Changing the deferral for men who have sex with men - an improved model to estimate HIV residual risk.

Abstract

Eight published studies modelled the impact of changing from a lifetime to time-limited deferral for men who have sex with men (MSM); each predicted greater risk impact than has been observed. This study uses these previous efforts to develop an 'optimized' model to inform future changes to MSM deferrals. HIV residual risk was calculated using observed HIV incidence/prevalence prior to the change in MSM deferral, then with the additional MSM expected under a 12-month deferral for five compliance scenarios, and finally using data observed after implementation of the deferral. Monte Carlo simulation calculated 95% confidence intervals (CI). The architecture of reviewed models was sound, and two were selected for combination into the optimized model. HIV risk estimated by this in the UK under MSM lifetime deferral was 0·102 (95% CI: 0·050-0·172) per million. The model predicted from a 27·8% decrease to a 47·6% increase depending upon compliance pre-implementation of the 12-month deferral. A decrease of 0·9% was observed post-implementation. For Canada, HIV risk under a 5-year deferral was 0·050 (95% CI: 0·00003-0·122) per million. Pre-implementation of the 12-month deferral, the model predicted from 30·2% decrease to 10-fold increase. A decrease of 47·0% was observed after implementation. The optimized modelpredicted HIV risk under 12-month MSM deferral in UK and Canada would remain low, and this was confirmed post-implementation. While the model is adaptable to other deferral scenarios, improved data quality would improve precision, particularly estimates of incidence in individuals likely to donate.