Computer versus physician identification of gastrointestinal alarm features

Area of Science:

• Gastroenterology
• Clinical Informatics
• Health Informatics

Background:

• Inquiring about "alarm features" is crucial for identifying patients at risk for organic gastrointestinal (GI) disease.
• Systematic collection of patient-reported symptoms and alarm features can aid in diagnostic workup.
• The Automated Evaluation of Gastrointestinal Symptoms (AEGIS) is a computer algorithm designed to collect GI symptoms and alarm features, translating them into a history of present illness (HPI).

Purpose of the Study:

• To compare the number of alarm features documented by physicians during usual care versus those collected by the AEGIS computer algorithm.
• To assess the efficacy of a computerized system in capturing critical patient-reported GI symptoms.

Main Methods:

• A cross-sectional study with a paired sample design was conducted in adult GI clinics.
• Participants underwent usual physician care followed by completion of the AEGIS system.
• Blinded physician reviewers enumerated positive alarm features from both physician-documented and AEGIS-generated HPIs.

Main Results:

• AEGIS identified significantly more patients with positive alarm features (53%) compared to physicians (27%) (p<.001).
• AEGIS documented a higher median number of positive alarms (1) than physicians (0) (p<.001).
• Physicians documented only 30% of the positive alarm features self-reported by patients via AEGIS.

Conclusions:

• Physicians under-report significant gastrointestinal alarm features compared to patient self-reporting via a computer algorithm.
• Computerized systems like AEGIS can serve as a valuable complement to standard HPIs, potentially enhancing clinical care and diagnostic accuracy.
• Utilizing technology to systematically collect alarm features may improve the identification of patients requiring further investigation for organic GI disease.