Smokers’ knowledge, attitude and practices toward cigarette butt disposal: a Bayesian network analysis of a cross-sectional study in Iran

Cigarette butt waste poses environmental and public health risks due to non-biodegradable filters and toxic chemicals. However, smokers' Knowledge and disposal practices remain understudied in Iran. This study assessed knowledge, attitude, and practices (KAP) regarding cigarette butt disposal among smokers in Isfahan and explored sociodemographic influences. A cross-sectional study was conducted (January-March 2024) with 636 adult smokers (mean age 37.9 ± 10.3; 70.8% male) using convenience sampling. A validated KAP questionnaire was administered. Bayesian network analysis with a Gaussian graphical model (GGM) examined conditional dependencies among ten variables (KAP, age, gender, education, income, occupation, marital status, cigarette consumption). Edges were retained if Bayes factor (BF₁₀) > 10 (strong evidence). Centrality indices (strength, closeness, betweenness, expected influence) with 95% bootstrapped confidence intervals were calculated, and correlation stability (CS) coefficients were reported. Most participants had low knowledge (45.0%) and negative attitudes (67.5%), while moderate practices (42.8%) were common. The Bayesian network and GGM were validated using bootstrap resampling (n = 1000), with CS coefficients above 0.5 for all centrality indices, confirming acceptable network stability. The network (10 nodes, 25 edges, sparsity = 0.444) showed practices and attitude as the most central nodes (betweenness: 1.489; expected influence: 1.280 for practices; betweenness: 1.489; expected influence: 1.182 for attitude). Strongest positive conditional dependencies were age-marital status (0.373), education-income (0.321), and cigarette consumption-knowledge (0.197). Knowledge and income were peripheral (strength = - 1.653 and - 0.973, respectively; standardized z-scores). All reported associations are conditional dependencies, not causal effects, due to the cross-sectional design. Attitude and practices are the most interconnected nodes in the conditional dependency network. Interventions targeting these constructs, especially regarding cigarette butt disposal behaviors, may reduce urban cigarette butt pollution. Findings are limited by cross-sectional design and convenience sampling.