Grading of Recommendations, Assessment, Development, and Evaluation guidance 44: strategies to enhance the utilization of randomized and nonrandomized studies in evidence syntheses of healthinterventions

BACKGROUND AND OBJECTIVES: Ideally, guideline developers and health technology assessment authors base intervention decisions on randomized controlled trials (RCTs). However, relying solely on RCTs is uncommon, especially for public health interventions and harms assessment. In these situations, nonrandomized studies of interventions (NRSIs) can provide valuable information. This article presents Grading of Recommendations Assessment, Development, and Evaluation (GRADE) guidance for integrating bodies of evidence RCT and NRSI in evidence syntheses of health interventions. METHODS: Following standard GRADE methods, we developed this guidance through iterative discussions and examples with experts from the GRADE NRSI project group in multiple dedicated meetings. We presented findings of the group discussions for feedback at GRADE Working Group meetings in September 2023 and May 2024. RESULTS: The resulting GRADE guidance outlines a structured approach: (1) assessing the certainty of evidence (CoE) after defining the number of decision thresholds and the target of the certainty rating; (2) evaluating congruency of effect estimates between RCTs and NRSIs; (3) identifying which GRADE domains are affected by certainty ratings to inform complementariness between RCTs and NRSIs and the overall CoE; and (4) deciding whether and how to use one or both types of studies. CONCLUSION: This GRADE guidance offers a structured and practical approach for integrating or not integrating RCTs and NRSIs in evidence syntheses. By addressing the interplay between affected GRADE domains and assessing the congruency of effects, it helps GRADE users determine when and how NRSIs can meaningfully complement or replace RCT evidence to inform certainty ratings and decision-making.