[Missing replication of study results in evidence assessment with GRADE]

The Institute for Quality and Efficiency in Health Care (IQWiG) can support guideline groups by producing evidence reports. These reports summarize the available evidence on specific issues.

To this end, the IQWiG uses, among other things, the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) approach in accordance with its mandate. In this approach, the reliability of the evidence is assessed on the basis of various factors. These factors encompass study limitations, indirectness of the evidence, imprecision of the effect estimate, publication bias, and inconsistency of the results. The latter is assessed on the basis of the heterogeneity of the effect estimates of different studies on an outcome. However, in the event that only a single study is available for a given body of evidence, the assessment of inconsistency is not possible. Consequently, a discrepancy emerges between the bodies of evidence from single and multiple studies with respect to the number of factors that can diminish the certainty of evidence and potentially result in an overly optimistic appraisal of the certainty of evidence from single studies.

This paper describes IQWiG's new methodological approach in evidence reports of assigning evidence from single randomized controlled trials (RCTs) a moderate certainty of evidence as a starting point at the beginning of the assessment. The evidence bodies from multiple RCTs continue to begin the assessment process with a high certainty of evidence, consistent with the GRADE approach. From this initial starting point, the certainty of evidence is then subjected to a comprehensive evaluation, incorporating additional factors, and is subsequently downgraded if deemed necessary.

The next step is to further develop the minimum requirements for studies that are used for replication.