Karlsson Omar, Kumar Akhil, Kim Rockli, Subramanian S V
Duke University, Durham, North Carolina, USA.
Lund University, Lund, Sweden.
BMJ Glob Health. 2025 Jun 16;10(6):e016732. doi: 10.1136/bmjgh-2024-016732.
Low birth weight is an important measure of the health of pregnant women and newborns. We investigated the prevalence of low birth weights in India over nearly three decades to assess trends and convergence across states.
Data came from five waves of the National Family Health Survey (1992-93 to 2019-21). The prevalence of low birth weight was estimated. To explore the sensitivity of our results to missing birth weight data-since the completeness of birth weight information has changed drastically-we also estimated prevalence from multiple imputation models, Heckman selection models, and by reweighting the data so that socioeconomic characteristics of children with birth weight data matched across surveys.
The overall prevalence of low birth weight in India declined from 26% to 18% during the period. The 2019-21 survey revealed that four states, Uttar Pradesh, Bihar, Maharashtra, and West Bengal accounted for almost half of all low-birth-weight births in India. The Pearson's correlation between the prevalence of low birth weight in 1992-93 and percentage point change across the period was -0.85, suggesting convergence between states, where states with greater prevalence in 1992-93 had faster declines. Convergence was robust across sensitivity specifications.
State-level convergence indicates a potential 'catch-up' phenomenon, where states with initially higher prevalence have experienced greater declines. This finding suggested a possible impact of interventions prompted by dire figures in the earliest surveys, yet also stresses the necessity for continued interventions across all states to maintain and further progress. Our analysis, however, warrants a cautious interpretation due to data limitations. However, we observed convergence in the prevalence of low birth weight across states in all sensitivity specifications.
低出生体重是衡量孕妇和新生儿健康状况的一项重要指标。我们调查了印度近三十年来低出生体重的患病率，以评估各邦之间的趋势和趋同情况。
数据来自五轮全国家庭健康调查（1992 - 1993年至2019 - 2021年）。对低出生体重的患病率进行了估计。由于出生体重信息的完整性发生了巨大变化，为了探究我们的结果对缺失出生体重数据的敏感性，我们还通过多重插补模型、赫克曼选择模型以及对数据重新加权来估计患病率，以便使有出生体重数据的儿童的社会经济特征在各次调查中相匹配。
在此期间，印度低出生体重的总体患病率从26%降至18%。2019 - 2021年的调查显示，北方邦、比哈尔邦、马哈拉施特拉邦和西孟加拉邦这四个邦的低出生体重儿数量几乎占印度低出生体重儿总数的一半。1992 - 1993年低出生体重患病率与该时期百分点变化之间的皮尔逊相关系数为 - 0.85，表明各邦之间存在趋同现象，即1992 - 1993年患病率较高的邦下降速度更快。在各种敏感性设定下，趋同情况都很稳健。
邦级趋同表明存在一种潜在的“追赶”现象，即最初患病率较高的邦下降幅度更大。这一发现表明最早调查中的严峻数据可能促使了干预措施的实施，但同时也强调了在所有邦持续开展干预措施以维持并进一步取得进展的必要性。然而，由于数据限制，我们的分析需要谨慎解读。不过，我们在所有敏感性设定下都观察到了各邦低出生体重患病率的趋同情况。
