Cost-effectiveness of dual maternal HIV and syphilis testing strategies in high and low HIV prevalence countries: a modelling study

Background: Dual HIV and syphilis testing might help to prevent mother-to-child transmission (MTCT) of HIV and syphilis through increased case detection and treatment. We aimed to model and assess the cost-effectiveness of dual testing during antenatal care in four countries with varying HIV and syphilis prevalence. Methods: In this modelling study, we developed Markov models of HIV and syphilis in pregnant women to estimate costs and infant health outcomes of maternal testing at the first antenatal care visit with individual HIV and syphilis tests (base case) and at the first antenatal care visit with a dual rapid diagnostic test (scenario one). We additionally evaluated retesting during late antenatal care and at delivery with either individual tests (scenario two) or a dual rapid diagnosis test (scenario three). We modelled four countries: South Africa, Kenya, Colombia, and Ukraine. Strategies with an incremental cost-effectiveness ratio (ICER) less than the country-specific cost-effectiveness threshold (US$500 in Kenya, $750 in South Africa, $3000 in Colombia, and $1000 in Ukraine) per disability-adjusted life-year averted were considered cost-effective. Findings: Routinely offering testing at the first antenatal care visit with a dual rapid diagnosis test was cost-saving compared with the base case in all four countries (ICER: –$26 in Kenya,–$559 in South Africa, –$844 in Colombia, and –$454 in Ukraine). Retesting during late antenatal care with a dual rapid diagnostic test (scenario three) was cost-effective compared with scenario one in all four countries (ICER: $270 in Kenya, $260 in South Africa, $2207 in Colombia, and $205 in Ukraine). Interpretation: Incorporating dual rapid diagnostic tests in antenatal care can be cost-saving across countries with varying HIV prevalence. Countries should consider incorporating dual HIV and syphilis rapid diagnostic tests as the first test in antenatal care to support efforts to eliminate MTCT of HIV and syphilis. Funding: WHO, US Agency for International Development, and the Bill & Melinda Gates Foundation. Using a Markov decision analytical model, we did a cost-effectiveness analysis of maternal HIV and syphilis testing using individual tests and dual rapid diagnostic tests in Kenya, South Africa, Colombia, and Ukraine—four countries that represent a range of HIV prevalence, syphilis prevalence, and geographical settings. Countries were classified as having low (<5%; Colombia and Ukraine), intermediate (5–15%; Kenya), and high (≥15%; South Africa) HIV prevalence.14 We compared four HIV and syphilis testing scenarios, varying the assay type and timing of testing (table 1). The primary analysis assessed the cost-effectiveness of dual HIV and syphilis testing compared with individual tests at the first antenatal care visit, and the secondary analysis evaluated the effect of retesting for HIV and syphilis during late antenatal care. We modelled the same test types across all four countries because of heterogeneity within countries and to enable comparison between countries. We modelled the use of dual rapid diagnostic test and individual HIV and syphilis tests followed by confirmatory testing. We assumed HIV diagnoses followed two consecutive reactive rapid diagnostic tests. Individual syphilis testing was modelled as a laboratory-based rapid plasma reagin test with confirmatory testing for reactive results with the Treponema pallidum haemagglutination assay. No confirmatory test for syphilis was modelled for dual rapid diagnostic test scenarios. Dual HIV and syphilis testing models scenarios Base case used independent tests at country-specific timepoints. Scenario one used dual HIV and syphilis testing with no retesting during late ANC. Scenario two used independent HIV and syphilis tests with retesting during late ANC. Scenario three used dual HIV and syphilis testing with retesting during late ANC. Dual test is a single, point-of-care rapid test for HIV and syphilis. Independent tests include rapid HIV tests and laboratory-based syphilis tests (RPR and TPHA to confirm reactive results). Late ANC occurred from 36 to 39 weeks’ gestation. ANC=antenatal care. RPR=rapid plasma reagin. TPHA=Treponema pallidum haemagglutination assay. In the base case scenario, we assumed that HIV and syphilis testing with individual tests was done at the first antenatal care visit, defined by country-specific average gestational age at the first visit. In scenario one, we modelled testing at the first antenatal care visit using a dual rapid diagnostic test. Additionally, because maternal HIV retesting is already recommended in many countries, we modelled two scenarios that incorporated maternal HIV retesting to address operational considerations. To estimate the relative health and economic effect attributed to HIV versus syphilis, we compared retesting with individual tests (scenario two) with retesting during late antenatal care and delivery with a dual rapid diagnostic test (scenario three). Retesting scenarios included women who missed HIV or syphilis testing at first antenatal care visit (ie, because they did not attend antenatal care before the gestational age when retesting would occur or because of gaps in test coverage). Retesting scenarios also allowed for HIV retesting at delivery for women not tested during late antenatal care (dual and individual tests), but not for syphilis retesting at delivery. For scenarios in which dual rapid diagnostic tests were modelled (scenarios one and three), we assumed that all women tested received the test regardless of previous HIV diagnosis or syphilis infection. Maternal HIV and syphilis were modelled using separate Markov models (appendix pp 1–2). The HIV model was adapted from a previous model of maternal HIV retesting, and reflects HIV prevalence, incidence, disease progression, and treatment during pregnancy and post partum using 54 Markov states (six maternal HIV states by nine antenatal and post partum periods; Meisner J and Roberts DA, unpublished). The congenital syphilis model reflects syphilis prevalence, incidence, testing, and treatment using eight maternal syphilis states with time-varying transitions. We assumed risk of MTCT of HIV and syphilis and associated infant health outcomes were independent. Risk of infant HIV was modelled for 1 year after birth. The time step in both models was 1 week, and the time horizon was 20 years after birth—a common time horizon that allows our results to be directly compared with other cost-effectiveness analyses. The model was built in Microsoft Excel (2016). Models were parametrised using values from the literature, in-country contacts, and expert opinion; country-specific estimates were used when available (table 2). Although HIV and syphilis models were constructed separately, they share common model parameter inputs. Both models incorporate disease prevalence, incidence, antenatal care attendance, test coverage, test sensitivity and specificity, maternal and infant mortality rates, and probability of infant infection. For each country, we modelled a population of pregnant women equal to the estimated number of annual pregnancies, or the number of livebirths if the number of pregnancies was unavailable. All pregnancies were assumed to be singleton. Fetal loss because of syphilis was modelled, but fetal loss because of HIV was not. Key model parameters Details and sources provided in appendix (pp 18–30). All costs reported in 2017 US$. ANC=antenatal care. ART=antiretroviral therapy. ARV=antiretrovirals. MCH=maternal and child health. MTCT=mother-to-child transmission. PrEP=pre-exposure prophylaxis. RDT=rapid diagnostic test. RPR=rapid plasma reagin. TPHA=Treponema pallidum hemagglutination assay. We modelled infant HIV infection, congenital syphilis, and infant death. Infant HIV infections and deaths were estimated on the basis of total exposure to HIV in utero and post partum, with probabilities dependent on new versus established maternal HIV infection (on the basis of Fiebag stages), maternal viral suppression, maternal antiretroviral therapy (ART), infant ART, and breastfeeding practices. Congenital syphilis outcomes (stillbirth, neonatal death, preterm birth and low birthweight, clinical congenital syphilis, and non-clinical syphilis) were estimated at birth and modified by receipt and timing of adequate maternal syphilis treatment during pregnancy.15 We followed WHO guidelines on the definition of adequate syphilis treatment as at least 30 days before birth (modelled as ≤32 weeks) to prevent fetal infection, and set probabilities of adverse infant outcomes for later treatment equal to those for no treatment. We considered all infants born to untreated and late treated mothers to have congenital syphilis, regardless of symptoms.6 We accounted for duplication of outcomes between the HIV and syphilis models by calculating infant co-infection as the joint probability of syphilis and HIV infection in infants, assuming independence. Syphilis-related stillbirths and neonatal deaths were subtracted from the population at delivery in the HIV model. We converted infant infections and adverse birth outcomes into disability-adjusted life-years (DALYs; appendix pp 16–17).16 Total DALYs were the sum of years lived with disability and years of life lost up to the age of 20 years. We assumed DALYs for each stillbirth and neonatal death were equal. We used disability weights from the Global Burden of Disease study and standard formulas for calculating co-infection disability weights.16, 17 For HIV-infected infants on ART at 1 year, the disability weight for HIV while on ART was applied over the 20-year time horizon. For HIV-infected infants not on ART at 1 year, we assumed 90% of time was spent with symptomatic HIV and 10% with AIDS over the 20-year time horizon. For infants with congenital syphilis, the disability weight was applied only to the first 3 years of life. DALYs were discounted at 3% annually.18 We used a health-care system perspective for the model. Costs of supplies, labour, and treatment were obtained from the literature, in-country contacts, and expert opinion (appendix pp 18–30). Empirical time-motion data collected in Kenya were used to calculate HIV testing costs in the country and estimates of time to testing were applied to other regions with country-specific labour and supply costs. Only incremental costs associated with testing and treatment were included. We excluded start-up costs (eg, training) and distribution costs. Costs given in local currency units were converted to US$ using exchange rates from June of the year they were collected. All costs were adjusted to 2017 US$ using the US gross domestic product deflator and discounted at 3% annually.19 The incremental cost-effectiveness ratio (ICER)—defined as the incremental costs divided by DALYs averted—was calculated by comparing scenario one (dual test at first antenatal care visit) with the base case. Retesting scenarios were compared with the next-best scenario (by DALYs). Dominated scenarios, those that were more costly and less effective than an alternative, were excluded from calculations.20 ICERs less than the country-specific cost-effectiveness threshold were considered cost-effective. We used the following thresholds based on estimated opportunity cost of health investment foregone: US$500 (Kenya),21, 22 $750 (South Africa),23 $1000 (Ukraine),24 and $3000 (Colombia)25 per DALY averted. Uncertainty analyses were done for scenario one. One-way sensitivity analyses were used to assess the effect of changing individual model parameters, including HIV and syphilis prevalence, test coverage, test characteristics, gestational age at first antenatal care visit, and costs. We used low and high parameter values from confidence interval bounds for test characteristics, and 20% relative changes for all other parameters. We also did scenario implementation analyses to estimate the effect of epidemic changes (declines in HIV and syphilis prevalence) or HIV test acceptance on cost-effectiveness, calculated using Microsoft Excel (2016). Collaborators from WHO were involved in study design, interpretation of results, and manuscript development; no other funders had a role in design, analysis, or interpretation. All collaborators had full access to all the data in the study. ALD had final responsibility for the decision to submit for publication, with concurrence from all authors.