Health SDGs are at risk from climate change: Evidence from India

PLOS One | https://doi.org/10.1371/journal.pone.0335529 November 26, 2025 4 / 17 India’s performance has been comparatively better in SDG 3, with an index score of 77 as compared to a target of 100 [31]. Significant advancements in maternal and child healthcare services have come from concerted efforts through pro- grams and interventions under the National Health Mission [56,57], and targeted initiatives from the government focusing on high quality, comprehensive ANC while alleviating financial burdens on women [58]. Some schemes incentivize institu- tional childbirth and mitigate direct expenses [59], while others focus on improving the quality of care and services during labor [60]. Child mortality rates have reduced significantly in India [61,62] due to some of these efforts, though disparities persist in some regions and for specific disadvantaged groups [63]. Institutional deliveries in India have significantly increased over the past three decades with most states having already surpassed SDG targets (92%) for 2030. However, challenges persist in some regions, with institutional delivery rates below 80%. A lack of ANC, poor socioeconomic status, locational disadvantages, and higher birth order are among the challenges associated with lower utilization of institutional delivery in certain regions [64–66], manifested in terms of poor performance in key indicators [67–69]. Access to healthcare is a crucial determinant for achieving good health and health

equity. Several studies in South Asia, including India, have identified access to health care as being critical for health promotion and health equity, and for addressing vulnerabilities across communities and overcoming gender disparities in health outcomes. Harsh geographic conditions, long distances to health centers, limited transportation, lack of services, absence of adequate and trained health care professionals, and financial barriers have been often noted as impacting access to equitable and quality health care [70–74]. Increasingly, climate change has become a major force disrupting health care delivery, undermining the social determinants of good health and reducing the capacity to provide health care, including universal health coverage [15], with the potential of significantly increasing health costs in LMICs [75,76]. Climate change is thus increasingly recognized globally as a major determinant of health, with differential and adverse impacts for vulnerable populations, including women and children. The need to address the impacts of climate change on health along with the other socio-economic determinants of health is being recognized [77]. In India, are preparing action plans on climate change and human health. However, the evidence base and understanding on the impacts due to climate change is weak. Prior to the current study, there have been no such attempt at quantifying the available evidence in terms of the impacts on human health in India at an All-India level, with sufficient granularity to understand differential impacts at sub-national levels.

2. Materials and methods Methods The study uses statistical and econometric methods to examine the relationship between climate change and health outcomes. A multivariate logistic regression is estimated. Such estimation techniques have been used elsewhere to examine the relationship between climate change variables and health outcomes, such as in USA and China [78,79]. The health-outcomes studied are stunting, wasting, and being underweight for children under five years of age, and non-institutional deliveries and problems in accessing health care for women aged 15–49 years. The role of climate change as a determinant for these health outcomes is captured by an explanatory variable on climatic vulnerability across districts. Other socio-economic variables are also included as determinants in the specification. The outcome in the logistic equation is specified as a dichotomous dependent variable taking a value of 0 or 1, depend- ing on whether the outcome is observed or not (e.g., 1 if stunted, 0 if otherwise). The model provides the probability (p) that the event (e.g., stunting) will occur relative to the probability of its non-occurrence (1-p), providing the odds of the event occurring, and the odds ratio, with a reference category [80]. In a multivariate specification, each odds ratio [81] provides an estimate of the change in the odds or chances (e.g., of a child being stunted) that is associated with the corresponding explanatory variable (xi, e.g., climatic vulnerability), while controlling for the other independent variables

(e.g., gender, age, mother’s education).Using a logarithmic transformation of the odds ratio (e.g., odds of being stunted), a multivariate specification generates the log odds ratio for each explanatory variable in the model. Equation 1 presents