Statistical Reporting Quality of Randomized Controlled Trials on Artificial Intelligence Chatbots for Promoting Healthy Lifestyle: Exploring the SAMPL Guideline

Abstract

Promoting healthy lifestyle in an efficient way is a major public health challenge. Artificial intelligence (AI) chatbots might substitute at least partly healthcare professionals in this work. Nonetheless, high-quality studies reported alongside high-quality standards are needed to prove their effectiveness and safety. We aimed to test the “Statistical analyzes and methods in the published literature” (SAMPL) guideline and assess the statistical reporting quality of available randomized controlled trials (RCTs) on AI chatbots promoting healthy lifestyles. We relied on a systematic literature review published in 2021, to identify relevant RCTs. Only one RCT fulfilled our eligibility criteria. Out of the 12 SAMPL categories six categories comprising 43 guideline items were applicable for this RCT, but only 20 were complied with and 23 were not found at all or only partially in this RCT. Further research is encouraged on how to expand statistical reporting guidelines to AI-specific issues and more attention is recommended from authors on following the reporting standards.

Ähnliche Arbeiten