A 10-Year Systematic Review of HIV/AIDS Mass Communication Campaigns: Have We Made Progress?

Published in the Journal of Health Communication (Vol. 14, pages 15-42), this study details a 10-year (1998–2007) systematic review of HIV/AIDS mass communication campaigns carried out around the world that focus on sexual behaviour, HIV testing, or both. The purpose is to compare the results with the last comprehensive review of such campaigns, which spanned 1986 to early 1998.

Early in the HIV/AIDS campaign literature, health communication researchers presented recommendations for how such campaigns should be carried out, identifying the following specific principles: (1) conducting formative research on and about the intended audience; (2) using theory as a conceptual foundation; (3) segmenting one's audience into meaningful subgroups; (4) using a message design approach that is targeted to the audience segment(s); (5) utilising effective channels widely viewed by and persuasive for the intended audience; (6) conducting process evaluation and ensuring high message exposure; and (7) using a sensitive outcome evaluation design that reduces threats to internal validity and allows causal inferences about campaign impact to be made. A question raised in the current study is the following: to what extent have recent HIV/AIDS campaigns in the literature adhered to such principles?

A comprehensive search strategy yielded 38 HIV/AIDS campaign evaluation articles published in peer-reviewed journals, representing 34 distinct campaign efforts conducted in 23 countries. The articles were coded on 10 dimensions of interest (country, sample, intended audience, formative research, use of theory, campaign channels/components, campaign slogan, message exposure, evaluation design, outcome measures) by two independent coders. Results indicated that compared with the previous systematic review (1986–1998 period), campaigns increasingly have employed the following strategies:

• Formative research: While the earlier review did not consider this element, in the current review, 47% (16 of 34) of campaigns reported conducting some type of formative research activities. Such research included a variety of methods, including focus groups, ethnographic interviews, quantitative surveys, and literature reviews. Among those reporting formative research, the most commonly reported activity was research about campaign messages, including pretesting messages or examining message preferences of members of the intended audience.
• Use of theory: In the current review, 15 of 34 campaigns, or 44%, reported using theory. This is in contrast to the earlier review, which reported that fewer than 20% of HIV/AIDS campaigns were theory-based. Some of the theories and models reported among the current set of campaigns included the Health Belief Model, Theories of Reasoned Action and Planned Behavior, Social Cognitive Theory (also known as Social Learning Theory), the Transtheoretical Model and Stages of Change, and the Information-Motivation-Behavioral Skills Model.
• Audience segmentation: All but two studies (94%) in the present review reported information related to an intended audience of interest. In addition, in only one case did this information suggest that a campaign was directed at the general public, and even in that case campaign activities directed toward particular groups were reported (suggesting some segmentation practices). This is in stark contrast to the earlier review, which found that 24 of 41 campaign evaluations, or 59%, were directed at the general public. Researchers note that this is likely indicative of the shift of HIV/AIDS campaigns from tools of awareness to tools of behaviour change among a number of high-risk groups.
• Message design: Although (as noted above) 44% of campaigns were theory-based, most were behavioural theories rather than theories specific to message design or message effects. Ten campaigns (30%) in the current review did use an entertainment-education strategy and in that manner used narratives for message design. Four of these 10 studies, however, did not list or describe any theory as informing the campaign. Also examined here were campaign slogans, which most campaigns (28 of 34, or 82%) did provide as an overall theme of the campaign. By contrast, the earlier review found that only 63% of campaigns reported a theme or slogan. It is notable that a number of campaign slogans in the current review included active references to the behaviours of interest, such as "HIV. Live With It. Get Tested!" and "Condoms: They Go Where You Go". In fact, 8 of 34 campaigns (24%) contained slogans with direct and clear messages related to behavioural change, including HIV testing, safer sex, or condom use. On the other hand, the earlier review found that a number of very general slogans were used, including "Understanding AIDS", "AIDS Awareness Week," and "America Responds to AIDS." Only 4 slogans (10%) in the earlier review contained direct reference to behavioural change.
• Channels: In the current review, all campaigns described channel use, with 7 of 34 campaigns (21%) using a single media channel, while the remainder (79%) employed multiple channels. This is similar to the finding in the earlier review, in which 93% reported channel use, with 26% using single and 74% using multiple media. Table 1 in the document reveals great diversity of channels and strategies used to "get the message out" in the multichannel and multi-component campaigns. Not surprisingly, a new trend in the campaigns area is the use of internet websites as an additional campaign channel. Campaigns also used a variety of creative materials, such as baseball cards, postcards, condom packs, and other promotional materials, as well as a variety of interpersonal strategies, including peer education, skill-building workshops, provider support, and hotlines. Further, a number of campaigns attempted to engage the larger community in the effort by including community partners, community coalitions, and community education/mobilisation.
• Process evaluation and campaign exposure: In the current review, 82% of campaigns reported on campaign exposure, which is a clear improvement from the earlier review, where only 62% reported such data. Using the highest exposure value reported for each campaign, the earlier review showed campaign exposure levels of between 25% and 94%, with a mean of 62%. In the current study, campaigns reported exposure levels of between 35% and 100%, with a mean of 77%. Thus, newer campaigns have achieved higher overall levels of exposure to campaign messages, although it should be noted that exposure often varies across channels within the same campaign, and exposure to one channel (e.g., billboard) may not be as important as exposure to another (e.g., radio).
• Outcome evaluation: The earlier review found the one-group pretest-post-test design to be the most often used outcome evaluation design. It also found that post-only designs were frequently used. In addition, only 7 of their 41 campaign studies (17%) used a quasi-experimental design, which included a control group, although 5 of these 5 evaluations were of differing phases of the same large-scale campaign. Among the current group of campaign evaluations, the one-group, pretest-post-test design was again the most commonly used design, being employed in 13 of 34 campaigns (38%). Some stronger designs did emerge among this group of campaign studies. While the earlier review found that only 17% of campaign evaluations used stronger quasi-experimental designs, the current review finds 30% of campaign evaluations using stronger designs with control groups.
• Campaign effects: The earlier review did not present any clear conclusions concerning the impact of campaigns on outcomes of interest. The present review looked at the 30% of studies with stronger outcome evaluation designs. While 2 of the 10 quasi-experimental studies did not find effects of the campaign, 8 of 10 demonstrated impact on key safer sexual behaviours such as condom use or HIV testing or impact on intentions to engage in these behaviours. "Despite these positive developments, most HIV/AIDS campaigns continue to use weak (i.e., pre-experimental) outcome evaluation designs."
  

Implications of these results for improved design, implementation, and evaluation of HIV/AIDS campaign efforts are discussed. For instance, "[a]t a minimum, researchers and practitioners in this area should begin a dialogue about more sophisticated approaches to evaluation of HIV/AIDS campaigns, and evaluations of national campaigns in other behavioral areas should be consulted....In addition, when weaker designs must be employed, newer sophisticated methodological innovations to control for confounding factors might be considered." Also, approaches that effectively combine both behavioural and message design theories are more likely to be successful than those that use one without the other.