NotesFAQContact Us
Search Tips
ERIC Number: ED547625
Record Type: Non-Journal
Publication Date: 2012
Pages: 84
Abstractor: As Provided
Reference Count: N/A
ISBN: 978-1-2674-9511-2
Improving Emergency Department Triage Classification with Computerized Clinical Decision Support at a Pediatric Hospital
Kunisch, Joseph Martin
ProQuest LLC, Ph.D. Dissertation, University of Colorado at Denver
Background: The Emergency Severity Index (ESI) is an emergency department (ED) triage classification system based on estimated patient-specific resource utilization. Rules for a computerized clinical decision support (CDS) system based on a patient's chief complaint were developed and tested using a stochastic model for predicting ESI scores. The CDS was implemented within the electronic health record (EHR) system used at a pediatric emergency department (ED). Methods: Retrospective data on chief complaints and resource use from 127,081 ED visits were examined to create the CDS rules for the electronic health record based on published ESI criteria. The performance of the CDS rules was compared to the triage staff's performance for accuracy using an independent test dataset consisting of 53,041 ED visits and stochastic simulation. A prospective study was done with the ESI probabilities displayed in a CDS to the triage staff. A six month study period utilizing an on-off evaluation design was used to identify time trends for increased accuracy. Results: The retrospective evaluation data set showed the triage staff accurately predicted the ED patients' ESI classification level based on resource use only 43.5% of the time. The stochastic model using the same retrospective data predicted the correct ESI classification level 71.6% of the time. The prospective arm of the study showed no change in accuracy of the ESI classification over the six-month study period ([x-bar] =40.5%). A change in accuracy was not statistically significant across the 6 month study period comparing the on-off periods (p = 0.9). In the prospective period, the stochastic model continued to perform a more accurate overall prediction of resource use ([x-bar] =78.6%). Survey responses indicated the staff performing triage did not use the CDS information. Conclusions: The probability model built using retrospective data demonstrated improvement in ESI classification using a stochastic model. The application of the probability model in an EHR-based CDS did not improve the triage staffs accuracy of ESI classification. Indifference of the staff towards the CDS may be one reason the accuracy did not change over time. Future studies to determine more effective methods of using the CDS probability model are recommended. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page:]
ProQuest LLC. 789 East Eisenhower Parkway, P.O. Box 1346, Ann Arbor, MI 48106. Tel: 800-521-0600; Web site:
Publication Type: Dissertations/Theses - Doctoral Dissertations
Education Level: N/A
Audience: N/A
Language: English
Sponsor: N/A
Authoring Institution: N/A