El Baz Noha, Middel Berrie, van Dijk Jitse P, Oosterhof Andre, Boonstra Piet W, Reijneveld Sijmen A
Department of Health Sciences, Subdivision Care Sciences, University Medical Center Groningen, University of Groningen, The Netherlands.
J Eval Clin Pract. 2007 Dec;13(6):920-9. doi: 10.1111/j.1365-2753.2006.00774.x.
To evaluate the validity of study outcomes of published papers that report the effects of clinical pathways (CP).
Systematic review based on two search strategies, including searching Medline, CINAHL, Embase, Psychinfo and Picarta from 1995 till 2005 and ISI Web of Knowledge SM. We included randomized controlled or quasi-experimental studies evaluating the efficacy of clinical pathway application. Assessment of the methodological quality of the studies included randomization, power analysis, selection bias, validity of outcome indicators, appropriateness of statistical tests, direct (matching) and indirect (statistical) control for confounders. Outcomes included length of stay, costs, readmission rate and complications. Two reviewers independently assessed the methodological quality of the selected papers and recorded the findings with an evaluation tool developed from a set of items for quality assessment derived from the Cochrane Library and other publications.
The study sample comprised of 115 publications. A total of 91.3% of the studies comprised of retrospective studies and 8.7% were randomized controlled studies. Using a quality-scoring assessment tool, 33% of the papers were classified as of good quality, whereas 67% were classified as of low quality. Of the studies, 10.4% controlled for confounding by matching and 59.1% adopted parametric statistical tests without testing variables on normal distribution. Differences in outcomes were not always statistically tested.
Readers should be cautious when interpreting the results of clinical pathway evaluation studies because of the confounding factors and sources of contamination affecting the evidence-based validity of the outcomes.
评估已发表的报告临床路径（CP）效果的论文研究结果的有效性。
基于两种检索策略进行系统评价，包括检索1995年至2005年的Medline、CINAHL、Embase、Psychinfo和Picarta以及ISI Web of Knowledge SM。我们纳入了评估临床路径应用效果的随机对照或准实验研究。对研究方法学质量的评估包括随机化、功效分析、选择偏倚、结果指标的有效性、统计检验的适当性、对混杂因素的直接（匹配）和间接（统计）控制。结果包括住院时间、费用、再入院率和并发症。两名评价者独立评估所选论文的方法学质量，并使用从Cochrane图书馆和其他出版物中提取的一组质量评估项目开发的评估工具记录研究结果。
研究样本包括115篇出版物。共有91.3%的研究为回顾性研究，8.7%为随机对照研究。使用质量评分评估工具，33%的论文被归类为高质量，而67%被归类为低质量。在这些研究中，10.4%通过匹配控制混杂因素，59.1%采用参数统计检验但未对变量进行正态分布检验。结果差异并非总是进行统计学检验。
由于存在影响结果循证有效性的混杂因素和污染来源，读者在解释临床路径评估研究结果时应谨慎。
