Grading the strength of a body of evidence when assessing health care interventions: an EPC update

详细信息    查看全文

• 作者：Nancy D. Berkman^a ; ^{berkman@rti.org" class="auth_mail" title="E-mail the corresponding author} ; Kathleen N. Lohr^a ; Mohammed T. Ansari^b ; Ethan M. Balk^c ; Robert Kane^d ; Marian McDonagh^e ; Sally C. Morton^f ; Meera Viswanathan^a ; Eric B. Bass^g ; Mary Butler^d ; Gerald Gartlehner^a ; ^h ; Lisa Hartlingⁱ ; Melissa McPheeters^j ; Laura C. Morgan^a ; James Reston^k ; Priyanka Sista^l ; Evelyn Whitlock^m ; Stephanie Changⁿ
• 关键词：Evidence-based practice ; Methods ; Strength of evidence ; Systematic reviews ; Clinical practice guidelines ; Health care delivery ; Health policy ; Minimally important differences ; Optimal information size
• 刊名：Journal of Clinical Epidemiology
• 出版年：2015
• 出版时间：November 2015
• 年：2015
• 卷：68
• 期：11
• 页码：1312-1324
• 全文大小：273 K

文摘

To revise 2010 guidance on grading the strength of evidence (SOE) of the effectiveness of drugs, devices, and other preventive and therapeutic interventions in systematic reviews produced by the Evidence-based Practice Center (EPC) program, established by the US Agency for Healthcare Research and Quality (AHRQ).

Study Design and Setting

A cross-EPC working group reviewed authoritative systems for grading SOE [primarily the approach from the Grading of Recommendations Assessment, Development and Evaluation (GRADE) working group] and conducted extensive discussions with GRADE and other experts.

Results

Updated guidance continues to be conceptually similar to GRADE. Reviewers are to evaluate SOE separately for each major treatment comparison for each major outcome. We added reporting bias as a required domain and retained study limitations (risk of bias), consistency, directness, and precision (and three optional domains). Additional guidance covers scoring consistency, precision, and reporting bias, grading bodies of evidence with randomized controlled trials and observational studies, evaluating single study bodies of evidence, using studies with high risk of bias, and presenting findings with greater clarity and transparency. SOE is graded high, moderate, low, or insufficient, reflecting reviewers' confidence in the findings for a specific treatment comparison and outcome.

Conclusion

No single approach for grading SOE suits all reviews, but a more consistent and transparent approach to reporting summary information will make reviews more useful to the broad range of audiences that AHRQ's work aims to reach. EPC working groups will consider ongoing challenges and modify guidance as needed, on issues such as combining trials and observational studies in bodies of evidence, weighting domains, and combining qualitative and quantitative syntheses.