CONSORT-AI and SPIRIT-AI: New Reporting Guidelines for Clinical Trials and Trial Protocols for Artificial Intelligence Interventions

Abstract

Purpose : The SPIRIT (Standard Protocol Items: Recommendations for Interventional Trials) and CONSORT (Consolidated Standards of Reporting Trials) statements are reporting guidelines for protocols of and completed randomized controlled trials. Their widespread endorsement has been instrumental in ensuring transparency for evaluating new interventions. It is increasingly recognized that interventions involving artificial intelligence (AI) may pose new challenges not adequately addressed by current guidelines. Therefore, the SPIRIT-AI and CONSORT-AI Steering Group is preparing international, consensus-based extensions to the SPIRIT and CONSORT statements for clinical trials involving AI interventions.

Methods : The guidelines were developed according to the EQUATOR (Enhancing Quality and Transparency of Health Research) Network methodological framework for guideline development. A literature search and an expert consultation generated a list of candidate items, which were presented to stakeholders in 2 online Delphi surveys, where they were asked to rate the importance of each item. A final consensus meeting was held in January 2020 to finalize the SPIRIT-AI and CONSORT-AI guidelines.

Results : The expert consultation identified 26 candidate items for SPIRIT-AI and 29 for CONSORT-AI. These included: description of the inputs and outputs of the algorithm, how the algorithm makes recommendations and fits into a clinical pathway, considerations for continuously improving algorithms, potential sources of bias and poor generalisability, and statements on algorithm ownership and access. The Delphi study included 98 stakeholders, representing an international group of clinicians, computer science/ML experts, methodologists, statisticians, funders, regulators, journals, patients and policy makers. The final guidelines resulting from the January 2020 consensus meeting will be presented.

Conclusions : This is an international consensus initiative to reflect the concerns of the wider community and to address an area of urgent need. Once the new guidance becomes available, we recommend that investigators planning clinical trials for AI interventions consult SPIRIT-AI and CONSORT-AI as early as possible. The new international consensus-based guidelines will provide a valuable resource to promote transparent design and reporting of AI trials.

This is a 2020 ARVO Annual Meeting abstract.