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

PLOS One | https://doi.org/10.1371/journal.pone.0335529 November 26, 2025 6 / 17 vulnerability levels were re-grouped to create a binary variable with two clearly distinct categories. The very high and highly vulnerable districts were clubbed together in one category (re-named and referred to hereafter as highly vulnerable districts), and the medium, low and very low districts in another category (re-named and referred to hereafter as other districts). Health data. The health outcome variables that have been selected for the analysis are specified as binary indicators. The data is extracted from the NFHS survey, which was conducted for most states in 2019, overlapping with the year for the dataset on climate vulnerability. The Covid-19 pandemic delayed the completion of the dataset to early 2021 [84]. Any limitations arising from these circumstances are deemed to have been taken care of as the study uses the finalized and corrected dataset approved by the Government and made available in the public domain. NFHS surveys are conducted under the stewardship of the Ministry of Health and Family Welfare, Government of India. We consider only data from rural areas to maintain consistency across the NFHS and the climate datasets. The NFHS covered more than 0.6 million (636000) households, including 726,828 women aged 15–49 and 102,266 men aged 15–54. It is also to be noted that the NFHS target populations were aligned with the SDG targets, further ensuring the suitability of the dataset for the analysis

in this study. This dataset is used by the Government of India for health policy planning [85] and for monitoring progress towards achieving SDG 2 and 3 [31]. The dependent variables measuring stunting, wasting, being underweight, having non-institutional deliveries and problems in accessing healthcare were directly available in the dataset. (34; the descriptions of the variables are pro- vided in S2 Table. A variable to measure access to health care was not directly available and was therefore constructed from four individual indicators on access to health care on which data had been collected from the respondents in the survey. Respondents provided information on ‘Distance to health care’; ‘availability of health care provider’; ‘availability of drugs’; and ‘availability of female health care provider’ at the nearest government health care facility. The responses were recorded in a categorical form as a ‘Big problem; ‘Small problem’; and ‘No problem’. The variable constructed from these responses assigns a value of 1 to responses of a “big problem” and clubs together responses of a ‘Small problem’ and ‘No problem’, assigning the latter a value of 0. Subsequently, all individuals who have responded as having a “big problem” Fig 1. State-wise distribution of districts by vulnerability levels. https://doi.org/10.1371/journal.pone.0335529.g001