如何阅读感控相关献,SR和META分析20120523.pptVIP

如何正确阅读感染相关的meta分析和系统评价文献 陈文森 江苏省疾控中心消毒与媒介生物防制所 Email:metrischen@126.com 电话18915999331 概述 相关基础知识 森林图快速入门 案例分析（一） 案例分析（二） 案例分析（三） 在用样本信息推断总体参数时，是存在抽样误差的，并且抽样误差的大小与样本量的大小有关 统计学用抽样分布的理论来描述样本统计量的变化规律 通过增大样本量可以增加估计精度 Meta分析的条件 数量要求：两个以上的研究 研究要求：研究因素要求相似或者接近 结果测量：结局变量采用相似的方法进行测量 资料要求：各研究的资料都可用 系统评价、Meta分析和综述的关系 Meta分析和系统评价属于综述类研究 Meta分析也并非一定要做系统评价，因为其本质是一种统计学方法 系统评价并非必须对纳入研究进行统计学合并（Meta分析），如定性系统评价 包含有对具同质性的多个研究进行Meta分析的系统评价称为定量系统评价 牢记四要素（PICO）： P: Participants ——研究对象 I: Intervention ——干预措施 C: Comparison ——比较 O: Outcome ——结局 森林图快速入门 发表偏倚 定义：有统计学意义的研究结果比无统计学意义的结果更容易被投稿或者发表。 -设计方案的选择 -研究者的主观愿望 -资助者的意愿 -编辑的倾向性 小结 案例分析（一） Study selection入选标准 We narrowed the list of publications obtained to studies meeting our predetermined inclusion criteria. Thereby, we included only randomized controlled trials of mechanically ventilated adult patients receiving oral care with chlorhexidine or povidone-iodine. We excluded studies in which antibiotics were used as experimental intervention for oral decontamination. We included standard oral care, use of a placebo, or another product for oral care as control interventions. We retained only studies reporting rates of ventilator-associated pneumonia as outcome. Two investigators (NB and KVdV) did a first broad selection based on study title, under close supervision of the principal investigator (SB) who is a content expert. To allow further narrowing, four independent reviewers (KVdV, SB, SL, NB) screened the selected abstracts, each masked to the results of the others’ selection. Mostly, all reviewers decided unanimously. In one case of disagreement, assessment of eligibility was done by mutual consideration. Data extraction数据提取 Categories of extracted data included author and year of publication, settings and study populations, inclusion and exclusion criteria, de?nitions and diagnosis of ventilator-associated pneumonia, intervention in the study and the control g