BACKGROUND

Community treatment orders are widely used in England. It is unclear whether their use varies between patients, places and services, or if they are associated with better patient outcomes.

OBJECTIVES

To examine variation in the use of community treatment orders and their associations with patient outcomes and health-care costs.

DESIGN

Secondary analysis using multilevel statistical modelling.

SETTING

England, including 61 NHS mental health provider trusts.

PARTICIPANTS

A total of 69,832 patients eligible to be subject to a community treatment order.

MAIN OUTCOME MEASURES

Use of community treatment orders and time subject to community treatment order; re-admission and total time in hospital after the start of a community treatment order; and mortality.

DATA SOURCES

The primary data source was the Mental Health Services Data Set. Mental Health Services Data Set data were linked to mortality records and local area deprivation statistics for England.

RESULTS

There was significant variation in community treatment order use between patients, provider trusts and local areas. Most variation arose from substantially different practice in a small number of providers. Community treatment order patients were more likely to be in the ‘severe psychotic’ care cluster grouping, male or black. There was also significant variation between service providers and local areas in the time patients remained on community treatment orders. Although slightly more community treatment order patients were re-admitted than non-community treatment order patients during the study period (36.9% vs. 35.6%), there was no significant difference in time to first re-admission (around 32 months on average for both). There was some evidence that the rate of re-admission differed between community treatment order and non-community treatment order patients according to care cluster grouping. Community treatment order patients spent 7.5 days longer, on average, in admission than non-community treatment order patients over the study period. This difference remained when other patient and local area characteristics were taken into account. There was no evidence of significant variation between service providers in the effect of community treatment order on total time in admission. Community treatment order patients were less likely to die than non-community treatment order patients, after taking account of other patient and local area characteristics (odds ratio 0.69, 95% credible interval 0.60 to 0.81).

LIMITATIONS

Confounding by indication and potential bias arising from missing data within the Mental Health Services Data Set. Data quality issues precluded inclusion of patients who were subject to community treatment orders more than once.

CONCLUSIONS

Community treatment order use varied between patients, provider trusts and local areas. Community treatment order use was not associated with shorter time to re-admission or reduced time in hospital to a statistically significant degree. We found no evidence that the effectiveness of community treatment orders varied to a significant degree between provider trusts, nor that community treatment orders were associated with reduced mental health treatment costs. Our findings support the view that community treatment orders in England are not effective in reducing future admissions or time spent in hospital. We provide preliminary evidence of an association between community treatment order use and reduced rate of death.

FUTURE WORK

These findings need to be replicated among patients who are subject to community treatment order more than once. The association between community treatment order use and reduced mortality requires further investigation.

STUDY REGISTRATION

The study was approved by the University of Warwick’s Biomedical and Scientific Research Ethics Committee (REGO-2015-1623).

FUNDING

This project was funded by the National Institute for Health Research (NIHR) Health Services and Delivery Research programme and will be published in full in ; Vol. 8, No. 9. See the NIHR Journals Library website for further project information.