Welch, Shari J. MD
Any physician, nurse, or technician can attest that the ED is not the same place at 1 a.m. as it is at 1 p.m. Yet most departments lack the data to quantify those differences.
As Todd Taylor, MD, observed, “The chances of surviving a heart attack now depend more on the time of day, day of week, and perhaps the type of insurance than any other factors.” (Personal communication, Aug. 15, 2005.) His comment was prophetic. A Journal of the American Medical Association study revealed that timeliness and outcomes are worse in patients presenting with acute myocardial infarction during off-hours. (2005;294:803.)
Armed with data, EDs can provide the services necessary to deliver quality care
This is not news to practitioners on the front lines. Every practitioner can provide testimony and anecdotes of patients whose care suffered because of services lacking on nights, weekends, and holidays. Whether it is difficulty in contacting a specialist or the inability to get certain diagnostic tests at night, we all have been frustrated by a system that mandates 24/7 emergency care but fails to provide the tools to deliver it.
This mismatch is amplified as the burdens on the emergency system are increased. Between 1992 and 2002, ED utilization increased by 20 percent, while the number of EDs decreased by 15 percent. (National Hospital Ambulatory Care Survey, 2001 ED Summary, CDC, 2003.) In an April 2002 national survey, 62 percent of all U.S. hospitals reported being at or over operating capacity, with 79 percent of urban EDs and a whopping 87 percent of Level I trauma centers in this predicament. (“Emergency Department Overload: A Growing Crisis.” Report prepared by Lewin Group for American Hospital Association, April 2002.)
Staff shortages, shrinking inpatient hospital capacity, and shortages of on-call specialists confound crowding. (“The U.S. Health Care Safety Net and Emergency Department Crowding,” Urgent Matters E-Newsletter, Aug. 21, 2005, www.urgentmatters.org/about/um_safety_net.htm.)
Homegrown Information System
Figure. Average Numb...Image Tools
At LDS Hospital in Salt Lake City, the flagship hospital for Intermountain Health Care, we are mining data to quantify what occurs hour by hour. IHC has a long commitment to quality in medicine, with Brent James, MD, pioneering some of the earliest writings on the subject. IHC has an almost 30-year investment in information technology to facilitate care. We are currently beta-testing a homegrown information system that integrates the information that affects ED patients. Databases from registration, triage, radiology, and other EDs within our system are integrated. Data are passively transferred from one database to another without manual entry. All data enters a repository or “ED Data Mart” for later data mining.
The charts show early data runs on the differences in the 24-hour ED cycle. These data were mined from the LDS Hospital ED Data Mart using 44,000 ED encounters between July 1, 2004, and June 30, 2005. Though they are institutionally specific, they may reflect similar patterns seen in U.S. EDs and others are recognizing the same arrival pattern curve. (Eitel D. “Workflow Analysis and Service Management: How is Your ED Working?” Presented at Urgent Matters conference, Las Vegas, Oct. 27, 2005.)
Figure. Average Acui...Image Tools
Census data show that arrival to the ED by hour is not constant. Peak utilization begins in late morning and remains steady until 10 p.m. The overall ED census lags slightly, reaching its peak at 3 p.m. and staying elevated until midnight.
Using a five-level triage acuity system, lower numbers equate with higher acuity. Acuity is highest between 1 a.m. and 5 a.m. and lowest from 7 a.m. to 1 p.m. There is another lower acuity period in the evening. Oddly, acuity varies inversely with the census, though it is not statistically significant. Yet isn't this the experience of physicians who prefer to work the graveyard shift? The census is lower, but patients are often more acute. Doctors in our group who favor nights do more intubations than anyone else, perhaps another indirect measure of acuity.
Radiology utilization almost directly correlates with the ED arrival curve. As the afternoon surge arrives, so does the need for radiology services. In fact, there are more x-rays ordered between noon and 8 p.m. than during the other 16 hours of the day! Is the radiology staff increased threefold to meet this demand?
Although the lab utilization curve also shows a clear correlation with the arrival curve, a subtlety was noted when the number of labs per patient arrival was calculated. That number is much higher in early morning and correlates with acuity. The average number of labs ordered between noon and 8 p.m. is higher than those ordered during the other 16 hours of the day. Again we ask, does lab staffing correspond to increased utilization?
Overall turnaround time correlates best with acuity: the higher the acuity, the longer the turnaround time. There was surprisingly no correlation with census.
Data for the Future
Figure. LDS Hospital...Image Tools
In clinical medicine, practitioners have begun to embrace information technology as the means for improving patient care, and emergency physicians see opportunities for improvement in ED operations and patient flow. (Acad Emerg Med 2004;11:1206; Acad Emerg Med 2004;11:1142; J Healthc Inf Manag 2001;15:155; and Manag Care Interface 2000;13:68.)
Any emergency practitioner can attest, the ED cycles and changes throughout the day. These crude data from one institution may be the first effort to use such technology to quantify the differences manifest in the 24-hour ED cycle.
Census and acuity vary over the day as do the needs of patients. From late morning until mid-evening, the flux of patients is almost three times the rate seen in the early morning. This arrival-by-time curve has been recognized across hospital cohort groups. (Augustine J. “ED Strategic Planning and Process Redesign” presented at ED Benchmarks conference, March 4–6, 2004, Orlando.) Even as arrivals fall after 9 p.m., the ED is still quite full.
This influx means more x-rays and lab tests. The increase in utilization of these services has a nearly 1:1 correlation with the curve of increase in arrivals by hour. Yet most hospitals do not triple staffing for late afternoon and early evening. In fact, scheduling for most departments falls off after 5 p.m., when need is highest. Many hospitals (including ours) fall short of the 58-minute benchmark suggested by VHA for turning around an x-ray. Though this may be tolerated on other shifts, it creates bottlenecks in the evening. Though the census falls during the after-midnight portion of the cycle, these patients utilize one-and-a-half times the radiology and laboratory services.
Patients seen during the after-midnight period have longer turnaround times, and this correlates nicely with acuity. The early-morning hours bring in the acuity, and the afternoon and early evening hours bring in the volume!
Data with this degree of granularity are new to EPs, and come from utilizing information technology to the fullest. With most EDs now seeing an average of 40,000 patients a year and generating physician fees of roughly $3 million, doesn't it seem odd that we have tried to practice medicine without it? Can you think of any business with as many customers and as much revenue that doesn't have these data at its fingertips?
The quest for these types of data, to define what goes on in our EDs in a 24-hour cycle and determine what we need to take care of our patients, should be a priority in the latter half of this decade. Armed with these data, we can provide the services necessary to deliver quality care.
© 2006 Lippincott Williams & Wilkins, Inc.