Project R-5556

Advanced development, testing and application of the Simpact Cyan infividual-base simulation tool to model preferential mixing among ART patients and heritability of HIV set point viral load. (Research)

A first application concerns the effect of preferential mixing among ART patients. By this we mean the situation in which a disproportionately high fraction of relations formed by ART patients is formed with other ART patients (for various social, psychological, safety and health reasons). The key question here is if/how the impact of TasP on HIV incidence could be modified as a result of preferential mixing among ART patients. Results from a simple risk equation model suggest that such non-random mixing pattern may substantially affect the impact of TasP on HIV incidence, and the size and direction of the effect depends on the local context in which the TasP programme is rolled out. In particular, key determinants of the effect are HIV prevalence, ART coverage and ART effectiveness (~ adherence). However, this simple model was only able to look at the impact-modifying effect of preferential mixing on the instantaneous hazard of HIV acquisition, not at the cumulative effect over several years. Further, it assumed that individual-level metrics of risk behaviour such as condom use and frequency of partner change remain the same after ART initiation. The new model application will allow for the investigation of the effect of preferential mixing among ART patients on HIV incidence. A second model is planned, to study the effect of heritable set point viral load on the impact of TasP. High set point viral load is associated with higher infectiousness, as well as with accelerated progression to AIDS and death, the latter of which reduces the infectious period and consequently the transmission potential. Further, people have a tendency to preferentially form relationships with partners who have a similar sexual activity level (assortative mixing), and set point viral load may be heritable from the infector to the newly infected individual. As a result of these behavioural and virological phenomena, the HIV set point viral load and sexual activity level of individuals may not be independent and this bio-behavioural correlation may have non-negligible effects on the impact of TasP on HIV incidence. Thus far, all population projection models have assumed that variation in infectiousness is random and non-heritable. However, it is unclear how sensitive model-based TasP impact projections are to the (false) assumptions around independence of set point viral load. The new model application will allow for the investigation of the effect of heritable HIV set point viral load on TasP impact (on HIV incidence).

01 January 2014 - 31 December 2016