A Population-Based Assessment of Emergency Department Observation Status for Older Adults With Cancer

Background: Hospitals' use of observation status for patients with cancer presenting to the emergency department (ED) is not well understood. This model of care delivery may be a viable alternative to inpatient admission for patients with cancer presenting with certain conditions. Our objective was to assess the use of observation status among Medicare beneficiaries with and without cancer. Methods: Population-based SEER-Medicare data were used to assess differences in the use of observation status between Medicare beneficiaries aged ≥66 years with and without cancer using a matched analysis (n=151,183 per cohort). We assessed the ratio of observation unit use to inpatient admission, between cancer and noncancer cohorts, and for patients diagnosed with breast, colon, lung, and prostate cancers. Poisson regression models were used to calculate observation rate estimates and 95% CIs while adjusting for selected patient characteristics. Results: When considering the volume of hospitalizations, observation status is used less frequently among beneficiaries with cancer than those without (43 vs 69 observation status visits per 1,000 inpatient admissions, respectively). The estimated observation rate per 1,000 inpatient admissions was higher for beneficiaries aged <75 years versus those aged ≥75 years, those with a Charlson comorbidity index of 0 vs 1 or ≥2, and those without a prior hospitalization versus those with ≥1 prior hospitalizations. Patients with breast and prostate cancers had higher adjusted and unadjusted observation rates per 1,000 inpatient admissions compared with those with colon and lung cancers. Conclusions: Observation status is used proportionately less for beneficiaries with cancer than those without. There may be opportunities to develop standards for ED staff to manage certain conditions for patients with cancer in observation status, and to reserve hospital resources for those who need it most.