Cochrane Child Health Field, Department of Pediatrics, University of Alberta, Edmonton, Canada.
PLoS One. 2010 Sep 30;5(9):e13106. doi: 10.1371/journal.pone.0013106.
The objective of this study was to describe randomized controlled trials (RCTs) and controlled clinical trials (CCTs) in child health published between 1948 and 2006, in terms of quantity, methodological quality, and publication and trial characteristics. We used the Trials Register of the Cochrane Child Health Field for overall trends and a sample from this to explore trial characteristics in more detail.
METHODOLOGY/PRINCIPAL FINDINGS: We extracted descriptive data on a random sample of 578 trials. Ninety-six percent of the trials were published in English; the percentage of child-only trials was 90.5%. The most frequent diagnostic categories were infectious diseases (13.2%), behavioural and psychiatric disorders (11.6%), neonatal critical care (11.4%), respiratory disorders (8.9%), non-critical neonatology (7.9%), and anaesthesia (6.5%). There were significantly fewer child-only studies (i.e., more mixed child and adult studies) over time (P = 0.0460). The proportion of RCTs to CCTs increased significantly over time (P<0.0001), as did the proportion of multicentre trials (P = 0.002). Significant increases over time were found in methodological quality (Jadad score) (P<0.0001), the proportion of double-blind studies (P<0.0001), and studies with adequate allocation concealment (P<0.0001). Additionally, we found an improvement in reporting over time: adequate description of withdrawals and losses to follow-up (P<0.0001), sample size calculations (P<0.0001), and intention-to-treat analysis (P<0.0001). However, many trials still do not describe their level of blinding, and allocation concealment was inadequately reported in the majority of studies across the entire time period. The proportion of studies with industry funding decreased slightly over time (P = 0.003), and these studies were more likely to report positive conclusions (P = 0.028).
CONCLUSIONS/SIGNIFICANCE: The quantity and quality of pediatric controlled trials has increased over time; however, much work remains to be done, particularly in improving methodological issues around conduct and reporting of trials.
本研究旨在描述 1948 年至 2006 年间发表的儿童健康随机对照试验（RCT）和对照临床试验（CCT），内容包括数量、方法学质量以及试验和研究特征。我们使用 Cochrane 儿童健康领域试验注册库来观察总体趋势，并抽取部分样本进行更详细的研究特征分析。
方法/主要发现：我们从随机抽取的 578 个试验中提取了描述性数据。96%的试验发表于英文期刊，仅针对儿童的研究占 90.5%。最常见的诊断类别为传染病（13.2%）、行为和精神障碍（11.6%）、新生儿重症监护（11.4%）、呼吸障碍（8.9%）、非重症新生儿学（7.9%）和麻醉（6.5%）。随着时间的推移，仅针对儿童的研究（即更多的儿童和成人混合研究）明显减少（P=0.0460）。RCT 与 CCT 的比例随时间显著增加（P<0.0001），多中心试验的比例也有所增加（P=0.002）。方法学质量（Jadad 评分）（P<0.0001）、双盲研究比例（P<0.0001）和适当的分配隐藏比例（P<0.0001）均随时间显著提高。此外，我们发现报告质量也有所提高：对退出和随访丢失的描述更加充分（P<0.0001），对样本量计算（P<0.0001）和意向治疗分析（P<0.0001）的描述更加充分。然而，许多试验仍然没有描述其盲法水平，并且在整个研究期间，大多数研究都没有充分报告分配隐藏。随时间推移，有行业资助的研究比例略有下降（P=0.003），这些研究更有可能报告阳性结论（P=0.028）。
结论/意义：儿童对照试验的数量和质量随时间推移而增加；然而，仍有许多工作要做，特别是在提高试验的实施和报告方法学问题方面。
