How well does decision support software perform in the emergency department?

Area of Science:

• Medical Informatics
• Clinical Decision Support
• Emergency Medicine

Background:

• Emergency departments (EDs) face unique challenges in diagnosis due to patient acuity and time constraints.
• General diagnostic decision support systems (DDSS) are increasingly available but their efficacy in ED settings requires evaluation.

Purpose of the Study:

• To assess the diagnostic performance of two general DDSS (Iliad and QMR) using data from an emergency department.
• To compare the accuracy of DDSS in the ED with their performance in other clinical environments.

Main Methods:

• A convenience sample of 25 ED patients with diagnostic questions was studied.
• Patient data, including audiotaped interactions, notes, labs, and EKGs, were abstracted.
• Data were input into Iliad and QMR; generated diagnoses were compared to final attending physician diagnoses.

Main Results:

• The correct ED diagnosis was present in Iliad's differential diagnosis 72% of the time and in QMR's 52% of the time.
• Correct diagnoses appeared in the top 10 for Iliad (51%) and QMR (44%), and top five for Iliad (36%) and QMR (32%).
• Performance in the ED was comparable to that reported in other clinical settings.

Conclusions:

• Diagnostic decision support software demonstrates consistent accuracy in the ED, similar to other clinical settings.
• Current DDSS accuracy is insufficient for independent diagnostic arbitration in individual cases.
• DDSS may serve as a valuable tool to prompt further investigation in challenging ED cases.