Hubei-MOST KLOS & KLOBM, School and Hospital of Stomatology, Wuhan University, Wuhan, China.
Center for Evidence-Based Stomatology, School and Hospital of Stomatology, Wuhan University, Wuhan, China.
J Dent Res. 2020 Dec;99(13):1453-1460. doi: 10.1177/0022034520946025. Epub 2020 Aug 3.
The objectives of this study were to assess the reporting quality and methodological quality of split-mouth trials (SMTs) published during the past 2 decades and to determine whether there has been an improvement in their quality over time. We searched the MEDLINE database via PubMed to identify SMTs published in 1998, 2008, and 2018. For each included SMT, we used the CONsolidated Standards Of Reporting Trials (CONSORT) 2010 guideline, CONSORT for within-person trial (WPT) extension, and a new 3-item checklist to assess its trial reporting quality (TRQ), WPT-specific reporting quality (WRQ), and SMT-specific methodological quality (SMQ), respectively. Multivariable generalized linear models were performed to analyze the quality of SMTs over time, adjusting for potential confounding factors. A total of 119 SMTs were included. The mean overall score for the TRQ (score range, 0 to 32), WRQ (0 to 15), and SMQ (0 to 3) was 15.77 (SD 4.51), 6.06 (2.06), and 1.12 (0.70), respectively. The primary outcome was clearly defined in only 28 SMTs (23.5%), and only 27 (22.7%) presented a replicable sample size calculation. Only 45 SMTs (37.8%) provided the rationale for using a split-mouth design. The correlation between body sites was reported in only 5 studies (4.2%) for sample size calculation and 4 studies (3.4%) for statistical results. Only 2 studies (1.7%) performed an appropriate sample size calculation, and 46 (38.7%) chose appropriate statistical methods, both accounting for the correlation among treatment groups and the clustering/multiplicity of measurements within an individual. Results of regression analyses suggested that the TRQ of SMTs improved significantly with time ( < 0.001), while there was no evidence of improvement in WRQ or SMQ. Both the reporting quality and methodological quality of SMTs still have much room for improvement. Concerted efforts are needed to improve the execution and reporting of SMTs.
本研究旨在评估过去 20 年发表的分口试验（SMT）的报告质量和方法学质量，并确定其质量是否随时间的推移而有所提高。我们通过 PubMed 检索 MEDLINE 数据库，以确定 1998 年、2008 年和 2018 年发表的 SMT。对于每一项纳入的 SMT，我们分别使用 CONSORT 2010 指南、CONSORT 用于个体内试验（WPT）扩展和新的 3 项检查表，评估其试验报告质量（TRQ）、WPT 特定报告质量（WRQ）和 SMT 特定方法学质量（SMQ）。采用多变量广义线性模型分析 SMT 随时间的质量变化，并调整潜在混杂因素。共纳入 119 项 SMT。TRQ（评分范围 0 至 32）、WRQ（0 至 15）和 SMQ（0 至 3）的平均总分分别为 15.77（SD 4.51）、6.06（2.06）和 1.12（0.70）。仅 28 项 SMT（23.5%）明确界定了主要结局，仅有 27 项（22.7%）提出了可重复的样本量计算。仅有 45 项 SMT（37.8%）提供了使用分口设计的理由。仅有 5 项研究（4.2%）报告了用于样本量计算的体部位相关性，4 项研究（3.4%）报告了统计结果的相关性。仅有 2 项研究（1.7%）进行了适当的样本量计算，46 项研究（38.7%）选择了适当的统计方法，两者均考虑了治疗组之间的相关性以及个体内测量的聚类/多重性。回归分析结果表明，SMT 的 TRQ 随时间显著提高（ < 0.001），而 WRQ 或 SMQ 并无改善的迹象。SMT 的报告质量和方法学质量仍有很大的改进空间。需要共同努力，提高 SMT 的执行和报告质量。
