Cognitive-behavioural therapy for severe mental disorders: back to the future?

Like recent medication studies, it appears that when cognitive–behaviouraltherapy is tested in pragmatic effectiveness trials involvingroutine clinical populations it does not fare as well as in efficacy trials. Given the multitude of factors that can `muddythe waters' in clinical trials, how do we best make sense ofthe findings?

INTRODUCTION

TOP

INTRODUCTION

View larger version (131K):
[in this window]
[in a new window]
[as a PowerPoint slide]

The basic aims of treatments for schizophrenia, and indeed allsevere mental disorders, are to reduce symptoms and distress,enhance functioning and prevent relapse. For many decades,individual variability in the nature and severity of symptomsand differing responses to available treatments suggested achievingthese aims would remain elusive. The introduction of second-generationatypical antipsychotics followed by developments such as cognitive–behaviouraltherapy (CBT) for medication-refractory delusions created renewedoptimism, especially as initial randomised controlled trials (RCTs) indicated that these new interventions produced significantlybetter clinical outcomes with fewer side-effects than all thepreviously established treatments. Presentations at researchmeetings began to echo the Dodo in Alice's Adventures in Wonderland:`Everybody had won and all must have prizes'.¹ Further efficacyRCTs of second-generation antipsychotics and CBT demonstrated benefits across the full range of severe mental disorders –indeed, a cursory review of the research literature at theturn of the century might easily leave the impression thatsecond-generation antipsychotics and CBT were psychiatry'sequivalent of steroids in general medicine.

Failures to replicate the initial impressive benefits in latertrials together with a series of meta-analyses of antipsychoticsand psychological therapies for schizophrenia suggested thateffects had been overestimated. Sadly, this information neitherreceived as much attention as the results of earlier studies,nor deterred those advocating for increased availability of these interventions. The National Institute for Health andClinical Excellence (NICE)² in the UK, and similar internationalorganisations elsewhere, published treatment guidelines recommendingtherapeutic doses of antipsychotic medication,³ with a nodof approval towards the newer second-generation antipsychotics,as the mainstay of clinical management, but also promoted wideraccess to adjunctive `empirically supported' therapies suchas family therapy and CBT. Providing CBT for psychosis acrossthe National Health Service (NHS) was not supported unanimously.⁴ However, the guidance was based on the balance of availableevidence at that time from RCTs, still the gold standard forjudging the relative benefits of treatments.⁵ Most of thesetrials involved selected patient populations and now we havenew evidence from larger pragmatic studies that the second-generation antipsychotics have, quite literally, been oversold,⁶ andthe trial of CBT reported in this issue by Garety et al,⁷seems to suggest that this treatment too is in a similar position.So the important question is whether we have truly been ledastray or whether these larger studies in the real world aredefective and have come up with the wrong answers?

Large-scale multicentre efficacy RCTs of (relatively) homogeneouspatient populations are critical in establishing the benefitsand tolerability of antipsychotic compounds, but most of thesehave the prime objective of meeting regulatory requirements.Narrow inclusion criteria dictate that only about 10% of individualsreceiving treatment for schizophrenia are eligible for recruitment.⁸The necessary post-marketing, large-scale effectiveness RCTsare seldom undertaken, but those available less commonly indicateclear advantages for new medications. It is this limited focusand lack of generalisability of efficacy trials that has promptedthe multicentre CATIE (Clinical Antipsychotic Trials of InterventionEffectiveness) and CUtLASS (Cost Utility of the Latest AntipsychoticDrugs in Schizophrenia Study) trials.^6,9 The benefits achievedwere considerably less than for highly selected patients receivingcarefully titrated doses of medication in controlled research environments.^6,8–12 The medication efficacy-effectivenessgap and failure to achieve significant recovery in schizophreniastimulated the introduction of adjunctive psychological treatmentfor those with severe mental disorders. Ironically, the datasupporting adjunctive therapy has virtually all come from efficacy RCTs, and it is important to acknowledge that it is not justBig Pharma that introduces bias into such studies;⁵ it is equally possible with psychological treatments, and the findingsare as much at risk of tendentious reporting as medicationstudies. The consequence of this now seems to be that whenindependent multicentre RCTs of effectiveness are carried outin `real world' secondary care settings the apparent additionalbenefits of new treatments disappear.^6,9,13–15 It isin this context that the trial of Garety et al⁷ needs to be judged.

Onken et al¹⁶ describe three progressive stages in the evolutionof evidence-based therapies: the development phase (feasibilitystudies, development of therapy manuals), efficacy RCTs andexploratory studies of mechanisms of action, and studies of generalisability and transportability. Cognitive–behaviouraltherapy for psychosis had a remarkably long gestation, withcase reports appearing nearly 40 years before the publicationof pilot RCTs and therapy manuals. There are now over 30 publishedefficacy trials and a recent comprehensive meta-analysis¹⁷ reported that, compared with usual treatment alone, CBT plususual treatment demonstrates statistically significant effectson depression, anxiety, positive and negative symptoms andfunctioning, but not relapse rates. It confirmed that individualswho show the most robust benefit from CBT for psychosis havestable, persistent but distressing, medication-refractory positivesymptoms. However, the meta-analysis also demonstrated thatas sample sizes and methodological rigour increase, effectsizes decrease by 50–100%.

Effectiveness trials and the exploration of generalisability

TOP

Effectiveness trials and the...

In light of this comprehensive evaluation of existing research,¹⁷the findings of Garety et al⁷ fall within the range of expectedoutcomes. The design of their trial also offers insights intofactors affecting the general application of CBT and familytherapy for psychoses. The findings that CBT is not effectivein reducing relapses in psychosis, yet has some benefit ondepression, delusional distress and social functioning areconsistent with the more modest effect sizes reported in larger-scaleefficacy RCTs. Also, Garety et al note that in CBT, for peoplewith psychosis, some of the therapy benefits are restrictedto those living with a significant other person (a factor knownto predict better outcome in CBT for depression). Unfortunately,the equivalence in group outcomes was not a consequence ofa good response to standard care in the controls but ratherto a disappointing lack of progress across all randomised groups.Although suboptimal prescribing³ and non-adherence were frequent,the clear message was that adjunctive therapy did not performwell and it is important to consider the reasons for this. First, in a sample predominantly comprised of single males, the acceptabilityof and adherence with therapy was no greater than for medication;the drop-out rates for CBT and family therapy (25–30%)are equivalent to those reported for medication, while thoseallocated to therapy usually attended only about 60% of thesessions offered. Furthermore, even though access to NHS therapyis still restricted and waiting times extensive, only 44% ofmore than 600 potential participants agreed to inclusion. Itis highly likely that, among the many reasons for this lowlevel of agreement, some declined because they did not wanttherapy. It seemed especially hard to get patients and their family to simultaneously consent to participate in family therapy,and there appeared to be under-representation of high `expressedemotion' families (those we might postulate would benefit mostfrom family therapy) among those recruited.

One barrier to the generalisability of CBT may be the levelof competence of local therapists and their compliance to themodel. Garety et al monitored quality as much as was feasible,but their comment that `therapy was competently delivered',may be an overestimate. The data provided only allow us toconclude that 44% of those in the CBT and 39% in the familytherapy groups received some sessions that met appropriatestandards. More tellingly, the authors discuss that, in theabsence of acute symptoms or distress, CBT therapists oftenfound it difficult to maintain a clear focus and instead covereda wide range of nonspecific problems and `adopted a generalapproach to emotional distress'. The lack of specific activeelements of CBT would suggest the approach had relatively fewfeatures distinguishing it from generic case management. Thismay have undermined the effectiveness of CBT in this RCT, butit is a realistic representation of the problems of delivering therapy in general clinical settings.

Every treatment has limits

TOP

Every treatment has limits

No single research trial provides the answers to all the questionsthat arise when we endeavour to employ research-proven treatmentsin clinical practice. Critics of efficacy RCTs are quick tosuggest we cannot use reported outcomes to predict responserates in heterogeneous clinical samples, but critics of effectivenesstrials justifiably argue that these RCTs just trade one setof problems for another. Broad-based, pragmatic trials withfew exclusion criteria often mean that moderators of groupoutcome (e.g. duration of illness, social support, comorbidities,quality of prescribing, compliance) produce so much `noise'that the chances of uncovering evidence of differential treatmentresponse are minimised.^7,13–15 However, RCTs reportgroup outcomes while clinicians treat individuals and we willcontinue to make selective rather than blanket referrals forCBT. Clinically meaningful information about who is more likelyto do well or badly with different treatment packages can beprovided from secondary and/or post hoc analyses of effectiveness trials.^6,13 These secondary studies, despite the perils ofpost hoc interpretation, can also pave the way for the nextstep in the research process. In Garety et al's study, thestandard deviations are very wide for most of the continuousmeasures, suggesting considerable interindividual variabilityin outcome. It may be possible to identify signals relatingto patient, therapy/therapist, or combinations of factors suggestingunder what circumstances adjunctive therapy should be employedor directing us to where further studies are warranted. Garetyet al's study should not be interpreted as a setback for CBTresearch, but it introduces some healthy realism about thelimits for the role of adjunctive therapy in severe mental disorders.It is also a timely reminder that, away from the `therapy forall' media hysteria, the world of routine psychiatric practicebrings us into contact with some patients who do not want ordo not respond optimally to antipsychotic medication, but whoalso do not always want or benefit from psychological therapieseither.

REFERENCES

TOP