Background: Seclusion (the isolation of a patient in a locked room) and transfer to a psychiatric intensive care unit (PICU; a specialised higher-security ward with higher staffing levels) are two common methods for the management of disturbed patient behaviour within acute psychiatric hospitals. Some hospitals do not have seclusion rooms or easy access to an on-site PICU. It is not known how these differences affect patient management and outcomes. Objectives: To (1) assess the factors associated with the use of seclusion and PICU care, (2) estimate the consequences of the use of these on subsequent violence and costs (study 1) and (3) describe differences in the management of disturbed patient behaviour related to differential availability (study 2). Design: The electronic patient record system at one trust was used to compare outcomes for patients who were and were not subject to seclusion or a PICU, controlling for variables, including recent behaviours. A cost-effectiveness analysis was performed (study 1). Nursing staff at eight hospitals with differing access to seclusion and a PICU completed attitudinal measures, a video test on restraint-use timing and an interview about the escalation pathway for the management of disturbed behaviour at their hospital. Analyses examined how results differed by access to PICU and seclusion (study 2). Participants: Patients on acute wards or PICUs in one NHS trust during the period 2008–13 (study 1) and nursing staff at eight randomly selected hospitals in England, with varying access to seclusion and to a PICU (study 2). Main outcome measures: Aggression, violence and cost (study 1), and utilisation, speed of use and attitudes to the full range of containment methods (study 2). Results: Patients subject to seclusion or held in a PICU were more likely than those who were not to be aggressive afterwards, and costs of care were higher, but this was probably because of selection bias. We could not derive satisfactory estimates of the causal effect of either intervention, but it appeared that it would be feasible to do so for seclusion based on an enriched sample of untreated controls (study 1).

Hospitals without seclusion rooms used more rapid tranquillisation, nursing of the patient in a side room accompanied by staff and seclusion using an ordinary room (study 2). Staff at hospitals without seclusion rated it as less acceptable and were slower to initiate manual restraint. Hospitals without an on-site PICU used more seclusion, de-escalation and within-eyesight observation. Limitations: Official record systems may be subject to recording biases and crucial variables may not be recorded (study 1). Interviews were complex, difficult, constrained by the need for standardisation and collected in small numbers at each hospital (study 2). Conclusions: Closing seclusion rooms and/or restricting PICU access does not appear to reduce the overall levels of containment, as substitution of other methods occurs. Services considering expanding access to seclusion or to a PICU should do so with caution. More evaluative research using stronger designs is required. Funding: The National Institute for Health Research Health Services and Delivery Research programme.