Despite early psychosis being proposed as the optimal phase for using exercise (12) the current study is among the first randomised controlled trial to assess feasibility of an exercise training program, compared to usual care in patients with first-episode psychosis undergoing specialised early intervention.
Our primary feasibility outcomes (screening and recruitment rates) indicate that our study set-up did not provide sufficient incentives and/or infrastructure to ensure consecutive screening and systematic promotion by the case managers and psychiatrists in charge of medical treatment. Despite a successful kick-off meeting, positive initial feedback from OPUS case managers in terms of their central role in screening and recruitment, and continuous encouragement and practical help from members of the research teams, we did not reach our recruitment target of 30 patients. More than half of screened patients were screened by less than 20% of case managers. While the case load in OPUS is 1:11 (i.e. each case manager covers the treatment of 11 patients), only 5 case managers (i.e. 17% of staff) screened more than 7 patients, whereas 14 (i.e. 47% of staff) screened three patients each or less. While staff motivation for active involvement in recruitment of patients remains unexplored in the current study, a recent qualitative study suggests an uncertainty among nurses concerning the benefits of physical activity as complementary treatment in patients with schizophrenia, indicating that hidden resistance exists in terms of the whole concept (28). Moreover, a recent study (29) indicated that exposing mental health staff to lifestyle interventions prior to targeting patients is critical in instigating culture change and improving patient outcomes. As such, closer collaboration with staff, including shared ownership and potentially staff-focused interventions, appear warranted to support recruitment of study participants in future exercise trials. Consequently, subsequent to trial completion, we invited participants and OPUS staff to an evaluation meeting, which resulted in the following additional suggestions for improved screening, recruitment, and retention rates in future trials: provide a trial exercise session to staff and potential participants, involve peers as part of recruitment strategy, focus on implementation of strategies for sustainment of exercise post-intervention, have flexible exercise schedules, and provide the option of choosing low intensity/relaxation exercises on days with high symptom burden/anxiety.
In this trial, secondary outcomes were selected based on the assumption that participants’ physical health would undergo measurable improvements attributable to our intervention. However, because the current trial was not designed or statistically powered to test for differences between treatment arms, no between-group analyses were performed, and the relatively small number of randomised patients and the large variation in outcomes make it impossible to derive an obvious candidate as the primary outcome for a large-scale randomised controlled trial based on the results from this feasibility trial. However, statistically significant within-group changes observed in sit-to-stand in INT from 0–8 weeks and in CON from 8–16 weeks suggest that the intervention may have had an impact on participants’ functional capacity, whereas no changes were observed in fitness, which may be related to the limited training frequency and the fact that only 19% of the intervention was performed at high intensity. Also, because participants had been on antipsychotic treatment for a minimum of 24 weeks ahead of enrolment, it is possible that metabolic changes had already installed, and that earlier initiation of exercise (i.e. before or concurrent with initiation of antipsychotic medication) would have been preferable. For example, attenuation of expected decreased fitness and prevention of weight gain, for which patients with first-episode psychosis are particularly susceptible, may be relevant candidates (30). As such, it should also be considered whether a more realistic goal besides improvements in physiological and functional outcomes should be chosen. Moreover, it is worth considering whether simultaneously targeting multiple lifestyle factors (i.e. poor diet and smoking) instead of focusing on one behavioural modification, i.e. increasing physical exercise, may be more appropriate as suggested by The Lancet Psychiatry Commission (11). Yet, the perceived changes in some outcomes (and the lack of changes in other others) in the current study should be interpreted with great caution and should not be ascribed a positive physiological effect (or the opposite). Furthermore, it is possible that the measures of mental health outcomes that are typically used in psychiatric treatment intervention studies may be equally relevant to physiological endpoints in a future large-scale trial. While subjective outcomes were not included in the current trial due to an expected lack of statistical power, our recently published qualitative sub-study (20) suggested that those who were included in the trial appreciated the intervention because of its potential to create an environment that was equally challenging and caring. Some participants also described the intervention as providing a relieving distraction from symptoms, while some reported improved sleep (20). This indicates that outcomes specifically related to self-efficacy and/or psychosocial well-being (i.e. stigmatisation, recovery, sleep quality, and loneliness) may be especially relevant to include in our future work.