Perceived risk versus objectively measured risk of HIV acquisition: a cross-sectional study among HIV-negative individuals in Serodiscordant partnerships with clients attending an Urban Clinic in Uganda

BACKGROUND: Acceptability of Pre-exposure Prophylaxis (PrEP) could be hampered by low self-perceived risk for HIV acquisition. Moreover, discordance between risk perception and actual risk of HIV acquisition is likely to occur. We assessed congruence between the level of self- perceived and that of objectively scored risk of HIV acquisition among HIV-negative individuals in discordant relationships. METHODS: This was a cross-sectional study among a representative sample of HIV-negative adult males and females whose partners were receiving antiretroviral therapy for at least 3 months from the Infectious Diseases Institute Clinic in Kampala, Uganda. Perceived risk was measured based on self-report using a numerical rating scale whereas objective risk was measured using a validated risk score tool. Congruence between perceived risk and objectively scored risk was evaluated using descriptive statistics and validity measures. Incongruence between the two phenomena was further evaluated using univariate and multivariate regression analyses. RESULTS: HIV-negative partners evaluated in this study were mostly male (64%) with a median age of 41 years (IQR 35 to 50). Majority (76.3%) of the partners perceived themselves as low risk for HIV acquisition. Similarly, most (93.8%) were objectively scored as low risk. However, nearly three quarters (72.7%) of partners who were objectively scored as high risk perceived themselves as being at low risk and all were men. The sensitivity and specificity of perceived risk for detecting the objectively measured risk was 27.3 and 76.5% respectively; area under ROC curve = 0.52; 95%CI (0.38, 0.66). The proportion of participants at high risk of HIV acquisition who perceived their risk as low was greater among those whose partners had detectable viral load compared to participants whose partners had undetectable viral load (PR = 0.51; 95%CI 0.29 to 0.90). CONCLUSION: Incongruence between perceived and objectively measured risk of HIV acquisition does occur especially among individuals whose partners had a detectable viral load. PrEP counselling for serodiscordant couples should focus on explaining the consequence of detectable viral load in the HIV-positive partner on HIV transmission risk.