Epidemiological studies using administrative databases have several advantages over other methodologies in studying the effectiveness of compulsory community treatment such as community treatment orders (CTOs). We compared patients placed on CTOs in Western Australia with controls drawn from both within the jurisdiction and from another without this measure (Nova Scotia). Although in different countries, the mental health services in both jurisdictions share common characteristics. Notably, we were able to control for forensic history in our comparison within Western Australia. We analysed predictors of admission and number of bed-days using multiple, logistic or Cox regression as appropriate. Of the 274 subjects placed on a CTO, we were able to find controls for up to 96% (n=265). CTO placement was not associated with reduced admissions or mean bed-days, although there was a threshold effect with a reduced risk of inpatient stays exceeding 100 days. Outpatient contacts were significantly greater for the CTO group. However, we do not know whether the intensity of treatment, or its compulsory nature, effected outcome.