Impact of Artificial Intelligence-based Patient Reinforcement on Quality of Colonoscopy

Overview

In order to improve bowel preparation for colonoscopy and consequently enhance detection rate of malignant and premalignant findings, a prospective, randomized and controlled three-arm study was developed. Patients who undergo ambulatory colonoscopy are randomly assigned into a control group with standard preparation, a phone call supported preparation group or a group supported by an artificial intelligence based chatbot. Primary endpoint is defined as quality of bowel preparation (Boston Bowel Preparation Score), secondary endpoints are patients satisfaction, comprehensiveness of bowel preparation, sedation dose, rate of coecal intubation and the rate of adenoma and polyp detection, anxiety referred to colonoscopy and patients satisfaction with preparation support.

Adequate bowel preparation is crucial for detection of adenoma and polyps of the colon, which can transform into malignant and premalignant lesions. In particular ambulatory performed colonoscopy is often insufficient due to inadequate preparation by patients at home. To focus this issue, a artificial intelligence based chatbot was developed to help patients who undergo bowel preparation by answering questions concerning bowel preparation. To compare the effect of this program, a three-arm randomized, controled multicentric clinical trial was developed. All patients who undergo ambulatory colonoscopy and meet the inclusion criteria (18 y.o., informed consent, WhatsApp access) were randomized in a standard preparation group (medical briefing and preparation brochure), a group receiving phone calls with instructions for bowel preparation and a third group with access to the chat bot answering questions concerning bowel preparation and examination conduct. As the primary end point, the quality of bowel preparation measured as Boston Bowel Preparation Scale was defined. As secondary endpoints patients satisfaction, comprehensiveness of bowel preparation, sedation dose, rate of coecal intubation, the rate of adenoma and polyp detection, anxiety referred to colonoscopy and patients satisfaction with preparation support were measured.