Article Text
PDF
Abstract
Background Condom promotion among female sex workers (FSWs) and men who have sex with men (MSM) is one of the key components of the Avahan HIV intervention currently underway in India. High rates of consistent condom use are reported across sites. However, condoms are only effective if they do not break or slip during intercourse, which occur frequently, especially in MSM.
Methods We use a deterministic compartmental mathematical model of MSM divided into high and low risk, with average condom use increasing over time, to investigate how breakage influences the course of the epidemic. High-risk MSM (HRMSM) sell sex or have many casual partners, and typically have mostly receptive anal sex. Prior parameter ranges (including condom use) for the model are based on one round of cross-sectional behavioural and biological (IBBA) surveys of HRMSM from four districts in Karnataka state in 2008. In these surveys 16.7% (82/491) of MSM report condom breakage in the last sex act. Condom use increased to 93% by 2008, and was assumed to be constant afterwards. The prior ranges are sampled using Latin hypercube sampling. Model runs that agree with measured HRMSM HIV prevalence (12.7–18.9%) provide the posterior parameter set fitting the data, and are used to make predictions of HIV prevalence with and without condom breakage after the start of the intervention in 2004.
Results Abstract P1-S5.15 figure 1 shows change in high/low risk prevalences with and without condom breakage. Even with high rates of condom breakage, HIV prevalence among MSM declines from 16.7% (95% CI 12.7 to 20.8%) in 2004 to 7.9% (5.9 to 9.8%) in HRMSM by 2015. Without condom breakage HIV prevalence declines faster, to 5.6% (4.2 to 6.9%) by 2015. Compared to no condom breakage, there will be 10.3% (10.7%) more cumulative infections with breakage in high-risk (low-risk) MSM from 2004 until 2015. The annual extra fraction of new infections due to condom breakage increases from 4.1% (4.3%), in 2004, to 158.9% (165.3%), in 2015 in high-risk (low-risk) MSM, although the absolute number of infections due to condom breakage decreases as the epidemic is declining and condom use increases.
HIV prevalence against time for high-risk MSM (upper curves)and low-risk MSM (lower curves). Shaded region shows 2.5–97.5 percentiles and IQR with breakage. The dashed lines show the 2.5, 97.5 percentiles and quartiles without condom breakage.
Conclusions HIV prevalence is projected to decline in MSM in Karnataka, given high rates of reported condom use, even with frequent breakage. However, there will be a large fraction of extra infections due to condom breakage. HIV intervention programmes should examine reasons for high rates of breakage and take steps to address this.