Abstract
BACKGROUND: Tobacco use among youth in low- and middle-income countries (LMICs), particularly India, remains a critical public health challenge. However, limited research has examined how different geographic determinants distinctly contribute to smoked and smokeless tobacco use among young men in India. METHODS: This study utilized data of 31,070 young men aged 15-24 years (urban = 7784, rural = 23,286) from the National Family Health Survey (NFHS-5, 2019-21). Four-level random intercept logistic regression models (individual, community, district, state) were employed separately for smoked tobacco, smokeless tobacco (SLT), and any tobacco use. Variance partitioning coefficients (VPC) with 95% confidence intervals quantified the proportion of variation attributable to each geographic level, adjusting for background characteristics. Sampling weights were applied for prevalence estimation. RESULTS: Overall prevalence was 22.5% for any tobacco, 15.2% for SLT, and 11.6% for smoked tobacco, with higher rates in rural areas (25.1%, 17.8%, 12.2% respectively). State-level variation was substantial, ranging from 1.4% (Puducherry) to 49.0% (Meghalaya) for any tobacco use. Variance partitioning revealed distinct geographic patterns: PSU-level factors dominated for smoked tobacco (VPC = 58.2%, 95% CI: 55.1-61.2%), while state-level factors predominated for SLT (VPC = 53.1%, 95% CI: 50.3-55.9%). Urban-rural differences showed stronger PSU influence in urban areas for smoking (VPC = 62.0% vs. 51.0% rural), and stronger state-level influence in rural areas for SLT (VPC = 55.4% vs. 48.2% urban). CONCLUSION: Tobacco use among young Indian men exhibits substantial multilevel geographic variation with distinct patterns for smoked versus smokeless products. Effective youth tobacco control requires coordinated multi-level strategies-community-based programs for smoking prevention and state-level policies for SLT control-rather than one-size-fits-all approaches.