Established treatments for obsessive compulsive disorder (OCD) include cognitive behaviour therapy (CBT) and selective serotonin reuptake inhibitor (SSRI) medication. Combined treatment may outperform monotherapy, but few studies have investigated this. A total of 49 community-based adults with OCD were randomly assigned to CBT, SSRI, or SSRI+CBT. Sertraline (50-200mg/day) was given as the SSRI for 52 weeks. A 16-h-manualized individual CBT was delivered over 8 weeks with four follow-up sessions. Assessors were blinded' to treatment allocation. A preliminary health economic evaluation was conducted. At week 16, combined treatment (n=13) was associated with the largest improvement, sertraline (n=7) the next largest and CBT (n=9) the smallest on the observed case analysis. The effect size (Cohen's d) comparing the improvement in Yale Brown Obsessive Compulsive Scale on CBT versus combined treatment was -0.39 and versus sertraline was -0.27. Between 16 and 52 weeks, the greatest clinical improvement was seen with sertraline, but participant discontinuation prevented reliable analysis. Compared with sertraline, the mean costs were higher for CBT and for combined treatment. The mean Quality Adjusted Life Year scores for sertraline were 0.1823 (95% confidence interval: 0.0447-0.3199) greater than for CBT and 0.1135 (95% confidence interval: -0.0290-0.2560), greater than for combined treatment. Combined treatment appeared the most clinically effective option, especially over CBT, but the advantages over SSRI monotherapy were not sustained beyond 16 weeks. SSRI monotherapy was the most cost-effective. A definitive study can and should be conducted.

Recruitment to target is a major challenge for randomised controlled treatment trials (RCTs) in psychiatry. Strong patient preference may hamper recruitment of protocol-eligible participants, thereby compromising statistical power and limiting the external validity and generalisability of study findings to target populations. The Optimal Treatment in OCD (OTO) is a UK non-placebo feasibility trial of CBT and sertraline (two standard treatments) and their combination, randomly allocated, designed to ascertain the optimal community-based treatment for OCD. OTO aims to provide information about relevant design issues, to optimise feasibility of the definitive trial. To identify the factors influencing recruitment to the RCT. The study was approved by the relevant NHS ethics committee. Adult OCD outpatients were recruited at three UK centres. They were identified from routine referrals, primary healthcare services, community mental health clinics, other clinics attracting high levels of OCD, trust databases of patients likely to be suffering with OCD and the use of promotional materials. Patients subjected to screening were given detailed information about the treatments. Flexible and responsive clinical support was offered for the trial duration. Those eligible candidates declining to consent to participate were asked for their reasons. We report the data from one centre. 86 individuals satisfied the trial eligibility criteria. Of these, only 24 (27.9%) agreed to participate in the study. Of the remaining 62 (72.1%) who refused to participate, 35 (56.5%) gave treatment-related reasons. The commonest treatment-related reason was refusal to take pharmacological treatment (N=17; 27.4% all refusers). 12 of these individuals refused medication in general, while 5 refused sertraline. Another 8 (12.9% all refusers), showed a clear preference for CBT. However, 10 (16.1% all refusers) declined because they were reluctant to discontinue ongoing medication prior to randomisation. Treatment–related preferences constituted a major obstacle to the recruitment of UK OCD patients to a non-placebo RCT of standard pharmacotherapy and CBT, to the extent that only 27.9% eligible patients were actually randomised. Over 40% (40.3%) of those refusing to enter the study gave as the reason a reluctance to receive pharmacotherapy or a bias toward receiving CBT, while 16.1% refused washout. These choices occurred despite patients being informed of the current evidence suggesting equivalent clinical effectiveness for CBT and SSRI and being offered clinical support during the washout phase. Trial designs that account for patient preference e.g. ‘patient-preference RCT’ ‘or ‘comprehensive cohort’ designs (Brewin & Bradley, 1989) or omit washout may increase recruitment but have methodological limitations. RCTs should gather information on the effect of pre-randomisation preference on outcomes to inform better trial design (Torgerson, Klaber-Moffett & Russell, 1996). New approaches to overcome public prejudice toward psychopharmacology are also needed to support robust trial design in psychiatry.