| Summary: | Abstract Background Few studies have investigated factors associated with smoking behaviors. In this population-based study, we investigated demographics and medical comorbid diseases to establish a prediction model for smoking behaviors by using the National Health Interview Survey (NHIS) and National Health Insurance Research Database (NHIRD). Methods We enrolled individuals aged ≥40 years who had participated in the NHIS in 2001, 2005, and 2009. We identified the smoking behaviors of the study participants in the NHIS. Smoking behaviors were divided into ever smokers (current smokers and ex-smokers) and nonsmokers (never smokers).We defined medical comorbid disorders of the study participants by using medical claim data from the NHIRD. We used multivariable logistic regression models to calculate the adjusted odds ratio and 95% confidence interval for variables associated with smoking. The significant variables in the multivariable model were included in the receiver operating characteristic curves (ROC) to predict the sensitivity and specificity of the model. Results In total, 26,375 participants (12,779 men and 13,596 women) were included in the analysis. The prevalence of smoking was 39.29%. The mean ages of the 16,012 nonsmokers were higher than those of the 10,363 smokers (57.86 ± 12.92 years vs. 53.59 ± 10.82 years). Men outnumbered women among smokers (68.18% vs. 31.82%). Male sex, young age and middle age, being insured categories, residence in suburban areas, and chronic obstructive pulmonary disease (COPD) were independent factors associated with smoking. The area under the ROC curve of these significant factors to predict smoking behaviors was 71.63%. Conclusion Sex, age, insured categories, residence in suburban areas, and COPD were associated with smoking in people.
|
|---|
