Estimating the range of potential epidemiological impact of pre-exposure prophylaxis: run-away success or run-away failure?

To investigate the influence of potential interactions between key aspects of a pre-exposure prophylaxis (PrEP) intervention on projections of epidemiological impact and cost-effectiveness.

A mathematical model representing the HIV epidemic and intervention context in Nyanza province in Kenya was developed. We consider a scenario whereby a fixed annual budget is allocated to a PrEP intervention. A standard projection of impact is generated, assuming that the unit cost of PrEP, adherence to PrEP and the ability of the programme to direct PrEP to those at high risk, all stay constant. The influence of dynamic assumptions and possible interactions between PrEP intervention assumptions is then assessed in comparison.

The cumulative impact of a PrEP intervention could be increased approximately two-fold, relative to the standard projection, if positive interactions (between coverage and cost, coverage and adherence, prioritization and time) are assumed, whereas negative interactions between these factors could almost entirely negate the preventive benefit of the PrEP intervention. The corresponding estimates of cost per infection averted span a wide range from $2060 to $36360.

Multiple potentially interacting factors will determine the impact of PrEP. Model forecasts should reflect that uncertainty and programmes should focus on these factors and measure them, to maximize the impact of programmes.