Costs and outcomes of observation units in triaging non-specific chest pain patients after their Emergency Room visits

Background: Recent studies cast doubt about the economic efficiencies of observation units. Objectives: We aimed to re-examine the cost savings and outcomes of observation units compared to inpatient. Methods: We retrospectively examined the costs of 15,851 patient visits for non-specific chest pain in Emergency Room (ER) and further classified them into inpatient and observation admission. Cost was calculated using the contractual allowed amount between BlueCross BlueShield of Texas (BCBS-TX) and providers. We also compared the utilization of advanced imaging procedures such as myocardial perfusion imaging (MPI), Coronary Computed Tomography Angiography (CCTA) and Computed Tomography (CT) chest scans between the two groups. Quantile regression was used to calculate the median adjusted cost difference between the two groups and the 95% confidence intervals. One to one propensity matching was further used to test the robustness of our findings. To test the differences in patients\u27 outcomes, we limited our sample to those who had one year continuous enrollment after their initial ER visit. Outcomes included the one year costs related to chest pain/cardiovascular diseases, primary and secondary outcomes. Primary outcomes included myocardial infarction, congestive heart failure, stroke and cardiac arrest while secondary outcomes included one year ER revisits for angina pectoris or chest pain, hospitalization due to circulatory disorders and revascularization procedures. Differences in the costs, primary and secondary outcomes over time between the two groups were measured using a Cox proportional hazard model and quantile regression. A propensity score matching was also used to validate the study findings. Results: Over thirty-seven percent (37.3% (n = 5,890)) of the sample was admitted to inpatient care vs. 62.7% (n= 9,961) in observation units. Patients admitted to the inpatient units stayed longer (1.5 days for inpatient vs. 21 hours for observation units), had higher comorbidities, and incurred higher costs. More than one third of patients admitted in observation units (37.5%) stayed longer than 24 hours. Inpatient admission had a higher per-patient cost of $1,308 (95% CI, $1,206 - $1,411) compared to observation units. Patients in observation units also received more MPI (35.8%) and CT scans (13.2%) compared to 31.5% and 10.4% of patients admitted to inpatient units. When comparing the outcomes between the two groups, only 7,549 patients had one year continuous enrollment following the initial ER visits in which 65.7% of the sample (n = 4,962) were admitted to observation units versus 34.3%(n= 2,587) of patients who were admitted to inpatient units wards. After adjusting for baseline characteristics, no differences were detected in the one year chest pain/cardiovascular diseases related costs, primary or secondary outcomes between the two groups except for the revascularization rates. Patients admitted to observation units were 77% (95%CI = 1.23-2.55) more likely to have revascularization procedures compared to those admitted to inpatient care. Conclusion: Observation units are still less costly compared to inpatient admission despite the increased proportion of patients exceeding the 24 hours length of stay that was initially allowed and the increased utilization of advanced imaging procedures. Observation units have similar outcomes compared to inpatient admission indicating that observation units provide comparable level of care compared to inpatient admission.