Recommendations for the Selection and Implementation of an Emergency Department Information System
July 4, 2006
Paul Postuma, BSc, MD, CCFP(EM), FCFP
Emergency medicine is a potential minefield for automation. Emergency Department Information Systems (EDIS) can aid or impair the delivery of patient care, can save lives or end them.
Health Care Information Systems are frequently compared with air traffic control (ATC): if such a critical industry can cut wait times with fewer staff, improve customer satisfaction, and above all realize significant gains in reliability and safety, then surely something as simple as a small urban Emergency Department can do the same.
The analogy soon falls apart:
"ATC is a well-bounded domain that follows prescribed procedures. When circumstances become marginal and overtly threaten safety, it is acceptable (and expected) that operations are suspended. The automated systems that support ATC are relatively well suited to such well-bounded problems. Even in this well-bounded setting, automation has not proven to be a panacea. Controllers frequently rely on paper artifacts, and an ambitious advanced automation system effort to upgrade the U.S. ATC computer system in the mid-1980s was 'sunk by unrealistic specifications and human factors difficulties among other problems.'
In contrast with ATC, the ED setting is highly variable, evanescent, contingent, uncertain, poorly bounded, resource constrained, and beholden to many external influences. For example, patients just "show up." In order to reconcile care resources with patient care demands, ED teams routinely engage in fluid, dynamic cognitive activities that require flexible, reliable artifacts to support them. They seek, track, digest, calibrate, probe, evaluate verify, and share information. They also plan, speculate, replan, and make trade-off decisions. No circumstances make it acceptable to suspend patient care. Because of this, practitioner teams have developed a range of sophisticated strategies to manage the balance between work and demand. In order to succeed, automation needs to be a team player in the ED setting."1
Large-scale failures abound. Adapting third-party EDIS solutions to a specific environment involves a large investment of time and money; ongoing costs often discourage the customization and updating required to make a system work. End-users are often far removed from the implementation, and as such, systems may not mirror or enhance work-flow, but frequently complicate it. Automated systems that work on smaller scales may buckle under the demands of a full-sized implementation handling hundreds of thousands of events, and thousands of data fields - some of which individually containing multiple megabytes of data - per event.
Automated systems are "brittle." They may decrease the frequency of small failures - but "when systems do fail, they do not degrade gracefully. Instead, they fail catastrophically and without warning."1
Additionally, patient presentations are infinitely varied, problems are often subtle, and difficult situations may force the adoption of less-than-ideal compromises. EDIS implementations frequently ignore these subtleties. Regimented structures can run counter-grain to the needs of a "live" department, especially under strain. Technology "can give the illusion of helping practitioners while at the same time creating a new layer of burdens and complexities."1 "Clumsy" automation creates extra steps: it wastes caregiver time and creates frustration, impedes patient care, and can endanger the patient.
Society of Academic Emergency Medicine Consensus Conference on Emergency Medicine Information Technology
The primary purpose of a clinical information system is straightforward: show all relevant information. Show it everywhere. Make it clear. Make it concise. Make it fast. Make it easy.2
Fortunately, the essential components of an effective EDIS have been addressed by the recent consensus conference on this subject,3 reported in a special issue of Academic Emergency Medicine: Informatics and Technology in Emergency Care (Volume 11, issue 11, available free of charge at http://www.aemj.org/content/vol11/issue11/).
This document follows their recommendations in basic outline (a few have been omitted, and the order rearranged, to better reflect our circumstances). Detailed requirements have been organized and itemized in the form of a Request for Proposal - see Emergency Department Information System: Request for Proposal.
Full article, PDF format
 Nemeth C, O'Connor M, et al. Crafting Information Technology Solutions, Not Experiments, for the Emergency Department. Acad. Emerg. Med. 2004;11(11):1114-7
 Feied CF, Handler JA, et al. Clinical Information Systems: Instant Ubiquitous Clinical Data for Error Reduction and Improved Clinical Outcomes. Acad. Emerg. Med. 2004;11(11):1162-9
 Gillam M, Rothenhaus T, et al. Information Technology Principles for Management, Reporting, and Research. Acad. Emerg. Med. 2004;11(11):1155-61