A controlled trial I. Impact on psychotic symptoms. British Journal ofPsychiatry, 169, 593–601.

COMH 36(5) #4344 08-08-00 08:13:27

Community Mental Health Journal500

Drury, V., Birchwood, M., Cochrane, R. & MacMillan, F. (1996b). Cognitive therapy and recoveryfrom acute psychosis: A controlled trial II. Impact on recovery time. British Journal of Psychia-try, 169, 602–607.

Ellis, A. (1970). The Essence of Rational Emotive Psychotherapy: A Comprehensive Approach toTreatment. New York: Institute for Rational Living.

Fenton, W. & McGlashen, T. (1997). We can talk: Individual psychotherapy for schizophrenia.American Journal of Psychiatry, 154, 1493–1494.

Frank, A. & Gunderson, J. (1990). The role of the therapeutic alliance in the treatment of schizophre-nia: Relationship to course and outcome. Archives of General Psychiatry, 47, 228–236.

Garety, P., Fowler, D., Kuipers, E., Freeman, D., Dunn, G., Bebbington, P., Hadley, C. & Jones,S. (1997). London-East Anglia randomised controlled trial of cognitive-behavioral therapy forpsychosis: II. Predictors of outcome. British Journal of Psychiatry, 171, 420–426.

Goodman, S., Sewell, D., Cooley, E. & Leavitt, N. (1993). Assessing levels of adaptive functioning:The role funtioning scale. Community Mental Health Journal, 29, 119–131.

Kingdom, D. & Turkington, D. (1991). The use of cognitive behavior therapy with a normalizingrationale in schizophrenia. Journal of Nervous and Mental Disease, 179, 207–211.

Kingdom, D. & Turkington, D. (1994). Cognitive Therapy of Schizophrenia. New York: GuilfordPress.

Kuipers, E., Garety, P., Fowler, D., Dunn, G., Bebbington, P., Freeman, D. & Hadley, C. (1997a).London-East Anglia randomised controlled trial of cognitive-behavioral therapy for psychosis:I. Effects of the treatment phase. British Journal of Psychiatry, 319–327.

Kuipers, E., Garety, P., Fowler, Chisholm, D., Freeman, D., Dunn, G., Bebbington, P. & Hadley,C. (1997b). London-East Anglia randomised controlled trial of cognitive-behavioral therapyfor psychosis: III. Follow-up and economic evaluation at 18 months. British Journal of Psychia-try, 173, 61–68.

Lowe, C. & Chadwick, P. (1990). Verbal control of delusions. Behavior Therapy, 21, 461–479.Perris, C. (1992). Integrating strategies in the treatment of young severely disturbed patients.

Journal of Cognitive Psychotherapy, 6, 205–213.Tarrier, N., Beckett, R., Harwood, S., Baker, B., Yusopoff, L. & Ugarteburu, I. (1993). A trial of

two cognitive-behavioural methods of treating drug-resistant residual psychotic symptoms inschizophrenic patients: I. Outcome. British Journal of Psychiatry, 162, 524–532.

Tarrier, N., Sharpe, L., Beckett, R., Harvood, S. & Yusopoff, L. (1993). A trial of two cognitivebehavioural methods of treating drug-resistant residual psychotic symptoms in schizophrenicpatients: II. Treatment specific changes in coping and problem-solving skills. Social Psychiatryand Psychiatric Epidemiology, 28, 5–10.

Wallace C., Liberman, R. MacKain, S., Eckman, T. & Blackwell. (1992). Effectiveness and replicabil-ity of modules for teaching social and instrumental skills to severely mentally ill. AmericanJournal of Psychiatry, 149, 654–658.

Wasylenski, D. (1992). Psychotherapy of Schizophrenia Revisited. Hospital & Community Psychia-try, 43, 123–127.

Wolpe, J. (1973). The Practice of Behavior Therapy. New York: Jason Aronson.

COMH 36(5) #4344 08-08-00 08:13:27