Diet, physical activity, sedentary time and demographic variables differentiate among drinkers and former drinkers who suffer or not depression: a CRT analysis.

Healthy lifestyles may have the potential to be protective factors for mental health problems. The aim was to classify drinkers and former drinkers in depression sufferers or non-sufferers according to demographic data and lifestyle habits, using a classification and regression tree (CRT). This cross-sectional study is based on the NHANES 2011-20. 3175 participants (2342 males and 833 females) drinkers and former drinkers formed the final sample. CRT analysis was used to classify them in depression sufferers and non-sufferers. Odds ratio (OR) and relative risks (RR) of suffering depression for each branch of the tree were calculated. The model was found to have an 82.0% accuracy. Independent variables were classified by their importance in the model, in this order: diet (0.007; 100.0%), gender (0.004; 60.4%), PA group (0.003; 45.4%), sedentary time (0.002; 32.6%), education level (0.001; 29.1%) and age (<0.001; 7.1%). Elevated OR and RR of suffering depression were found for participants with: fair and poor diet compared to good, very good or excellent diet (OR = 2.02, CI95%: 1.69-2.43, RR = 1.77, CI95%: 1.52-2.10), for inactive, walkers and low PA, compared to moderate PA and high PA (OR = 1.99,CI95% = 1.53-2.60, RR = 1.68, CI95%:1.37-2.07), for females compared to males (OR = 2.00, CI95%:1.54-2.66, RR = 1.81, CI95%:1.45-2.26), for people with a '9-11th grade' or below education level compared to those with a higher level of education (OR = 1.50, CI95%:1.05-2.15, RR = 1.32, CI95%:1.04-1.67) in branch beneath node 3; and for people with a high school/GED or below education level, compared to those with higher education (OR = 1.99, CI95%:1.27-3.13, RR = 1.71, CI95%:1.21-2.43), in branch beneath node 6. Diet, PA group, sedentary time, gender, education level and age allow us to classify drinkers and former drinkers in sufferers or non-sufferers of depression. The CRT model combined with OR and RR allow us to identify which participants have greater odds of suffering depression based on lifestyle and demographic data.