Optimizing infant HIV diagnosis with additional screening at immunization clinics in three sub-Saharan African settings: a cost-effectiveness analysis

Introduction: Uptake of early infant HIV diagnosis (EID) varies widely across sub-Saharan African settings. We evaluated the potential clinical impact and cost-effectiveness of universal maternal HIV screening at infant immunization visits, with referral to EID and maternal antiretroviral therapy (ART) initiation. Methods: Using the CEPAC-Pediatric model, we compared two strategies for infants born in 2017 in Côte d’Ivoire (CI), South Africa (SA), and Zimbabwe: (1) existing EID programmes offering six-week nucleic acid testing (NAT) for infants with known HIV exposure (EID), and (2) EID plus universal maternal HIV screening at six-week infant immunization visits, leading to referral for infant NAT and maternal ART initiation (screen-and-test). Model inputs included published Ivoirian/South African/Zimbabwean data: maternal HIV prevalence (4.8/30.8/16.1%), current uptake of EID (40/95/65%) and six-week immunization attendance (99/74/94%). Referral rates for infant NAT and maternal ART initiation after screen-and-test were 80%. Costs included NAT ($24/infant), maternal screening ($10/mother–infant pair), ART ($5 to 31/month) and HIV care ($15 to 190/month). Model outcomes included mother-to-child transmission of HIV (MTCT) among HIV-exposed infants, and life expectancy (LE) and mean lifetime per-person costs for children with HIV (CWH) and all children born in 2017. We calculated incremental cost-effectiveness ratios (ICERs) using discounted (3%/year) lifetime costs and LE for all children. We considered two cost-effectiveness thresholds in each country: (1) the per-capita GDP ($1720/6380/2150) per year-of-life saved (YLS), and (2) the CEPAC-generated ICER of offering 2 versus 1 lifetime ART regimens (e.g. offering second-line ART; $520/500/580/YLS). Results: With EID, projected six-week MTCT was 9.3% (CI), 4.2% (SA) and 5.2% (Zimbabwe). Screen-and-test decreased total MTCT by 0.2% to 0.5%, improved LE by 2.0 to 3.5 years for CWH and 0.03 to 0.07 years for all children, and increased discounted costs by $17 to 22/child (all children). The ICER of screen-and-test compared to EID was $1340/YLS (CI), $650/YLS (SA) and $670/YLS (Zimbabwe), below the per-capita GDP but above the ICER of 2 versus 1 lifetime ART regimens in all countries. Conclusions: Universal maternal HIV screening at immunization visits with referral to EID and maternal ART initiation may reduce MTCT, improve paediatric LE, and be of comparable value to current HIV-related interventions in high maternal HIV prevalence settings like SA and Zimbabwe. We used the Cost‐Effectiveness of Preventing AIDS Complications (CEPAC)‐Pediatric model, a validated Monte Carlo simulation model of paediatric HIV acquisition, disease progression, diagnosis and treatment [10, 11, 12, 13, 14]. Children were simulated from birth until death. In the model, children who are HIV‐exposed have a risk of intrauterine or intrapartum HIV acquisition dependent on maternal ART use during pregnancy (reflecting prevention of mother‐to‐child transmission [PMTCT] coverage) and CD4 count (reflecting disease stage). Children who are HIV‐exposed but uninfected face a monthly risk of postnatal HIV acquisition based on maternal ART use, infant antiretroviral prophylaxis and maternal disease stage (including acute infection during breastfeeding) until cessation of breastfeeding, and no risk thereafter. All simulated children face monthly risks of non‐HIV‐related mortality. CWH face additional risks of opportunistic infections (OIs), and OI‐ and HIV‐related mortality based on their CD4% (age <5 years) or CD4 count (age ≥5 years), retention of care and ART use. Details of HIV disease progression; ART regimens, monitoring and outcomes; and loss to follow‐up and return to care, are provided in the Appendix and at https://mpec.massgeneral.org/cepac‐model/. We simulated all infants born in CI, SA and Zimbabwe in 2017 (the most recent year for which complete data were available), including infants born to women with and without HIV. These countries represent variation in key characteristics of the HIV epidemic, including maternal HIV prevalence (low/high/medium), EID uptake (low/moderate/high), maternal ART coverage (low/high/high) and income level (low/low/middle). Country guidelines recommend EID at six weeks of age for most infants in CI and Zimbabwe, and at birth and 10 weeks of age in SA; both are consistent with World Health Organization recommendations. We modelled six‐week EID for all countries to permit us to isolate the most influential parameters (EPI uptake, screening costs and care costs). We compared two strategies in each setting: (1) six‐week NAT for infants with known HIV exposure (EID), and (2) EID plus universal maternal rapid diagnostic testing (RDT) at six‐week infant immunization visits, with positive result leading to referral for infant NAT and maternal ART initiation (screen‐and‐test). In screen‐and‐test, HIV status‐aware mothers who did not attend an EID visit at six weeks with their infant, but presented to an EPI visit, could be referred back to an EID clinic for NAT. In both strategies, we simulated a confirmatory NAT algorithm before ART initiation. Children who develop an OI and are not in care experience a probability of presenting to care, undergoing HIV testing and linking to care. Women with HIV were not directly simulated in either strategy; rather, maternal characteristics were reflected in changes in infant HIV acquisition risk over time. We derived clinical data to inform cohort characteristics, MTCT risks, assay characteristics and treatment outcomes from published trials and cohort studies in sub‐Saharan Africa (Table 1, Sections I‐III) [15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35]. We used Ivoirian, South African and Zimbabwean programmatic data for three separate country‐specific base‐case analyses (Table 1, Section IV), and varied these parameters from sensitivity analyses. We used national estimates of maternal HIV prevalence during pregnancy (4.8%, 30.8%, 16.1%) and postpartum maternal HIV incidence (0.4/100 person‐years (PY), 2.9/100PY, 1.5/100 PY) [36, 37, 38, 39]. Maternal knowledge of HIV status during pregnancy (86%, 89%, 84%) reflected the product of antenatal care (ANC) attendance (91%, 94%, 93%) and HIV testing coverage during ANC or at delivery (95%, 95%, 90%) [36, 37, 38, 40]. Maternal ART coverage during both pregnancy and breastfeeding (70%, 95%, 95%) and uptake of existing EID programmes (40%, 95%, 65%) were from UNAIDS data [36]. In screen‐and‐test, the probability of maternal screening was the product of six‐week immunization coverage (99%, 74%, 94%), which was derived from UNICEF country‐specific data, and a 90% probability of offer and acceptance of testing [5, 6, 7, 8, 9, 41]. At EPI visits, newly diagnosed mothers and HIV status‐aware mothers who missed an EID visit at six weeks had a modelled 80% probability of successful referral to existing EID programmes [42]. We assumed that the probability of maternal linkage to HIV care and ART after the EID visit was equal to country‐specific maternal ART coverage. Once diagnosed through any mechanism, including detection after OI, infants had a 71% probability of linking to HIV care and ART in all countries [43, 44]. Selected base‐case input parameters for the CEPAC‐Paediatric model analysis of EID and screen‐and‐test ANC, antenatal care; ART, antiretroviral therapy; EID, early infant diagnosis; EPI, expanded programme on immunization; IP, intrapartum; IU, intrauterine; MTCT, mother‐to‐child transmission; NAT, nucleic acid test; OI, opportunistic infection; PP, postpartum; PY, person‐years; RDT, rapid diagnostic test; SD, standard deviation. We derived country‐ and sex‐ specific mortality rates for HIV‐unexposed children from UNAIDS HIV‐deleted life tables, and mortality rates for HIV‐exposed/uninfected infants from pooled UNAIDS analyses [45, 46]. Therefore, life expectancy (LE) projections are not expected to be directly comparable across country settings. Risks of disease progression without ART were calibrated to survival data for African children and adults [11, 47, 48, 49, 50, 51, 52]. Survival and OI risks for children and adults on ART were calibrated to clinical trial data [11, 33, 34, 35, 49, 53]. We modelled costs of HIV testing and clinical care in 2018 USD (Table 1, Section V). Costs specific to each country included routine HIV care (e.g., laboratory monitoring, personnel, facilities) and acute OI care [54, 55, 56, 57, 58]. ART costs ranged by age and weight ($5 to 31/month) and were derived from Clinton Health Access Initiative price lists and World Health Organization weight‐based dosing [59, 60]. Assay costs were modelled as “fully loaded,” including personnel time and training, and were derived from Global Fund and published HIV testing reports: NAT ($24/assay), maternal HIV screening ($10/mother–infant pair: $3 for RDT plus personnel and training costs; Appendix p3) and ART monitoring (CD4: $5 to 12/assay; HIV RNA: $17 to 32/assay) [56, 61, 62, 63, 64]. In screen‐and‐test, per‐person lifetime costs included the cost of maternal ART during breastfeeding for mothers diagnosed and linked to care through screening. Primary model outcomes were MTCT proportion at six weeks and after weaning, incremental yield of the screening programme (the additional number of infants identified with HIV divided by the number of women reached by the screen‐and‐test programme), proportion of all CWH identified and linked to care, 1‐ and 2‐year survival, LE (years) and average per‐person lifetime costs from a healthcare system perspective (2018 USD). We projected outcomes for both CWH and the complete birth cohort (including CWH, uninfected children with HIV exposure and HIV‐unexposed children), but not for mothers. Using discounted (3%/year) birth cohort outcomes, we calculated incremental cost‐effectiveness ratios (ICERs) in USD per year‐of‐life saved ($/YLS). In the absence of consensus about country‐specific cost‐effectiveness thresholds, we compared ICERs to (1) the 2018 per‐capita GDP in each country (CI: $1720/YLS, SA: $6380/YLS, Zimbabwe: $2150/YLS), and (2) the CEPAC‐generated ICER of a paediatric HIV programme offering 2 versus 1 lifetime ART regimens (e.g., offering second‐line ART; CI: $520/YLS, SA: $500/YLS, Zimbabwe: $580/YLS) [65, 66, 67, 68]. This ICER can be used to estimate the health benefits that would be foregone by diverting resources from an existing programme to a novel intervention, as a reasonable proxy for the value of alternative claims upon limited resources for HIV services. We varied cost‐effectiveness thresholds in sensitivity analyses. We examined the impact of a birth and 10‐week EID schedule in SA. Modelled EID coverage was 67% at birth and 80% at 10 weeks [69]. HIV status‐aware mothers who presented to an EPI visit at six weeks, but whose infant did not receive a test at birth, could be referred back to an EID clinic in the next month. We followed international guidance on uncertainty analysis and reported extensive univariate and multivariate uncertainty analyses, using literature‐based estimates of the uncertainty around key parameters [70]. For each country, we varied key epidemic‐specific parameters, uptake at each care “cascade” step for EID and screen‐and‐test, and costs of diagnostics and HIV care. We first varied these parameters through their published ranges, where available, to identify the impact of data uncertainty on results, including: maternal HIV prevalence, knowledge of HIV status, HIV incidence and ART coverage during pregnancy and breastfeeding; uptake of existing EID programmes; immunization coverage; and cost of infant NAT. We next evaluated wider ranges for remaining parameters where data‐informed ranges were unavailable (e.g. linkage to EID after screening and screening costs) in order to identify the threshold values at which screen‐and‐test would reach each cost‐effectiveness threshold. Table S1 shows the ranges through which parameters were varied. In multivariate sensitivity analyses, we varied the most influential individual parameters simultaneously.