INSERM UMR 1153, Epidemilogy ans statistics Sorbonne Paris Cité, 1 place du Parvis Notre Dame, 75181 Paris, France; INSERM CIC 1415, 2 boulevard Tonnelé, 37044 Tours cedex 9, France; Université François-Rabelais de Tours, faculté de médecine, 10 boulevard Tonnelé, 37000 Tours, France; Département Universitaire de Médecine Générale, Faculté de Médecine - Université François-Rabelais de Tours, PRES Centre-Val de Loire Université, 10 Boulevard Tonnellé, B.P. 3223, 37032 Tours Cedex 1, France.
INSERM UMR 1153, Epidemilogy ans statistics Sorbonne Paris Cité, 1 place du Parvis Notre Dame, 75181 Paris, France; INSERM CIC 1415, 2 boulevard Tonnelé, 37044 Tours cedex 9, France; Université François-Rabelais de Tours, faculté de médecine, 10 boulevard Tonnelé, 37000 Tours, France; CHRU de Tours, 2 boulevard Tonnelé, 37044 Tours cedex 9, France.
J Clin Epidemiol. 2014 Oct;67(10):1070-5. doi: 10.1016/j.jclinepi.2014.05.015.
To assess the reporting of the unequal randomization ratio in reports of trials with this design and to identify the justification for the design.
Systematic review of reports of trials with unequal randomization. We selected all original reports of two parallel-group randomized controlled trials with unequal randomization, which were published in 2009 and 2010 in core clinical journals in MEDLINE on the basis of the abstract (and full text, if necessary). Additional information was collected by an author survey.
We retrieved 106 reports (prevalence, 4.7%). The randomization ratio was not stated in 8.5% of reports and 51.9% of abstracts. Sample size calculation was reported in 70 reports, with unequal randomization not taken into account in 25.7% (n = 18). Justification for unequal randomization was not reported in 77.4% (n = 82) of reports. Combining information from reports and author surveys, we had justification for 41 trials. The main justification was safety issues for 20 trials. In 11 of those latter 20 reports, adverse events were not fully reported.
A better reporting of the randomization ratio, sample size calculation, and justification to unequal randomization could help readers appraise the quality and risk of bias of such trials.
评估报告中具有这种设计的试验中不均衡随机化比例的报告情况，并确定该设计的合理性。
对不均衡随机分组试验报告的系统评价。我们根据摘要（必要时为全文），从 MEDLINE 核心临床期刊 2009 年和 2010 年发表的所有两平行组随机对照试验的原始报告中，选择了具有不均衡随机化的试验报告。通过作者调查收集了额外的信息。
我们检索到 106 篇报告（患病率为 4.7%）。8.5%的报告和 51.9%的摘要未说明随机化比例。70 篇报告报告了样本量计算，但 25.7%（n=18）未考虑不均衡随机化。77.4%（n=82）的报告未报告不均衡随机化的理由。根据报告和作者调查的综合信息，我们对 41 项试验有理由。主要理由是 20 项试验的安全问题。在这 20 项报告中的 11 项中，未充分报告不良事件。
更好地报告随机化比例、样本量计算和不均衡随机化的理由，可以帮助读者评估此类试验的质量和偏倚风险。
