Small area variations in low birth weight and small size of births in India.

The states and districts are the primary focal points for policy formulation and programme intervention in India. The within‐districts variation of key health indicators is not well understood and consequently underemphasised. This study aims to partition geographic variation in low birthweight (LBW) and small birth size (SBS) in India and geovisualize the distribution of small area estimates. Applying a four‐level logistic regression model to the latest round of the National Family Health Survey (2015–2016) covering 640 districts within 36 states and union territories of India, the variance partitioning coefficient and precision‐weighted prevalence of LBW (<2.5 kg) and SBS (mother's self‐report) were estimated. For each outcome, the spatial distribution by districts of mean prevalence and small area variation (as measured by standard deviation) and the correlation between them were computed. Of the total valid sample, 17.6% (out of 193,345 children) had LBW and 12.4% (out of 253,213 children) had SBS. The small areas contributed the highest share of total geographic variance in LBW (52%) and SBS (78%). The variance of LBW attributed to small areas was unevenly distributed across the regions of India. While a strong correlation between district‐wide percent and within‐district standard deviation was identified in both LBW (r = 0.88) and SBS (r = 0.87), they were not necessarily concentrated in the aspirational districts. We find the necessity of precise policy attention specifically to the small areas in the districts of India with a high prevalence of LBW and SBS in programme formulation and intervention that may be beneficial to improve childbirth outcomes. Key messages: The small areas contribute the highest share of the total geographic variance of low birth weight (LBW) and small birth size (SBS) in India.A high burden of LBW is found mostly in the central‐western part of India and Odisha. The prevalence of SBS is high across the district of northern‐western regions and the north‐eastern regions of India.The mean prevalence and standard deviation are strongly correlated in the case of both LBW (r = 0.88) and SBS (r = 0.87) in India. It indicates that the districts which have a higher prevalence of LBW and SBS also have a higher between small area disparity within the districts.We find a similar pattern of distribution in LBW and SBS between the policy‐focused aspirational districts and other districts of India.Findings indicate reprioritizing the policy intervention, focusing on the small areas of India for better childbirth outcomes. [ABSTRACT FROM AUTHOR]