JMIR Publications

ABSTRACT

Background:

Despite immense progress in artificial intelligence (AI) models, there has been limited deployment in healthcare environments. The gap between potential and actual AI applications is likely due to the lack of translatability between controlled research environments (where these models are developed) and clinical environments for which the AI tools are ultimately intended.

Objective:

We have previously developed the Translational Evaluation of Healthcare AI (TEHAI) framework to assess the translational value of AI models and to support successful transition to healthcare environments. In this study, we apply the TEHAI to identify areas where translatability could be improved in recently emerged COVID-19 literature.

Methods:

A systematic literature search for COVID-AI studies published between December 2019-2020 resulted in 3,830 records. A subset of 102 papers that passed inclusion criteria were sampled for full review. Nine reviewers assessed the papers for translational value and collected descriptive data (each study was assessed by two reviewers). Evaluation scores and extracted data were compared by a third reviewer for resolution of discrepancies. The review process was conducted on the Covidence software platform.

Results:

We observed a significant trend for studies to attain high scores for technical capability but low scores for the areas essential for clinical translatability. Specific questions regarding external model validation, safety, non-maleficence and service adoption received failed scores in most studies.

Conclusions:

Using TEHAI, we identified notable gaps in application of AI models in the COVID-19 clinical sphere. These gaps in areas crucial for clinical translatability could, and should, be considered already at the model development stage to increase translatability into real healthcare environments.