Viral load dynamics in HIV-1 infected sex workers in the Sex Worker Outreach Project (SWOP)-Kenya after art initiation

Abstract

Models have been developed to simplify complex processes for purposes of understanding them in order to solve a problem and make decisions. They are designed for a specific situation/process in manner that they reflect the reality. The main aim of this study was to model the viral load dynamics in HIV infected sex workers at SWOP after initiation of Antiretroviral Therapy. This was a stochastic model using continuous time. The model structure comprised of four compartments- initiated into care, virally suppressed, virologically failed and viral load missing. A set of differential equations were defined to explain the transition process between the model compartments. HIV data for sex workers was obtained from SWOP city clinic to estimate the rates of transition between the compartments. For demographic rates, information from recent HIV reports in Kenya was used to define the parameters. Simulations were done to project the changes in model compartments within a period of 5 years. The net reproductive number was determined using the defined rate parameters, to set the threshold to control the incidence of virological failure among HIV positive sex workers in care. The two year projection showed that after first two years (2019), for each patient initiated into care/ART, we have <1 patient in care achieving viral load suppression. In the five year projections, it was noted that after 2 years and 8 months, each patient initiated into care generates >1 virally suppressed patient. A similar trend was observed among male and female patients; however for the four year projection curve for male, the initiated into care curve and virally suppressed curve cross later than the female curves (after 3.5 years in males vs 2.5 years in females). To control the number of virological failure; it was determined that for one patient x initiated into care at least three patients in care should achieve viral suppression. From the simulation, it was noted that it would take the project at least 7 years to achieve this ratio. The NHAS 90/90/90 model was successfully adapted; however the model can be improved further by adjusting the rates using accruing data from the project. The rates of initiation to care and viral suppression was lower in male than female sex workers, therefore SWOP should come up with tailored strategies to targeting undiagnosed HIV positive male sex workers to increase the number of HIV positive male sex workers with suppressed viral load. This will in turn reduce the rate of HIV transmission from infected male sex workers to susceptible people.