Delivered at the United States (US) President's Emergency Plan for AIDS Relief (PEPFAR) Expert Consultation on Unresolved Issues in HIV Prevention Programming in Generalized Epidemics, this presentation reviews the evidence for behaviour change communication (BCC) interventions that use mass media in an effort to prevent HIV through these behavioural outcomes: knowledge, condom use, and reduction in the number of sexual partners.

First, Jane Bertrand explains why randomised controls cannot be used to evaluate mass media interventions. Namely, full scale, comprehensive mass media potentially reach all segments of the intended population, so it is not possible to randomly allocate part of the population to experimental groups. Also, the wedge design violates a key principle of BCC: mutually reinforcing channels and messages. Furthermore, comparison groups are rarely comparable on all socio-demographic characteristics.

Next, she reviews the existing research, which includes over 90 single studies or evaluations since 1986, synthesis reviews, and a large, yet-to-be-published meta-analysis by Dr. Leslie Snyder that includes research on health communication interventions from 1986-2006 that included a media component. Dr. Bertrand then explains calculation of effect sizes for the behavioural sciences. Subsequent slides show the results of Dr. Snyder's research - e.g., improved effects of communication campaigns on condom use over time (condom use effect sizes) for interventions that used mass media alone (radio, television, newspapers, billboards, and the internet), small media (posters, pamphlets, and in-clinic videos) combined with interpersonal communication (IPC), and mass media combined with IPC. (See slide #9 for a graph). Selected other findings: Interventions in less developed countries have greater effect on condom use, especially in a combined mass media and interpersonal approach; a greater effect size is seen in longer interventions.

Citing a recent Lancet article (Wakefield, M.A., Loken, B. and Hornik, R.C. 2010. Use of mass media campaigns to change health behaviour), Dr. Bertrand asserts that Dr. Snyder's findings also show that mass media campaigns can produce positive changes or prevent negative changes in health-related behaviours across large populations. Factors that contribute to these outcomes: concurrent availability of required services and products, availability of community-based programmes, and policies that support behaviour change.

Slide #16 shows an ideation model for effects of knowledge on HIV behavior change. Dr. Bertrand asserts that communication can affect all factors including attitudes, self-image, emotions, perceived risk, self-efficacy, social support, social norms, and personal advocacy.

There is also a dose effect, as one 2006 study cited here found with regard to the percent of condom use by exposure to 19 AIDS communication programmes in South Africa: the percentage increased as the number of AIDS communication programmes seen or heard increased.

Based on a review of all of this evidence, Dr. Bertrand concludes that:

• There is consistency of positive findings for effects of mass media on knowledge and condom use, but a lack of effect on reducing the number of sexual partners.
• More recent studies show greater effects, possibly reflective of improved programming and/or better analytic techniques for evaluation.
• It has been found that key strategies for improving HIV mass communication programming and evaluation include: Target messages using audience segmentation, focus on behaviour change (rather than knowledge), use behavioral theories, achieve high message exposure, use stronger research designs for outcome assessment, and include measures of behaviour in outcome assessments.
• "Mass media are essential to future HIV programming" because they can changes behaviours outside the clinic (condom use), change community norms, and increase service utilisation by creating demand and enhancing compliance through counselling.

PEPFAR website, July 13 2011.