Machine Learning for Depression Severity and Suicide Risk Assessment Across Generations: A Cross Sectional Study
Year: 2026
Type: article
Study design: Observational Study
Abstract: ABSTRACT Background: Demographic, psychological and social factors affect the symptoms of depression and risk of suicide. Structured questionnaire-based assessment can be used to support assessment and machine learning can be used to investigate the multivariable patterns between age groups. Any model output should be used in addition to but not in lieu of, clinical assessment. Objective: Questionnaires-based psychological measures in Generation Z, Millennials and Generation X will be different and regression analysis of total scores on the Suicide Behaviors Questionnaire Revised (SBQ-R) will be different. Methods: This cross-sectional observation study has 500 participants between the ages of 13-59. Out of 175 reported Generation Z participants, 200 reported Millennials, and 125 reported Generation X participants. Data were analyzed by using a demographic and the scores on the following scales: PHQ-9 total, the verified SSQ-6 total, and BSMAS total, which were used to estimate the SBQ-R total scores. The ten regression algorithms were compared based on the following criteria: MAE, MSE, RMSE, R², cross-validation R² and training time. A revised pipeline that deals with leakage was used for the final comparison supplied. Results: The Stacking Regressor model proved to be the best in this dataset and using the above mentioned evaluation procedure to estimate SBQ-R questionnaire scores. The results are indicative of prediction, and not diagnostic or clinically useful. The model should be transparent in reporting the architecture and the way the model was validated, and then should be independently validated by an external group and validated in the clinical setting. Conclusion: The results of this study is a combination of questionnaire-based measures with comparative machine learning methods which examine symptoms related to depression and scores of the questionnaires in relation to suicide across generational groups. The system suggested is a prototype of research only and needs to be validated by external validation and clinical governance for its use. Keywords: depression severity; suicide-related risk; machine learning; PHQ-9; SSQ-6; BSMAS; SBQ-R; generational differences; Stacking Regressor,clinical decision support
Institutions: GLS University, B.J. Medical College
Language: English
Subfield: Health Information Management
Field: Health Professions
Domain: Health Sciences
Open Access status: diamond
OpenAlex ID: W7225375907