Bhandari Mohit, Richards Robin R, Sprague Sheila, Schemitsch Emil H
Department of Clinical Epidemiology and Biostatistics, McMaster University Health Sciences Centre, Hamilton, Ontario, Canada.
J Bone Joint Surg Am. 2002 Mar;84(3):388-96. doi: 10.2106/00004623-200203000-00009.
The purpose of this study was threefold: (1) to determine the scientific quality of published randomized trials in the American Volume of The Journal of Bone and Joint Surgery from 1988 through 2000, (2) to identify predictors of study quality, and (3) to evaluate inter-rater agreement in the scoring of study quality with use of a simple scale.
Hand searches of The Journal of Bone and Joint Surgery were conducted in duplicate to identify randomized clinical trials. Of 2468 studies identified, seventy-two (2.9%) met all eligibility criteria. Two investigators each assessed the quality of the study under blinded conditions and abstracted relevant data.
The mean score (and standard error) for the quality of the seventy-two randomized trials was 68.1% plus minus 1.6%; 60% (forty-three) scored <75%. Drug trials had a significantly higher mean quality score than did surgical trials (72.8% compared with 63.9%, p < 0.05). Regression analysis revealed that cited affiliation with an epidemiology department and cited funding were associated with higher quality scores. Failure to conceal randomization, to blind outcome assessors, and to describe why patients were excluded resulted in significantly lower quality scores (p < 0.05), more than the 5% decrease expected by removal of each item. A priori calculations of sample size were rarely performed in the reviewed studies, and only 2% of the studies with negative results included a post hoc power analysis. The Detsky quality scale met accepted standards of interobserver reliability (kappa, 0.87; 95% confidence interval, 0.70 to 0.95).
Few studies published in The Journal of Bone and Joint Surgery were randomized trials. More than half of the trials were limited by a lack of concealed randomization, lack of blinding of outcome assessors, or failure to report reasons for excluding patients. Application of standardized guidelines for the reporting of clinical trials in orthopaedics should improve quality.
本研究目的有三个:(1)确定1988年至2000年《骨与关节外科杂志》美国卷中发表的随机试验的科学质量;(2)确定研究质量的预测因素;(3)使用简单量表评估研究质量评分中的评分者间一致性。
对《骨与关节外科杂志》进行了两次手工检索,以确定随机临床试验。在检索出的2468项研究中,72项(2.9%)符合所有纳入标准。两名研究者在盲态条件下各自评估研究质量并提取相关数据。
72项随机试验质量的平均评分(及标准误)为68.1%±1.6%;60%(43项)评分低于75%。药物试验的平均质量评分显著高于手术试验(分别为72.8%和63.9%,p<0.05)。回归分析显示,被引用的与流行病学系的关联以及被引用的资金与较高的质量评分相关。未进行随机化隐藏、未对结果评估者设盲以及未描述排除患者的原因导致质量评分显著降低(p<0.05),降幅超过剔除每个项目预期的5%。在纳入综述的研究中,很少进行样本量的预先计算,且只有2%的阴性结果研究进行了事后效能分析。德茨基质量量表符合观察者间可靠性的公认标准(kappa值为0.87;95%置信区间为0.70至0.95)。
《骨与关节外科杂志》上发表的研究中随机试验较少。超过半数的试验因缺乏随机化隐藏、缺乏对结果评估者设盲或未报告排除患者的原因而受到限制。应用骨科临床试验报告的标准化指南应能提高质量。