Abstract Randomised trials are often viewed as an important source of evidence for informing health services commissioning. This paper examines whether three trials of HIV prevention interventions were initiated to inform commissioning decisions, and whether this actually occurred. Referring to various theories that focus on the research/policy interface, the study concludes that: trials were generated via complex processes involving several agencies but that HIV commissioners were rarely important initiators of trials; and that various motivations were implicated in initiating trials. Determining the effectiveness of interventions was not a central motivation. Other more important motivations included: asserting/subverting trials as an appropriate evaluation technology; maintaining organisational/occupational identities; and supporting existing decisions. Trial evidence did not greatly inform planning decisions, many of which were made by providers rather than commissioners.