Machine Learning-Based Analysis of BMI, Tobacco Use and Oral Health Outcomes Among Chennai Construction Workers: Insights for Targeted Interventions

Background: This study investigates the complex relationship between Body Mass Index (BMI), tobacco consumption and oral health status among construction workers in Chennai, India. Using a neural network model, the study aims to identify significant predictors of oral health outcomes. Methods: A cross-sectional study was conducted among 812 construction workers in Chennai. Data on BMI, tobacco usage patterns and oral health indicators such as the Oral Hygiene Index-Simplified (OHI-S), Decayed, Missing and Filled Teeth (DMFT) index, Gingival Index and loss of attachment (LOA) were collected. A pretested, interviewer-administered questionnaire was used to gather demographic and behavioral information. BMI was calculated using participants' height and weight. Data were analyzed using SPSS software, with a neural network model employed to identify complex interactions among variables. Results: The neural network analysis revealed significant positive correlations between BMI and oral health indicators such as DMFT, Gingival Index and missing teeth. Additionally, a negative correlation was identified between BMI and tooth brushing frequency. Notably, education emerged as the most influential predictor, followed by age, gender, BMI and tobacco consumption. The neural network model demonstrated superior predictive accuracy of 87%, outperforming traditional statistical methods in identifying complex, non-linear relationships. Conclusion: The study highlights the intricate interplay between BMI, tobacco use and oral health outcomes, underscoring the need for targeted interventions. Emphasis should be placed on improving nutrition, promoting tobacco cessation and enhancing oral health awareness among construction workers. The findings demonstrate the value of machine learning in public health research and emphasize the need for future longitudinal studies to establish causal links and explore additional factors such as occupational stress, dietary patterns and healthcare access.