The impact of psychiatric decision units on mental health crisis care pathways: a synthetic control study

Psychiatric crisis care is under great pressure, with the number of psychiatric presentations to emergency departments increasing and inpatient wards operating with occupancy rates above recommended levels. Internationally, hospital-based short-stay crisis units (named Psychiatric Decision Units; (PDU) in the UK) have been introduced to address these challenges, but the current evidence for their effectiveness is limited. We estimated the effects of PDUs in four geographic locations in England, linked to three National Health Service (NHS) mental health trusts and six NHS acute hospital trusts. Using national data sets to create synthetic controls from areas without PDUs (following the generalised synthetic control method), we estimated trust-wide changes to the primary outcomes of psychiatric inpatient admissions and psychiatric presentations to emergency departments (ED), compared to the synthetic controls, alongside secondary outcomes. We used meta-analysis to robustly combine outcomes. We analysed NHS hospital activity data for adults aged between 18 and 75 years covering 24 months preceding and following the introduction of each PDU (November 2012 to January 2021). We found no significant impacts of PDUs on primary outcomes, except at Sheffield Teaching Hospitals NHS Foundation Trust with 1.5 fewer psychiatric presentations to ED per 10,000 trust population per month (relative difference: 24.9%, p = 0.034) than the synthetic control. We found mixed effects of the opening of PDUs on secondary outcomes. Meta-analyses indicated a significantly lower mean length of stay for psychiatric admissions (-6.4 days, p < 0.001) for patients in mental health trusts with a PDU compared to the synthetic control and no significant effects on other outcomes. Heterogeneity of effect across sites probably reflects variation in PDU configuration and implementation. Further research should explore the intended aims of PDUs alongside how they operate in practice. Registration: The study is registered with the ISRCTN (ISRCTN77588384).