Mira Johri

BackgroundReaching zero-dose children, defined here as children who have not received a first DTP-containing dose, is a global priority. However, zero-dose counts do not capture children who start vaccination and later drop out. In India, many children remain under-immunised, meaning that they miss one or more routine vaccine doses. We assessed whether the risk profile differs between non-initiation and later dropout, across geography, and beyond what additive regression can represent.Methods We analysed NFHS-5 child records for three outcomes: zero-dose (43{,}181 children; 2{,}976 positive), within-series dropout (DTP1--DTP3), defined as failure to complete the three-dose DTP-containing series after initiation (40{,}186; 2{,}578 positive), and cross-series dropout (BCG--MCV1), defined as receipt of BCG but not a first measles-containing vaccine dose across the first-year schedule (40{,}985; 3{,}510 positive). We fitted penalised logistic regression and feed-forward neural networks at pooled India, five macro-regions, and state level. To compare additive and non-additive risk structure, we used Shapley additive explanations (SHAP) and interaction analyses.FindingsAt pooled India level, the neural network improved accuracy and Macro-F1 for all three outcomes; the largest gain was for cross-series dropout (BCG--MCV1). On a common SHAP scale, zero-dose risk was driven more strongly by deprivation, home delivery, and rural residence, whereas the two dropout outcomes gave greater weight to child age, birth order, and maternal educational disadvantage. Interaction analyses showed moderate non-additive structure centred on deprivation, maternal education, birth order, and service-contact variables. Regionally, Macro-F1 improved in all 15 region-outcome cells. At state level, mean gains remained positive, but minority-class detection was inconsistent; the main contribution at that scale was localisation of heterogeneous drivers and interaction pairs.InterpretationChildhood under-immunisation in India is not a single process. The risk profile differs between children who never start routine vaccination and those who start but do not complete it, and it varies across regions and states. Flexible models added value mainly by clarifying that structure rather than by simple algorithmic superiority.

Management response processes following evaluations foster accountability, evidence-informed decision-making and organisational learning. This article examines trends, good practices and remaining challenges in management response systems across multilateral organisations within the United Nations system or in close collaboration with it. It draws on a rapid evidence assessment – combining interviews with key informants from 14 multilateral organisations and a review of over 60 documents (evaluation policies, guidelines and annual reports) from these organisations – complemented by insights from a dissemination workshop with 26 United Nations Evaluation Group member organisations. Findings indicate substantial variation in management response practices, explained by structural and operational differences such as reporting lines and levels of decentralisation. Persistent gaps include diffuse ownership, limited validation and infrequent public reporting. This article highlights effective and transferable practices, such as integrating management response actions into corporate planning, linking implementation to key performance indicators and using dashboards to track progress.

Introduction: Intensified efforts are needed to achieve the Immunization Agenda 2030 (IA2030) target to halve the number of children not reached by routine vaccination (‘zero-dose’ (ZD) children) by 2030. This scoping review maps and synthesises the scientific literature on ZD children in low-income and middle-income countries since the IA2030 launch. Methods: Our protocol, developed per the PRISMA-ScR (Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews) guidelines, was registered on the Open Science Framework prior to execution, and our search strategies were designed by a research librarian. We searched MEDLINE (Ovid), CINAHL Complete (EBSCOhost), EBM Reviews (Ovid), EMBASE (Ovid), LILACS and Google Scholar for peer-reviewed studies with quantitative evidence on ZD children published between January 2020 and January 2024. Using Covidence, we screened citations in a two-stage process by two independent reviewers and conducted data charting using a pretested form. Disagreements were resolved by consensus or consultation with a senior reviewer. Results: Of the 82 articles included, 73 provided evidence on prevalence, and/or distribution, 24 on barriers to vaccination services and 44 on deprivations faced by ZD children, their households and communities, with the risk of deprivations varying importantly across contexts. The barriers to vaccination most reported related to the intent to vaccinate (n=20), followed by community access (n=10). Deprivations mainly concerned access to health services, such as low utilisation of maternal health services (n=22), and other key development indicators, such as low socioeconomic status (n=29) and poor maternal education/literacy (n=28). We found no studies (n=0) on interventions focused on reaching ZD children with routine immunisation services. Conclusion: Our findings highlight the imperative to generate evidence on interventions delivering routine immunisation to ZD children and missed communities. Results also underscore the need for transformative approaches that address multiple deprivations using carefully selected, integrated services, tailored to context-specific needs.

Background Infectious diseases in childhood and adolescence are significant and often preventable causes of hearing loss, especially in low- and middle-income countries. We conducted a scoping review to examine the extent, range and nature of available evidence on the role of vaccination for prevention of hearing loss worldwide. Methods We reviewed the published scientific literature to identify studies providing quantitative information on the relationship between vaccination and hearing loss. We searched four databases: MEDLINE, EMBASE, Cochrane Library and Global Index Medicus. No date, language, or geographical restrictions were imposed. Two independent reviewers assessed eligibility and charted data. Results Here we show that vaccination may be a key, underexploited strategy for primary prevention of child and adolescent hearing loss. Although the important contributions of rubella and meningitis vaccinations to hearing loss prevention are widely recognised, we identify 26 distinct known or potential infectious causes of hearing loss that are preventable or possibly preventable through vaccination. Notwithstanding, we find a dearth of empirical evidence on the impacts of vaccination on hearing loss prevention. In addition, the review identifies no research from low- and middle-income countries, which bear the overwhelming burden of child and adolescent hearing loss. Finally, it shows that numerous vaccines that address priority infectious diseases relevant to hearing loss are in development and could be brought into use. Conclusions We recommend strategic investment in research concerning vaccination as a strategy for primary prevention of child and adolescent hearing loss.