BACKGROUND
Background: The low incidence of suicide in oncology patients often leads clinicians to not take it seriously. However, conducting observational studies on these patients is extremely challenging. To overcome this, it is crucial to digitally analyze these patients using data from large public platforms.
OBJECTIVE
Objective: The aim of this study was to identify the factors influencing suicide in patients with digestive system tumors (DST) and construct a clinically applicable risk prediction model.
METHODS
Methods: The data of 173,804 patients diagnosed with DST between 1998 and 2015 was retrieved from the Surveillance, Epidemiology and End Results (SEER) database. The standard mortality rate (SMR) of suicide among the DST patients was compared to that of the general population of USA. The predictors of suicide in DST patients were identified by Lasso regression and logistic regression, and visualized using a nomogram. Receiver operating characteristic (ROC) curves and consistency-index were used to determine the accuracy and discriminative ability of the nomogram. Calibration curve was plotted to evaluate the agreement between the predicted and actual probabilities. Furthermore, decision curve analysis (DCA) and clinical influence curve were used to assess the clinical utility of the nomogram.
RESULTS
Results: Age, sex, grade, stage, surgery, chemotherapy, marital status, and residence were identified as independent predictors of suicide in patients with DST. The nomogram based on these independent predictors showed high predictive accuracy, with area under the ROC curve of 0.777. Furthermore, the calibration curve showed that the predicted probability of the model was in high agreement with the actual probability. The DCA and clinical influence curve showed that the model can be applied in the clinical setting.
CONCLUSIONS
Conclusion: We identified factors affecting suicide in DST patients and established a reliable prediction model for suicide to guide clinical strategy.