UCLA

Introduction: Emergency Department (ED) is an important component of the US healthcare system but lacks appropriate usage. To incentivize high-value care in ED, transparency of ED utilization is critical as it offers comparability across providers.Data Source & Study Population: Data sources include the 2018 All-Payer Claims Database in Colorado and American Hospital Association Annual Survey. Study population contains eligible beneficiaries who have received 30-day episodes of care for condition-specific cohorts in 2018.

Methods: We construct the 30-day window and standardize price to estimate ED-level, unadjusted variation in episode cost and hospitalization. We design risk-adjustment model and measures of expected-to-predicted ratios to capture ED-level variation in cost and hospitalization, which reflects ED practice differences. We conduct an improved signal-to-noise analysis for measures’ reliability assessment. Descriptive and factor analyses examine ED’s pattern consistency and correlation among risk-adjusted measures. Lastly, we evaluated the effect of systemic factors on risk-adjusted measures using regression analysis.Results: Unadjusted variations for ED-level episode cost and hospitalization range from $675 to $4,589 and 1.85% to 32.54% across condition-specific cohorts, respectively. Adjusted variations in RAPRs and RAARs range from 11% to 18%, and 15% to 48% respectively. Risk adjustment models explain 20% variations on average in episode cost and hospitalization. Average signal-to-noise ratio of episode cost and hospitalization surpassed 0.7, indicating good reliability. 43 out of 55 EDs exhibit coherent patterns of care of ED utilization, 12 EDs demonstrate mixed patterns, and the rest do not show explicit patterns. Higher variation in episode cost is associated with freestanding EDs, urban location. Higher hospitalizations are observed for not-for-profit hospital-based EDs, minor teaching responsibilities, and urban locations. Conclusion: Sizeable ED-level variations in episode cost and hospitalization are captured. Patient and systemic factors are partial contributors to these variations. Coherent patterns found in majority of EDs and strong correlation in measures illustrate that EDs inclined to maintain similar levels of cost or hospitalization across conditions. This study expands the design of the risk-adjustment measures to all-payer beneficiaries, facilitates the profiling of hospital value in ED care, and provides information enabling hospitals to optimize and coordinate care.