A field test of time-based emergency department quality measures.

STUDY OBJECTIVE We examine practical aspects of collecting time-based emergency department (ED) performance measures. METHODS Seven measures were implemented in 6 hospitals during 1 year. Structured interviews were used to assess the benefits and burdens of reporting. In 2 hospitals, Centers for Medicare & Medicaid Services (CMS) sample size requirements for 3 measures were compared to a reasonable sample size estimate (in which 95% of samples fell within 15 minutes of the population median). RESULTS ED performance data on 29,587 admitted patients and 127,467 discharged patients were reported. Median throughput time for admitted patients ranged from 327 to 663 minutes and for discharged patients ranged from 143 to 311 minutes. Other performance measures varied similarly (2- to 3-fold between hospitals). In general, ED throughput was longer at academic sites and those with higher volume. Several benefits of reporting were identified, including promoting ED quality improvement, accountability, and practice standardization. The burdens included having to access multiple information technology systems and difficulties setting up the data collection. Most respondents found great value in the throughput measures and time to pain medication but less value in time to chest radiograph. The human capital required to implement measures varied by hospital and staff demonstrated a learning curve. Our empirically derived minimum reliable sample sizes were different from CMS recommendations. CONCLUSION There is great variation in performance between EDs in time-based ED measures. There are multiple reporting benefits. Reporting burdens seemed to lessen after data systems were established. The CMS sample size requirements for throughput measures may not be optimal compared with actual ED throughput data.