Journal of Mental Health (1999) 8, 6, 569 - PDF document

Journal of Mental Health (1999) 8, 6, 569±582ISSN 0963-8237print/ISSN 1360-0567online/99/060569-14 € Shadowfax Publishing and Taylor & Francis Ltd Address for Correspondence: Professor Nicholas Tarrier, Department of Clinical Psychology School ofPsychiatry and Behavioural Sciences, Ground Floor, Research and Teaching Building, Withington Hospital, WestDidsbury, Manchester M20 8LR, UK. Tel: 0161 291 4319; Fax: 0161 291 3814; E-mail: ntarrier@fs1.with.man.ac.uk The dissemination of innovative cognitive±behaviouralpsychosocial treatments for schizophreniaNICHOLAS TARRIER, CHRISTINE BARROWCLOUGH, GILLIAN HADDOCK &JOHN McGOVERNDepartment of Clinical Psychology School of Psychiatry and Behavioural Sciences,University of Manchester, UKAbstractThere has been considerable research in recent years that has suggested that non-drug psychosocialinterventions have considerable benefits to patients suffering from psychoses. These interventionsinclude family interventions, individual cognitive±behaviour therapy and early signs monitoring. Inspite of these research findings the dissemination of these interventions into routine practice has beenslow and patchy. This paper briefly reviews these research studies and investigates reasons whydissemination of such evidence-based practice has not progressed. The absence of skills in the mentalhealth workforce is one reason for the slow implementation of treatment innovations. The attemptsto skill sections of the workforce are described and the relative success of the various training projectsis described. The difficulties and limitations of these attempts are discussed.IntroductionThe question of how mental health servicesare improved is enigmatic. In theory, re-search should provide information aboutwhich new treatments are efficacious. Theseshould then be disseminated into clinicalpractice and services be organised orconfigured to accommodate them and facili-tate their delivery. However, the `provisionof mental health services bears little relationto research' whereas `psychiatry incorpo-rates many commonly used treatments whosecomparative effectiveness is equivocal atbest.' (Anderson & Adams, 1996). Why areempirically validated treatments not becom-ing widely available while less well sup-ported treatments remain common practice?Historically, treatments for severe mentalillness such as schizophrenia have beenthought of as being solely a biological pre-serve. Over the last decade there have beenvery positive advances in the development ofnon-drug, or psychosocial treatments. De-spite these developments the implementationof psychosocial treatments as routine withinmental health services has been patchy. Whatthen is necessary for these research-validatedtreatments to become established in clinicalpractice and be available to those who wouldbenefit from them? There may be a numberof barriers to this process such as: a lack of a 570Nicholas Tarrier et al.partnership between researcher and clinician(Goldfried & Wolfe, 1996; Barlow, 1996),absence of appropriate knowledge and/orclinical skills within the relevant workforce(Lancashire et al., 1996); characteristics ofthe organisation, institution or workplace inconstraining such new developments(Bernstein, 1982; Corrigan & McCracken,1995a,1995b) or the fact that clinical tech-niques are more difficult to learn and imple-ment than is generally realised (Wilson, 1997).Furthermore, the time allowed for post-quali-fication training in combination with prob-lems in existing levels of theoretical under-standing can result in staff only acquiringslight or partial knowledge of the completetreatment that has been validated by research.Partial knowledge and incomplete skills willbe less resistant to these barriers to imple-mentation. To progress from innovation topractice it is necessary that new treatments beseen as clinically relevant, be empiricallyvalidated, meet prioritised needs and can betaught in way that is accessible enough forstaff to readily and faithfully acquire theknowledge and skills (Liberman & Eckman,1989).In this paper we aim to review briefly therecent innovations in non-pharmacologicaltreatments of psychotic disorders and theissues surrounding their dissemination intoroutine clinical practice. We will specificallydeal with the area of training of professionalstaff as a method of dissemination and wewill concentrate on the work of the Manches-ter group with which we have been involved.Recent developments in psychosocialtreatments of schizophreniaRecent interest in psychosocial treatmentsfor psychotic disorders has been the result ofa number of different influences. Theseinclude: the wider acceptance of the vulner-ability±stress model of schizophrenia; theresearch on Expressed Emotion (EE) of rela-tives and its effect on the course of schizo-phrenia; the failure of medication to totallyeradicate psychotic symptoms and preventrelapse; and a general rise in the consumermovement in mental health and a desire formore needs led services (Tarrier &Barrowclough, 1990). Psychosocial treat-ments can be divided into three general cat-egories which have a separate set of clinicalprocedures and aims. These are: (1) familyintervention; (2) cognitive-behaviour therapyfor psychotic symptoms; and (3) early signsmonitoring and early intervention. Each ofthese areas and the empirical research sup-porting their efficacy will be reviewed briefly.Family interventionFamily interventions were developed fromthe research on EE in which studies consist-ently reported that patients who returned tolive with relatives rated as high on EE had amuch greater relapse rate than those whoreturned to live with relatives assessed as lowEE (Kavanagh, 1992). It was reasoned that,if EE, which was assumed to reflect behav-iour of the relative towards the patient, couldbe changed through intervention then the riskof relapse would be reduced. Interventionswere designed independently by a number ofgroups with the aim of reducing stress in thehome environment. These interventions, al-though differing in some aspects, have manycommon features, such as including an edu-cational component to provide the familywith information about schizophrenia, adopt-ing a practical problem oriented approachand assisting the family to better cope withthe difficulties of living with a family mem-ber who has schizophrenia.Rigorous controlled trials of family inter-ventions have been carried out (see Mari &Streiner, 1994; Penn & Mueser, 1996 for Cognitive±behavioural treatments for schizophrenia 571reviews). Typically these studies have re-cruited families while the patient is experi-encing an acute episode and the interventioncommenced at hospital discharge and contin-ued over a 9±12-month follow-up periodwith the principle aim of preventing relapse.Relapse rates are then compared betweenfamilies who received the family interven-tion as an adjunct to routine care with thosewho received routine care alone. Routinecare would include the use of prophylacticmedication. The consistent finding of theseclinical trials was that relapse rates weresignificantly reduced by family interventionover the follow-up period. Some studieshave also indicated that after 2 years therewas still a significant reduction in relapserates. In our own trial in Salford, patientswere followed up at 5 and 8 years; althoughrelapses did accumulate in all groups overthis period, patients who had received a 9-month family intervention still had a signifi-cantly lower relapse rate compared to theappropriate control group (Tarrier et al., 1994)The effects of family interventions on otheroutcomes beside relapse are less clear. Thereis some evidence from the studies by Falloonand colleagues and Tarrier and colleaguesthat significant but modest improvements inthe patient's social functioning were achieved(Falloon et al., 1984; Boyd & McGill, 1985;Barrowclough & Tarrier, 1990). Falloon etal. (1985) also reported a decrease in theburden of care of family members. TheCalifornia and Salford studies also reportedthat family interventions result in a financialsaving (Cardin et al., 1986; Tarrier et al.,1991).In summary, there is good and consistentevidence from methodologically sound clini-cal trials that family intervention in conjunc-tion with prophylactic medication reducesrelapse rates in schizophrenia at least in theshort and medium term when families arerecruited during an acute crisis. Attemptingto recruit families while the patient is inremission and when the family has a reason-ably stable situation may be more problem-atic (McCreadie et al., 1991; Tarrier, 1991)with non-engagement rates of over 50% re-ported in service settings (Smith, 1992).Cognitive±behavioural therapy forpsychotic symptomsA number of case studies and case serieshave been published over the last few dec-ades on the use of various cognitive andbehavioural interventions in the treatment ofhallucinations and delusions but it was notuntil the late 1980s that there was a prolifera-tion of research into the efficacy of cogni-tive±behavioural treatments for psychotic dis-orders (see Haddock & Slade, 1996; Had-dock et al., 1998 for reviews). These treat-ments were initially developed to treat pa-tients suffering from a chronic psychotic con-dition, mainly schizophrenia, who were ex-periencing persistent delusions and halluci-nations despite medication.A number of methods have been describedsuch as: cognitive therapy with a normalisingrationale (Kingdon & Turkington, 1991);focusing (Bentall et al., 1994); coping strat-egy enhancement (Tarrier et al., 1993); andcomprehensive cognitive±behaviouraltherapy (Garety et al., 1994).Three research groups in the UK have, orare in the process of, reporting results of largeand rigorous controlled trials (Kuipers et al.,1997; Kingdon, 1997; Tarrier, 1987; Tarrieret al., 1998). So far these studies have pro-duced considerable interest and optimismand there appears to be a consistent findingthat chronic psychotic and affective symp-toms can be reduced using cognitive±behav-ioural treatments. Their impact on reducingnegative symptoms or improving social func-tioning has been limited, although Tarrier 572Nicholas Tarrier et al.and colleagues (Tarrier, 1987; Tarrier et al.,1998) reported significant improvements innegative symptoms in patients receiving CBT.The available evidence from limited follow-up studies indicates that the clinical benefitsare durable, at least in the short term of 12 to18 months (Kuipers et al., 1998; Tarrier et al.,1999). Thus, the evidence for individualcognitive±behavioural treatments is currentlyweaker than for family interventions and thelong-term benefits unknown.A management problem being encounteredincreasingly is the dual diagnosis patient.These are patients diagnosed as sufferingfrom a psychotic illness who also have aserious substance abuse problem. Although,a project is under way in Manchester evaluat-ing a combination of psychosocial treatments,family intervention, cognitive±behaviourtherapy and motivational interviewing, littlecan be said about the established effective-ness of these techniques at present(Barrowclough et al., 1996).Research into CBT for acute recent onsetpsychosis has been limited, although a recentstudy suggested that CBT could be effectivein acutely ill patients by speeding recoveryand reducing time spent in hospital (Drury etal., 1996a; 1996b). The Manchester groupfailed to replicate this finding with the effectsize and sample predicted from the Drurystudy (Haddock et al., 1999)To further evaluate the effects of cognitive-behaviour therapy used during an acute epi-sode a multi-site trial of cognitive±behaviourtherapy for recent onset schizophrenia, termedthe SOCRATES Trial (Lewis et al., 1996) iscurrently under way. This is to evaluatewhether cognitive±behaviour therapy is su-perior to supportive counselling and routinecare alone in speeding recovery and prevent-ing subsequent relapse in recent onset schizo-phrenic patients hospitalised for an acuteillness episode.Early interventionThere has been considerable enthusiasmfor the general principle of early interven-tion, with some service projects organisedspecifically for this purpose, notably theBuckingham project (Falloon et al., 1966)and the EPPIC project in Melbourne(McGorry et al., 1996). Clinical interventionimplemented on the detection of prodromalsigns of relapse with a view to preventing oraborting relapse has been described withschizophrenic patients (Birchwood et al.,1989) with a reduction in relapse and re-hospitalisation (Birchwood, 1996). A simi-lar strategy has been evaluated in a randomisedcontrolled trial with patients suffering frommanic depressive psychosis which resultedin significant benefits, especially in reduc-tion of manic episodes and resultant days inhospital (Perry et al., 1999). Although earlyintervention is a plausible strategy for relapseprevention, robust empirical evidence for itsefficacy is currently lacking.Dissemination and training inpsychiatric rehabilitationInnovations face many barriers to imple-mentation. Dissemination through the tradi-tional methods such as publication and pro-fessional meetings may have little impactupon practitioners. There is little evidencethat attendance at professional meetings orshort educational courses have resulted inany change in clinical practice or services(Liberman & Eckman, 1989). There is con-siderable evidence to suggest that the acqui-sition of clinical skills requires active andpractical training in those skills rather thanlecture-style didactic teaching.Bernstein, in her influential 1982 concep-tual review of training, suggested a func-tional eco±behavioural framework for inves- Cognitive±behavioural treatments for schizophrenia 573tigating training effectiveness in which abehavioural±analytical approach was inte-grated into a perspective which had socialand ecological validity (Bernstein, 1982).She was interested in the dissemination ofskills in behaviour modification in general,but many of the points she made have rel-evance to the implementation of new devel-opments in schizophrenia. Bernstein sug-gests that the traditional interest in producingchange in an individual patient was to focuson the staff who are in contact with thepatient. Thus, the questions of interest werehow to teach use of behavioural change skillsor how to maintain use or generalisation ofthose skills. This assumes that the target staffcould be taught those skills and would be ableto implement them to a satisfactory standard.This approach also ignores the interactionwith various ecological systems that wouldfacilitate or inhibit the use of these skills.Difficulties arising from these issues are stillfaced in the area of staff training today andhave been incompletely addressed. Bernsteinproposed a model that would enhance thedelivery of behavioural skills, termed theBehavioural Service Delivery Model. Sheexamined the function of various roles throughhow they interacted with different parts of theecological system. Functions such as `pro-gramme implementers', `programme de-signer' and `trouble shooter/resource pro-vider' could be described and assigned, al-though depending on the situation these dif-ferent functions would not necessarily corre-spond with different individuals. The impor-tant aspect is that the system could be ad-dressed at different levels so as to maximisethe implementation of the programme. Dis-semination of a particular procedure wouldbe more probable if it was possible to nego-tiate access through the various systems. Thusto deliver a new psychological treatment to apatient suffering from psychosis requires notjust the ability to deliver that treatment but tobe able to negotiate with those who willbenefit from the treatment and those whowork within the mental health service at alllevels.Milne (1984) addressed some of these is-sues in a short in-service course on behaviourtherapy for psychiatric nurses on a long-stayward. Paying attention to the ward environ-ment as a setting event in which the newlyacquired skills would occur, he attempted tocompensate for particular difficulties in thatenvironment that would impede the use ofthese skills. However, implementation ratesat 26% were still disappointingly low.A similar ecological theme can be seen inwriting of Corrigan and colleagues (Corrigan& McCracken, 1995a, 1995b; Corrigan et al.,1997) who distinguish between an educa-tional and an organisational model of staffdevelopment. They provide evidence thatskills based on social learning principles couldbe successfully taught to staff in psychiatricrehabilitation settings but that follow-up stud-ies indicated that organisational barriers wereimpeding their introduction and maintenance.They propose a set of organisational develop-ment strategies to overcome these barriers.Although these strategies were aimed tochange the organisation as a whole, the Cor-rigan group's main point appears to be thatchange to the practice of clinical teams is bestassured by conducting training with the wholeteam. This training will most probably suc-ceed if it targets methods the team perceivesas relevant to their clinical goals, and there isgood social support between members (Cor-rigan et al., 1997).Training in psychosocialinterventionsThere has been a long tradition in Man-chester of developing programmes for train- 574Nicholas Tarrier et al.ing staff in behavioural methods and in evalu-ating the effectiveness of such programmes.For example, the Hester Adrian ResearchCentre at Manchester University had devel-oped the EDY project to train educationalpsychologists to train teachers of people withlearning disabilities in behavioural methodsas a means of disseminating the behaviouralapproach (McBrien & Foxton, 1982). Simi-larly Barrowclough & Fleming have devel-oped training packages for training staff work-ing with elderly people (Barrowclough &Fleming, 1986a) and have undertaken evalu-ation of such training methods (Barrowclough& Fleming, 1986b). From this training tradi-tion, programmes of psychosocial interven-tions for mental health professional staff de-veloped in the Manchester area and an inter-est and a demand for training began to grow.The Preston projectThe first training programme was devel-oped to train social workers in Preston infamily intervention (Tarrier et al., 1988).Experience has suggested that although the`one-off workshop' may be a useful way ofcreating interest in a new therapeutic ap-proach it would be ineffective in disseminat-ing clinical skills and in changing clinicalpractice. Thus to maximise the acquisition offamily intervention skills by the trainees thetraining programme was based on clinicalsupervision of case work. The initial trainingconsisted of 5-day workshops over a 6-weekperiod, with one-and-a-half days of follow-up sessions over the subsequent 2 months and10 days over the subsequent 9 months, con-cluding with a 2-day summary workshop. Anumber of conditions were requested beforetrainees enrolled on the course; these were:(1) that each participant had a co-workerfrom the same workplace, so that all interven-tions were carried out in pairs; (2) that therebe agreement from the managers of eachtrainee-pair that the course case-work wouldbe given high priority so that adequate timecould be given to carry out the family inter-vention, hence avoiding the problem of pres-sure of work not allowing sufficient time forthe intervention; (3) that each trainee-pairshould have identified and attempted to en-gage a patient suffering from schizophreniaand their relative prior to the commencementof training; (4) that each trainee-pair informedtheir professional colleagues about the natureof the intervention so as to facilitate theirsupport for their work.The content of the training course wasbased on the Salford Family Interventionprogramme (see Barrowclough & Tarrier,1992) and followed a modular format inwhich module training was followed by su-pervised practice. Training began with back-ground material followed by instruction inthe use of assessments with patients andfamilies. Subsequent modules consisted ofproviding education and knowledge aboutschizophrenia, stress management, copingskills and goal setting with relatives andpatients. Following each didactic module thetrainees were required to implement thatmodule in the clinical setting and presentcase material for supervision. Each modulewas taught through instruction and then role-play with attention given to the trainees pre-senting their case work and to using the largertrainee group as a process whereby issues andproblems were discussed so as to generatesolutions. No formal evaluation of the train-ing was carried out but written feedback fromtrainees was generally very positive and fami-lies indicated that they had been helped andreceived benefit from the new approach be-ing used by their social workers. We weresufficiently satisfied with the training courseand the feedback that had been received toregard this as a good model for trainingprogrammes. Cognitive±behavioural treatments for schizophrenia 575The Manchester community psychiatricnurse training programmesWorkers at the University of Manchesterset up two training programmes of familyinterventions that were evaluated in moredetail. The drift by community psychiatricnurses away from working with patients withserious mental illness towards treating pa-tients with much more minor and transientconditions within a primary care setting wasviewed as resulting from a feeling amongCPNs that there was little that could be donepsychotherapeutically for patients withschizophrenia (Brooker, 1990). Thus it wasargued that helping CPNs acquire skills infamily interventions might motivate them toreturn to working with patients with schizo-phrenia (Brooker & Butterworth, 1993).In the initial study (Brooker et al., 1992)nine CPNs were selected onto the trainingcourse, each was matched with a colleague atthe same workplace on a number of variablesincluding age, gender, length of experienceas a CPN and post-basic training undertaken.The matched CPNs acted as a control groupin a quasi-experimental design. All CPNsreceived training in assessments but only theexperimental group received training in fam-ily interventions. Each CPN aimed to recruitthree schizophrenic patients and their fami-lies into the study for which they would act asa key worker. Initially 87% of the targetnumber of 54 families were recruited, al-though 17 families dropped out before the 1-year follow-up. The 30 families who com-pleted the trial represented 64% of the re-quired sample. Comparison of the samplewith other English family intervention stud-ies indicated that they were comparable. Thestudy drop-outs had significantly more ad-missions and days in hospital than the studycompleters. Comparison between the pa-tients of the experimental and those of thecontrol CPNs did not reveal significant dif-ferences in demographic or clinical variablesnor were there differences between charac-teristics of their relatives. The training pro-gramme was delivered by Tarrier &Barrowclough and followed a similar butextended version of the Preston training pro-gramme.The patients in the experimental groupshowed significant improvements on four ofthe eight subscales of the KGV (Krawiecka etal., 1977); depression, anxiety, delusions andpsychomotor retardation at post-treatmentand 1-year follow-up. Control patientsshowed a significant improvement at post-treatment only. There were significant im-provements in social functioning, which weremaintained at follow-up in the experimentalgroup patients but no change in the controlgroup patients. Comparisons between pa-tients on medication dosage and admissionsto hospital indicated no difference betweenthe groups. Comparisons of variables associ-ated with the relatives indicated that therewere significant improvements in the rela-tives' satisfaction with the patient's personalfunctioning, improvements in the relatives'mental health and improvements in a meas-ure of consumer satisfaction with services.No significant changes were found in any ofthese measures in families treated by thecontrol CPNs. The relatives treated by thecontrol CPNs indicated a significant decreasein their satisfaction with the emotional sup-port provided by the CPNs.In a second study, Brooker and colleagues(Brooker et al., 1994) adopted a prospectivequasi-experimental design in which each CPNacted as their own control. Ten CPNs wereselected onto the training course of whichtwo subsequently dropped out. Each CPNaimed to identify three experimental familiesand three control families. Initially a sample 576Nicholas Tarrier et al.of 60 families was anticipated but with theloss of two CPNs to training the potentialsample size was reduced to 48. Forty-onefamilies, 85% of the target sample were re-cruited to the trial of which 34 (83%) re-mained in the study until the 12-month fol-low-up. A within-subject design was used sothat patients were assessed and then managed`as usual' by the CPN for 6 months to providea baseline period. After 6 months baselinethe experimental group received family in-tervention while the control group continuedto receive treatment as usual. After the studyperiod of 12 months the control group werethen offered family intervention. Of the 24control families one had dropped out beforethe end of the 12-month study period and 15of the remainder took up the offer of familyintervention. The remaining eight consti-tuted a control subgroup who never receivedfamily intervention. There were no signifi-cant differences between the control and ex-perimental patients or their relatives. Thetraining was organised by Falloon and hiscolleagues and follows their manual (Falloonet al., 1984) and consisted of organisinghousehold meetings, family education, com-munication training and family problem solv-ing. In a similar way to the first trial thetraining course was orientated towards ac-quisition of clinical skills which were taughtthrough role play and clinical supervision.The results indicated that during the base-line period there were no changes in positiveor negative symptoms of psychosis. Theexperimental group showed a significant re-duction of positive and negative symptomsover the time that they received family inter-vention. Similarly, the control group showedno change over the first 12 months of thestudy but those families who accepted familyintervention at 12 months showed a signifi-cant decrease in both positive and negativesymptoms at 12- month follow-up. A similarpattern was seen in improvements in themeasure of social functioning. In families inboth the experimental group and the controlgroup who went on to receive family inter-vention, there was a dramatic decrease in themean number of days spent in hospital whenthe period that they did not receive familyintervention was compared with the periodthat they did (experimental group: no-FI 18.4days, FI 1.8 days; control group: no-FI 24.2,FI 3.1 days). In contrast the control subgroupthat did not receive the offered family inter-vention showed a marked increase in meandays in hospital (23.5±73.5 days). It could beargued that the latter group was highly self-selected and their poor outcome owes moreto some characteristic of treatment refusersrather than solely a result of not receivingfamily intervention. The poor outcome inthis group of treatment refusers closely re-sembles those reported for treatment refusersin other family intervention studies (Tarrier,1991).Both these studies of the Brooker groupsuffer from considerable methodologicalshortcomings. There was neither blind norindependent assessment of patients or rela-tives; and families were not randomly allo-cated to treatment but could be selected intothe study thus introducing a bias in favour ofthe experimental group. In neither of thestudies did the statistical analysis adequatelyaddress the cluster design (Simpson et al.,1995; Bland & Kerry, 1997). Moreover, inthis second study it is assumed that CPNswould treat the control families in the usualmanner while they learnt and implementednew treatments with the experimental fami-lies, thus ignoring the potential for transfer ofbenefit to the control families. Because ofthese severe methodological limitations thesestudies are better thought of as preliminaryindications of the benefits of training ratherthan substantial supportive evidence. Cognitive±behavioural treatments for schizophrenia 577The New South Wales studyIt is worth examining another study carriedout in Australia by Kavanagh and his col-leagues (Kavanagh et al., 1993) which had avery different outcome to the Brooker stud-ies. Kavanagh trained mental health workersin the Sydney area in a cognitive±behav-ioural approach to family intervention. How-ever, despite considerable effort by the train-ing team the training appeared to have littleimpact upon clinical practice with few of thetrainees actually attempting to engage fami-lies for any length of time. Trainees receiveddidactic and workshop training of approxi-mately 30±35 hours duration after which theywere asked to participate in a controlled evalu-ation of family intervention by acting as trialtherapists. Initially, 160 therapists receivedtraining but only 44 of these elected to takepart in the treatment trial and 28 of these sawonly one family. In fact, 57% of the familiesin the study were seen by only six of thetherapists. Therapists reported particular dif-ficulty in integrating the family work withtheir other duties and interests. In an evalu-ation of the training programme only 4% ofthe sample reported that their knowledge ofcognitive±behavioural approaches was a sig-nificant problem, but in a written test mosttherapists failed to demonstrate even the mini-mum recall of the material they had beentaught (Kavanagh et al., 1993). They con-clude: `as a demonstration that the structuredfamily intervention could be disseminatedinto routine community health practice, theproject clearly was of limited success'. Thissuggests that skills in psychosocial treat-ments may be considerably more difficult toacquire than was anticipated and that `front-line staff' may be unaware of the complexi-ties of what they are being taught.Beside the possible differences in attitudeand practice between Manchester and NSWthere are two possible reasons why the train-ing programme of Kavanagh and colleaguesyielded such poor results. Successful train-ing requires continuity and progressive clini-cal supervision, and time-limited didacticand workshop teaching is unlikely to result inskill acquisition in the absence of guidedpractice. It is probable that without supervi-sion within a structured teaching programmetrainees will very quickly abandon the newapproach, failing either to acquire the skill ornot perceiving it as priority for meeting im-portant clinical goals. The second importantfactor is the necessity for management com-mitment for the training and practice of thenew approach. In Kavanagh's programmethere appears to have been little in the way ofmanagement commitment to the reorganisa-tion of workloads so that family interventioncould be given a high priority. It appears thatfor most trained therapists in this study fam-ily work was just one more activity that theyhad to compete in an already overloadedschedule. It is perhaps not surprising thatvery few therapists attempted to implementwhat they had been taught without manage-ment prioritising of their family work, andclinical supervision to shape their practice.The Thorn Nurse Training ProjectIn 1991 the Sir Jules Thorn CharitableTrust commenced supporting specialist men-tal health nurse training in London and Man-chester. The broad aim of the project was totrain CPNs in problem oriented psychosocialand psychological interventions with the se-riously mentally ill. The project was fundedby the Trust for 3 years and subcommitteeswere constituted to cover the development ofthe curriculum and to carry out an evaluationof the effectiveness of the training. Thecontent of the course followed a broadlysimilar content at both sites with the reflec-tion of local strengths and interests. At theend of the 3-year period funding was secured 578Nicholas Tarrier et al.from health service sources and at the time ofwriting a tendering process for training inpsychosocial interventions was under way.There was also considerable interest fromother professions besides nursing and since1994 the training has become open to traineesfrom other professional backgrounds and therehas been a general move to multi-disciplinarytraining. In the next section the Thorn train-ing at Manchester will be described in greaterdetail.Thorn training in ManchesterThe University of Manchester Thorn train-ing consisted of a 1-year diploma coursewhich covered three main subject areas: casemanagement and assessment, cognitive±be-havioural family interventions and individualcognitive±behavioural interventions. Train-ees received 36 whole days formal teachingover the year and carried out equivalentnumber of days in clinical practice with se-lected patients who suffer from serious men-tal illness and their families. The first mod-ule, case management, consists of: literatureand government legislation surrounding casemanagement; an up-to-date review of theliterature relating to schizophrenia and otherforms of serious mental illness; training inassessment of the multiple needs of psychoticpatients and their families and monitoring ofmedication and its side effects. Modulesfollow this on family and individual cogni-tive±behavioural interventions, which runconcurrently. Teaching includes didacticteaching, group work and clinical supervi-sion.The Thorn project has been subject to evalu-ation since its inception. Initial results showedthat trainees increased their knowledge andskills during training and patient outcomeimproved following the Thorn intervention(Lancashire et al., 1996). However, thisevaluation is limited in that patients wereselected by trainees for the intervention andtheir assessment was neither independent norblind. In this published report there was nocomparison or control group with which tocontrast the patients receiving the trainees'intervention. However, these limitations areunderstandable in what was an evaluation ofprofessional training and not a clinical trial.Further, analysis of audio-taped therapysessions indicated that cognitive±behaviourtherapy skills improved over training but didnot change over time in a control group ofpsychiatric nurses who did not receive train-ing. However, the level of skill achieved inthe nurses who did receive training was stillquite modest (Haddock et al.,submitted).Training in situThere are considerable advantages of for-mal and structured training courses such asThorn, such as disseminating research intoclinical practice and creating a better trainedand relevantly skilled workforce. However,there are also a number of weaknesses in thismode of training where the majority of theteaching takes place away from the workplace. Not the least of these is that staffreceiving training have to leave their work-place to attend the training thus requiring thattheir service input is somehow covered or notreplaced. The argument is that this is a shortterm loss which must be balanced by thelong-term gain of better-trained personnelwho will be able to implement interventionsthat will improve clinical and social out-comes and reduce service costs. Moreover,although the personnel who become trainedhave increased skills there may be impedi-ments to them utilising these skills such aschange in jobs, increased seniority resultingin more administration and less clinical time;and varying clinical and business priorities.An alternative method of dissemination is Cognitive±behavioural treatments for schizophrenia 579to investigate the patient and carer needsrather than the needs of the general workforce.We are currently carrying out an evaluationof this method of dissemination in Manches-ter (Barrowclough et al., 1995). The projectattempts to address the general question ofwhether empirically validated family inter-ventions can be delivered in a service setting.Can research be translated into practice? Theproject aims to assess the psychosocial inter-vention needs of families of schizophrenicpatients, to appraise how far these needs arebeing met by existing services and to identifythe shortfall. To achieve this Barrowcloughet al. (1998) have developed the RelativesCardinal Needs Schedule, based on the meth-odology for patients needs assessment(Marshall et al., 1995), to assess the psycho-social needs in carers of schizophrenic pa-tients. Relatively unskilled assessors can usethe schedule reliably and there are data sup-porting its concurrent validity (Barrowcloughet al., 1998). Once needs have been identi-fied an intervention can be delivered mainlythrough existing key workers who receivetraining in this through supervision, instruc-tion and workshops. Thus training is given inspecific skills to meet the individual needs ofa patient and carer with the assumption thatthese will generalise to their work with otherfamilies. This is in contrast to a structuredtraining programme such as Thorn, whichworks from the general to the specific. Arandomised controlled trial of this individu-alised training approach has demonstrated itseffectiveness. Relapse outcomes were supe-rior for patients in families treated in this waycompared to standard care (Barrowclough etal., 1999).ConclusionsIt is clear that the issue of translating re-search into clinical practice will be with usfor a long time and it is an issue that is likelyto be attributed increasing importance andattention. It is also clear that it is a complexissue and difficult to resolve. We have at-tempted to discuss how to improve mentalhealth services to patients who have psy-chotic disorders by making available to themnew innovative psychosocial treatments. Thishas raised issues concerning whether thesenew treatments have been validated suffi-ciently, how is the best way to teach them andhow should training be organised, how cantraining be assessed and evaluated and beconsidered good value for money? It isessential that new interventions be evaluatedproperly before they are disseminated andbecome established clinical practice. Refer-ence to the early part of this paper will indi-cate that insufficient evaluation has beencarried out in some areas, for example thetreatment of dual diagnosis or acutely illpatients.If time and money is invested in trainingprogrammes how can we assure that a skilledworkforce implements these skills to the ben-efit of the patient and their carer and that theseskills do not fall into disuse or that organisa-tional factors conspire against their use? Atpresent our understanding on these topics ispatchy and the methodology for investigat-ing them undeveloped. There has also beensome lack of clarity in the goals of dissemina-tion. Originally training in cognitive±behav-ioural psychosocial interventions was pros-ecuted with the aim of generally improvingknowledge and skills within the workforce,mainly CPNs, to encourage them to workwith psychotic patients. Other than the ap-praisal requirements of the training coursesthere was no clear idea of the level of knowl-edge and skills which dissemination wouldmake available. An easily extrapolated viewwas that training and dissemination wouldproduce a workforce of adequately qualified 580Nicholas Tarrier et al.and independent practitioners of these thera-peutic methods. This latter view is clearlydifferent from the former and may not bewarranted. If cognitive±behaviour therapy isthe application of cognitive and behaviouralsciences to clinical problems then it would beunrealistic to expect those who do not have athorough grounding and education in thecognitive and behavioural sciences to be in-dependent and unsupervised practitioners ofthese clinical applications. It may be that toomuch has been expected. The limitations ofthe dissemination of psychosocial treatmentsin schizophrenia need to be highlighted. Themethodological quality of the evaluation oftraining in psychosocial management is suchthat we must be cautious in our interpreta-tions. Further high quality evaluation isclearly necessary. Notwithstanding, devel-opments have been made and that there hasbeen a generation of interest in these topicsand an encouraging willingness to addressthe difficulties.ReferencesAnderson, J. & Adams, C. (1996). Family interventionsin schizophrenia: An effective but underused treat-ment. British Medical Journal, 313, 505.Barlow, D.H. (1996). Health care policy, psychotherapyresearch and the future of psychotherapy. Ameri-can Psychologist, 51, 1050±1058.Barrowclough, C. & Fleming, I. (1986a). Goal plan-ning with elderly people. How to make plans tomeet an individual's needs: A manual of instruc-tion. Manchester: Manchester University Press.Barrowclough, C. & Fleming, I. (1986b). Trainingdirect care staff in goal planning with elderly peo-ple. Behavioural Psychotherapy, 14, 192±209.Barrowclough, C. & Tarrier, N. (1990). Social func-tioning in schizophrenic patients. I: The effects ofexpressed emotion and family intervention. SocialPsychiatry and Psychiatric Epidemiology, 25, 125±129.Barrowclough, C. & Tarrier, N. (1992). Families ofschizophrenic patients: Cognitive behavioural in-terventions. London: Chapman & Hall.Barrowclough, C. Tarrier, N. & Lewis, S (1995). Anovel family support and psychoeducational serv-ice in schziophrenia: effectiveness and training.NHS Executive North West: Innovation Develop-ment Fund grantBarrowclough, C., Haddock, G., Tarrier, N. & Lewis,S. (1996). Evaluation of family support and cogni-tive-behavioural treatment service for schizophre-nia sufferers with substance misuse. London: NHSResearch and Development Grant.Barrowclough,C., Marshall, M., Lockwood, A., Quinn,J. & Sellwood, W. (1998). Assessing relatives'needs for psychosocial intervention in schizophre-nia: a relatives' version of the Cardinal NeedsSchedule. Psychological Medicine, 28, 531±542.Barrowclough, C., Tarrier, N., Lewis, S., Sellwood,W., Mainwaring. J., Quinn, J. & Hamlin, C. (1999).A randomised controlled effectiveness trial of aneeds based psychosocial intervention service forcarers of schizophrenic patients. British Journal ofPsychiatry, 174, 506±511.Bentall, R., Haddock, G. & Slade, P. (1994). Cogni-tive±behavioural therapy for persistent hallucina-tions: from theory to therapy. Behavior Therapy,25, 51±66.Bernstein,G. (1982). Training behavioural changeagents: A conceptual review. Behavior Therapy,13, 1±23.Birchwood, M. (1996). Early intervention in psychoticrelapse. In G. Haddock & P. Slade (Eds.), Cogni-tive-behavioural interventions for psychotic disor-ders. London: Routledge.Birchwood, M., Smith, J., Macmillan, F., Hogg, B.,Prasad, R., Harvey, C. & Bering, S. (1989). Predict-ing relapse in schizophrenia: The development andimplementation of an early signs monitoring sys-tem using patients and their families as observers:A preliminary investigation. Psychological Medi-cine, 19, 649±656.Bland, J.M. & Kerry, S.M. (1997). Trials randomised inclusters. British Medical Journal, 315, 600.Brooker, C. (1990). A description of clients nursed bycommunity psychiatric nurses while attending ENBclinical course number 811: A clarification of cur-rent roles. Journal of Advanced Nursing, 15, 155±166.Brooker, C. & Butterworth, A. (1993). Training inpsychosocial intervention: The impact on the roleof community psychiatric nurses. Journal of Ad-vanced Nursing, 18, 583±590.Brooker, C., Tarrier, N., Barrowclough, C., Butterworth,A. & Goldberg, D. (1992). Training communitypsychiatric nurses for psychosocial intervention:Report of a pilot study. British Journal of Psychia-try, 160, 836±844.Brooker, C., Falloon, I., Butterworth, A., Goldberg, D., Cognitive±behavioural treatments for schizophrenia 581Graham-Hole, V. & Hillier, V. (1994). The out-come of training community psychiatric nurses todeliver psychosocial intervention. British Journalof Psychiatry, 165, 222±230.Cardin, V.A., McGill, C.W. & Falloon, I.R.H. (1986).An economic analysis: Costs, benefits and effec-tiveness. In I.R.H. Fallooon (Ed.), Family manage-ment of schizophrenia. Baltimore: John HopkinsUniversity Press.Corrigan, P.W. & McCracken, S.G. (1995a). Psychiat-ric rehabilitation and staff development: Educa-tional and organisational models. Clinical Psy-chology Review, 15, 699±719.Corrigan, P.W. & McCracken, S.G. (1995b). Refocusingthe training of psychiatric rehabilitation staff. Psy-chiatric Services, 46, 1172±1177.Corrigan, P., McCracken, S., Edwards, M., Brunner, J.,Garman, A., Nelson, D. & Leary, M. (1997). Col-legial support and barriers to behavioural programsfor severely mental illness. Journal of BehaviorTherapy and Experimental Psychiatry, 28, 193±202.Drury. V., Birchwood, M., Cochrane, R. & Macmillan,F. (1996a). Cognitive therapy and recovery fromacute psychosis. I. Impact on psychotic symptoms.British Journal of Psychiatry, 169, 593±601.Drury. V., Birchwood, M., Cochrane, R. & Macmillan,F. (1996b). Cognitive therapy and recovery fromacute psychosis. II. Impact on recovery time. Brit-ish Journal of Psychiatry, 169, 602±607.Falloon, I., Boyd, J. & McGill, C. (1984). The familycare of schizophrenia. London: Guilford.Falloon, I., Kydd, R., Coverdale, J., & Laidlaw, T.(1966). Early detection and intervention for initialepisodes of schizophrenia. Schizophrenia Bulletin,22, 271±282Garety, P., Kuipers, l., Fowler, D., Chamberlain, F. &Dunn, G. (1994). Cognitive±behaviour therapy fordrug resistant psychosis. British Journal of Medi-cal Psychology, 67, 259±271.Goldfried, M.R. & Wolfe, B.E. (1996). Psychotherapy,practice and research, repairing a strained aliance.American Psychologist, 51, 1007±1016.Haddock, G. & Slade, P. (1996). Cognitive±behav-ioural interventions for psychotic disorders. Lon-don, Routledge.Haddock, G., Tarrier, N., Spaulding, W., Yusupoff, L.,Kinney, C. & McCarthy, E (1998). Individual cog-nitive-behavioural therapy in the treatment of hal-lucinations and delusions: A review. Clinical Psy-chology Review, 18, 821±838.Haddock, G., Tarrier, N., Morrison, A. P., Hopkins, R.,Dake, R. & Lewis, S. (1999). A pilot study evalu-ating the effectiveness of individual in-patient cog-nitive behavioural therapy in early psychosis. So-cial Psychiatry and Psychiatric Epidemiology, 34,254±258.Haddock, G., Devane, S., Bradshaw, T., McGovern, J.,Tarrier, N., Kinderman, P., Baguley, I., Lanca-shire, S. & Harris, N. (submitted). The acquisitionof cognitive±behavioural therapy skills by mentalhealth professional working with patients with se-vere mental health problems.Kavanagh, D. (1992). Recent developments in ex-pressed emotion and schizophrenia. British Jour-nal of Psychiatry, 160, 601±620.Kavanagh, D., Clark, D., Piatkowska, O.,Manicasvasagar, V., Rosen, A. & Tennant, C.(1993). Application of cognitive±behavioural fam-ily intervention for schizophrenia in multidisciplin-ary teams: What can the matter be? AustralianPsychologist, 28, 1±8.Kingdon, D. (1997). The Wellcome study of cognitivebehaviour therapy for psychotic symptoms. Paperpresented at the Second International Conferenceon Psychological Treatments for Schizophrenia,Oxford, October, 1997. Abstract published in theProceedings.Kingdon, D. & Turkington, D. (1991). Preliminaryreports: The use of cognitive behaviour therapy anda normalising rationale in schizophrenia. Journalof Nervous and Mental Disease, 179, 207±211.Krawiecka, M., Goldberg, D. & Vaughan, M. (1977). Astandardised psychiatric assessment scale forchronic psychotic patients. Acta PsychiatricaScandinavica, 55, 299±308.Kuipers, E., Garety, P., Fowler, D., Dunn, G.,Bebbington, P., Freeman, D. & Hadley, C. (1997).London±East Anglia randomised controlled trial ofcognitive±behavioural therapy for psychosis. I:Effects of the treatment phase. British Journal ofPsychiatry, 171, 319±327.Kuipers, E., Fowler, D., Garety, P., Chisholm, D.,Freeman, D., Dunn, G., Bebbington, P. & Hadley,C. (1998). London±East Anglia randomised con-trolled trial of cognitive behaviour therapy forpsychosis. III: Follow-up and economic evaluationat 18 months. British Journal of Psychiatry, 173,61±68.Lancashire, S., Haddock, G., Tarrier, N., Baguley, I.,Butterworth, C.A. & Brooker, C. (1997). The im-pact of training community psychiatric nurses touse psychosocial interventions with people whohave serious mental health problems: The Thornnurse training project. Psychiatric Services, 48,39±41.Lewis, S., Tarrier, N., Haddock, G., Bentall, R.,Kinderman, P. & Kingdon, D (1996). A multi-centre, randomised, controlled trial of cognitive± 582Nicholas Tarrier et al.behaviour therapy in early schizophrenia. London:Medical Research Grant (# G9519373)Liberman,R.P. & Eckman,T. (1989). Dissemination ofskills training modules to psychiatric facilities.Overcoming obstacles to the utilisation of a reha-bilitation innovation. British Journal of Psychia-try, 155, 117±122.McBrien, J. & Foxton, T. (1982). Education for thedevelopmentally young (EDY) training manual.Manchester: Manchester University Press.McCreadie, R.G., Phillips, L.A., Harvey, J.A., Waldron,G., Stewart, M. & Baird, D. (1991). The Nithsdaleschizophrenia survey, VIII: Do relatives want fam-ily intervention ± and does it help. British Journalof Psychiatry, 158, 110±113.McGorry, P., Edwards, J., Mihalopoulos, C., Harrigan,S.& Jackson, H. (1996). EPPIC: An evolving systemof early detection and optimal management. Schizo-phrenia Bulletin, 22, 305±326.Mari, J.J. & Streiner, D. (1994). An overview of familyinterventions and relapse on schizophrenia: Meta-analysis of research findings. Psychological Medi-cine, 24, 565±578.Marshall, M., Hogg, L.I., Gath, D.H. & Lockwood, A.(1995). The Cardinal Needs Schedule ± a modifiedversion of the MRC Needs for Care AssessmentSchedule. Psychological Medicine, 25, 605±617.Milne, D. (1984). The development and evaluation ofa structured learning format introduction to behav-iour therapy for psychiatric nurses. British Journalof Clinical Psychology, 23, 175±185.Penn, D.L. & Mueser, K.T. (1996). Research update onthe psychosocial treatment of schizophrenia. Ameri-can Journal of Psychiatry, 153, 607±617.Perry, A., Tarrier, N., Morriss, R., McCarthy, E. &Limb, K. (1999). A randomised controlled trial ofteaching bipolar disorder patients to identify earlysymptoms of relapse and obtain early treatment.British Medical Journal, 318, 149±153.Simpson, J.M., Klar, N. & Donner, A. (1995). Account-ing for cluster ramdomization: A review of primaryprevention trials. 1990 through 1993. AmericanJournal of Public Health, 85, 1378±1383.Smith, J. (1992). Family interventions: Service impli-cations. In M. Birchwood & N. Tarrier (Eds.),Innovations in the psychological treatment of schizo-phrenia. Chichester: Wiley.Tarrier, N. (1987). An investigation of residual psy-chotic symptoms in discharged schizophrenic pa-tients. British Journal of Clinical Psychology, 26,141±143.Tarrier, N. (1991). Some aspects of family interven-tions in schizophrenia: I. Adherence to interventionprogrammes. British Journal of Psychiatry, 159,475±480.Tarrier, N. (1996). A psychological approach to themanagement of schizophrenia. In M. Moscarelli,A. Rupp & N. Sartorius (Eds.), Handbook of Men-tal Health Economics and Health Policy, Volume1. Schizophrenia. Chichester: Wiley.Tarrier, N. (1997). Coping and problem solving in thetreatment of persistent psychotic symptoms. Paperpresented at the Second International Conferenceon Psychological Treatments for Schizophrenia,Oxford, October, 1997. Abstract published in theProceedings.Tarrier, N., Barrowclough, C. & D'Ambrosio, P. (1988).A training programme in psychosocial interventionwith families with a schizophrenic member. TheBehavioural Psychotherapist, 27, 2±4.Tarrier, N., Lowson, K. & Barrowclough, C. (1991).Some aspects of family interventions in schizo-phrenia. II: Financial considerations. British Jour-nal of Psychiatry, 159, 481±484.Tarrier, N., Beckett, R., Harwood, S., Baker, A.,Yusupoff, L. & Ugarteburu, I. (1993). A trial of twocognitive±behavioural methods of treating drugresistant psychotic symptoms in schizophrenic pa-tients. I: Outcome. British Journal of Psychiatry,162, 524±532Tarrier, N., Barrowclough, C., Porceddu, K. &Fitzpatrick, E. (1994). The Salford Family Inter-vention Project for schizophrenic relapse preven-tion: Five and eight year accumulating relapses.British Journal of Psychiatry, 165, 829±832.Tarrier, N., Yusupoff, L., Kinney, C., McCarthy, E.,Gledhill, A., Haddock, G. & Morris, J. (1998). Arandomised controlled trial of intensive cognitivebehaviour therapy for chronic schizophrenia. Brit-ish Medical Journal, 317, 303±307.Tarrier, N., Wittkowski, A., Kinney, C., McCarthy, E.,Morris, J. & Humphreys, L. (1999). The durabilityof the effects of cognitive±behaviour therapy in thetreatment of chronic schizophrenia: Twelve monthsfollow-up. British Journal of Psychiatry, 174, 500±504.Wilson, G.T. (1997). Behaviour therapy at centuryclose. Behavior Therapy, 28, 449±458.