Article type
Year
Abstract
Background: BackInfo (www.rygginfo.no) presents the results of Cochrane reviews to back pain sufferers. It is a collaboration between the Norwegian Health Services Research Centre, the Norwegian Back Pain Association and the Cochrane Back Group.
Objectives: To describe the challenges we encountered when attempting to extract and present information from reviews that is relevant, consistent, and easy to understand, and to discuss the implications of these challenges for review authors and review groups.
Methods: An international, multidisciplinary team selected data from 18 reviews. We evaluated the quality of each outcome using the GRADE approach, and used the GRADE evidence profiles to develop standardised information. This information was pilot tested among back pain sufferers and was refereed by the review authors and by Norwegian clinicians.
Results: We encountered a number of challenges, including large amounts of information, particularly large numbers of comparisons and outcomes; missing information about the treatment and about side effects; and variations in the manner in which effect was measured and presented.
Our solutions to these problems involved the selection of comparisons and outcomes according to principles of patient-relevance, quality, and redundancy; the development of additional information about the treatment and about side effects; the consistent framing (positive or negative) of specific types of outcomes; and the standardised presentation of magnitude of effect.
Conclusions: While we succeeded in decreasing the information amount, our attempts to increase the consistency and comprehension of this information by re-framing and re-calculating results may have led to some distortion of the original data. Our attempt to compensate for missing information by collecting information from other sources may also have led to less reliable information.
These problems could be minimised if review groups could agree upon standard outcomes for use across reviews. The incorporation of more information about the treatment and its possible side effects should also be encouraged. In addition, the Cochrane Collaboration as a whole should continue their work to develop methods for incorporating information about the likelihood of side effects, and should develop clearer guidelines for the presentation of results.
Objectives: To describe the challenges we encountered when attempting to extract and present information from reviews that is relevant, consistent, and easy to understand, and to discuss the implications of these challenges for review authors and review groups.
Methods: An international, multidisciplinary team selected data from 18 reviews. We evaluated the quality of each outcome using the GRADE approach, and used the GRADE evidence profiles to develop standardised information. This information was pilot tested among back pain sufferers and was refereed by the review authors and by Norwegian clinicians.
Results: We encountered a number of challenges, including large amounts of information, particularly large numbers of comparisons and outcomes; missing information about the treatment and about side effects; and variations in the manner in which effect was measured and presented.
Our solutions to these problems involved the selection of comparisons and outcomes according to principles of patient-relevance, quality, and redundancy; the development of additional information about the treatment and about side effects; the consistent framing (positive or negative) of specific types of outcomes; and the standardised presentation of magnitude of effect.
Conclusions: While we succeeded in decreasing the information amount, our attempts to increase the consistency and comprehension of this information by re-framing and re-calculating results may have led to some distortion of the original data. Our attempt to compensate for missing information by collecting information from other sources may also have led to less reliable information.
These problems could be minimised if review groups could agree upon standard outcomes for use across reviews. The incorporation of more information about the treatment and its possible side effects should also be encouraged. In addition, the Cochrane Collaboration as a whole should continue their work to develop methods for incorporating information about the likelihood of side effects, and should develop clearer guidelines for the presentation of results.
PDF