Abstract: Objective To systematically evaluate the efficacy of problem-based learning (PBL) and traditional lecture-based learning (LBL) in cerebrovascular disease clinical teaching in China. Methods Randomized controlled trials (RCTs) of PBL and LBL in cerebrovascular disease clinical teaching were retrieved from Chinese databases (China National Knowledge Infrastructure, Wanfang Data Knowledge Service Platform, VIP Information Chinese Journal Service Platform) from inception to May 2021. Revman 5.4 and Stata 15.0 were used for meta analysis and publication

bias detection. By using standard mean difference (SMD) or RR and 95%CI , the teaching efficacy

of PBL and LBL were compared through examination scores (theoretical knowledge, case analysis and skill operation), questionnaire scoring items (medical knowledge development, doctor-patient communication ability, clinical practice ability, clinical thinking ability, team cooperation ability, autonomous learning ability) and teaching satisfaction. Results A total of 25 RCTs involving 1493 participants were included, of them, 747 in PBL group and 746 in LBL group. The examination scores of theoretical knowledge (SMD 1.41, 95%CI 0.79- 2.02, P <0.001), case analysis (SMD 1.74, 95%CI 0.84-2.64, P <0.001) and skill operation (SMD 3.35, 95%CI 1.09-5.60, P =0.004) in PBL group were superior to those in LBL group (P <0.001). The questionnaire scores of medical knowledge development (SMD 1.08, 95%CI 0.53-1.62, P <0.001), doctor-patient communication ability (SMD 3.07, 95%CI 2.07-4.07, P <0.001), clinical practice ability (SMD 2.55, 95%CI 1.33-3.78, P <0.001), clinical thinking ability (SMD 3.02, 95%CI 0.99- 5.06, P =0.004), team cooperation ability (SMD 1.64, 95%CI 1.19-2.08, P <0.001), autonomous learning ability (SMD 2.35, 95%CI 1.20-3.50, P <0.001) in PBL group were superior to that in LBL group. The teaching satisfaction (RR 1.18, 95%CI 1.10-1.26, P <0.001) in PBL group was also superior to that in LBL model. Except for the homogeneity of teaching satisfaction (P =0.23, I 2=29%), the other outcome indicators were heterogeneous (P ≤0.1 or I 2≥50%). There was no statistical significance of publication bias for theoretical knowledge score, case analysis score, skill operation score, teaching satisfaction and questionnaire score of medical knowledge development, while there was statistical significance of publication bias for other outcome indicators. Conclusions Based on the existing evidence, PBL method was more effective than LBL method in cerebrovascular diseases clinical teaching in China. However, the risk of bias and heterogeneity of included studies should not be ignored, and more higher quality multicenter studies are needed to further confirm the above conclusions.

Key words: Problem-based learning; Lecture-based learning ; Cerebrovascular disease; Clinical teaching; Meta analysis

摘要： 目的 系统评价基于问题学习（problem-based learning，PBL）和传统的授课式学习（lecture-based learning，LBL）教学模式在我国脑血管病学临床教学中的应用效果。 方法 检索中文数据库（中国期刊全文数据库、万方数据知识服务平台、维普资讯中文期刊服务 平台）中关于脑血管病临床教学应用PBL和LBL的随机对照试验，时间范围为自建库至2021年5月。使 用Revman 5.4和Stata 15.0进行meta分析和发表偏倚检测。采用标准化均数差（standardized mean difference，SMD）或RR及其95%CI来评价PBL和LBL的考核成绩（理论知识、病例分析和技能操作）， 问卷评分项目（医学知识拓展、医患沟通能力、临床实践能力、临床思维能力、团队协作能力、自主学 习能力）和教学满意度等教学效果。 结果 共纳入25项随机对照试验，学员1493人，其中PBL组747人，LBL组746人。PBL组的理论知 识（SMD 1.41，95%CI 0.79～2.02，P <0.001）、病例分析（SMD 1.74，95%CI 0.84～2.64，P <0.001）、技 能操作（SMD 3.35，95%CI 1.09～5.60，P =0.004）的考核成绩均优于LBL组。PBL组的医学知识拓展 （SMD 1.08，95%CI 0.53～1.62，P <0.001）、医患沟通能力（SMD 3.07，95%CI 2.07～4.07，P <0.001）、临 床实践能力（SMD 2.55，95%CI 1.33～3.78，P <0.001）、临床思维能力（SMD 3.02，95%CI 0.99～5.06， P =0.004）、团队协作能力（SMD 1.64，95%CI 1.19～2.08，P <0.001）、自主学习能力（SMD 2.35，95%CI 1.20～3.50，P <0.001）的问卷评分均高于LBL组。PBL组的教学满意度（RR 1.18，95%CI 1.10～1.26， P <0.001）也优于LBL组。除教学满意度（P =0.23，I 2=29%）具有同质性，其余结局指标均具有异质性 （P≤0.1或I 2≥50%）。理论知识成绩、病例分析成绩、技能操作成绩、教学满意度、医学知识拓展问卷 评分的发表偏倚无统计学意义，其余结局指标的发表偏倚有统计学意义。 结论 基于现有证据，在我国脑血管病临床教学中，PBL比LBL的教学效果更优，但纳入研究的偏倚 风险和异质性不可忽视，需更高质量的多中心研究来进一步证实上述结论。

关键词: 基于问题学习; 授课式学习; 脑血管病; 教学; 荟萃分析