Return To Abstract Listing

Efficacy, Education, Administration, Informatics

E3168. Analyzing Delays in the ED: How to Optimize Processes in a Methodical Way

Spence S,  Sharma K,  Phan A,  Hasapes J. University of Houston, Houston, TX

Address correspondence to S. Spence (

Background Information: Our emergency department (ED) colleagues came to us with a proposal to improve wait times in our county ED by optimizing workflow through all services. Currently, the number of patients “left without being seen” was well above national average. Patients sometimes sat in the waiting rooms for hours, and improving wait times is considered a priority by the administration. The purpose of this presentation is to demonstrate how we logically established sources of delay and targeted intervention. The ED physicians initially thought that transport and queries from technicians about protocol issues were a major source of delay, while the radiologist had no concrete idea of what occurred before the scan was complete. Our aim was to meet the TAT goal as defined by the ED physicians. TAT was defined by our ED colleagues as the time from the order to the time of the first actionable report (preliminary or final). They wanted that TAT within 2 hours, 80% of the time.

Educational Goals/Teaching Points: Methods we used were to calculate the types and numbers of studies performed. We calculated which of the examinations go quickly and which take longer and mapped the 13 most common studies using Pareto graphs to break down the examination time into technician and radiologist components. We found that CT head and C spine are very common but performed very quickly (no intervention needed).CT abdomen pelvis (CT A/P) is the next most common, but taking enormous amounts of time to complete. Plotting the times against the TAT goals showed that we were not meeting our goal of a 2-hour TAT for 80% of cases on CT A/P cases and needed to target those cases. We broke down the workflow for CT A/P with time periods assigned to each component wherever data was available. Using this, we discovered that oral contrast was an enormous source of delay. Radiologists had not been aware of this, although it was a vague concern for our ordering clinicians. Prior attempts to alter the contrast protocol had been met with resistance from the radiologists; however, these results prompted a literature review and allowed creation of a new oral contrast protocol removing oral contrast for all but suspected bowel perforation and fistula.

Key Anatomic/Physiologic Issues and Imaging Findings/Techniques: Our results show median time to perform a CT of the abdomen and pelvis decreased from 160.5 mins to 116.6 minutes (difference of 43.9 minutes). These studies are ~32% of total volume and total time for all CT cases also decreased by 32.8 minutes. The percent of cases read within 2 hours increased from 57 to 63% for all CT cases, and from 33 to 52% in the CT abdomen and pelvis group. Adding the total time it takes all patients in 1 month to get a CT examination, it took 167.6 days in the preintervention group and 134.9 in the postintervention group. Therefore, now on average, our physicians are waiting ~ 32.7 fewer days per month to have these studies performed and read.

Conclusion: There are always more interventions to undertake. We already have more in progress based on these data. We look forward to continuing evidence-based interventions in this case, but wanted to demonstrate the power of a solid preintervention analysis in directing our energies in the right direction.