Exploring the Feasibility of the Meta-Analysis of Randomized Controlled Trials on Artificial Intelligence Chatbots for Use in Healthcare Based on a Published Systematic Review

Abstract

Usage of artificial intelligence (AI) based chatbot systems has been increasing not only in many industrial fields but also in the health care system. The effectiveness of AI chatbot systems as therapeutic tools has to be confirmed by clinical trials. Meta-analysis, a statistical method that synthetizes results of multiple studies thus increases the power of the findings. In this paper, our focus is on randomized controlled trials (RCTs) for the purpose of studying the effectiveness of AI chatbots used for healthcare purposes and analyzing step by step their applicability for meta-analysis. This article is based on a systematic literature review that identified eight RCTs. Only two RCTs (in the field of mental health) were feasible for meta-analysis. Standardization of the RCTs, development of points to consider as guidelines for conducting clinical trials with AI chatbots in diverse clinical areas could efficiently increase the strength of the studies and enable meta-analyses.

Ähnliche Arbeiten