Popular opinion leaders in London: a response to Kelly.

Controlled trials conducted in the USA provide clear evidence that peer education can bring about a reduction in high risk sexual behaviour among gay men. HIV prevention interventions that systematically identified, recruited, trained and engaged popular opinion leaders (POLs) made a significant impact on sexual behaviour at a community level in small US towns. However, recent trials conducted in the UK have failed to replicate these findings. A POL intervention in London made no significant impact at a community level on the risk behaviours of gay men. Jeffrey Kelly, one of the authors of the US studies, has identified nine core elements central to the popular opinion leader model. In Kelly's view 'the UK projects were not tests of the popular opinion leader model because they did not employ most of these POL core elements'. The absence of any significant impact of the UK programmes on sexual risk behaviour at a community level was not, therefore, surprising. In fact, the London POL project incorporated all the core elements into its design and succeeded in employing seven out of nine in its delivery. Attempts to employ all the core elements, however, were hampered by problems in recruiting popular opinion leaders as well as barriers to communication. Process evaluation revealed that it was these obstacles which limited diffusion. This in turn explained the absence of any impact of the London POL project on sexual risk behaviour at a community level. The obstacles to successful diffusion in London have provided a valuable opportunity for examining the processes that underlie the POL model. Our study raises the question as to whether social interventions shown to be effective in one setting, place or moment in time can be replicated in another.