Risk profiles for smoke behavior in COVID-19: a classification and regression tree analysis approach

Area of Science:

• Public Health
• Epidemiology
• Behavioral Science

Background:

• The COVID-19 pandemic significantly impacted global health, with established links to smoking behaviors.
• Previous research noted changes in smoking during the pandemic but did not identify key influencing factors or high-risk groups.

Purpose of the Study:

• To identify the primary factors influencing smoking behavior during the COVID-19 pandemic.
• To determine high-risk populations for increased smoking during this period.

Main Methods:

• A nationwide survey of 21,916 participants was conducted.
• Logistic regression analyzed factors (sociodemographics, social support, mental health, self-efficacy) related to smoking.
• Classification and Regression Tree (CART) identified high-risk groups and key predictors.

Main Results:

• Acceptance of passive smoking was the strongest predictor of smoking behavior.
• A high-risk subgroup (34%) was identified: those with high passive smoking acceptance, no smokers nearby, and ≥30 years of smoking history.
• The CART model achieved 87% accuracy.

Conclusions:

• The acceptance of passive smoking is the principal factor affecting smoking behavior during the pandemic.
• Promoting awareness of secondhand smoke harm is crucial.
• Leveraging the 'mask protection' effect can aid smoking cessation in high-risk individuals.