Diagnostic test guidelines based on high-quality evidence had greater rates of adherence: a meta-epidemiological study

Objectives: To determine the association between the quality of guidelines for diagnostic tests (both the quality and reporting and the quality of the evidence underpinning recommendations) and nonadherence.

Study design and setting: We conducted a meta-epidemiological study. We previously published a systematic review that quantified the percentage of test use that was nonadherent with guidelines. For the present study, we assessed these guidelines using the Appraisal of Guidelines for Research & Evaluation (AGREE) II tool. We then assessed the quality of evidence underpinning recommendations within these guidelines using Grading of Recommendations Assessment, Development, and Evaluation (GRADE). Linear models were then constructed to determine the association between guideline nonadherence and (1) AGREE II score and (2) GRADE score.

Results: There was no significant association between AGREE II score and nonadherent testing (P = 0.09). There was a significant association between GRADE score and nonadherence: recommendations based on low-quality and very low-quality evidence had 38% (P < 0.01) and 24% (P = 0.02) more nonadherent testing, compared with recommendations based on high-quality evidence.

Conclusion: Diagnostic test guideline recommendations based on high-quality evidence are adhered to more frequently.