Time Series Analysis of Variables Associated With Daily Mean Emergency Department Length of Stay
Study objective
We measure the effect of various input, throughput, and output factors on daily emergency department (ED) mean length of stay per patient (daily mean length of stay).
Methods
The study was a retrospective review of 93,274 ED visits between April 15, 2002, and December 31, 2003. The association between the daily mean length of stay and the independent variables was assessed with autoregressive moving average time series analysis (ARIMA). The following independent variables were measured per 24-hour period: number of elective surgical admissions, ED volume, number of ED admissions, number of ED ICU admissions, number of ED clinical attending hours, hospital medical-surgical occupancy (hospital occupancy), and day of the week.
Results
Three factors were independently associated with daily mean length of stay in time series analysis: number of elective surgical admissions, number of ED admissions, and hospital occupancy. The daily mean length of stay increased by 0.21 minutes for every additional elective surgical admission, 2.2 minutes for every additional admission, and 4.1 minutes for every 5% increase in hospital occupancy. Elective surgical admissions were associated with a maximum of 35 hours of additional ED dwell time. The model accounted for 31.5% of the variability in daily mean length of stay. The final model parameters for the ARIMA analysis were autoregressive term (1) moving average (1).
Conclusion
Hospital occupancy and the number of ED admissions are associated with daily mean length of stay. Every additional elective surgical admission prolonged the daily mean length of stay by 0.21 minutes per ED patient. Autocorrelation exists between the daily mean length of stay of the current day and the previous day.
To access this article, please choose from the options below
Supervising editor: Brent R. Asplin, MD, MPH
Author contributions: NKR, JC, JO, SDM, SC, BM, LF, and EL conceived the study and designed the trial. NKR, JC, JO, KD, and EL obtained the research funding. DO, SDM, TR, and AC provided statistical advice and analyzed the data. NKR, SC, BM, KD, RS, KL, JMB, LF, LG, and AC supervised the conduct of the trial and data collection. JC, DO, TR, RS, KL, JMB, and LG managed the data for analysis. All authors contributed substantively to the revision of the article. NKR drafted the article. NKR takes responsibility for the paper as a whole.
Funding and support: This study was funded by the “Urgent Matters” Technical Assistance Grant (#048555) and Demonstration Grant (#048556) from the Robert Wood Johnson Foundation.
Reprints not available from the authors.
PII: S0196-0644(06)02528-5
doi:10.1016/j.annemergmed.2006.11.007
© 2006 American College of Emergency Physicians. Published by Elsevier Inc. All rights reserved.
