193 Emergency Department Patients With Cancer: How Do Outcomes at National Cancer Institute-Designated Hospitals Compare to the General Population?

Abstract

There is a growing recognition of the unique challenges encountered while providing emergency care to cancer patients. We have recently presented data describing why these patients present to the ED and have explored patient characteristics that are most likely to require a repeat ED visit within 7 days. Here, we designed a study to investigate if there is a difference in outcomes when patients are seen at an emergency department affiliated with a National Cancer Institute (NCI)-designated hospital versus the general community. We performed a retrospective cohort study to investigate cancer-related ED visits from all licensed non-military acute care hospitals in California using a non-public statewide database. We queried cancer-related ED visits for patients ≥18 years of age during the 2016 calendar year and recorded ED admission and discharge diagnoses, cancer types, and medical comorbidities. Bounce-back visits were defined as ED return visits within 7 days of a preceding ED discharge. Visits were categorized based on whether the ED was in the general community, or affiliated with 1 of 8 NCI-designated cancer centers in California (City of Hope, UC Irvine, UCLA, UCSD, USC, UCSF, Stanford, or UC Davis). Cancer types were defined and categorized using primary and secondary ICD-10 codes as designated by the NCI. In 2016 there were 103,523 cancer-related ED visits that met our inclusion criteria. At NCI-designated hospital EDs (NCI EDs), 4,297 patients comprised 5,501 index visits. Interestingly, the 7-day bounce-back rate at NCI EDs was slightly higher than at general community EDs (GC EDs), 18.1% versus 17.6%. There was no significant difference in the types of cancers that were seen at GC EDs versus NCI EDs. The most common diagnoses among all ED bounce-backs were septicemia, abdominal symptoms, and other cancer-related pain. One stark difference was the rate of septicemia, which was present in 6.1% of GC ED revisits versus 3.8% of NCI ED revisits. The rate of admission upon bounce-back to a GC ED was 36.6% b 40.3% at an NCI ED, while the rate of mortality during that admission was higher for GC EDs (8.7% versus 6.5%). Interestingly, the Charlson Comorbity Index Score, which accounts for cardiopulmonary disease, diabetes, and liver disease in addition to cancer, was significantly higher among patients at GC EDs compared to NCI EDs (CMI Score of 3+ in 53.3% versus 42.8%). These data highlight interesting differences in the characteristics of bounce-backs among cancer patients seen at NCI EDs versus GC EDs. The slightly higher rate of bounce-backs at NCI EDs may be due to a combination of treating more aggressive cancers and being more comfortable with discharging potentially sick patients. The fact that sepsis was less common on repeat visits at NCI EDs perhaps suggests that these EDs are more proficient at identifying sepsis on the index visit. Although the rate of admission upon bounce-back was higher at NCI EDs, they had a lower mortality during that admission, suggesting either a higher level of care provided at those NCI hospitals, or that those sicker patients were more appropriately identified during the index visit. We hope that by further studying these differences in outcomes at NCI EDs, we can identify appropriate strategies for treating these cancer patients in the general community on their index visit.