Pragmatic Implementation Study of a Computer-Aided Detection Device on Colonoscopy Outcomes at a Trainee-Dominated...

Area of Science:

• Gastroenterology
• Medical Devices
• Clinical Implementation Science

Background:

• Randomized controlled trials demonstrate real-time computer-aided detection (CADe) efficacy in increasing adenoma and sessile serrated polyp detection.
• Pragmatic implementation studies have raised questions about the real-world utility of CADe devices in routine clinical practice.

Purpose of the Study:

• To evaluate the impact of pragmatic computer-aided detection (CADe) device implementation on colonoscopy metrics within a trainee-dominated academic county hospital.
• To assess the effectiveness of CADe in improving adenoma detection rates (ADR) and sessile serrated lesion detection rates (SSLDR) in a real-world clinical setting.

Main Methods:

• A retrospective analysis of colonoscopy data from one year before and one year after CADe implementation (1665 vs. 2301 procedures).
• Exclusion criteria included incomplete colonoscopies and inadequate bowel preparation (Boston Bowel Prep Score < 6, or < 2 in any segment).
• CADe usage was not mandated, allowing for assessment of pragmatic, non-compulsory adoption.

Main Results:

• Baseline adenoma detection rate (ADR) was 31.5% and sessile serrated lesion detection rate (SSLDR) was 1.3%.
• Following CADe implementation, ADR increased by 7.6% (p < 0.001) and SSLDR by 2.9% (p < 0.001).
• Adjusted odds ratios indicated higher adenoma detection (aOR 1.29) and sessile serrated lesion detection (aOR 3.10) post-implementation, controlling for covariates.

Conclusions:

• Implementation of a CADe device significantly improved both ADR and SSLDR in a trainee-dominated academic county hospital, regardless of device usage frequency.
• The benefits of CADe devices appear pronounced in settings with suboptimal baseline ADR and SSLDR.
• Consideration of CADe device implementation is recommended for practices seeking to enhance polyp detection rates.