A Practical Guide to Cluster Randomized Trials in School Health Research
A Practical Guide to Cluster Randomized Trials in School Health Research
复制标题
DOI：
10.1111/josh.12826
复制
发表时间：
2019-11-01
影响因子：
2.2
通讯作者：
Goesling, Brian
中科院分区：
文献类型：
--
作者：
Goesling, Brian
如果设计和实施良好，群集随机试验可以满足联邦机构和其他资助者越来越多地要求学校健康计划和服务有效性证据的高标准。然而，设计和实施这些研究可能会带来比最初看起来更多的挑战。方法：我回顾了关于整群随机试验的方法学文献的摘要。然后，我进行了搜索，以确定这些方法在学校健康研究中的实际应用。结果：该综述确定了学校卫生研究人员在设计、实施和分析随机分组试验数据时必须解决的6个关键问题或决策：（1）使用分组设计的原因，（2）样本量计算，（3）匹配或分层的使用，（4）学校和学生样本的定义，（5）知情同意收集，（6）分析方法。结论：学校卫生研究人员可以采取几个切实可行的步骤，以确保高质量的研究证据的可用性，并满足日益增长的需求，证据和问责制的教育政策和规划。这些步骤包括为干预和评估环境选择正确的研究设计，确定适当的样本定义和分析方法，并制定适当的程序收集家长的许可和学生的同意。
BACKGROUND When well-designed and implemented, cluster randomized trials can meet the high standards federal agencies and other funders increasingly require for evidence on the effectiveness of school health programs and services. However, designing and implementing these studies can present more challenges than at first appears. METHODS I reviewed summaries of the methodological literature on cluster randomized trials. I then conducted a search to identify practical applications of these methods in school health research. RESULTS The review identified 6 key issues or decisions school health researchers must address when designing, conducting, and analyzing data from a cluster randomized trial: (1) reasons to use a clustered design, (2) sample size calculation, (3) the use of matching or stratification, (4) definition of the school and student samples, (5) consent gathering, and (6) analysis methods. CONCLUSIONS School health researchers can take several practical steps to ensure the availability of high-quality research evidence and meet the growing demands for evidence and accountability in education policy and programming. These steps include selecting the right research design for the intervention and evaluation setting, identifying appropriate sample definitions and analysis methods, and developing appropriate procedures for gathering parental permission and student assent.