High dimensional predictions of suicide risk in 4.2 million US Veterans using ensemble transfer learning

• Published in: Scientific reports Vol. 14; no. 1; p. 1793
• Main Authors: Dhaubhadel, Sayera, Ganguly, Kumkum, Ribeiro, Ruy M., Cohn, Judith D., Hyman, James M., Hengartner, Nicolas W., Kolade, Beauty, Singley, Anna, Bhattacharya, Tanmoy, Finley, Patrick, Levin, Drew, Thelen, Haedi, Cho, Kelly, Costa, Lauren, Ho, Yuk-Lam, Justice, Amy C., Pestian, John, Santel, Daniel, Zamora-Resendiz, Rafael, Crivelli, Silvia, Tamang, Suzanne, Martins, Susana, Trafton, Jodie, Oslin, David W., Beckham, Jean C., Kimbrel, Nathan A., McMahon, Benjamin H.
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 20-01-2024; Nature Publishing Group; Nature Portfolio
• Subjects: 631/114/1305; 692/308/409; Carcinoma, Renal Cell; Cross-Sectional Studies; Electronic medical records; Humanities and Social Sciences; Humans; Hypoxia; Kidney Neoplasms; Machine Learning; multidisciplinary; outcomes research; Overdose; Predictions; Prospective Studies; Retrospective Studies; Science; Science (multidisciplinary); Statistical analysis; Substance use; Suicide; Suicide, Attempted; Suicides & suicide attempts; Transfer learning; Veterans; Veterans - psychology
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-024-51762-9

We present an ensemble transfer learning method to predict suicide from Veterans Affairs (VA) electronic medical records (EMR). A diverse set of base models was trained to predict a binary outcome constructed from reported suicide, suicide attempt, and overdose diagnoses with varying choices of study design and prediction methodology. Each model used twenty cross-sectional and 190 longitudinal variables observed in eight time intervals covering 7.5 years prior to the time of prediction. Ensembles of seven base models were created and fine-tuned with ten variables expected to change with study design and outcome definition in order to predict suicide and combined outcome in a prospective cohort. The ensemble models achieved c-statistics of 0.73 on 2-year suicide risk and 0.83 on the combined outcome when predicting on a prospective cohort of
∼
 4.2 M veterans. The ensembles rely on nonlinear base models trained using a matched retrospective nested case-control (Rcc) study cohort and show good calibration across a diversity of subgroups, including risk strata, age, sex, race, and level of healthcare utilization. In addition, a linear Rcc base model provided a rich set of biological predictors, including indicators of suicide, substance use disorder, mental health diagnoses and treatments, hypoxia and vascular damage, and demographics.