Systematic reviews identify important methodological flaws in stroke rehabilitation therapy primary studies: review of reviews

详细信息	查看全文 | 推荐本文 |

• 作者：Pasqualina Santaguida^a ; Mark Oremus^a ; Kathryn Walker^a ; Laurie R. Wishart^b ; Karen Lohmann Siegel^c ; Parminder Raina^a ; ^{praina@mcmaster.ca}
• 关键词：Methods ; Systematic review ; Clinical trials ; Stroke ; Rehabilitation ; Reporting
• 刊名：Journal of Clinical Epidemiology
• 出版年：2012
• 期刊代码：148_08954356
• 类别：med
• 出版时间：April, 2012
• 卷：65
• 期：4
• 页码：358-367
• 文件大小：248 K

摘要

Objective

A 鈥渞eview of reviews鈥?was undertaken to assess methodological issues in studies evaluating nondrug rehabilitation interventions in stroke patients.

Study Design and Setting

MEDLINE, CINAHL, PsycINFO, and the Cochrane Database of Systematic Reviews were searched from January 2000 to January 2008 within the stroke rehabilitation setting. Electronic searches were supplemented by reviews of reference lists and citations identified by experts. Eligible studies were systematic reviews; excluded citations were narrative reviews or reviews of reviews. Review characteristics and criteria for assessing methodological quality of primary studies within them were extracted.

Results

The search yielded 949 English-language citations. We included a final set of 38 systematic reviews. Cochrane reviews, which have a standardized methodology, were generally of higher methodological quality than non-Cochrane reviews. Most systematic reviews used standardized quality assessment criteria for primary studies, but not all were comprehensive. Reviews showed that primary studies had problems with randomization, allocation concealment, and blinding. Baseline comparability, adverse events, and cointervention or contamination were not consistently assessed. Blinding of patients and providers was often not feasible and was not evaluated as a source of bias.

Conclusions

The eligible systematic reviews identified important methodological flaws in the evaluated primary studies, suggesting the need for improvement of research methods and reporting.