Gay men and other men who have sex with men (MSM) suffer particular sexual health inequalities and are the principal group at greatest risk of acquiring HIV in the UK . They are also disproportionately affected by a range of sexually transmitted infections (STIs), particularly syphilis and gonorrhoea . Responding to the sexual health needs of MSM represents a key public health priority and there is increasing global demand for evidence-based policy and practice with regard to the best approach to sexual health promotion for this group. Although there has been a trend to frame, and inevitably fund, HIV prevention in biomedical terms , there remains a clear role for behavioural scientists and health psychologists in informing intervention design and development, in contributing to understanding the process of behaviour change, and, as reported here, in evaluating interventions focussed upon sexual health promotion.
There is global consensus regarding the need for ‘combination’ HIV prevention [4–6]. This is usually conceptualised as a combination of different types of intervention (typically categorised as biomedical, psychological/behavioural and structural) straddling a spectrum of different levels of delivery (e.g., somatic, individual, dyad, group, community, societal, cultural). The evidence for the success of such interventions is governed not only by effectiveness, but also by existing standards of research methods and designs . Thus compared to biomedical interventions, fewer trials concerning either psychological/behavioural or structural interventions get funded and their potential to contribute to the evidence base is systematically negated . In relation to the high quality evidence that does exist, questions remain about key choices in the measurement of effectiveness.
One longstanding and arguably central tenet of HIV prevention, working at the psychological/behavioural and structural levels and delivered to the community, has been the mass media campaign or, more recently with nuanced targeting, the social marketing approach [8, 9]. There is some evidence across a range of populations and settings that supports the effectiveness of these approaches with respect to HIV prevention; however, effect sizes tend to be small to moderate, and short lived [10, 11]. A dose–response effect to mass media messages has also been demonstrated in various settings and with various international populations, with increased exposure to mass media resulting in increased positive behavioural change . These findings are echoed across the wider literature regarding mass media interventions targeting health behaviours within a range of other health conditions .
Mass media approaches represent particularly complex interventions and are notoriously difficult to evaluate [7, 14–16]. One central re-occurring problem within this field is the lack of studies featuring randomised control trials (RCTs) or other experimental designs involving control groups. In part, these limitations relate to the demand characteristics of mass media campaigns themselves: by design they aim to maximise spread and saturation within a population. Further, evaluations of campaigns are often subject to compromise from the outset, given problems with funding and external pressures to implement or roll out campaigns within a few months, leaving researchers insufficient time to obtain baseline pre-exposure measures. Noar et al.  highlighted that only 30% of the published mass media campaign evaluations targeting HIV used robust evaluative designs, and most used either within-group pre-test-post-test designs (38%) or post-test only designs (32%). The key problem with these designs is that it is difficult to disentangle campaign effects from other explanations such as, for example, reverse causality, differences within sample characteristics, secular trends and/or historical events . A further systematic review of social marketing interventions to increase HIV/STI testing uptake (specifically amongst MSM and male-to-female transgender women) showed that multimedia social marketing campaigns had a significant impact on HIV testing uptake: odds ratio (OR) = 1.58, 95% Confidence Intervals (CI) = 1.40 to 1.77 . This supports the results of a previous systematic review addressing mass media and HIV testing in various populations . However, the campaigns identified were not found to be effective in increasing STI testing uptake (OR = 0.94, 95% CI = 0.68 to 1.28) and, in only including RCTs and other controlled research designs, it identified only three studies [20–22], none of which explicitly utilised theoretical frameworks.
Given the strict inclusion criteria of the systematic reviews outlined above, it is worth examining the findings in relation to evaluations of mass media interventions that are similar to that reported here. Martínez-Donate et al. , for example, report on a social marketing campaign, ‘Hombres Sanos’, which targeted behaviourally bisexual Latino men in the USA. It used a variety of print materials in a variety of locations with radio ads and local activities in clubs. Behaviourally bisexual men were recruited through multiple cross-sectional surveys. Of these, 87% reported exposure to the campaign. Details of self-reported behavioural changes as a result of the campaign were assessed. Behavioural changes resulting from exposure to the campaign included around 20% claiming it had made them use condoms. A further 10.8% reported that as a result of the campaign they had sought an HIV test, while 9.2% reported that they had sought an STI test. Similarly, Plant et al. , in an evaluation of a social marketing campaign addressing syphilis screening, found that 27% of their convenience sample spontaneously mentioned the campaign with no prompting and an additional 44% remembered the campaign when prompted. With another campaign addressing syphilis screening, Stephens et al.  indicated that 33% of their convenience sample reported some recall of the campaign, with around half of these recalling the campaign spontaneously, and 44% recalled the campaign when prompted with a visual aid.
Against this background of promising but partial evidence (and with a recognition of the common design challenges which accompany evaluating mass media campaigns), we report an outcome evaluation of the Make Your Position Clear (MYPC) mass media campaign. In this paper, we focus on our two principal research questions: 1. What was the extent of self-reported exposure to the MYPC campaign among men frequenting venues for gay men and MSM? 2. Did sexual health related behaviours (i.e., unprotected anal intercourse (UAI), HIV testing and STI testing and use of appropriate lubricant) vary by degree of exposure to the campaign?