Quantifying and modeling respirable dust and crystalline silica exposure in rice mills using the CART algorithm.

This study aimed to quantify respirable dust (RD) and respirable crystalline silica (RCS) exposure in rice mills of Western India, and, evaluating their association with rice varieties and operational activities using CART algorithm to handle missing data. Fifty RD samples were collected from breathing zones of workers across mills processing different rice varieties, post which corresponding RCS levels were quantified. Statistical analyses in Minitab and R, incorporating CART-imputed datasets, revealed that operational activities significantly influenced RD (22.55%) and RCS (10.59%) variability (p < 0.05), while rice varieties showed minimal impact. Dehusking operations exhibited highest RD and RCS concentrations with notable variability, while sieving operations reported lowest exposures. Amongst rice varieties, Kolam reported highest RD levels, whereas Parimal the lowest. RCS levels were higher in Kolam and Parimal varieties as compared to IR-8, but within occupational exposure limits. ANOVA confirmed the model's statistical robustness. Overall, as compared to existing literature, RD and RCS levels were reportedly lower in this study, possibly due to advanced processing technologies and regional soil variations. The study emphasized need for enhanced dust control, particularly in high-exposure tasks like dehusking. It was also suggestive towards refining predictive accuracy with additional environmental data for rice mill settings.