Determinants of patient trust in gastroenterology televisits: Results of machine learning analysis

Abstract

Background: The introduction of telemedicine into gastroenterology practice has been a major change over the past decade. Particularly during the COVID-19 pandemic, it has been very helpful for patients with chronic gastrointestinal disease as it has allowed continued healthcare delivery. Patient acceptance of televisits is key for its implementation in usual clinical practice, but lack of patient trust may limit its adoption. During the COVID-19 pandemic, we have embraced televisits instead of the traditional in-person medical examinations.The aim of the study was to evaluate the feasibility of televisits and factors influencing patient trust. Methods: Patient trust in televisits was assessed through a validated questionnaire (PATAT). We employed machine learning (decision trees and random forests) in order to clearly understand the relationships between covariates influencing patient trust. Results: Most televisits were successfully performed (186/218, 86.2%) and highly trusted (155/163, 95.2%). According to the decision tree, ‘The video service is easy to use’ in the parent node had the most influence on patient trust. Trust in the care organization, in the treatment, and in guaranteed data protection policies were the other factors influencing patient trust. In the random forest analysis, the use of known and user-friendly video services (12.8%IncMSE) and confidence in the data protection policies (12.4%IncMSE) were the two variables contributing most to trust in televisits. Conclusion: Most patients with chronic gastrointestinal disease agreed to receive a televisit and trusted it. Knowledge of factors determining patient trust is essential to improve patient–doctor communication in order to increase the use of telemedicine in gastroenterology.

Keywords

• Telemedicine
• Televisits
• Gastroenterology
• Machine learning
• COVID-19