Publications

Background: Long-acting injectable HIV pre-exposure prophylaxis (PrEP), including Lenacapavir, has the potential to accelerate HIV incidence declines in eastern and southern Africa (ESA). However, high product and delivery costs and constrained budgets necessitate efficient prioritization strategies to maximize impact and achieve cost-effectiveness. Methods: We used district-level HIV incidence estimates published by UNAIDS to estimate the direct health and economic impact of prioritizing Lenacapavir delivery according to geography, age, and sex across 837 districts in 11 high-burden ESA countries. Infections and disability-adjusted life years (DALY) averted, number needed to treat (NNT), cost per DALY averted, and price thresholds to achieve cost-effectiveness were estimated across geographic prioritization scenarios. Cost-effectiveness was assessed against a $500 per DALY averted threshold, assuming $5,000 discounted lifetime HIV treatment costs and 10 DALYs per HIV infection. Sensitivity analyses varied Lenacapavir costs (commodities + delivery) per person per year (pppy) ($125 versus $55), DALYs per HIV infection (7.5), and the risk differentiation among those who uptake long-acting PrEP. Results: HIV incidence varied substantially across ESA, with 50% of new infections in districts containing less than 20% of at-risk adults. Lenacapavir cost-effectiveness varied accordingly, with high-incidence districts exhibiting substantially lower NNT and higher price thresholds for cost-effective delivery. In high-incidence districts, [>5/1,000 person-years (py)], of South Africa, Mozambique, Lesotho, and eSwatini, Lenacapavir would be cost-effective at $50-100 pppy. In South Africa, at annual cost $55 pppy, Lenacapavir was cost-effective in all 52 districts when provided to women aged 15-24 years with incidence exceeding twice the district average and could reach approximately 18-20% of new infections while covering 4% of the full HIV-negative adult population aged 15-49 years. Geographically optimized prioritization in South Africa with minimal age and risk-group stratification achieved efficiency comparable to country-level prioritization to high-risk groups and key populations (~20% incidence reduction with 3-5% coverage). Impact and cost-effectiveness were sensitive to assumptions about risk heterogeneity. Conclusions: Lenacapavir impact and cost-effectiveness varies substantially across geographic settings, driven primarily by variation in HIV incidence. Simple incidence-based models can identify where universal provision to certain demographic groups is both impactful and cost-effective, particularly in high-incidence districts and age groups.

Background: Zimbabwe’s national guidelines for sexually transmitted infection (STI) management recommend that high-risk women presenting with vaginal discharge syndrome (VDS) are prescribed antibiotics for gonorrhoea (Neisseria gonorrhoeae (NG)), chlamydia (Chlamydia trachomatis (CT)), trichomoniasis (Trichomonas vaginalis (TV)) and bacterial vaginosis (BV). The performance of this approach depends on its clinical interpretation and implementation. Here, we investigate the potential relative impact of an NG/CT/TV point-of-care (POC) test on undertreatment, overtreatment and disease burden in the context of different implementations of syndromic management of women with VDS.
Methods: We created an agent-based model with an age- and risk-stratified sexual network and modelled co-circulation of NG, CT and TV along with HIV and BV. We estimated symptomatic proportions and care-seeking rates under three different scenarios around the implementation of treatment guidelines, corresponding to all, most or half of women being treated for NG+CT upon presentation with VDS. For each implementation scenario, we estimated disease burden and over/undertreatment rates assuming continuation of the standard of care with/without a POC NG/CT/TV test available over 2027–2040.
Results: Under a treat-all interpretation of the syndromic management guidelines, we estimate that 70%–80% of antibiotics for NG/CT would currently be given to women without these infections. Overtreatment would fall to less than 5% if a sensitive POC test for NG/CT/TV were available. However, if the implementation of the guidelines implies that only half of women seeking care for VDS are treated, then a POC test would also reduce undertreatment and disease burden, with >500 000 additional women correctly treated for NG and ~1.5 million correctly treated for CT and TV, and 24%/15% reductions in the number of women with NG/CT by 2040.
Conclusion: Improved data on the functioning of syndromic management in practice would help refine the estimates of the health impact and the overall value proposition of a highly sensitive POC diagnostic for NG/CT/TV. However, even without such data, our analysis demonstrates the potential for such a diagnostic to reduce overtreatment by >90% relative to plausible assumptions regarding the standard of care.

Background: Over 2023 and 2024, 19 of the countries that were supported by Gavi to purchase HPV vaccines adopted a single-dose HPV vaccination schedule. The goal of this study is to estimate the impact on vaccination access and the number of cervical cancers averted compared to a two-dose schedule.
Methods: We estimated the population that could be targeted in countries supported by Gavi to purchase HPV vaccines. We used UNICEF shipment plans to identify the number of HPV doses shipped to each country in 2023 and 2024, plus information supplied by Gavi on the dose schedule implemented in each country and year, adjusting for vaccine wastage. We computed the number of girls that could have been reached, first assuming complete utilization of all shipped doses under a single-dose schedule, and second assuming a counterfactual scenario where all countries would have used a 2-dose schedule. We then compared this to country-reported data on the number of girls actually vaccinated. For each of the three scenarios we modeled the number of cervical cancers averted using HPVsim, a microsimulation model calibrated to each country.
Findings: We calculate that the introduction of single-dose HPV vaccination in Gavi-supported countries would have allowed these countries to target 23.3M additional girls if all supply was utilized. Reported data on girls vaccinated indicates that in actuality an additional 18.5M girls were reached due to adoption of single-dose. We estimate that the use of single-dose schedule in 2023 and 2024 could have averted up to 370,000 (356,000–376,000) additional future cervical cancers if all supply had been utilized, and 297,000 (222,000–369,000) given actual utilization.
Interpretation: The single-dose HPV vaccination strategy has had a substantial positive impact on cervical cancer elimination in context of supply constraints affecting low and middle-income countries.

Outbreaks of vaccine-derived poliovirus type 2 (cVDPV2) have become a major threat to polio eradication. However, variations in spatiotemporal spread have not been quantified. Here we analysed cVDPV2 cases and wild poliovirus type 1 sequences to uncover spatiotemporal patterns and drivers of poliovirus spread. Between 1 May 2016 and 29 September 2023, 3,120 cVDPV2 poliomyelitis cases were reported across 75 outbreaks in 39 countries. Outbreaks had a median observed circulation of 202 (range 0–1,905) days and a median maximum distance of 231 (range 0–4,442) km. Wavefront velocity analysis of large outbreaks revealed a median velocity of spread of 2.3 (5th–95th percentile 0.7–9.2) km per day. International borders were associated with a slower velocity of spread (P < 0.001), in periods with high estimated population immunity. Phylogeographic analysis of 1,572 global wild poliovirus 1 sequences revealed that historic spread resembles recent cVDPV2 patterns and that international spread is largely sustained by unidirectional movement between neighbouring countries. Our findings offer insights for enhancing the geographical scope of vaccination response in the final phases of poliovirus eradication.

Background: Malaria control in sub-Saharan Africa faces significant challenges from biological threats, such as insecticide resistance and adaptive vector behaviours, as well as increasing financial constraints, which necessitate strategic intervention planning to maximize impact. This study assesses the effectiveness of combining vector control methods, case management, and immunoprevention to reduce malaria in Tanzania, considering varying intensities of insecticide resistance in the main vector species. Methods: A compartmental model was developed to simulate malaria transmission, incorporating the dominant vectors: Anopheles funestus (anthropophilic and endophilic) and Anopheles arabiensis (zoophilic and exophilic). The model was used to analyse the impacts of insecticide-treated nets (ITNs), indoor residual spraying (IRS), and biolarvicides, used singly or in combinations, under varying intensities of pyrethroid resistance. The analysis was further expanded to explore the impacts of adding case management (treatment using artemisinin-based combinations) and immunization (RTS,S/AS01 and R21/Matrix-M vaccines). Results: At moderate levels of pyrethroid resistance (50%), achieving at least 71% ITN coverage combined with either 50% IRS or 32% biolarvicide coverage reduces the effective reproduction number ( ) to below 1. However, at high resistance levels (exceeding 75%), the effective reproduction number ( ) consistently remains above 1, irrespective of the type or combination of vector control interventions. Adding immunization ( 40% coverage) to ITNs (80% coverage), along with effective treatment (80% coverage), can further reduce the proportion of infectious individuals to <20% and below 1, even under high resistance intensities. Conclusions: Compared to ITNs alone, combining ITNs with IRS and/or biolarvicides greatly improves malaria control at low to moderate intensities of pyrethroid resistance but yields no additional benefits at high resistance intensities. However, integrating these vector control strategies with immunization and effective case management using artemisinin-based combination therapy (ACT) further enhances impact by reducing both parasite transmission and the infectious reservoir.

Despite concerted global vaccination efforts, wild poliovirus remains endemic in two countries in 2024, Pakistan and Afghanistan. This study uses phylogeographic analysis of poliovirus genetic and epidemiological data from clinical and wastewater surveillance to identify the causes of poliovirus persistence and routes of spread over the last decade (2012 to 2023). Poliovirus genetic diversity declined after 2020, with one of two major genetic clusters dying out, and recent detections are now closely related genetically. High-risk and hard-to-access regions have sustained polio transmission over the past decade, even when interrupted elsewhere. Karachi, one of the most densely populated cities globally, has acted as a hub for the amplification and spread of poliovirus to other regions, many of which we show to be dead-end for onwards transmission despite frequent virus detection. Phylogenetic analysis has long been central to the polio surveillance network, and advancing the approaches used can provide critical epidemiological insights to accelerate eradication efforts.

Background: Intermittent preventive treatment (IPT) of school-aged children with antimalarial drugs decreases rates of infection, anaemia, and clinical malaria. Since school-aged children are a major transmission reservoir, we estimated the effect of IPT for this group on Plasmodium falciparum transmission to younger children and adults across three epidemiological settings. Methods: Using an established malaria transmission model, three epidemiological archetypes (Sahelian, Central, and Southern African) were developed and the effect of IPT of school-age children was estimated across transmission levels (P. falciparum parasite rate in children aged 2–10 years [PfPR2–10]: 5–40%). Baseline interventions included long-lasting insecticide-treated nets and clinical case management. Scenarios compared three drug options (dihydroartemisinin–piperaquine, artesunate–amodiaquine, sulfadoxine–pyrimethamine–amodiaquine) with different delivery options. Findings: With frequent administration of long-acting drugs (monthly dihydroartemisinin–piperaquine), modelled IPT averted 70–90% of cases in school-aged children and 20–60% in younger children and adults, with greater benefit at lower transmission levels. Shorter-acting drugs administered with various schedules averted 40–60% of cases in school-aged children and 15–50% in other ages. Interpretation: Our model suggests that adding IPT of school-age children to current control tools could decrease malaria burden in this group and reduce P. falciparum transmission.

Rubella is a leading cause of vaccine-preventable birth defects. Rubella virus infection during early pregnancy can result in miscarriage, fetal death, stillbirth, or a constellation of birth defects known as congenital rubella syndrome (CRS). This report describes current and future estimated CRS incidence in countries that have not yet introduced rubella-containing vaccine (RCV) into their national childhood immunization schedules and the estimated effect of implementing a recent recommendation to introduce RCV into these programs even if population coverage with measles-containing vaccine is <80%. During 2000–2022, the number of countries that introduced RCV increased from 99 (52%) of 191 in 2000 to 175 (90%) of 194 in 2022. By the end of 2023, 19 lower- and middle-income countries had not yet introduced RCV. In 2019, an estimated 24,000 CRS cases occurred in these countries, representing 75% of the estimated 32,000 cases worldwide. In a modeling study estimating the effect of RCV introduction in these countries during 2025–2055, an estimated 1.03 million CRS cases are projected to occur without RCV. In contrast, fewer than 60,000 cases are estimated if RCV is introduced with catch-up and follow-up supplementary immunization activities, averting more than an estimated 986,000 CRS cases over 30 years. Based in part on these estimates, in September 2024, the World Health Organization Strategic Advisory Group of Experts on Immunization recommended removing the ≥80% coverage threshold and instituting universal RCV introduction in these countries. RCV introduction in these 19 countries during 2025–2030 could rapidly accelerate progress toward rubella and CRS elimination worldwide.

Agent-Based Models (ABMs) have gained popularity over the COVID-19 epidemic, but their efficient calibration remains challenging. Here we propose a novel calibration architecture by investigating the role of pruning in ABM calibration. We use a recently developed model for human papillomavirus (HPV) transmission and focus on its integrated calibration framework, Optuna. Simulating six synthetic datasets of various temporal skewness, with six pruners, we show that more aggressive pruners perform best (in terms of loss function at end of calibration) for very-back-heavy datasets, while median pruners are better for more-front-heavy datasets. For more balanced datasets most of the pruners perform similarly to no pruning. However, across all datasets pruning notably sped up calibration, in many cases without compromising on – or even improving upon – the optimal found parameter set. We validate our results through application to real-life data. Finally, we discuss approaches for improving “bad pruners” for balanced datasets. Our proof-of-principle study shows that pruners can improve ABMs’ calibration. As ABMs are becoming more widely used in epidemiological modelling, designing the next level of pandemic preparedness strategies will need to address efficient calibration; we believe pruning is a cornerstone for this.

Measles is one of the most contagious vaccine preventable diseases, causing severe complications and deaths globally. While vaccination with a measles-containing vaccine (MCV) has prevented millions of measles deaths, recent trends, especially from low- and middle-income countries, are discouraging. Measles cases have increased since 2021 as MCV coverage has decreased; and an estimated 107,500 measles deaths, mostly in children under-five years, occurred in 2023. Thus, a renewed focus on proven and innovative strategies to control measles is needed. The World Health Organization (WHO) recommends a first MCV dose administered at 9–15 months of age (routine MCV1), however MCV1 below 9 months of age (early MCV1) may increase vaccination coverage because uptake of all vaccines tends to be higher the younger the child, and this might protect vulnerable infants earlier in life. However, due to concerns about possible reduced vaccine performance, early MCV1 is not routinely recommended by WHO. WHO hosted an informal technical consultation on December 6–7, 2023, in Geneva, Switzerland to evaluate recent evidence on early MCV1 and identify evidence gaps for policy making. The recent evidence suggests a robust humoral immune response shortly after early MCV1 at 5–8 months of age. Immune blunting of a routine second MCV dose (e.g., MCV2) after early MCV1 was not demonstrated in the presented data. However, 3–7 years after MCV1, children receiving early MCV1 had lower measles antibodies than children receiving routine MCV1, suggesting faster waning of immunity. The totality of evidence on immune blunting remains inconsistent. Meeting participants thought more data are needed before revisiting WHO’s current recommendation for a potential revision. Evidence gaps include: understanding measles disease burden and severity in infants; early MCV1 effectiveness and duration; vaccine-induced cellular immunogenicity; whether measles in infants is acquired from other infants or older children or adults; and blunting of routine MCV2. Addressing evidence gaps through targeted studies and measles outbreak investigations, as well as evaluations of country-level introductions of early MCV1 are warranted. Ensuring high MCV1 and MCV2 coverage remains the priority in measles control.