Background: Risk perception is central to behaviour change theories that informed HIV prevention interventions, but mixed evidence exists on associations between risk perception and sexual behaviour. This may be because risk perception is essentially a bi-directional longitudinal concept which is difficult to capture in cross-sectional studies. We use longitudinal data to examine whether changes in sexual behaviour lead to changes in risk perception, and whether changes in risk perception lead to changes in protective behaviour.
Methods: Data on HIV-negative adults (15-54 years) participating in ≥2 of 4 surveys (2003-2013) in a general-population open-cohort study in Zimbabwe were analysed. Individuals not reporting risk perception in the first of two surveys were examined regarding whether changes in sexual risk (having multiple, casual, or concurrent sexual partnerships) between surveys were associated with risk perception at follow-up. Individuals not reporting condom use during last sex in the first survey were examined regarding whether developing risk perception between surveys was associated with condom use at follow-up. Data were analysed as panel and logistic regressions were adjusted for socio-demographic characteristics.
Results: Among respondents not reporting HIV risk perception in the first of two surveys (n=5259/N=10089), increasing sexual risk between surveys was associated with developing risk perception (odds ratio [OR]=1.56, 95% confidence interval=1.09-2.24) (Table). Despite this higher relative risk among those increasing sexual risk, 86.0% of all developing risk perception (993/1155) did not change their sexual risk. Among respondents not using condoms in the first survey (n=7234), developing risk perception was associated with starting condom use (OR=1.52 [1.19-1.94]), controlling for changes in sexual risk, marital status, and other socio-demographic factors (Table). 63.5% of all starting to use condoms (374/590) did not change their risk perception.
Conclusions: This is the first longitudinal study confirming expected associations between changes in sexual behaviour and HIV risk perception, and between changes in risk perception and protective behaviour. The results support prevention programmes to educate about risky behaviour and increase risk perception. However, only a small proportion of changes in condom use could be attributed to changes in risk perception, underlining the importance of other factors that determine HIV prevention behaviour, including factors beyond individual control.

Download the e-Poster