Quantifying relative health impact across Gavi, the Vaccine Alliance's portfolio

# Quantifying relative health impact across Gavi, the Vaccine Alliance's portfolio in 117 countries at the subregional level: a modelling study *The Lancet 2026* Articles Quantifying relative health impact across Gavi, the Vaccine Alliance’s portfolio in 117 countries at the subregional level: a modelling study Katy A M Gaythorpe , Xiang Li, Manjari Shankar, Anna-Maria Hartner, Zoë Gibney, Kaja Abbas, Romesh Abeysuriya, Christina Alam, Megan Auzenbergs, Andrew S Azman, Edwine Barasa, Alan Costello, Matthew J Ferrari, Keith Fraser, Han Fu, Lydia Haile, Romain Glèlè Kakaï, Andromachi Karachaliou-Prasinou, Elizabeth C Lee, Esther Nyadzua Katama, Jong-Hoon Kim, Mark Jit, Yang Liu, Josephine Malinga, Sean Moore, Shevanthi Nayagam, Gemma Nedjati-Gilani, Lucy C Okell, Akindele Akano Onifade, Timos Papadopoulos, Melissa A Penny, T Alex Perkins, Virginia E Pitzer, Allison Portnoy, Simon R Procter, Chitra Maharani Saraswati, Nick Scott, Chris Seaman, Andrew J Shattock, So Yoon Sim, Quan Tran, Emilia Vynnycky, Amy K Winter, Wes Hinsley, Neil M Ferguson, Caroline L Trotter Summary Background Estimates of vaccine impact have typically been used to quantify the effects of, and inform, immunisation Lancet 2026; 407: 1941–52 strategies. Given the growing resource constraints on health systems worldwide, robust estimates of vaccine impact See Comment page 1895 that allow comparison across different vaccines are now more crucial for decision making than ever. Building on MRC Centre for Global previous modelling studies, we aimed to estimate vaccine impact ratios for an expanded portfolio of Gavi, the Vaccine Infectious Disease Analysis, Alliance-supported vaccination programmes against 14 vaccine-preventable diseases across 117 low-income and School of Public Health, Imperial College London, middle-income countries using multiple models. London, UK (K A M Gaythorpe PhD, X Li PhD, Methods In this modelling study, we have presented Vaccine Impact Modelling Consortium estimates of vaccine M Shankar MSc, impact ratios, defined as deaths or disability-adjusted life-years averted per 1000 vaccinations, for the Gavi portfolio of A-M Hartner MSc, Z Gibney MSc, M Auzenbergs PhD, K Fraser PhD, vaccines. Modelling groups used standardised inputs for demographic data and vaccination coverage assumptions, L Haile MPH, S Nayagam PhD, including a no-vaccination counterfactual, and accounted for structural, parameter, and stochastic uncertainty to G Nedjati-Gilani PhD, produce burden estimates. These estimates were then compared to calculate vaccine impact ratios, disaggregated by L C Okell PhD, W Hinsley PhD, immunisation activity type and geographical subregions for vaccinations given between 2000 and 2030 (or 2000 and Prof N M Ferguson DPhil, Prof C L Trotter PhD); Digital 2040 for cholera). Health and Biomedical Engineering, Department of Findings Overall, we observed human papillomavirus (11·24 [95% uncertainty interval 10·88–11·64]) and measles Electronics and Computer (6·09 [4·90–7·07])vaccines averting a higher number of deaths per 1000 vaccinations than others. For other vaccines, Science, University of Southampton, UK the impact ratios varied across subregions and activity types. Due to parameter, structural, and stochastic uncertainty, (A A Onifade PhD); Department the ranges of these ratios often overlap. of Computer Science and Mathematics, Mountain Top Interpretation Decisions around which vaccines to use are increasingly important in the context of Gavi’s country University, Ibafo, Nigeria (A A Onifade); University of vaccine budgets. Robust metrics that allow comparison between vaccines are thus essential to inform discussions. Notre Dame, Notre Dame, IN, The vaccine impact ratios presented in this study can be used to complement other evidence to support effective USA (A Costello MSc, planning and prioritisation in national immunisation programmes. S Moore PhD, T A Perkins PhD, Q Tran PhD); School of Public Health; Department of Global Funding Gavi, the Vaccine Alliance, Gates Foundation, and Wellcome Trust. Health, Boston University, Boston, MA, USA Copyright © 2026 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 (A Portnoy ScD); Center for the license. Ecology of Infectious Diseases, Department of Epidemiology and Biostatistics, University of Introduction Gavi, the Vaccine Alliance, was formed in 2000 with a Georgia, Athens, GA, USA Vaccines are effective tools for preventing infectious mission to save lives, reduce poverty, and protect the (A K Winter PhD); Swiss Tropical and Public Health Institute, diseases and improving population health. Since its world against the threat of epidemics and pandemics.6 Allschwil, Switzerland inception in 1974, the Expanded Programme on Immuni­ Between 2000 and 2023, Gavi supported the vaccination (A J Shattock PhD); The Kids, zation has saved an estimated 154 million lives and of more than 1·1 billion children through routine University of Western expanded from six universally recommended vaccines to immunisation programmes.7 The impact of this Australia, Perth, WA, Australia (J Malinga PhD, 13 across the life course, alongside 17 additional context­ investment is assessed through health and economic Prof M A Penny PhD, dependent recommen dations.1,2 However, this expansion outcomes, with health impacts measured by reductions C M Saraswati MSc, A J Shattock), has increased pressures on country resources and in mortality and morbidity. Mathematical models are University of Cambridge, budgets. For low­income and middle­income countries essential tools for estimating the impact of a vaccination. Cambridge, UK (Prof C L Trotter, A Karachaliou-Prasinou PhD), (LMICs), these have been further exacerbated by The mathematical models can examine the impact of an Burnet Institute, Melbourne, population growth and the recent large reductions in intervention over time horizons that could not be VIC, Australia overseas aid spending by many high­income countries.3–5 captured in trials or demonstration studies and can (R Abeysuriya PhD, N Scott PhD, Articles C Seaman PhD); Health Economics Research Unit, Research in context KEMRI–Wellcome Trust Evidence before this study Added value of this study Research Programme, Nairobi, Kenya (E Barasa PhD); Statistics We searched PubMed on Nov 24, 2025, using the search terms To our knowledge, this study is the first comprehensive modelling Modelling and Economics ((“vaccine impact” OR “vaccine-preventable diseases” OR analysis to date that uses impact ratios to characterise the Department, UK Health “immunization impact”) AND (“multi-country” OR “global” morbidity and mortality averted per course of vaccination, Security Agency, London, UK OR “multiple countries”) AND (“burden” OR “cases” OR covering diseases in 117 LMICs. This analysis builds on the (T Papadopoulos PhD, E Vynnycky PhD); London “deaths” OR “DALYs”) AND (“averted” OR “prevented” OR previous work by the Vaccine Impact Modelling Consortium by School of Hygiene & Tropical “reduced” OR “reduction”)). We searched studies of all types incorporating updated coverage data after catch-up efforts post- Medicine, London, UK published on or before Nov 24, 2025, in the English language. COVID-19, improving upon previous modelling efforts by the (K Abbas PhD, M Auzenbergs, Original research articles were included if they met the Vaccine Impact Modelling Consortium, and by modelling H Fu PhD, Y Liu PhD, S R Procter DPhil, E Vynnycky); following inclusion criteria: (1) focused on the impact of additional vaccines, including for malaria and COVID-19. We have International Vaccine Institute, vaccine-preventable diseases; (2) included two or more contributed to the existing literature on global vaccination Seoul, South Korea countries; and (3) included two or more vaccines. We found impact modelling by adding a subregional focus, in which the (J-H Kim PhD); School of Tropical Medicine and Global 139 studies, of which 14 met the inclusion criteria. Of these, impact of different vaccines can be compared at a more granular Health, Nagasaki University, five studies were published by the Vaccine Impact Modelling scale. These estimates can be used by stakeholders such as Gavi, Nagasaki, Japan (K Abbas); Consortium, and seven studies used data from the Vaccine the Vaccine Alliance, alongside other considerations, such as cost, Institute of Tropical Medicine, Impact Modelling Consortium. Five of these studies had a to evaluate and optimise vaccination strategy. Nagasaki University, Nagasaki, Japan (K Abbas); Public Health global focus, whereas an additional two studies Implications of all the available evidence Foundation of India, New predominantly examined high-income countries, and the final Vaccination continues to be one of the most effective public Delhi, India (K Abbas); seven studies examined low-income and middle-income Department of Global and health interventions worldwide. However, impact varies across countries (LMICs). Five studies had an additional focus on the Environmental Health, School vaccines. We have estimated that the highest vaccine impact of Global Public Health, economic impact averted or programmatic costs, and ratios are from the human papillomavirus vaccine and measles- New York University, New York, one additional study focused on the implications of equity. containing vaccines. For other vaccines, the relative impact NY, USA (Prof M Jit PhD); Two studies included the impact of COVID-19 on the burden Laboratoire de ratio varies by WHO subregion, with uncertainty distributions averted by immunisation. Most studies examined the deaths Biomathematiques et often overlapping. There are also differences in vaccine impact d’Estimations Forestieres, averted by vaccination, although impact measures by dose and by immunisation activity type (ie, routine versus University of Abomey-Calavi, additionally included years of life lost, disability-adjusted life- campaign). These findings highlight the complexity of Cotonou, Benin years, and cases. No studies explicitly reported the impact (Prof R Glèlè Kakaï PhD); World prioritising vaccination efforts across a broad portfolio and ratios (ie, the burden averted per course of vaccination). Full Health Organization, Geneva, provides one metric to consider within a suite of factors. Switzerland (S Y Sim MA MSPH); details of the studies considered have been summarised in the Department of Epidemiology appendix (p 2). of Microbial Diseases, Yale School of Public Health, New Haven, CT, USA provide insights in which empirical evidence from between models are discussed, and standardised outputs (Prof V E Pitzer ScD); Centre for Artificial Intelligence in Public surveillance is limited. Although mathematical are shared across modelling groups. Health Research, Robert Koch modelling is no substitute for surveillance, it is the only Policy decisions for immunisation are often made on a Institute, Berlin, Germany available tool to describe both the direct and indirect vaccine­by­vaccine basis. Until Gavi 5·0 (2021–25),13 (A-M Hartner); National effects of vaccination and project disease burden under sufficient financial resources meant prioritisation across Institute of Infectious Diseases, Japan Institute for Health different vaccination coverage assumptions. vaccine programmes was rarely needed. However, in the Security, Tokyo, Japan The Vaccine Impact Modelling Consortium is an context of both an expanding suite of recommended (K Abbas); Center for Infectious international community of modellers established to vaccines and funding constraints, looking more holistically Disease Dynamics, produce reliable estimates of the impact of vaccines across at immunisation programmes is essential. Additionally, Pennsylvania State University, State College, PA, USA the portfolio of vaccines that Gavi supports.8 Modelling Gavi and country decision­making processes could (Prof M J Ferrari PhD); Institute assumptions are developed in collaboration with consider differential use depending on epidemiological of Global Health, University of stakeholders to ensure alignment with Gavi’s prospective and funding contexts. Comparisons across vaccine Geneva, Geneva, Switzerland resource allocation strategy and national immunisation programmes can be challenging in the absence of (A S Azman PhD); Geneva Centre for Emerging Viral programme planning. Using these assumptions and standardised vaccine impact estimates because differences, Diseases, Geneva University standardised inputs, modelling groups from different such as outcome definitions or population denominators, Hospitals, Geneva, Switzerland institutions produce estimates of disease burden that are can lead to incommensurable estimates (appendix p 2). (A S Azman); Department of Epidemiology, Johns Hopkins processed by the Vaccine Impact Modelling Consortium Since the Vaccine Impact Modelling Consortium adopts a Bloomberg School of Public secretariat to provide estimates of vaccine impact using standardised approach for estimating vaccine impact, it is Health, Baltimore, MA, USA established methods (appendix p 8).9 An important feature well placed to facilitate such comparisons. (C Alam PhD, A S Azman, of this pipeline is that it is traceable and reproducible In this study, we aimed to present our latest estimates E C Lee PhD); Division of Tropical and Humanitarian from data input to vaccine impact output.10–12 Throughout of deaths and disability­adjusted life­years (DALYs) Medicine, Geneva University this process, structural, parameter, and stochastic averted per vaccination, referred to as vaccine impact Hospitals, Geneva, Switzerland uncertainties are estimated, similarities and differences ratios, for 117 LMICs. These ratios span the Gavi portfolio 1942 Articles of vaccines, including newer vaccines such as malaria, the model standards.8 To capture structural uncertainty, a (A S Azman); KEMRI-Wellcome and represent a standardised set of metrics that minimum of two modelling groups were included Trust Research Programme, incorporate uncertainty. The vaccine impact ratios can be per vaccine (except meningitis and COVID­19 with Kilifi, Kenya (E Nyadzua Katama BSc) used as part of a decision­making toolkit to compare one group each). To incorporate parameter uncertainty, Correspondence to: vaccine programmes and prioritise scarce funding, each Vaccine Impact Modelling Consortium model was Dr Katy A M Gaythorpe, MRC alongside economic impact measures and broader expected to produce 200 estimates per set of vaccination Centre for Global Infectious societal benefits of vaccines. coverage assumptions using a different sample of input Disease Analysis, School of Public parameters and propagating stochasticity, where Health, Imperial College London, London W12 0BZ, UK Methods applicable. Estimates were presented by the antigen and/ k.gaythorpe@imperial.ac.uk Vaccination assumptions and demographic inputs or vaccine: for instance, Penta (Diptheria, tetanus, See Online for appendix This modelling study includes vaccine impact ratios for pertussis [DTwP]­HepB­Hib) was split into its constituent Gavi­supported vaccine programmes against 14 infectious antigens. diseases. We have presented new Vaccine Impact Estimates of burden per model per set of coverage Modelling Consortium impact estimates for COVID­19, assumptions were submitted to the Vaccine Impact malaria, and cholera, and have updated Vaccine Impact Modelling Consortium secretariat for processing. Before Modelling Consortium estimates for hepatitis B (HepB), calculating the health impact of the vaccination, the human papillomavirus (HPV), measles, meningitis burden estimates went through automated and human­led (MenA from 2010 with replacement by MenACWYX diagnostic checks to ensure that they were complete and from 2027), rubella, typhoid, and yellow fever. These met the Consortium standards.8 These estimates were estimates were supple mented with existing estimates for then processed to calculate the impact of vaccination, with Japanese encephalitis, Haemophilus influenzae type b further diagnostics conducted on the resulting impact (Hib), rotavirus (Rota), and pneumococcal conjugate estimates. Once assessed by the Vaccine Impact Modelling vaccine (PCV).12 Consortium secretariat, these estimates were then shared We have summarised the coverage assumptions used with the respective modelling groups for their review. per vaccine, with these described in detail in the appendix Subsequently, a model review process was conducted (pp 10–20). The assumed course of vaccination follows among vaccine­specific modelling groups to discuss Strategic Advisory Group of Experts on Immunisation outputs and examine any notable differences or (SAGE) guidance and has been detailed in the appendix similarities. These estimates were then revised where (p 9). Vaccine coverage up to the last year of data necessary. The central processing continued based on availability (2022) before the model runs was informed by iterative feedback from the modelling groups until both estimates from WHO and UNICEF Estimates of National the burden and impact estimates were approved and Immunisation Coverage, the WHO immunisation signed off. We focused on deaths and DALYs as burden repository, and the Gavi data repository.11,12,14,15 From 2023 outcomes in this study, as these generally have a onwards, we projected coverage for all vaccines except standardised definition across the diseases included.8 In malaria based on recommended campaign frequency contrast, cases varied substantially in their definition and (from regional and WHO guidance), 2030 zero­dose estimation by disease. coverage endpoints informed by the WHO Immunisation Agenda 2030, and expert consultation.16 Future malaria Statistical analysis of impact ratio vaccine coverage was informed by Gavi’s demand We focused on one main outcome—the impact ratio, forecasting projections as of 2023. which is defined as the burden averted per All Vaccine Impact Modelling Consortium models 1000 vaccinations. The inverse can be considered as the used age­stratified, standardised demographic number needed to vaccinate to avert a burden outcome information derived from the UN World Population during the modelled time period. Prospects released in 2022.17 Standardised inputs are The approach to calculating impact compared modelled available to download from the corresponding GitHub estimates of burden with and without vaccination.9 The For more on the GitHub repository. Countries were grouped into subregions time horizon of vaccination considered was from respository see https://github. according to the UN statistics division M49 codes for 2000 to 2030 for all vaccines except cholera, which was com/vimc/paper4 statistical use (appendix p 4–7).18 extended up to 2040 to reach an equilibrium state of All data used were from secondary sources and did not immunisation. We estimated impact across different require ethical approval. vaccination activity types: routine immunisation only, campaigns only, and overall. When estimating the total Vaccine Impact Modelling Consortium models impact of routine immunisation programmes, we Each model contributing to the Vaccine Impact Modelling considered the difference in burden compared with a Consortium has been described in detail in the appendix counterfactual without vaccination (no vaccination) (pp 35–59). Modelling groups joined the Vaccine Impact scenario for all birth cohorts vaccinated in the years of Modelling Consortium through a request for proposals in vaccination and modelled up to the year 2100 to which the applicants justified how they met or will meet accommodate burden during their lifespans. When Articles estimating the total impact of vaccination campaigns, we MenACWYX, we divided the impact between cohorts that considered the burden for all ages in all years modelled were born before and after the projected switch to up to the year 2100. To avoid major perturbations in MenACWYX (appendix pp 22, 23). COVID­19 vaccines future disease dynamics, we assumed routine have been presented together as there was not sufficient immunisation activities continue up to 2100 in our information to disaggregate by product. Malaria vaccines vaccine coverage assumptions, but there were no (RTS,S and R21) have also been presented together as additional campaigns beyond the year 2030 (2040 for information on their comparative characteristics is still in cholera) as these have been planned during a shorter development. We sampled 1000 times with replacement timescale according to resource availability and disease from the stochastic model estimates from all models for burden. each vaccine and calculated the mean between the In the primary analyses, vaccination (both routine and modelling groups. As such, we arrived at 1000 estimates of campaign) refers to the full course of vaccination as the modelling group average impact ratio per geography, recommended by WHO SAGE guidelines (appendix p 9), vaccine, and activity combination. Where we present an but where these were ambiguous, we used the country­ impact ratio for multiple countries, these have been given specific vaccination schedule. We calculated additional by the following equation across the entire geography: impact ratios for some vaccines by vaccine dose (eg, measles­containing vaccine first dose [MCV1]). The Total burden averted denominator of the impact ratio is the difference in the Total vaccinated people number of vaccinated individuals between the with­ vaccination and no­vaccination scenarios for the years of An extended description of this approach has been vaccination.9 detailed in the appendix (pp 21–25). All analyses for this Impact ratios have been presented separately for each study have been conducted in R (version 4.3.1 or later) vaccine and can be stratified by subregion, country, Vaccine with R packages called orderly and vimpact among Impact Modelling Consortium model, and immunisation others. activity type (that distinguishes between vaccines given during routine immunisation and vaccination campaigns). Role of the funding source To disaggregate and compare impact between MenA and The funders of the study provided feedback on the vaccination coverage assumptions and viewed this Article before publication. The funders of the study had no role Deaths averted per DALYs averted per 1000 vaccinations 1000 vaccinations in study design, data collection, data analysis, data (95% uncertainty (95% uncertainty interpretation, or writing of the report. interval) interval) Cholera 0·20 (0·13–0·56) 6·13 (4·81–12·61) Results COVID-19 0·12 (0·08–0·18) 4·03 (2·86–6·19) In this modelling study, we have presented impact ratios Hib 2·22 (1·81–2·61) 150·37 (124·74–176·12) per 1000 vaccinations for each vaccine programme HepB 5·00 (4·47–5·58) 141·94 (127·28–157·71) aggregated across all immunisation activity types and all HPV 11·24 (10·88–11·64) 523·04 (499·92–547·11) subregions (figure 1). The variation between modelling Japanese 0·18 (0·10–0·30) 80·57 (55·55–106·80) groups has been presented in the appendix (pp 32–34). encephalitis This figure highlights that while there is some relative Malaria 2·78 (2·16–3·53) 203·04 (156·99–259·97) ordering in the mean impact ratio, the uncertainty ranges Measles 6·09 (4·90–7·07) 411·01 (331·63–476·14) for each vaccine programme can overlap due to the Meningitis 0·66 (0·49–0·85) 42·78 (31·90–54·91) stochastic, parameter, and structural uncertainty PCV 1·53 (0·87–2·09) 104·91 (61·47–142·93) captured in the estimates. HPV, measles, and HepB Rota 0·80 (0·49–1·04) 46·07 (25·78–61·78) consistently had the highest mean number of deaths Rubella 0·22 (0·12–0·35) 16·78 (11·91 23·51) averted per vaccination whereas HPV, measles, and Typhoid 0·68 (0·21–1·72) 26·57 (7·96–70·69) malaria had the highest mean number of DALYs averted Yellow fever 1·86 (0·52–4·23) 72·26 (20·06–165·63) per vaccination. We observed HPV (11·24 [95% Diseases have been ordered alphabetically. Deaths for rubella reflect deaths uncertainty interval 10·88–11·64]) and measles (6·09 related to Congenital Rubella Syndrome. Malaria refers to the combined value of [4·90–7·07]) averted deaths per 1000 vaccinations RTS,S and R21 vaccine effect. COVID-19 refers to an average of different COVID-19 (table 1). In comparison, cholera, Japanese encephalitis, vaccine product effects informed by the most prevalent vaccine product per and COVID­19 had the lowest mean number of deaths country. Meningitis refers to the combined effect of MenA and MenACWYX vaccination. DALYs=disability-adjusted life-years. Hib=Haemophilus influenzae averted per vaccination, and rubella, cholera, and type b. HepB=hepatitis B. HPV=human papillomavirus. PCV=pneumococcal COVID­19 had the lowest mean number of DALYs conjugate vaccine. Rota=rotavirus. averted per vaccination. Given the overlapping Table 1: Mean and 95% uncertainty interval of bootstrap mean impact uncertainty in these estimates, the results should be ratio of deaths and DALYs averted per 1000 vaccinations for all interpreted with caution. Additionally, it is worth noting activities the difference in ranking when examining deaths 1944 Articles HPV Measles HepB Malaria Hib Yellow fever PCV Rota Figure 1: Bootstrap mean impact ratios defined as deaths or disability-adjusted Typhoid life-years (DALYs) averted per 1000 vaccinations across all subregions and activity types Meningitis Diseases have been arranged by mean impact ratio value for the deaths averted. Deaths for rubella reflect deaths related to Congenital Rubella Rubella Syndrome. Malaria refers to the combined value of RTS,S and R21 vaccine effect. COVID-19 refers to an average of different COVID-19 vaccine Cholera product effects informed by the most prevalent vaccine product per country. Meningitis refers to the Japanese combined effect of MenA and encephalitis MenACWYX vaccination. Each histogram contains 1000 bootstrapped mean estimates per vaccine for all COVID-19 relevant countries. Hib=Haemophilus influenzae type b. HepB=hepatitis B. 10–1 10–0·5 100 100·5 101 HPV=Human papillomavirus. PCV=pneumococcal conjugate vaccine. Rota=Rotavirus. esaesiD Deaths DALYs 100·5 101·0 101·5 102·0 102·5 Impact ratio (per thousand vaccinated, log scale) Impact ratio (per thousand vaccinated, log scale) 10 10 Articles Cholera COVID-19 HepB Hib HPV Japanese encephalitis Malaria Measles Meningitis PCV Rota Rubella Figure 2: Bootstrap mean impact ratios and 95% CIs defined as deaths averted per 1000 vaccinations by subregion and disease for all activities Dots indicate bootstrap mean impact ratios, and lines indicate 95% CIs (when the 95% CI is smaller than the length of the dot, it is shown as a white colour). Deaths for rubella reflect deaths related 10^–510–4 10–3 10–2 10–1 100 101 10^2 10^–510–4 10–3 10–2 10–1 100 101 10^2 to Congenital Rubella Impact ratio (deaths averted Impact ratio (deaths averted Syndrome. Malaria refers to per thousand vaccinated, log scale) per thousand vaccinated, log scale) 10 10 the combined value of RTS,S and R21 vaccine effect. Typhoid Yellow fever COVID-19 refers to an average Oceania of different COVID-19 vaccine Latin America and the Caribbean product effects informed by Eastern and southern Europe the most prevalent vaccine Eastern and southeastern Asia product per country. Central and southern Asia Meningitis refers to the Southern Africa combined effect of MenA and Western Africa Eastern africa MenACWYX vaccination. Northern Africa and western Asia Results for disability-adjusted Middle Africa life-years and results split by vaccination activity can be found in the appendix (pp 25–27). Hib=Haemophilus influenzae type b. HepB=hepatitis B. 10^–510–4 10–3 10–2 10–1 100 101 10^2 10^–510–4 10–3 10–2 10–1 100 101 10^2 HPV=human papillomavirus. Impact ratio (deaths averted Impact ratio (deaths averted PCV=pneumococcal conjugate per thousand vaccinated, log scale) per thousand vaccinated, log scale) vaccine. Rota=rotavirus. 10 10 1946 Articles Middle Africa Northern Africa and western Asia Eastern Africa Western Africa Southern Africa Central and southern Asia Eastern and southeastern Asia Eastern and southern Europe Latin America and the Caribbean Figure 3: Bootstrap mean impact ratios and 95% CIs defined as deaths averted per 1000 vaccinated by disease, subregion, and activity 10–5 10–4 10–3 10–2 10–1 100 101 102 10–5 10–4 10–3 10–2 10–1 100 101 102 Dots or triangles indicate Impact ratio (deaths averted Impact ratio (deaths averted bootstrap mean impact ratios, per thousand vaccinated, log scale) per thousand vaccinated, log scale) and lines indicate 95% CIs. 10 10 Deaths for rubella reflect Oceania deaths related to Congenital Activity type Yellow fever Rubella Syndrome. Malaria Campaign Typhoid refers to the combined value Routine Rubella of RTS,S and R21 vaccine Rota effect. COVID-19 refers to an PCV average of different COVID-19 Measles Malaria vaccine product effects Japanese encephalitis informed by the most HPV prevalent vaccine product per Hib country. Results for disability- HepB adjusted life-years can be COVID-19 found in the appendix Cholera (pp 25–27). Hib=Haemophilus influenzae type b. HepB=hepatitis B. 10–5 10–4 10–3 10–2 10–1 100 101 102 HPV=human papillomavirus. PCV= pneumococcal Impact ratio (deaths averted per thousand vaccinated, log scale) conjugate vaccine. 10 Rota=rotavirus. Articles compared with DALYs; this is motivated by age profile of outbreak response, further to commitments to control, impact, the likelihood and duration of associated eliminate, or eradicate these vaccine­preventable diseases. morbidity, and the types of associated morbidity. For Under the Gavi 6·0 framework (2025–30), the example, we see a change in ranking for malaria (has a introduction of country vaccine budgets shifts more higher ranking for DALYs), which largely affects very control to national governments. Although this control young children, leading to a higher estimate of years of increases country­level ownership of immunisation life lost compared with HepB and Japanese encephalitis, strategy, it requires rigorous prioritisation because which can cause severe neurological damage in survivors current funding levels are insufficient to cover every (appendix pp 23–24). discretionary vaccine available. In this context, having a The relative ordering and magnitude of impact ratios standardised and consistent measure of the health impact has been shown in figure 2 and by immunisation activity of a range of vaccines by country could be very useful. type in figure 3. Although there is some overlap in The strength of our approach is that the Vaccine Impact uncertainty ranges between subregions, the impact ratios Modelling Consortium has a systematised methodology.8 do vary by subregion for some vaccines, either due to Key inputs such as coverage assumptions and targeting, baseline burden, or other factors. As a result, demography were standardised. Vaccine Impact in some circumstances, impact ratios should be Modelling Consortium models conform to a set of considered at a more granular programmatic scale, such criteria but have freedom to appropriately describe and For more on the national-level as sub­nationally. National­level estimates are available to parametrise the transmission and vaccine in question. extimates see https://vaxviz. explore in the associated data visualisation tool. The use of two models per disease accounted for vaccineimpact.org In table 1, we have incorporated all vaccines. We have structural uncertainty, and parameter uncertainty and highlighted that although there are uncertainties and stochastic variation were captured by each model overlaps with current vaccination programmes, malaria submitting 200 runs per set of coverage assumptions, (2·78 [95% uncertainty interval 2·16–3·53]) falls between which included capturing uncertainty in projecting HepB (5·00 [4·47–5·58]) and Hib (2·22 [1·81–2·61]) in burden in the underlying no­vaccination counterfactual. terms of deaths averted per course and between measles There is a systematic method for processing burden (411·01 [331·63–476·14]) and Hib (150·37 [124·74–176·12]) projections and estimating vaccine impact, including in terms of DALYs averted per course. In contrast, extensive diagnostic checks. These factors facilitate COVID­19 (0·12 [0·08–0·18]) falls after Japanese comparability across vaccine programmes. encephalitis (0·18 [0·10–0·30]) vaccination in terms of Mathematical models are often complex, and variations deaths averted per course and after cholera (6·13 in disease epidemiology and data quality mean that [4·81–12·61]) in terms of DALYs averted per course there are nuances that might not be reflected in (table 1, figure 1). Note, although the mean has been one summary metric of health impact. Sometimes, the presented in table 1, given the skewed distributions of Vaccine Impact Modelling Consortium model inputs are the impact ratios for some vaccines, the range should be themselves model estimates and so inherited the considered in interpretation. We have shown the uncertainty of those inputs. Although we generally differences in impact ratio between MenA versus included multiple models per disease, we did not MenACWYX; although the uncertainty ranges overlap, necessarily capture all structural uncertainty, and where generally MenACWYX has a higher impact ratio for we had only one model per disease (COVID­19 and deaths averted and this is true regardless of which meningitis), uncertainty was underestimated (appendix method has been used to calculate the impact (appendix p 31). The impact ratios presented in this study were pp 31–32). appropriate for the coverage assumptions used; outside of this context, they might not be accurate.9 For example, Discussion the benefit of an improvement in measles coverage from We have presented vaccine impact ratios to quantify the 25% to 40%, (below the herd immunity threshold for health impacts of 14 vaccine programmes in the Gavi measles) would have a different overall impact of moving portfolio across 117 countries, implemented between from 80% to 95%. For such an assessment, and 2000 and 2030 (or 2040 for cholera). We found that across questions around elimination of diseases specifically, immunisation activities and subregions, HPV and bespoke dynamic modelling would be required. measles programmes had the highest impact ratios for Additionally, even though we have presented antigen­ deaths and DALYs. Although there is some clear ordering specific results, in reality, some of the vaccines were among other vaccines, we note that the ranges of the more commonly combined. We also did not consider impact ratios often overlap due to parameter and how impact ratios could have changed by year or cohort; structural uncertainty in addition to true heterogeneity in this is particularly important as future projections of the impact ratio between locations and targeting coverage could be influenced by the changing global strategies. It is therefore important to contextualise the health funding landscape.3,4 We have presented estimates vaccine impact ratios and consider other elements of by subregion in the main text so as not to imply false vaccine value, such as cost­effectiveness or use in precision in the overall conclusions. However, 1948 Articles country­level estimates can be explored in the Model Code link corresponding data visualisation tool and can have use for informing discussions around vaccine prioritisation Cholera Cholera-IVI-Kim https://github.com/kimfinale/cholera_typhoid_ vimc_ivi or highlighting national evidence gaps in which they are Cholera Cholera-JHU-Lee https://github.com/HopkinsIDD/gavi_vimc_cholera appropriately tailored to the country and data sources. COVID COVID-LSHTM-Liu Results not presented in the current publication There are disease­specific considerations that have (appendix p 60) affected the interpretation of the impact ratios. The HPV COVID COVID-IC-Ghani https://mrc-ide.github.io/safir results have shown the impact per two­dose schedule; HepB HepB-Burnet-Scott https://github.com/Burnet-Modelling/burnet-hbv however, many countries are moving towards a one­dose HepB HepB-IC-Nayagam https://github.com/mrc-ide/icl-hbv schedule, which significantly improves the per­dose HepB HepB-Goldstein No publicly available code impact. Further, the results focus on averting cervical HPV HPV-BU-Portnoy Modelling outputs are based on three separate cancer only and do not consider other outcomes, such as modelling approaches (appendix pp 43–44) anal cancer. Regarding HepB, chronic infection is HPV HPV-LSHTM-Jit https://github.com/lshtm-vimc/prime lifelong and can include years of asymptomatic infection Malaria Malaria-UAC-Glele-Kakai https://github.com/RomainGleleKakai/Sub-national- before developing liver cancer or cirrhosis, which can malaria-model incur disease monitoring costs. As such, vaccinations Malaria Malaria-IC-Okell https://github.com/mrc-ide/VIMC_malaria have economic benefits beyond just the prevention of Malaria Malaria-TKI-Penny https://github.com/SwissTPH/openmalaria morbidity and mortality that have not been captured Measles Measles-PSU-Ferrari https://github.com/bwlambert/pfilter_plus here in this study. The high impact ratios for measles­ Measles Measles-LSHTM-Jit https://github.com/lshtm-vimc/dynamice containing vaccines indicate strong indirect effects and Meningitis MenA-Cambridge-Trotter https://github.com/andromachi889/MMCV-model are consistent with its high transmissibility compared Rubella Rubella-UKHSA-Vynnycky https://github.com/EmiliaVynnycky/Rubella-model- with other infectious diseases. Although campaigns for VIMC-runs-2023-24 the measles vaccine in general had lower impact ratios Rubella Rubella-UGA-Winter https://github.com/UGA-IDD/ than routine immunisation, adjusting campaign MRTransmissionModel; https://github.com/UGA- IDD/globalRubellaIgG strategies and target age groups to reach more Typhoid Typhoid-IVI-Kim https://github.com/kimfinale/cholera_typhoid_ unvaccinated children could increase the impact ratio.19 vimc_ivi The current oral cholera vaccine programme is divided Typhoid Typhoid-Yale-Pitzer https://github.com/vepitzer/typhoidVIMC into preventive and reactive use cases, and the impact Yellow fever YF-IC-Gaythorpe https://github.com/mrc-ide/YF_VIMC_Burden_ ratios presented here pertain only to preventive oral orderly cholera vaccine. Previous work20 has highlighted the Yellow fever YF-UND-Perkins https://github.com/TAlexPerkins/yf_ensemble importance of geographical targeting to cholera­endemic Hib Hib-LSHTM-Clark https://github.com/lshtm-vimc/UNIVAC_model areas, as burden is spatially clustered and characterised Hib Hib-PCV-JHU-Tam https://list.spectrumweb.org by unpredictable and shifting geographical patterns.21 Japanese encephalitis JE-OUCRU-Clapham https://github.com/tranquanc123/JE_burden_ Although we limited the oral cholera vaccine coverage estimates assumptions to countries with high cholera burden, the Japanese encephalitis JE-UND-Moore https://github.com/mooresea/JEV_estimation heterogeneity in impact ratios across subregions PCV PCV-LSHTM-NUS No publicly available code but building on UNIVAC underscores the need for efficient subnational model geographical targeting to maximise the impact of PCV PCV-JHU-Tam https://list.spectrumweb.org campaigns. Further, while the long­term effects of acute Rota Rota-Emory-Lopman https://github.com/lopmanlab/VIMC_public diarrhoea events in adults are poorly understood, data Rota Rota-JHU-Tam https://list.spectrumweb.org among children suggest the potential for long­term Rota Rota-LSHTM-Clark https://github.com/lshtm-vimc/UNIVAC_model disability consequences and heightened risk of mortality Hib= Haemophilus influenzae type b. HepB=hepatitis B. HPV=human papillomavirus. PCV=pneumococcal conjugate after the occurrence of a moderate­to­severe diarrhoea vaccine. Rota=rotavirus. event; DALYs and deaths attributed to cholera could be Table 2: Links to publicly available code for each modelling group underestimated due to these measurement challenges.22 The COVID­19 vaccination is a new addition to programmes and is estimated to have averted Rubella Syndrome, the estimates do not include approximately 14 million deaths in 185 territories in the infections during pregnancy that could lead to first year of the pandemic.23 We have presented COVID­19 spontaneous or induced abortions and stillbirths. This impact ratios with other vaccines for the first time and number could be high; for example, in one study, 77% found the overall estimates to be similar in magnitude to (163 of 211 participants) and 6% (13 participants) of rubella across the 117 LMICs. However, the focus of this confirmed infections during the first 16 weeks of study will not capture the impact in high­burden pregnancy had a therapeutic or spontaneous abortion, countries. The burden of rubella is measured through respectively.24 However, estimates varied. As a result, the the number of deaths from Congenital Rubella impact ratios estimated here could be considered Syndrome, which could have implications on the conservative for the effect of the rubella vaccination. magnitude of impact. By focusing only on Congenital Finally, for all vaccines included in the study, our focus Articles on deaths and DALYs averted did not reflect the potential and editing. A­MH: data curation, formal analysis, methodology, reductions in health­care burden. software, validation, visualisation, writing—original draft, and writing— Stakeholder engagement has played a key role in review and editing. ZG: data curation, formal analysis, methodology, software, validation, visualisation, writing—original draft, and writing— defining and estimating these impact ratios. The coverage review and editing. KA: conceptualisation, formal analysis, methodology, assumptions were developed with input from disease supervision, and writing—review and editing. RA: data curation, formal focal points and representatives of WHO and Gavi to analysis, software, visualisation, and writing—review and editing. KAMG, XL, MS, A­MH, ZG, and WH: accessed and verified the data. ensure they align with current guidance and future CA: data curation, formal analysis, software, visualisation, and writing— planning considerations. The estimates were used as review and editing. MA: data curation, formal analysis, software, inputs for global target setting and evaluation.8 Previous visualisation, and writing—review and editing. ASA: conceptualisation, estimates of deaths and DALYs averted from the Vaccine formal analysis, methodology, supervision, and writing—review and editing. EB: investigation, resources, and writing—review and editing. Impact Modelling Consortium have informed the WHO AC: data curation, formal analysis, software, visualisation, and writing— Immunisation Agenda 2030 estimates of vaccine review and editing. MJF: conceptualisation, formal analysis, impact.1,16 The impact ratios are a component of calculating methodology, supervision, and writing—review and editing. the impact by year of vaccination, attributing health KF: data curation, formal analysis, software, visualisation, and writing— review and editing. HF: data curation, formal analysis, software, impact back to the year in which the immunisation visualisation, and writing—review and editing. LH: data curation, formal occurred, which is a methodology that underpins not only analysis, software, visualisation, and writing—review and editing. previous Vaccine Impact Modelling Consortium estimates RGK: conceptualisation, formal analysis, methodology, supervision, but also WHO Immunisation Agenda 2030 target and writing—review and editing. AK­P: data curation, formal analysis, software, visualisation, and writing—review and editing. tracking.11,12,16 Further, impact by year of vaccination was ECL: conceptualisation, formal analysis, methodology, supervision, used to determine health impact targets for Gavi’s and writing—review and editing. ENK: investigation, resources, 5·0 and 6·0 investment opportunities25 and by the Gates and writing—review and editing. J­HK: conceptualisation, formal Foundation for internal strategy development and analysis, methodology, supervision, and writing—review and editing. MJ: conceptualisation, formal analysis, methodology, supervision, and reporting. As more vaccines become available, and writing—review and editing. YL: conceptualisation, formal analysis, funding envelopes remain static or reduce in size, methodology, supervision, and writing—review and editing. rigorous and comparable metrics, such as impact ratios, JM: data curation, formal analysis, software, visualisation, and writing— will increasingly be used to ensure vaccine strategy is review and editing. SM: data curation, formal analysis, software, visualisation, and writing—review and editing. SN: conceptualisation, optimised. formal analysis, methodology, supervision, and writing—review and Health impact informs major decisions around vaccine editing. GN­G: data curation, formal analysis, software, visualisation, investment and prioritisation, but it is one among a suite and writing—review and editing. LCO: conceptualisation, formal of considerations (appendix p 1). Cost­effectiveness, analysis, methodology, supervision, and writing—review and editing. AAO: investigation, resources, and writing—review and editing. equity, market challenges, and economic returns on TP: data curation, formal analysis, software, visualisation, and writing— investment of immunisation all contribute to decision review and editing. MAP: conceptualisation, formal analysis, making.25 Other studies focusing on LMICs found the methodology, supervision, and writing—review and editing. highest cost of illness averted for measles, Hib, and TAP: conceptualisation, formal analysis, methodology, supervision, and writing—review and editing. VEP: conceptualisation, formal HepB,26 and the highest medical impoverishment cases analysis, methodology, supervision, and writing—review and editing. averted for HepB, measles, and MenA.27 Yet, both AP: conceptualisation, formal analysis, methodology, supervision, estimates were sensitive to the assumed underlying and writing—review and editing. SRP: investigation, resources, and health burden. Further, the epidemic potential of a writing—review and editing. CMS: data curation, formal analysis, software, visualisation, and writing—review and editing. disease, and whether that could be exacerbated by climate NS: conceptualisation, formal analysis, methodology, supervision, change, can also affect decisions around how vaccines are and writing—review and editing. CS: data curation, formal analysis, prioritised.28 software, visualisation, and writing—review and editing. AJS: data curation, formal analysis, software, visualisation, and writing—review Given the reality of scarce resources for vaccinations, and editing. SYS: investigation, resources, and writing—review and prioritisation must be rooted in robust evidence. The editing. QT: data curation, formal analysis, software, visualisation, Vaccine Impact Modelling Consortium impact ratios offer and writing—review and editing. EV: conceptualisation, formal analysis, a standardised, reproducible, and transparent framework methodology, supervision, and writing—review and editing. AKW: conceptualisation, formal analysis, methodology, supervision, for comparing the health impacts of a range of vaccines in and writing—review and editing. WH: data curation, resources, the Gavi portfolio. By combining these ratios with cost­ software, and writing—review and editing. NMF: conceptualisation, effectiveness data and broader contextual evidence, formal analysis, funding acquisition, investigation, methodology, stakeholders can ensure that vaccine investments are both supervision, validation, writing—original draft, and writing—review and editing. CLT: conceptualisation, formal analysis, funding evidence­based and optimised for public health impact. acquisition, investigation, methodology, supervision, validation, Contributors writing—original draft, and writing—review and editing. We confirm all KAMG: conceptualisation, formal analysis, funding acquisition, authors had access to all the data in the study and the final responsibility investigation, methodology, project administration, software, for the decision to submit for publication. supervision, validation, writing—original draft, and writing—review and Declaration of interests editing. XL: data curation, formal analysis, methodology, software, KAMG declares speaker fees from Sanofi. A­MH declares grants and validation, visualisation, writing—original draft, and writing—review contracts from German Federal Ministry of Research, Technology and and editing. MS: data curation, formal analysis, methodology, software, Space. KA declares grants and contracts from WHO, International validation, visualisation, writing—original draft, and writing—review 1950 Articles Vaccine Institute, Save the Children UK, Japan AMED, and Gates 2 WHO. 50th anniversary of the Expanded Programme on Foundation; consulting fees from WHO; and meeting and travel support Immunization (EPI). https://www.who.int/news­room/events/ from International Vaccine Institute. ASA declares participation on a detail/2024/01/01/default­calendar/50th­anniversary­of­the­expanded­ data safety monitoring board and advisory board for Gavi. MJF declares programme­on­immunization­(epi) (accessed July 31, 2025). grants and contracts from US National Science Foundation and Gates 3 WHO. WHO’s Strategic Group of Experts charts bold path to Foundation. ECL declares grants and contracts from Gates Foundation; strengthen global immunization amid new challenges. https:// www.who.int/news/item/31­03­2025­who­s­strategic­group­of­ consulting fees from MSF Epicentre, and is a member of the Global experts­charts­bold­path­to­strengthen­global­immunization­amid­ Task Force on Cholera Control Epidemiology Working Group. new­challenges (accessed July 31, 2025). MJ declares grants and contracts from National Institute for Health and 4 Sheikh K, Schneider H. USAID withdrawal and the erosion of Care Research (NIHR), Research Councils UK, Gates Foundation, Gavi, development assistance for health: considerations for health system Wellcome Trust, European Commission, InnoHK, the Taskforce for leadership in LMICs. SSM Health Syst 2025; 5: 100107. Global Health (TFGH), and the US Centers for Disease Control. 5 Apeagyei AE, Bisignano C, Elliott H, et al. Tracking development YL declares grants and contracts from Ministry of Health Singapore, assistance for health, 1990–2030: historical trends, recent cuts, and InnoHK, and Wellcome Trust. LCO declares grants from Merck, Gates outlook. Lancet 2025; 406: 337–48. Foundation, Wellcome Trust, Malaria Consortium, and UK Royal 6 Gavi, the Vaccine Alliance. About our alliance. https://www.gavi. Society; and consulting fees from NIH. SN declares grants from Gavi. org/about­us (accessed July 31, 2025). RGK declares support from European & Developing Countries Clinical 7 Gavi, the Vaccine Alliance. Gavi progress report 2024. https://www. Trials Partnership, AvH/German Foreign Office, and Delta Africa gavi.org/progress­report (accessed July 31, 2025). (Wellcome Trust). VEP declares grants and contracts from Wellcome 8 Gaythorpe KAM, Li X, Clapham H, et al. Estimating the impact of Trust, Gates Foundation, US Centers for Disease Control and vaccination: lessons learned in the first phase of the Vaccine Impact Prevention, National Institutes for Health (NIH)/National Institute of Modelling Consortium. Gates Open Research 2024; 8: 97. Allergy and Infectious Diseases, and NIHR; meeting support from 9 Echeverria­Londono S, Li X, Toor J, et al. How can the public health WHO; and participation on a data safety monitoring board and advisory impact of vaccination be estimated? BMC Public Health 2021; board for WHO. AP declares grants and contracts from NIH, Gates 21: 2049. Foundation, Open Philanthropy, Boston University–Boston Medical 10 Li X, Mukandavire C, Cucunubá ZM, et al, and the Vaccine Impact Center Cancer Center, and Gavi. SRP declares grants and contracts from Modelling Consortium. Estimating the health impact of vaccination NIHR, Gates Foundation, Gavi, TFGH, and WHO. NS declares grants against ten pathogens in 98 low­income and middle­income countries and contracts from Gilead Sciences. CS declares grants and contracts from 2000 to 2030: a modelling study. Lancet 2021; 397: 398–408. from Gilead Sciences. AKW declares grants and contracts from Gates 11 Jaspreet Toor, Susy Echeverria­Londono, Xiang Li, et al. Lives saved Foundation, Gavi, and The Council of State and Territory with vaccination for 10 pathogens across 112 countries in a Epidemiologist. NMF declares grants and contracts from NIH, and pre­COVID­19 world. Elife 2021; 10: e67635. consulting fees from World Bank Group. CLT declares grants from 12 Hartner AM, Li X, Echeverria­Londono S, et al. Estimating the NIHR, and is the Co­chair of WHO Technical Taskforce on Defeating health effects of COVID­19­related immunisation disruptions in Meningitis by 2030. All other authors declare no competing interests. 112 countries during 2020–30: a modelling study. Lancet Glob Health 2024; 12: e563–71. Data sharing 13 Gavi, the Vaccine Alliance. Phase V (2021–2025). https://www.gavi. The code can be accessed from the following repository: https://github. org/our­work/strategy/phase­5­2021­2025 (accessed Feb 3, 2026). com/vimc/paper4. Data can be explored and downloaded from https:// 14 WHO. Immunisation dashboard. https://immunizationdata.who. vaxviz.vaccineimpact.org. Specific model code can be accessed from the int (accessed April 10, 2026). repositories listed in table 2. 15 WHO. Immunization analysis and insights. https://www.who.int/ teams/immunization­vaccines­and­biologicals/immunization­ Acknowledgments analysis­and­insights/global­monitoring/immunization­coverage/ This work was carried out as part of the Vaccine Impact Modelling who­unicef­estimates­of­national­immunization­coverage(accessed Consortium, but the views expressed are those of the authors and not April 10, 2026). necessarily those of the Consortium or its funders. The funders were 16 Carter A, Msemburi W, Sim SY, et al. Modeling the impact of given the opportunity to review this paper before publication, but the vaccination for the Immunization Agenda 2030: deaths averted due final decision on the content of the publication was taken by the to vaccination against 14 pathogens in 194 countries from 2021 to authors. This work was supported, in whole or in part, by Gates 2030. Vaccine 2024; published online Aug 1. doi.org/10.1016/j. Foundation, via the Vaccine Impact Modelling Consortium (grant vaccine.2023.07.033. number: INV­034281), previously (OPP1157270 / INV­009125) and Gavi, 17 UN. Department of Economic and Social Affairs. World population the Vaccine Alliance. The conclusions and opinions expressed in this prospects 2022. https://www.un.org/development/desa/pd/sites/ work are those of the authors alone and shall not be attributed to the www.un.org.development.desa.pd/files/wpp2022_summary_of_ Foundation. Under the grant conditions of the Gates Foundation, a results.pdf (accessed Aug 1, 2025). Creative Commons Attribution 4.0 License has already been assigned to 18 UN. Department of Economic and Social Affairs. Standard country the Author Accepted Manuscript version that might arise from this or area codes for statistical use. https://unstats.un.org/unsd/ submission. Please note works submitted as a preprint have not methodology/m49/overview (accessed Aug 20, 2025). undergone a peer review process. The Gates Foundation UK­funded 19 Auzenbergs M, Fu H, Abbas K, Procter SR, Cutts FT, Jit M. Health award has been carried out in the frame of the Global Health EDCTP3 effects of routine measles vaccination and supplementary Joint Undertaking. So Yoon Sim works for WHO. The contents of this immunisation activities in 14 high­burden countries: a Dynamic Measles Immunization Calculation Engine (DynaMICE) modelling Article are solely the responsibility of the authors and do not represent study. Lancet Glob Health 2023; 11: e1194–204. the official views of their affiliated organisations. Where authors are 20 Lee EC, Azman AS, Kaminsky J, Moore SM, McKay HS, Lessler J. identified as personnel of WHO, the authors alone are responsible for The projected impact of geographic targeting of oral cholera the views expressed in this Article, and they do not necessarily represent vaccination in sub­Saharan Africa: a modeling study.PLoS Med the decisions, policy, or views of WHO. We acknowledge the 2019; 16: e1003003. contributions of David Mears, and we would like to thank all groups 21 Perez­Saez J, Zheng Q, Kaminsky J, et al. Geographical shifting of who have participated in the Vaccine Impact Modelling Consortium 1.0 cholera burden in Africa and its implications for disease control. (https://www.vaccineimpact.org/vimc­models­1). Nat Med 2025: 31: 3380–87. References 22 Kotloff KL, Nataro JP, Blackwelder WC, et al. Burden and 1 Shattock AJ, Johnson HC, Sim SY, et al. Contribution of aetiology of diarrhoeal disease in infants and young children in vaccination to improved survival and health: modelling 50 years of developing countries (the Global Enteric Multicenter Study, the Expanded Programme on Immunization. Lancet 2024; GEMS): a prospective, case­control study. Lancet 2013; 403: 2307–16. 382: 209–22. Articles 23 Watson OJ, Barnsley G, Toor J, Hogan AB, Winskill P, Ghani AC. 26 Ozawa S, Clark S, Portnoy A, et al. Estimated economic impact of Global impact of the first year of COVID­19 vaccination: a vaccinations in 73 low­and middle­income countries, 2001–2020. mathematical modelling study. Lancet Infect Dis 2022; 22: 1293–302. Bull World Health Organ 2017; 95: 629–38. 24 Grillner L, Forsgren M, Barr B, Böttiger M, Danielsson L, and De 27 Chang AY, Riumallo­Herl C, Perales NA, et al. The equity impact Verdier C. 1983. Outcome of rubella during pregnancy with special vaccines may have on averting deaths and medical impoverishment reference to the 17th–24th weeks of gestation. Scand J Infect Dis 1983; in developing countries. Health Aff 2018; 37: 316–24. 15: 321–25. 28 Mahmud AS, Martinez PP, He J, et al. The impact of climate 25 Gavi, the Vaccine Alliance. The Gavi investment opportunity change on vaccine­preventable diseases: insights from current 2021–25. https://www.gavi.org/gavi­investment­ research and new directions. Curr Environ Health Rep 2020; opportunity­2021­2025 (accessed Jan 15, 2024). 7: 384–91. 1952 --- [PDF原文](https://sci-net.xyz/storage/7932541/c85e30f858594ad027aa68b1d2a46cb244dcd77b5884ad476e919c7d82f349ea/Quantifying-relative-health-impact-across-Gavi-the-Vaccine-Alliance-s-portfolio-in-117-countries.pdf) DOI: 10.1016/S0140-6736(26)00555-6