Can primary mental health services impact levels of involuntary admissions? A cluster-RCT of the ReCoN intervention.

Purpose: Internationally, policies and legal changes seek to reduce the use of involuntary psychiatric admissions. Usually directed towards specialist services, these initiatives show little sustained progress. We tested whether an intervention at the level of primary mental health services has potential to reduce the use of involuntary admissions.

Methods: We conducted a two-arm cluster-RCT following Zelen's design (ClinicalTrials.gov:NCT03989765). Each cluster included the primary mental health service and their local collaborators in mid-sized Norwegian municipalities with rates of involuntary admissions above the national average. Five clusters were randomised to co-create and implement a comprehensive intervention. These could not be blinded, but the five control clusters were. Our primary hypothesis was that rates of involuntary admissions would be lower in the intervention arm when comparing change over time between arms, and that this would sustain. Secondary hypotheses were that rates of referrals for involuntary admissions and rates of referrals confirmed for involuntary status after the second statutory assessment, would be lower in the intervention arm.

Results: Data obtained from the Norwegian Patient Registry included all events in the study period. The difference between trial arms in changes of rates of involuntary admissions from baseline to the intervention period was 6.8 (95% CI 1.8 to 11.7; effect size (EC) 2.7), and reduced to 3.0 (95% CI -3.8 to 9.7; ES 0.9) between baseline and the post-intervention period. The difference between arms regarding changes in referral rates between the baseline and intervention period was 1.7 (95% CI -4.6 to 8.1; ES 0.5), and for changes in the rate of referrals resulting in involuntary status it was 1.3 (95% CI -3.4 to 6.0; ES 0.8).

Conclusion: We found a clear difference between trial arms in our primary outcome of involuntary admissions during the intervention period, but not beyond that period, and not regarding referrals for involuntary admissions, although the consistent direction of change favoured the intervention. We interpret the results to constitute 'proof of concept' that adequately resourced primary mental health services might contribute to policy aims of reducing involuntary care. Further rigorous studies in heterogeneous contexts are required.