Computer-Aided Colonoscopy Alert Fatigue and Its Effect on Adenoma Detection

Area of Science:

• Gastroenterology
• Medical Imaging
• Endoscopy

Background:

• Computer-aided detection (CADe) systems aim to enhance adenoma detection rates (ADR) during colonoscopies.
• Potential alert fatigue from continuous CADe use may decrease its efficacy over time.
• Understanding the impact of alert fatigue on ADR is crucial for optimizing CADe implementation.

Purpose of the Study:

• To investigate whether alert fatigue occurs with CADe use during colonoscopies.
• To determine how alert fatigue influences ADR throughout an endoscopist's daily schedule.
• To evaluate the real-world effectiveness of CADe in a pragmatic setting.

Main Methods:

• Secondary analysis of a prospectively collected cohort of 1,700 colonoscopies (880 with CADe, 820 without).
• Colonoscopies were analyzed based on their order in the endoscopist's daily schedule.
• Primary outcome: ADR adjusted for withdrawal time (WT); Secondary outcomes: Adenomas per colonoscopy (APC), ADR by time of day.

Main Results:

• CADe significantly increased ADR in the first three procedures (49.9%) compared to unassisted procedures (39.4%).
• ADR in CADe-assisted procedures dropped in later cases (39.9%), showing no benefit over unassisted procedures (40.9%).
• Adenomas per colonoscopy (APC) gains with CADe were significant early but disappeared in later procedures, suggesting reduced effectiveness due to potential alert fatigue.

Conclusions:

• CADe initially boosts ADR but its effectiveness diminishes in later-day colonoscopies.
• The observed decline in ADR suggests that endoscopist alert fatigue may limit responsiveness to CADe prompts.
• Optimizing CADe alert strategies may be necessary to sustain its benefits throughout the entire colonoscopy procedure list.