METHODS: This secondary dataset analysis used data from the National Health and Morbidity Survey (NHMS) 2018. Data from 3914 participants were collected on elderly health in the Malaysian population. Sociodemographic characteristics were recorded. Smoking status was grouped as current smokers, former smokers, and non-smokers. A validated Malay language version of the Geriatric Depression Scale (M-GDS-14) was used to screen for depression among the elderly.

RESULTS: There was a significant association between smoking status with location, gender, employment status, marital status, ethnicity, education level, income, and depression. Current smokers are significantly higher in rural than urban areas. Among depressed participants, 65.7%, 17.1% and 17.2% were non-smokers, former smokers and current smokers, respectively. Multiple logistic regression showed that single (unmarried/separated/ divorced/widowed) participants were more likely to be depressed compared to married participants (AOR=1.68; 95% CI: 1.16-2.43). Whilst unemployed participants were more likely to be depressed than those who were employed (AOR=1.72; 95% CI: 1.22-2.44). Other Bumiputras were more likely to have depression compared to Malay, Chinese and Indian participants. Participants without formal education were more likely to be depressed compared to those having tertiary education. These participants have a 2-fold increased risk of depression (AOR=2.13; 95% CI: 1.02-4.45). Participants whose monthly salaries were <2000 MYR (AOR=3.67; 95% CI: 1.84-7.31) and 1000-1999 MYR (AOR=2.71; 95% CI: 1.23-5.94) were more likely to have depression compared with those who had received ≥3000 MYR. Ever smokers were more likely to be depressed than non-smokers (AOR=1.68; 95% CI: 1.23-2.29).

PATIENTS AND METHODS: The dataset encompassed patient data from a tertiary cardiothoracic center in Malaysia between 2011 and 2015, sourced from electronic health records. Extensive preprocessing and feature selection ensured data quality and relevance. Four machine learning algorithms were applied: Logistic Regression, Gradient Boosted Trees, Support Vector Machine, and Random Forest. The dataset was split into training and validation sets and the hyperparameters were tuned. Accuracy, Area Under the ROC Curve (AUC), precision, F-measure, sensitivity, and specificity were some of the evaluation criteria. Ethical guidelines for data use and patient privacy were rigorously followed throughout the study.

RESULTS: With the highest accuracy (88.66%), AUC (94.61%), and sensitivity (91.30%), Gradient Boosted Trees emerged as the top performance. Random Forest displayed strong AUC (94.78%) and accuracy (87.39%). In contrast, the Support Vector Machine showed higher sensitivity (98.57%) with lower specificity (59.55%), but lower accuracy (79.02%) and precision (70.81%). Sensitivity (87.70%) and specificity (87.05%) were maintained in balance via Logistic Regression.