Article type
Year
Abstract
Background: Systematic reviews that find no studies eligible for inclusion, commonly known as ‘empty reviews’, may be especially problematic for clinicians and other decision-makers. The reporting of implications for practice in particular has been suggested as potentially sustaining a risk for bias.
Objectives: This research examines the reporting of ‘Implications for Practice’ of empty reviews in The Cochrane Database of Systematic Reviews (The CDSR).
Methods: All empty reviews within the CDSR as of August 15, 2010 were identified, extracted, and coded for analysis. Descriptive characteristics and the complete ‘Implications for Practice’ section were extracted from each empty review. Thematic content analysis was performed by two authors (JY & LS) and refereed by the third author (PM).
Results: 376 (8.7%) active reviews in the CDSR report no studies eligible for inclusion. Of these, 117 (31.1%) contained a one sentence ‘Implications for Practice’ section, generally concluding that no studies were identified for inclusion. 59 (15.7%) appeared to contain recommendations to use the reviewed intervention (39.0%), restrict the use of the intervention (22.0%), not use the intervention (25.4%), or use something else (13.6%). 26 (6.9%) were written in such tentative language, that it was not possible to come to an agreement as to whether or not the section made recommendations. Around half of ‘Implications for Practice’ sections made reference to evidence, but only 16% of these cited its source. At least one Cochrane Review Group (Schizophrenia Group) provided implications for multiple audiences, though resulting in rather lengthy discussions given the absence of eligible studies.
Conclusions: The reporting of ‘Implications for Practice’ sections appears to differ widely across empty reviews. A considerable number of these reviews were shown to simply conclude no studies were included or to issue recommendations, discuss excluded studies, use vague language, and/or fail to cite referenced studies.
Objectives: This research examines the reporting of ‘Implications for Practice’ of empty reviews in The Cochrane Database of Systematic Reviews (The CDSR).
Methods: All empty reviews within the CDSR as of August 15, 2010 were identified, extracted, and coded for analysis. Descriptive characteristics and the complete ‘Implications for Practice’ section were extracted from each empty review. Thematic content analysis was performed by two authors (JY & LS) and refereed by the third author (PM).
Results: 376 (8.7%) active reviews in the CDSR report no studies eligible for inclusion. Of these, 117 (31.1%) contained a one sentence ‘Implications for Practice’ section, generally concluding that no studies were identified for inclusion. 59 (15.7%) appeared to contain recommendations to use the reviewed intervention (39.0%), restrict the use of the intervention (22.0%), not use the intervention (25.4%), or use something else (13.6%). 26 (6.9%) were written in such tentative language, that it was not possible to come to an agreement as to whether or not the section made recommendations. Around half of ‘Implications for Practice’ sections made reference to evidence, but only 16% of these cited its source. At least one Cochrane Review Group (Schizophrenia Group) provided implications for multiple audiences, though resulting in rather lengthy discussions given the absence of eligible studies.
Conclusions: The reporting of ‘Implications for Practice’ sections appears to differ widely across empty reviews. A considerable number of these reviews were shown to simply conclude no studies were included or to issue recommendations, discuss excluded studies, use vague language, and/or fail to cite referenced studies.