Abstract Body

HIV disease progresses rapidly in neonates with mortality peaking in the first 2-3 months of life. Late diagnosis causes delays in access to antiretroviral treatment (ART), often past the mortality peak. This study assessed the cost of HIV early infant diagnosis (HEID) at birth using point-of-care nucleic acid tests (PoC-NATs) and its effect on ART initiation.

We conducted a microcosting study from a provider perspective at 28 primary health facilities in Mozambique and Tanzania. Cost data were collected for set-up and operations, reflecting fixed and variable costs incurred by the health system. Resource use and outcome data were derived from the LIFE Study. Intervention sites implemented PoC-NAT at birth with follow-up testing at 4-6 weeks. Comparison sites followed the standard HEID algorithm with first PoC-NAT at 4-6 weeks. Mozambique used the Abbot mPIMA platform as true bedside testing in maternity wards for HEID and maternal HIV viral load. Tanzania used the Cepheid GeneXpert platform with on-site laboratory support across several programs.

Of 5895 HIV-exposed infants (HEI) tested, 120 were diagnosed with HIV by three months of age. HEID and ART initiation at birth was associated with a median additional cost of $37.42/infant in Mozambique and $22.69/infant in Tanzania. Consumables contributed the largest share to PoC-NAT unit costs (73-89%). Equipment costs were 6-23% of total costs in both countries, dependent on testing volume. Fewer repeat tests in Tanzania reduced costs on average by 35% compared to Mozambique. Cost sharing of the GeneXpert platform across programs in Tanzania further reduced costs by 13% overall and up to 38% at low volume sites (<10 HEID tests/month). Infants offered birth testing initiated ART a median of 4.29 weeks (95% CI: 4.00, 4.43; p<0.001) earlier compared to the standard of care group.

For perinatally infected neonates, birth testing increases costs but results in significantly earlier ART initiation, potentially to before the period of peak mortality. On-site laboratory support may decrease PoC-NAT error rates and thus reduce repeat testing costs while maintaining same-day delivery of results afforded by PoC testing. Primary health facilities with established PoC HEID programs serving large numbers of HEI or with sufficient demand for other PoC assays run on the same platform could consider birth testing to identify and link HIV-infected neonates to care earlier and potentially reduce HIV-related infant mortality.