Article type
Year
Abstract
Background: Conflict of interest (COI) among clinical practice guideline (CPG) authors is an important potential source of bias for recommendations. A recent report of the Institute of Medicine (2009) suggested that updated descriptive information is needed on these conflicts.
Objectives: To describe the extent and predictors of COI among CPG authors in the National Guideline Clearinghouse.
Methods: A random sample of 250 CPGs was obtained on 11/22/2010. For each CPG, data were abstracted on characteristics of the CPG and author disclosures. Regression techniques were used to predict the presence of COI among authors.
Results: Of 114 CPGs analyzed to date, over a third were developed and funded by medical specialty societies, 3% were funded by industry and 34% by government agencies. The median number of authors per CPG was 14; 21% of guidelines had ≥25 authors. The most common focus for CPGs was disease management, followed by diagnosis and therapy. Author disclosures were reported in 76% of CPGs. Of these, 30% reported one or more COIs. An average of two authors (range 0-17) disclosed any COI for each CPG, the most frequently reported disclosure was paid consultancy for industry. Although 87% of organizations provided a COI policy, 24% of guidelines reported no information on COIs. Journal publication was a significant predictor of a guideline reporting any authors with a COI (p<0.0001). The year the CPG was released, the focus of the CPG, and the country originating the CPG were not significant predictors of COI reporting.
Conclusions: Funding of CPGs by private specialty organizations is common, conflicts of interest among CPG authors is frequent, and a significant percentage of CPGs do not provide disclosures. To be trustworthy, efforts are needed to achieve accurate, relevant, and readily accessible author disclosures of conflicts of interest in all clinical practice guidelines.
Objectives: To describe the extent and predictors of COI among CPG authors in the National Guideline Clearinghouse.
Methods: A random sample of 250 CPGs was obtained on 11/22/2010. For each CPG, data were abstracted on characteristics of the CPG and author disclosures. Regression techniques were used to predict the presence of COI among authors.
Results: Of 114 CPGs analyzed to date, over a third were developed and funded by medical specialty societies, 3% were funded by industry and 34% by government agencies. The median number of authors per CPG was 14; 21% of guidelines had ≥25 authors. The most common focus for CPGs was disease management, followed by diagnosis and therapy. Author disclosures were reported in 76% of CPGs. Of these, 30% reported one or more COIs. An average of two authors (range 0-17) disclosed any COI for each CPG, the most frequently reported disclosure was paid consultancy for industry. Although 87% of organizations provided a COI policy, 24% of guidelines reported no information on COIs. Journal publication was a significant predictor of a guideline reporting any authors with a COI (p<0.0001). The year the CPG was released, the focus of the CPG, and the country originating the CPG were not significant predictors of COI reporting.
Conclusions: Funding of CPGs by private specialty organizations is common, conflicts of interest among CPG authors is frequent, and a significant percentage of CPGs do not provide disclosures. To be trustworthy, efforts are needed to achieve accurate, relevant, and readily accessible author disclosures of conflicts of interest in all clinical practice guidelines.