Predicting suicide risk in patients with digestive system tumors: A retrospective cohort study

Suicide in patients with digestive system tumors has been a concern, yet relevant studies remain limited. This study aimed to identify factors influencing suicide in patients with digestive system tumors using a large sample size from public databases and to develop a clinically applicable risk prediction model, thereby providing a reference for clinical interventions.

Data for 173,804 patients diagnosed with digestive system tumors between 1998 and 2015 were obtained from the Surveillance, Epidemiology, and End Results database. The standard mortality rate of suicide among digestive system tumor patients was compared with that of the general US population. Predictors of suicide in patients with digestive system tumors were identified using lasso regression and logistic regression, and subsequently visualized with a nomogram. Receiver operating characteristic curve analysis was used to determine the predictive accuracy of the nomogram. A calibration curve was plotted to assess the concordance between predicted and observed probabilities. Additionally, decision curve analysis and clinical influence curves were used to evaluate the clinical utility of the nomogram.

A total of 131,354 patients from the Surveillance, Epidemiology, and End Results database were used to develop the model; 652 (0.5%) of those patients died by suicide. The cohort was followed for a cumulative duration of 1,277,281.75 person-years. Age, sex, tumor grade, staging, surgical intervention, chemotherapy, marital status, and place of residence were identified as independent predictors of suicide in patients with digestive system tumors. The nomogram constructed based on these predictors demonstrated high predictive accuracy, with an area under the receiver operating characteristic curve of 0.78. Additionally, the calibration curve indicated a strong concordance between the predicted and observed probabilities. The decision curve analysis and clinical influence curves further validated the clinical applicability of the nomogram.

We identified factors influencing suicide in patients with digestive system tumors and developed a robust prediction model to guide decision making.