Evaluating the cost-effectiveness of pre-exposure prophylaxis (PrEP) and its impact on HIV-1 transmission in South Africa

Abstract

Background: Mathematical modelers have given little attention to the question of how pre-exposure prophylaxis (PrEP) may impact on a generalized national HIV epidemic and its cost-effectiveness, in the context of control strategies such as condom use promotion and expanding ART programs.

Methodology/principal findings: We use an age- and gender-structured model of the generalized HIV epidemic in South Africa to investigate the potential impact of PrEP in averting new infections. The model utilizes age-structured mortality, fertility, partnership and condom use data to model the spread of HIV and the shift of peak prevalence to older age groups. The model shows that universal PrEP coverage would have to be impractically high to have a significant effect on incidence reduction while ART coverage expands. PrEP targeted to 15-35-year-old women would avert 10%-25% (resp. 13%-28%) of infections in this group and 5%-12% (resp. 7%-16%) of all infections in the period 2014-2025 if baseline incidence is 0.5% per year at 2025 (resp. 0.8% per year at 2025). The cost would be $12,500-$20,000 per infection averted, depending on the level of ART coverage and baseline incidence. An optimistic scenario of 30%-60% PrEP coverage, efficacy of at least 90%, no behavior change among PrEP users and ART coverage less than three times its 2010 levels is required to achieve this result. Targeting PrEP to 25-35-year-old women (at highest risk of infection) improves impact and cost-effectiveness marginally. Relatively low levels of condom substitution (e.g., 30%) do not nullify the efficacy of PrEP, but reduces cost-effectiveness by 35%-40%.

Conclusions/significance: PrEP can avert as many as 30% of new infections in targeted age groups of women at highest risk of infection. The cost-effectiveness of PrEP relative to ART decreases rapidly as ART coverage increases beyond three times its coverage in 2010, after which the ART program would provide coverage to more than 65% of HIV(+) individuals. To have a high relative cost-effective impact on reducing infections in generalized epidemics, PrEP must utilize a window of opportunity until ART has been scaled up beyond this level.

Figure 1. Simple schematic of the model.

S-susceptible, I-infected, P-receiving PrEP, I′-infected while on PrEP, A-receiving ART.

Figure 2. PrEP and UTT: impact and coverage at 2025.

(a) Current % of adults on ART (solid line), expanded ART coverage under UTT (dashed line) and PrEP coverage (dotted line). (b) Adult HIV incidence (age ) in the model and 2008 UNAIDS (diamond) and ASSA (square) estimates . (c) HIV prevalence in antenatal clinics (the upperpoints) as reported in and in the population aged 15–49 (upper solid lines and vertical data points) as reported in . The lower solid line shows the prevalence of HIV people without ART in the population aged 15–49. (d) Crude death rate. Data from Statistics South Africa () , Anderson et al. (), and the US Census Bureau (◊).

Figure 3. Percentage of new infections (cumulative between 2014 and 2025) averted due to PrEP in targeted group, in addition to infections averted due to ART and condom use.

Baseline incidence 0.5% per year at 2025. Contours in intervals of 0%, 5%, 10%, and so on. Vertically spaced points depicts reduction in incidence among all adults. The ordinate corresponds to the 10% intervals of targeted PrEP coverage and the co-ordinate to ART coverage in 2025 reaching twice its 2010 level.

Figure 4. Percentage of new infections (cumulative between 2014 and 2025) averted due to PrEP in targeted group, in addition to infections averted due to ART and condom use.

Baseline incidence 0.8% per year at 2025. Contours in intervals of 0%, 5%, 10%, and so on. Vertically spaced points depicts reduction in incidence among all adults. The ordinate corresponds to the 10% intervals of targeted PrEP coverage and the co-ordinate to ART coverage in 2025 reaching twice its 2010 level.

Figure 5. Incremental cost-effectiveness ratio of PrEP: (incremental cost of PrEP)/(additional infections averted due to PrEP).

Baseline incidence 0.5% per year at 2025. Contours in intervals of $10,000. Contours above $50,000 are grayed out.

Figure 6. Incremental cost-effectiveness ratio of PrEP: (incremental cost of PrEP)/(additional infections averted due to PrEP).

Baseline incidence 0.8% per year at 2025. Contours in intervals of $10,000. Contours above $50,000 are grayed out.