Predicting the risk of subsequent suicide among multiple myeloma patients in US: A population-based cohort study

Abstract

Suicide risk is higher in patients with multiple myeloma (MM) than in the general population. We used LASSO and Cox regression to identify factors associated with suicide in MM and to build a predictive nomogram. Sex, race, marital status, radiotherapy, year of diagnosis, and Rural-Urban Continuum Codes were independent predictors. Female sex, Black race, receipt of radiotherapy, more recent diagnosis (2010–2014, 2015–2020), and living in a metropolitan area were associated with lower suicide risk, whereas being single increased risk. Model performance was evaluated using Kaplan–Meier curves, calibration plots, decision curve analysis, internal validation, and a competing-risk model, all of which showed good discrimination, calibration, and robustness. The resulting nomogram provides an accurate and clinically useful tool to identify MM patients at elevated risk of suicide and may help guide targeted psychological and supportive interventions.