Changing the South African national antiretroviral therapy guidelines: The role of cost modelling

Abstract

Background: We were tasked by the South African Department of Health to assess the cost implications to the largest ART programme in the world of adopting sets of ART guidelines issued by the World Health Organization between 2010 and 2016.

Methods: Using data from large South African ART clinics (n = 24,244 patients), projections of patients in need of ART, and cost data from bottom-up cost analyses, we constructed a population-level health-state transition model with 6-monthly transitions between health states depending on patients' age, CD4 cell count/ percentage, and, for adult first-line ART, time on treatment.

Findings: For each set of guidelines, the modelled increase in patient numbers as a result of prevalence and uptake was substantially more than the increase resulting from additional eligibility. Under each set of guidelines, the number of people on ART was projected to increase by 31-133% over the next seven years, and cost by 84-175%, while increased eligibility led to 1-26% more patients, and 1-17% higher cost. The projected increases in treatment cost due to the 2010 and the 2015 WHO guidelines could be offset in their entirety by the introduction of cost-saving measures such as opening the drug tenders for international competition and task-shifting. Under universal treatment, annual costs of the treatment programme will decrease for the first time from 2024 onwards.

Conclusions: Annual budgetary requirements for ART will continue to increase in South Africa until universal treatment is taken to full scale. Model results were instrumental in changing South African ART guidelines, more than tripling the population on treatment between 2009 and 2017, and reducing the per-patient cost of treatment by 64%.

Fig 1. Health-state transition model.

with ar: age rate, tp: transition probability, fr: failure rate, cr: rate of initiation of second-line ART, dr: default rate, mr: mortality rate. Blue arrows and boxes represent drug choices and transitions under the Old Guidelines scenario, red arrows and boxes represent the New Guidelines and Full WHO Guidelines scenarios. Drugs and transitions that are the same in all scenarios are represented in grey and white. For better legibility, rates are represented for the first row or column only; small black arrows represent movements in both directions; for drug choices, colours in the first row are representative of all rows; and only those drugs that change between scenarios are mentioned. *15% and **16% for children in the age group 6–13.

Fig 2. Development of HIV Conditional Grant amounts (2009/10 to 2011/12), in comparison to total patient numbers and total ART cost as calculated by NACM.

Fig 3. Government expenditure on HIV (2003/04 to 2018/19) [2017 ZAR].

Data updated based on [18], sources: Estimates of Provincial Expenditure/ Estimates of National Expenditure 2004/5 to 2014/15, Medium Term Policy Statements, Division of Revenue Bill/ Acts; all from National Treasury.