Authors' reply

Hotopf et al make essentially the same point that we stated in the article ‘… the study found no significant difference in the prospective rate of violence between the two randomly assigned groups: 32.3% in the OPC group v. 36.8% in the control group (Fisher's exact test, one-tailed: P=0.292; two-tailed: P=0.567)’ (Swanson et al, 2000).

Critics of OPC policy might wish we had left it at that, but straightforward analysis of randomised controlled trials does not tell the whole story. In this case it excluded people with a documented history of serious violence (n=64), since the court did not permit us to randomise these to the control group. However, variability in the real-world application of OPC allowed us to examine whether longer periods of court-ordered treatment were associated with lower rates of violence over the study year. They were.

Hotopf et al are rightly concerned about the possibility of favourable selection bias, but we think this is an unlikely explanation for our findings. Indeed, people with a history of treatment non-adherence were more than twice as likely to receive an extended period of OPC (40.0 v. 18.75%). If anything, this should have stacked the deck against finding an effect for long-term OPC.

Hotopf et al recalculated the post-randomisation effect for longer-term OPC in what they refer to as our ITT sample, rather than the sample we actually used. They say the effect is not significant but their calculation excludes the historically violent subgroup.

For hospital outcomes, unlike violence, we obtained follow-up information on the entire ITT sample through admission records. Here we found a statistically significant experimental result. For any month during the study year, the randomly assigned OPC group had a lower risk of readmission than the control group (OR=0.64, P<0.01). Hotopf et al do not mention this finding.

About one-third of the OPC group had their court orders expire very early in the study – during the first or second month – and more of these individuals were rehospitalised than those remaining on OPC, which explains the early separation of the lines in the figures from Swartz et al (1999).