90-90-90 by 2020? Estimation and Projection of the HIV epidemic and ART Programme in Zimbabwe - 2017-2020

Introduction

The 90-90-90 targets set by the United Nations aspire to 73% of people living with HIV (PLHIV) being virally suppressed by 2020. Using the HIV Synthesis Model, we aim to mimic the epidemic in Zimbabwe and make projections to assess whether Zimbabwe is on track to meeting the 90-90-90 targets and assess whether recently proposed UNAIDS HIV transition metrics are likely to be met.

Methods

We used an Approximate Bayesian Computation approach to identify model parameter values which result in model outputs consistent with observed data, evaluated using a calibration score. These parameter values were then used to make projections to 2020 to compare with the 90-90-90 targets and other key indicators.

We also calculated HIV transition metrics proposed by UNAIDS (percentage reduction in new HIV-infections and AIDS related mortality from 2010 to 2020, absolute rate of new infections and AIDS related mortality, incidence-mortality ratio and incidence-prevalence ratios).

Results

After calibration, there was general agreement between modelled and observed data. The median predicted outcomes in 2020 were: proportion of PLHIV (aged 15-65) diagnosed 0.91 (90% uncertainty range 0.87, 0.94) [0.84 men, 0.95 women]; of those diagnosed, proportion on treatment 0.92 (0.90, 0.93); of those receiving treatment, proportion with viral suppression 0.86 (0.81, 0.91). This results in 72% of PLHIV having viral suppression in 2020. We estimated a percentage reduction of 36.5% (13.7% increase to 67.4% reduction) in new infections from 2010 to 2020, and of 30.4% (9.7% increase to 56.6% reduction) in AIDS-related mortality (UNAIDS target 75%). The modelled absolute rates of HIV incidence and AIDS-related mortality in 2020 were 5.48 (2.26, 9.24) and 1.93 (1.31, 2.71) per 1000py respectively. The modelled incidence-mortality ratio and incidence-prevalence ratios in 2020 were 1.05 (0.46, 1.66) and 0.009 (0.004, 0.013) respectively.

Conclusions

Our model was able to produce outputs that are simultaneously consistent with an array of observed data and predicted that whilst the 90-90-90 targets are within reach in Zimbabwe, increased efforts are required in diagnosing men in particular. Calculation of the HIV transition metrics suggest increased efforts are needed to bring the HIV epidemic under control.