Optimizing HIV Pre-Exposure Prophylaxis Implementation among Men Who Have Sex with Men in Toronto: A Dynamic Modelling Study
Methods: We developed a dynamic, stochastic compartmental model of HIV transmission among the ~57,000 MSM in Toronto, Canada, stratified by HIV infection (CD4), serostatus (known/unknown diagnosis), and sexual behaviour. Parameterization was performed using local epidemiologic data. We calibrated the model to observed data on annual HIV diagnoses (300-400), annual HIV-attributable deaths (40-65), and ART coverage among MSM (50-70%). Baseline annual HIV testing was 22%. Strategies examined included: a) uniform PrEP delivery vs. targeting the highest-risk decile of MSM (avg. 36 partners/year); b) varying PrEP efficacy (44% to 99%); c) increasing HIV test frequency (Q3 to Q24 months). Outcomes included HIV infections averted and the incremental cost (in $CAD) per incremental quality-adjusted-life-years (QALYs) gained over 10 years.
Results: Use of PrEP among all HIV-uninfected MSM at 25, 50, 75, and 100% coverage prevented 832, 1387, 1693, and 1736 infections respectively, with costs per QALY increasing from $230,000 to $300,000. Targeted PrEP for the highest-risk MSM at 25, 50 75, and 100% coverage prevented 241, 452, 555, and 590 infections respectively, with costs per QALY ranging from $45,000-60,000 CAD. Maximizing PrEP efficacy prevented 424 infections with a cost per QALY of $32,000 (assuming PrEP in 25% of high-risk). HIV testing alone (Q3 months) averted 50% of infections with a cost per QALY of less than $10,000. However, increasing HIV testing frequency had minimal prevention impact. Maximizing PrEP efficacy increased the number of infections prevented. Assuming PrEP use in 25% of high-risk MSM with 99% efficacy and with Q3 month testing, a $32,000 cost per QALY was achieved.
Conclusion: Among those examined, the optimal implementation strategy for PrEP over the next 10 years in Toronto is to target high-risk MSM with strategies to maximize efficacy. Frequent HIV testing alone in high-risk individuals provides a substantial benefit.
S. Mishra, None