Risk of mortality in adult cancer febrile neutropenia patients with a machine learning approach.

e13562Background: Febrile neutropenia (FN) is a major dose-limiting toxicity of myelosuppressive therapy in cancer patients and impose substantial morbidity and mortality risks. We analyzed National Inpatient Sample (NIS) database to understand the trend in FN and predict mortality in cancer patients who experienced FN using a machine learning (ML) approach. Methods: We used the Healthcare Cost and Utilization Project (HCUP) Nationwide Inpatient Sample (NIS) data. Cancer discharges were identified using the HCUP Clinical Classifications Software (CCS) categorizations (11-45), within which FN patients were identified with International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes (ICD-9-CM: 7806 and 288, 2841 or 1125). We examined trends in FN discharges and mortality rates using during 2003 – 2015, and then built various models on NIS 2013 data to predict mortality. Results: The percentage of discharges secondary to cancer-related FN kept increasing from 0.2% in 2003 ...