Article type
Year
Abstract
Introduction:
Objectives:
Methods:
Results:
Discussion: In this paper we argue that systematic reviews are actively interpreted syntheses rather than passive reflections of available research evidence. The framework for undertaking systematic reviews of randomised controlled trials is now quite clearly established. The "recipe" requires us to formulate clear research questions, undertake comprehensive literature searches, apply explicit criteria to determine which primary studies will be included, assess the quality of included studies, extract data from them in a systematic way and use explicit approaches to combine the data to arrive at our conclusion. When using this recipe, however, reviewers have to make value-laden judgements, w nether they recognise it or not, their decisions about, for example, now to phrase research questions, winch inclusion criteria to apply, what data to extract and how quality assessments influence data synthesis can all affect the overall summary. We draw on our experience of reviewing evidence about the effects of providing people with cancer with tape-recordings of written summaries of their consultations with health professionals to discuss how.
1) widely accepted criteria for assessing the quality of primary studies can sometimes lead us to "miss the point"; the relative merits of either explanatory or pragmatic trials may be underplayed in systematic reviews that include both; it may be inappropriate to use the same decisions about review methods for all the outcomes of interest;
2) when the available data is limited, mere may be a tendency to give more emphasis to the outcomes mat nave actually been studied despite attempts to indicate which outcomes are deemed a priori to be the "most important";
3) reviews may perpetuate the assumptions that underly primary studies.
The widespread acceptance of a methodological framework for systematic reviews has many advantages, however, there is a danger that confidence in the recipe may lead people to neglect important conceptual and interpretive issues and produce formulaic reviews. It may also lull readers into a false sense of security about the lack of bias in the conclusions drawn.
Objectives:
Methods:
Results:
Discussion: In this paper we argue that systematic reviews are actively interpreted syntheses rather than passive reflections of available research evidence. The framework for undertaking systematic reviews of randomised controlled trials is now quite clearly established. The "recipe" requires us to formulate clear research questions, undertake comprehensive literature searches, apply explicit criteria to determine which primary studies will be included, assess the quality of included studies, extract data from them in a systematic way and use explicit approaches to combine the data to arrive at our conclusion. When using this recipe, however, reviewers have to make value-laden judgements, w nether they recognise it or not, their decisions about, for example, now to phrase research questions, winch inclusion criteria to apply, what data to extract and how quality assessments influence data synthesis can all affect the overall summary. We draw on our experience of reviewing evidence about the effects of providing people with cancer with tape-recordings of written summaries of their consultations with health professionals to discuss how.
1) widely accepted criteria for assessing the quality of primary studies can sometimes lead us to "miss the point"; the relative merits of either explanatory or pragmatic trials may be underplayed in systematic reviews that include both; it may be inappropriate to use the same decisions about review methods for all the outcomes of interest;
2) when the available data is limited, mere may be a tendency to give more emphasis to the outcomes mat nave actually been studied despite attempts to indicate which outcomes are deemed a priori to be the "most important";
3) reviews may perpetuate the assumptions that underly primary studies.
The widespread acceptance of a methodological framework for systematic reviews has many advantages, however, there is a danger that confidence in the recipe may lead people to neglect important conceptual and interpretive issues and produce formulaic reviews. It may also lull readers into a false sense of security about the lack of bias in the conclusions drawn.