Effect of increased compulsion on readmission to hospital or disengagement from community services for patients with psychosis: follow-up of a cohort from the OCTET trial

Background Community treatment orders (CTOs) have not been shown in randomised trials to reduce readmission to hospital in patients with psychosis, but these trials have been short (11-12 months). We previously investigated the effect of CTOs on readmission rates over 12 months in a randomised trial (OCTET). Here, we present follow-up data for a cohort of individuals recruited to our original trial to examine the long-term effect of CTOs on readmissions and the risk of patients disengaging from mental health services temporarily or enduringly. Methods For OCTET, an open-label, parallel, randomised controlled trial, we recruited patients aged 18-65 years involuntarily admitted to mental health hospitals in 32 trusts in England, with a diagnosis of psychosis and deemed suitable for CTOs by their clinicians. Between Nov 10, 2008, and Feb 22, 2011, we recruited and randomly assigned 336 eligible patients (1:1) to be discharged on either a CTO (n=167) or to voluntary status via Section 17 leave (control group; n=169). For the analysis presented in this report, we assessed data at 36 months for 330 of these patients. We tested rates of readmission to hospital, time to first readmission, number of readmissions, and duration of readmission in patients assigned to CTO versus those assigned to control, and in all patients with CTO experience at any time in the 36 months versus those without. We also tested whether duration of CTO affected readmission outcomes in patients with CTO experience. We examined discontinuation (>= 60 days between clinical contacts) and disengagement from services (no clinical contact for >= 90 days with no return to contact) in the whole cohort. OCTET is registered with isrctn.com, number ISRCTN73110773. Findings We obtained data for 330 patients in the relevant period between Nov 10, 2008 and Feb 22, 2014 (36 months after the last patient was randomly assigned to OCTET). We identified no difference between the randomised groups in the numbers of patients readmitted (100 [61%] of 165 CTOs vs 113 [68%] of 165 controls; relative risk 0.88 [95% CI 0.75-1.03]), number of readmissions (mean 2.4 readmissions [SD 1.91] vs 2.2 [1.43]; incident density ratio [IDR] 0.97 [95% CI 0.76-1.24]), duration of readmissions (median 117.5 days [IQR 63-303] vs 139.5 days [63.0-309.5]; IDR 0.84 [95% CI 0.51-1.38]), or time to first readmission (median 601.0 days [95% CI 387.0-777.0] vs 420.0 days [352.0-548.0]; hazard ratio [HR] 0.81 [95% CI 0.62-1.06]). The CTO experience group had significantly more readmissions than the group without (IDR 1.39 [95% CI 1.07-1.79]) and we noted no significant difference between groups in readmission rates, duration of readmission, or time to first readmission. We did not identify a linear relationship between readmission outcomes and duration of CTO. 19 (6%) patients disengaged from services (12 [7%] of 165 CTOs vs 7 [4%] of 165 controls). Longer duration of compulsion was associated with later disengagement (HR 0.946 [95% CI 0.90-0.99, p=0.023). 187 (57%) experienced no discontinuities, and we noted no significant difference between the CTO and control groups for time to disengagement or number of discontinuities. Levels of discontinuity were associated with compulsion (IDR 0.973 [95% CI 0.96-0.99, p<0.0001]. We identified no effect of baseline characteristics on the associations between compulsion and disengagement. Interpretation We identified no evidence that increased compulsion leads to improved readmission outcomes or to disengagement from services in patients with psychosis over 36 months. The level of persisting clinical follow-up was much higher than expected, irrespective of CTO status, and could partly account for the absence of CTO effect. The findings from our 36-month follow-up support our original findings that CTOs do not provide patient benefits, and the continued high level of their use should be reviewed.