Article type
Abstract
Background: Clinical practice guidelines (CPG) provide a critical link in the implementation of best evidence in clinical practice. GRADE is a methodologically rigorous and transparent system for rating the confidence in the estimates of effect and moving from evidence to recommendations. It is being increasingly used by guideline developers across the globe including the United States (US).
Objectives: To assess the uptake of GRADE among guideline developers in the US and the extent to which suggested criteria for the use of GRADE system are implemented and reported.
Methods: We conducted a protocol-driven search for CPG by US-based guideline developing organizations listed in the National Guideline Clearinghouse published between 2005 and 2015. In order to attribute every CPG to one organisation, we excluded CPG that were developed jointly by more than one organization. Using a piloted data abstraction form applied independently and in duplicate we assessed the reported use of the suggested GRADE criteria for up to two CPG (the most recent) per organisation. We performed descriptive statistics using SPSS Vs. 24.
Results: Of 315 guideline documents, 135 by 33 organisations met inclusion criteria. We formally assessed 49 CPG. A majority of documents (87.8%) defined the certainty in the evidence consistent with GRADE. Only less than a third of CPG (32.7%) addressed all 5 domains for downgrading RCT evidence and few (6.1%) addressed all 3 domains for downgrading evidence from observational studies. All but one document (98.0%) used 3 or 4 categories for the overall certainty of evidence. Less than half (44.9%) provided a full evidence profile summarising the body of evidence. Approximately half of CPG (51.0%) addressed all 5 GRADE criteria for determining the direction and strength of recommendations. About 4 out of 5 (81.6%) of documents defined the strength of recommendations as strong and weak/conditional consistent with GRADE.
Conclusions: 1 in 3 evidence-based CPG originating in the US is developed using GRADE. Reporting of suggested criteria for the use of GRADE though is inconsistent with much room for improvement.
Objectives: To assess the uptake of GRADE among guideline developers in the US and the extent to which suggested criteria for the use of GRADE system are implemented and reported.
Methods: We conducted a protocol-driven search for CPG by US-based guideline developing organizations listed in the National Guideline Clearinghouse published between 2005 and 2015. In order to attribute every CPG to one organisation, we excluded CPG that were developed jointly by more than one organization. Using a piloted data abstraction form applied independently and in duplicate we assessed the reported use of the suggested GRADE criteria for up to two CPG (the most recent) per organisation. We performed descriptive statistics using SPSS Vs. 24.
Results: Of 315 guideline documents, 135 by 33 organisations met inclusion criteria. We formally assessed 49 CPG. A majority of documents (87.8%) defined the certainty in the evidence consistent with GRADE. Only less than a third of CPG (32.7%) addressed all 5 domains for downgrading RCT evidence and few (6.1%) addressed all 3 domains for downgrading evidence from observational studies. All but one document (98.0%) used 3 or 4 categories for the overall certainty of evidence. Less than half (44.9%) provided a full evidence profile summarising the body of evidence. Approximately half of CPG (51.0%) addressed all 5 GRADE criteria for determining the direction and strength of recommendations. About 4 out of 5 (81.6%) of documents defined the strength of recommendations as strong and weak/conditional consistent with GRADE.
Conclusions: 1 in 3 evidence-based CPG originating in the US is developed using GRADE. Reporting of suggested criteria for the use of GRADE though is inconsistent with much room for improvement.