本课题通过对病毒性肝炎后肝癌患者伴门静脉高压及脾功能亢进病人等临床观察性数据的提取、整理实践，得出结论：如果要获得高质量的临床观察性数据，需要①在医院推广结构化电子病历系统，使临床数据可以被记录得更全，更准。②在进行观察性临床研究之前，首先应进行研究设计和并根据试验设计进行模板设计。③应详细记录病人的联系方法，以便于出院后随访。基于以上理念，我们与西京医院神经内科合作开发的癫痫病历收集系统已经收集了超过1000份高质量的病人数据。.本课题通过将Logistic模型中的"分组变量"和"结局变量"依次作为应变量的二次建模法，建立了二分类和三分类倾向指数数据模型。并利用模拟数据比较了倾向指数二分组匹配分析法与Logistic回归分析法的I类错误、检验效能的差异，比较了"标准化差异法"与假设检验的P值法在评估样本基线因素均衡性方面的差异。本研究探索了三分组条件下的基线因素均衡性的评价标准，编写了三分组匹配的程序。对三分组模拟数据分别采用不同的卡钳值进行匹配，以"标准化差异、匹配比例、平均误差均方和相对偏倚"为评价标准，探索得出了三分组匹配最优卡钳值的设置值。

From the practice of extracting clinical data for research purpose from Hospital information system, we recommend that ① Structured EMR (Electronic Medical Record) should be implemented to record accurate and completed clinical data. ② Study design and EMR template should be developed before the start of the research. ③ The contact information of patients should be recorded in detail for following up. Based on the view above, cooperating with Neurology department of XiJing hospital, we developed a data acquisition system for falling sickness, more than 1000 cases have been collected. .This subject constructed two and three groups of propensity score simulation data model by taking the treatment variables and outcome variables as the dependent variable in logistic models in turn. And with the data model, we compared the power and type I error rates between propensity score and logistic regression. Then Monte Carlo simulations were used to compare standardized differences and hypothesis test that were commonly employed in evaluating the balance of baseline variables. We compared propensity score matching methods using different calipers and to choose the optimal caliper width for use with multiple treatment groups. We used caliper widths from 0.1 to 0.8 of the pooled standard deviation of the logit of the propensity score, in increments of 0.1. The balance in baseline variables was assessed by standardized difference. The matching ratio, relative bias, and mean squared error (MSE) of the estimate between groups in different propensity score-matched samples were also simulated. The results indicate that matching using a caliper width of 0.2 of the pooled standard deviation of the logit of the propensity score affords superior performance in the estimation of treatment effects. This subject provides a theoretical basis for the application of propensity score matching of three treatment groups.

本课题通过对病毒性肝炎后肝癌患者伴门静脉高压及脾功能亢进病人等临床观察性数据的提取、整理实践，得出结论：如果要获得高质量的临床观察性数据，需要①在医院推广结构化电子病历系统，使临床数据可以被记录得更全，更准。②在进行观察性临床研究之前，首先应进行研究设计和并根据试验设计进行模板设计。③应详细记录病人的联系方法，以便于出院后随访。基于以上理念，我们与西京医院神经内科合作开发的癫痫病历收集系统已经收集了超过1000份高质量的病人数据。本课题通过将Logistic模型中的"分组变量"和"结局变量"依次作为应变量的二次建模法，建立了二分类和三分类倾向指数数据模型。并利用模拟数据比较了倾向指数二分组匹配分析法与Logistic回归分析法的I类错误、检验效能的差异，比较了"标准化差异法"与假设检验的P值法在评估样本基线因素均衡性方面的差异。本研究探索了三分组条件下的基线因素均衡性的评价标准，编写了三分组匹配的程序。对三分组模拟数据分别采用不同的卡钳值进行匹配，以"标准化差异、匹配比例、平均误差均方和相对偏倚"为评价标准，探索得出了三分组匹配最优卡钳值的设置值。