This review of efficacy measures is intended to highlight the huge difficulties in applying simple measurements to the evaluation of HIV prevention programmes. From the examples cited above, it should be evident that the more robust measures of effectiveness are those which form part of an interlinked sequence of measurements which test the various assumptions underlying a programme. For example, it would be desirable in assessing the success of a nationwide HIV prevention programme targeted at gay men to include the following measurements as performance indicators:

• measures of awareness of particular interventions.
• measures of understanding of messages.
• process efficiency measures of numbers reached.
• baseline incidence.
• a large cohort to measure changes in incidence over time.
• a qualitative research project which looked at the sexual behaviour of cohort seroconverters, such as self–reported reasons for unprotected sex.
• proof of concept or mediator studies which tested assumptions such as: increased availability of condoms translates into increased use of condoms by people who would otherwise have engaged in unprotected anal intercourse; or, knowledge of own or partner's HIV status influences condom use.

Can randomised, controlled trials be used in HIV prevention research?

A number of the proof of concept studies identified above appear to demand controlled study. It has been argued in the past that randomised controlled studies in HIV prevention are extremely difficult, particularly if HIV incidence is the outcome measure used. However, there is no reason why studies which examine particular elements of a prevention strategy, and which use carefully chosen outcome measures, should not be conducted successfully.

The major difficulty in using such studies as `proof of concept' investigations lies not so much in the element of randomisation or the selection of a valid control group, but in the resources needed to mount such studies. Relatively large studies will be needed to draw valid conclusions, and national research networks do not yet exist in the genitourinary setting to mirror those developed by the Medical Research Council for its anti–retroviral studies. Such a network could be developed alongside efforts to establish a national gay men's vaccine cohort (See Developing prevention technologies: HIV & AIDS vaccines).

Further reading

Outcome measures are also discussed in:

• Risk, intervention and change, by Tim Rhodes, HEA, 1994.
• Outcomes in HIV prevention by Chris Bonell, HIV Project, 1996 and Using outcomes in HIV prevention: a how–to guide by Chris Bonell and Paul Devlin, HIV Project, 1997.