大型公共卫生项目评价研究设计的有关问题探讨

Issues in the study design of large scale public health program evaluation

对源于临床医疗领域的随机对照试验(RCT)在大型公共卫生干预尤其是社区干预中拓展应用存在的外部效度和内部效度方面的局限性进行了重点阐释,认为外部效度受到社会和行为效应修饰问题的严重制约,且其传统的内部效度优势也因公共卫生项目因果路径长等特征而不能充分发挥,并指出一系列限制其现实可行性的因素。建议在大型公共卫生干预项目中,主要根据评价结论的用途并结合现实条件来确定合适的评价研究设计。针对意在检验因果关系的探索性项目,建议采用更广泛意义上的实验设计,包括整群随机试验设计、分阶渐进随机试验设计和N-of-1设计等,以改善传统RCT的现实可行性问题,同时,应发挥定性研究方法的优势,尽可能开展过程评价和项目监测以弥补内部效度方面存在的缺陷;对于因果关系已得到验证的示范项目或应用型项目,建议采用准实验设计,如时间序列设计或多重基线设计等,以及流行病学观察性研究方法,以保证充分的外部效度。

This paper explored the internal and external validity limitations of extending RCT design to evaluate large-scale public health interventions.The authors believed that the external validity of RCT design would be restricted by social and behavioral modification,and the advantage of RCT design on internal validity would not be demonstrated due to the long and complex causal pathway in public health interventions.It was recommended that the selection of the study design for evaluating large-scale public health programs be based on the use of the program evaluation and actual field conditions.For exploratory programs,which aim to establish causal inference,experimental designs such as cluster-randomized trials,stepped wedge design,N-of-1 design,were recommended to improve the feasibility of traditional RCTs and to make up for the deficiency in internal validity.For demonstration programs in which the cause-effect relationship has been clarified,quasi-experimental designs were recommended to ensure external validity.