Background: Although cancer patients (CPs) are increasingly likely to visit emergency department (ED), no population-based study has compared the characteristics of CPs and non-cancer patients (NCPs) who visit the ED and examined factors associated with hospitalization via the ED. In this study, we (1) compared characteristics and diagnoses between CPs and NCPs who visited the ED in a cancer center or general hospital; (2) compared characteristics and diagnoses between CPs and NCPs who were hospitalized via the ED in a cancer center or general hospital; and (3) investigated important factors associated with such hospitalization.
Methods and findings: We analyzed patient characteristic and diagnosis [based on International Classification of Diseases-9 (ICD-9) codes] data from the ED of a comprehensive cancer center (MDACC), 24 general EDs in Harris County, Texas (HCED), and the National Hospital Ambulatory Medical Care Survey (NHAMCS) from 1/1/2007-12/31/2009. Approximately 3.4 million ED visits were analyzed: 47,245, 3,248,973, and 104,566 visits for MDACC, HCED, and NHAMCS, respectively, of which 44,143 (93.4%), 44,583 (1.4%), and 632 (0.6%) were CP visits. CPs were older than NCPs and stayed longer in EDs. Lung, gastrointestinal (excluding colorectal), and genitourinary (excluding prostate) cancers were the three most common diagnoses related to ED visits at general EDs. CPs visiting MDACC were more likely than CPs visiting HCED to be privately insured. CPs were more likely than NCPs to be hospitalized. Pneumonia and influenza, fluid and electrolyte disorders, and fever were important predictive factors for CP hospitalization; coronary artery disease, cerebrovascular disease, and heart failure were important factors for NCP hospitalization.
Conclusions: CPs consumed more ED resources than NCPs and had a higher hospitalization rate. Given the differences in characteristics and diagnoses between CPs and NCPs, ED physicians must pay special attention to CPs and be familiar with their unique set of oncologic emergencies.