目的 分析基于奇异谱分析(singular spectrum analysis,SSA)的自回归移动平均模型(Autoregressive integrated moving average,ARIMA)模型预测流感样病例(influenza like illness,ILI)发病趋势的可行性,为流感防控工作提供合理的预测方法.方法 利用山西省2010年第14周-2017年第13周的流感监测资料以不同长度配比的训练集、测试集构建SSA-ARIMA模型,并与ARIMA、BP神经网络(Back propagation neural network,BPNN)、广义回归神经网络(General Regression Neural Network,GRNN)模型进行比较.采用平均绝对误差(Mean Absolute Error,MAE)、均方误差(Mean Squared Error,MSE)、均方根误差(Root Mean Squared Error,RMSE)比较各模型预测效果.结果 模型拟合方面,SSA-ARIMA模型在预测未来一个月发病趋势时的MAE、MSE、RMSE分别为0.163、0.061、0.248;预测六个月时分别为0.161、0.061、0.248;预测一年时分别为0.168、0.066、0.256;均低于ARIMA、BPNN、GRNN.模型预测方面,在预测未来一个月发病趋势时的MAE、MSE、RMSE分别为0.056、0.005、0.068;预测六个月时分别为0.189、0.081、0.285;预测一年时分别为0.210、0.075、0.273;也均低于ARIMA、BPNN、GRNN.结论 SSA-ARIMA模型对山西省ILI的预测效果优于ARIMA、BPNN、GRNN,可为流感预测提供科学依据.

目的 应用随机生存森林模型探讨肺癌患者预后影响因素的重要性并对预测结果进行评价.方法 对山西省某三甲医院342例确诊的肺癌患者进行随访研究,建立随机生存森林模型,并与传统的Cox回归模型进行比较.结果 342例肺癌患者中226例患者发生死亡,中位生存时间为28.23月.治疗方式、肿瘤大小、临床分期等变量是影响肺癌患者预后的重要因素,淋巴结转移、分化程度、病理分型、年龄是中度预测因素,并分析了变量之间的交互作用.二者的模型比较结果显示随机生存森林模型预测错误率以及预测误差均低于Cox回归模型.结论 随机生存森林模型拟合效果好,可用于右删失生存数据的分析,不但能发现重要的影响因素,还能发现变量之间的交互作用,为肺癌患者预后状况的改善,提升生命质量提供科学依据.

Objective:To investigate the characteristics of lymphocyte subsets and the correlation between the immune imbalance of T helper 17 cell/regulatory T cell (Th17/Treg) and cytokines in peripheral blood of patients with Beh?et's disease(BD).Methods:From January 2018 to November 2019, 82 outpatient and inpatient with BD of the Department of Rheumatology and Immunology of the Second Hospital of Shanxi Medical University with complete data were enrolled. The disease activity was evaluated according to BD Current Activity Form(BDCAF), and 66 age and matched healthy people were selected as the control group. The absolute numbers of lymphocyte subsets and T cell subsets dominated by Th17 and CD4 +CD25 +Foxp3 + Treg in patients with BD and healthy controls were detected by Flow cytometry. The levels of interleukin (IL-2), IL-4, IL-6, IL-10, IL-17, Interferon (IFN)-γ and Tumor necrosis factor-α (TNF)-α in patients with BD were measured by Cytometric Beads Array (CBA). The correlations between the ratio of Th17/Treg with inflammatory index, the number of organ involved and the levels of cytokines were analyzed. Data were analyzed by Mann-Whitney U test, Kruskal Wallis H test, Spearman correlation analysis and multiple linear regression. Results:①The absolute numbers of Th17 cells in peripheral blood of patients with active BD [13.9(7.7, 21.1) cells/μl] and patients with stable BD [8.7(6.1, 14.0) cells/μl] were higher than those of healthy controls [6.8(4.4, 8.5) cells/μl] ( P<0.01); The ratio of Th17/Treg was significantly increased ( P<0.01). The absolute co unts of Treg cells in BD group [24.79(15.64, 37.91) cells/μl] were sign-ificantly lower than those in healthy controls [30.59(23.04, 42.08) cells/μl], the difference was statistically significant ( P=0.016). ② The ratio of Th17/Treg was positively correlated with BDCAF ( r=0.298, P=0.007) and the number of organ involved ( r=0.304, P=0.006) was negatively correlated with age ( r=-0.254, P<0.05), and not correlated with the duration of disease and ESR ( P>0.05) in patients with BD. In addition, multiple linear stepwise regression showed that the ratio of Th17/Treg was positively correlated with BDCAF ( β=0.228, P=0.036) and negatively correlated with age ( β=-0.219, P=0.043), R2=0.101. ③ The levels of IL-2, IL-6, IL-10, IFN-γ in patients with BD were stati-stically higher than those of healthy controls ( P<0.01). The ratio of Th17/Treg was positively correlated with the levels of IL-2 ( r=0.307, P<0.01) and IL-4 ( r=0.301, P<0.01) in patients with BD. Conclusion:There is immune imbalance of Th17/Treg in patients with BD, which is closely related to disease activity, the number of organ involved, and the levels of cytokines such as IL-2 and IL-4. IL-2 and IL-4 may play an important role in the immune imbal-ance of Th17/Treg in patients with BD.

目的 探讨联合脆弱Copula模型在肺癌患者反复住院、死亡的影响因素分析.方法 通过收集确诊的非小细胞肺癌患者反复住院的随访资料,构建联合脆弱Copula模型,采用极大惩罚对数似然函数进行参数估计.结果 联合脆弱Copula模型利用Copula函数描述了由于未测量的复发水平上的协变量所导致的残差相依性,利用脆弱项引入了个体水平上的复发事件与终止事件间及复发事件间的相关性,客观地评价了肺癌患者疾病进展及死亡的影响因素,结果解释合理,软件实现方便.结论 联合脆弱Copula模型能够深入地分析和解释肺癌患者疾病进展及结局的随防资料所蕴含的信息,进一步拓展联合脆弱模型在含终点的临床复发事件数据研究中的应用,为医学实践中肿瘤患者疾病进展的随访研究提供方法学支持.

在目前研究生医学基础统计学实践教学学时压缩50％的环境下,如何保证教学质量,并且让学生的综合应用能力得到提升,是每一位教育工作者应该深思的问题.文章通过调整实践教学大纲、目标和内容,并采用随机对照的方法,比较CBL与TBL结合的新教学模式与传统教学模式的两个教学班的成绩,探讨教改效果,并为贯彻教育部研究生教育改革的精神提供一套行之有效的教学改革方案.

目的 探讨联合脆弱模型在含有终止事件的临床复发事件数据分析中的应用及R软件的实现.方法 收集肺癌患者复发数据(多次住院),构建联合脆弱模型,拟采用最大惩罚似然估计(MPnLE)进行模型的参数估计,并评估肺癌患者个体内多次复发间的相关性以及复发事件与终止事件(死亡)间的相关性.结果联合脆弱模型分别评估了协变量对疾病复发进程与死亡进程的效应,同时也考虑了多次复发与死亡的相关性,结果 解释合理,软件实现方便.结论 联合脆弱模型可以充分挖掘含终止事件的肺癌患者复发数据所蕴含的信息,也可用于其他肿瘤患者预后因素的分析,为临床诊断和治疗提供统计学支持.

Objective To observe the change in IL-1β,IL-13mRNA expression in drowning rat lungs and serum,so as to investigate the significance of IL-1β and IL-13 mechanism in the development of drowning.Methods SD rats were randomly divided into control group,drowning group.Then using TaqMan probe method to determine the expression of IL-1β and IL-13 mRNA in Right lower lobe of lung tissue and the serum of right ventricle,which were extracted respectively from each group of rats.Results (1) The lung tissue morphological changes:Typical appearance signs and anatomy of drowning group meet ante-mortem drowning feature.(2) The expression of IL-1β,IL-13 in lung tissue:compared with the control group,the expression of IL-1β and IL-13 were slightly decreased,which has no statistical significance.(3) The expression of IL-1β and IL-13 in serum:compared with the control group,the expression of IL-1β and IL-13 were significant increased,both of which has statistical significance.Conclusion (1)The expression of IL-1β and IL-13 were decreased in lung tissue may be due to drowned rats present compensatory anti-inflammatory response syndrome which causes immune incompetent performance.(2) The expression of IL-1β and IL-13 were significant increased in serum may be relate to drown stress and drowning associated acute lung injury after traumatic stress.

目的 探讨通过Bootstrap抽样的方法来解决决定系数的可信区间估计.方法 通过实例分析,重复从原始数据中进行有放回的抽样得到Bootstrap样本,然后计算每个Bootstrap样本的决定系数的Bootstrap估计量.结果 Bootstrap估计量的BCa可信区间不仅矫正了非对称性,还对原始数据的异常值给予了矫正.结论 Bootstrap抽样不需要任何理论推导,使用方便,其估计量的BCa可信区间能够自动校正统计量的非对称性,因此更具优良性.

目的 探讨条件脆弱模型在癫痫复发事件数据中的应用及软件实现.方法 利用癫痫复发数据构建条件脆弱模型,采用惩罚偏似然函数进行参数估计、并与共享脆弱模型和Cox比例风险回归模型进行对比分析.结果 条件脆弱模型分析复发数据既考虑复发时间的非独立性也考虑了异质性问题,可以用来评价癫痫复发临床疗效,结果解释合理,软件容易实现.结论 条件脆弱模型可以较好分析复发数据,值得推广应用.

Objective To explore the applications of conditional frailty models in recurrent event data about epileptic seizures and software implement.Methods Collecting the epileptic seizures data,establishing the conditional frailty models,parameters of the estimate are calculated by penalized partial likehood function,and comparing with shared frailty models and Cox proportional hazard models.Results The conditional frailty models are best account for the heterogeneity and event dependence,we can evaluate the clinical effect of epileptic seizures data using conditional frailty models,the explain is reasonable,the implement of software is easy.Conclusion Conditional frailty models are useful for analyzing recurrent events data,it deserves to be widely applied.

Objective To investigate the stability of rRNA subunits expression levels in human breast muscle tissue after death, the application value in estimation of time since death was analyzed. Methods Left breast muscle in three adults and three infants were collected respectively in one day until the ten day, sample tissues were preserved in RNA fixative solutions, Real-time quantitative RT-PCR were used to detect rRNA subunits(5srRNA, 5. 8srRNA, 18srRNA, 28srRNA) expression levels. The relationship between age, cause of death and postmortem interval were also analyzed. Results The expresion level of rRNA subunits showed no significant degradation, changes of Ct values showed no correlation with postmortem interval after death (P>0. 05), and showed no correlation with the cause of death, age. Ct values of 5s rRNA was in the range of 18. 503±2. 655∼20. 937±2. 340, Ct values of 5. 8 srRN A was in the range of 17. 687±3. 011∼20. 617± 2. 204, Ct values of 18 srRN A was in the range of 16. 457±3. 920∼22. 330±2. 571, Ct values of 28s rRNA was in the range of 15. 077±6. 051∼22. 207±2. 68. Conclusion The expression of rRNA subunits showed good stability suitable as inferred reference gene in corrupt corpse. Age cause of death has no relation with the expression level of rRNA subunits.

马尔科夫蒙特卡洛算法(MCMC)的应用使得贝叶斯统计在过去20年得以迅速发展,已经越来越多地应用到医学和社会学等领域,本文主要非技术性地介绍MCMC算法的有关历史和涉及到的个人,以了解MCMC算法是如何进入统计学家的视线,以及MCMC算法是如何解决贝叶斯统计计算问题的.

Objective To investigate the knowledge,attitude and practice(KAP) of calcium and iron nutrition among preschool children's parents in Taiyuan,and thereby provide a basis for the further nutrition publicity and health intervention.Methods The survey was conducted through KAP questionnaire among 260 preschool children's parents in four kindergartens of Taiyuan by clustering sampling.Results The passing rates of nutrition knowledge,attitude and practice were 52.8%,92.8%,78.6%,respectively,and the excellent rates were 25.0%,82.6%,37.7%.And there was a significant difference in knowledge and practice scores among parents with different levels of education.The scores of parents with the high educational level were significantly higher than that of the parents with the low education level.In addition,the practice scores of parents with different income levels were also significantly different,but not in knowledge and practice scores.Conclusion Parents with the low education level apparently lack the knowledge to improve the health of preschool children,and the relevant departments should take targeted measures to enrich the parents' related knowledge.

ObjectiveTo explore the applications of shared gamma frailty models in recurrent event data about epileptic seizures,and implement of R software. MethodsCollecting the epileptic seizures data,establishing the shared frailty models,explaining the correlation between event times using gamma frailty. Results We can evaluate the clinical effect of epileptic seizures data using shared gamma frailty models,the explain is reasonable,the implement of R software is easy. ConclusionShared frailty models are useful for handling dependence in recurrent events data.

在数理统计中,Poisson分布有着悠久的历史,最早可追溯到1838年.对当时广泛研究的二项分布,在事件的发生概率p很小、试验次数n很大的情况下,法国数学家Poisson(1)推导出了二项分布的极限分布,为了纪念他而称为Poisson分布.其早期一个著名的应用例子是Bortkiewicz(1898)观察到普鲁士的骑兵部队中每年被马踢死的士兵数服从Poisson分布(2).

Objective To explore the methods of the parameter estimation and probability estimation in rare event logistic regression model.Methods Examples in this article,the following methods were used,including logistic regression prior correction,logistic regression weighted correction and logistic regression MCN correction.And the Vuong test was used to compare among different models.Results The logistic regression MCN weighted correction was fit model respectively.Conclusion The rare event logistic regression was superior to the classical logistic regression in the rare event analysis,it was worthy of promoted and applied for the rare disdeases.

Objective:To introduce Hurdle model and its application in medical count data with zero extra.Methods:We employed the logit-Poisson Hurdle model to the the data of the number of doctor visits.Results:The gender and age of the residents decided whether or not to visit doctor.The income,the times of illnesses in past two weeks and the number of days of reduced activity in the past two weeks due to illness or injury were the factors of how many the number of the resident visited doctor in the past two weeks.Conclusion:Hurdle model takes the zero-extra into account and could analyze zero-inflated count data efficiently.

Objective To introduce the theory of Bayesian estimation for receiver operating characteristic curve regression model.Methods To estimate related parameter of regression model using the software WinBUGS,then introduce its application in practical problems.Results The impact of covariates on diagnostic test assessment could be detected by adopting the methods of Bayesian estimation of ROC curve regression analysis.Moreover,the areas under the ROC curve with different values of covariates could be calculated.Results of the method were more stable to prior distribution in a certain range.The method was also treated as the proof of statistical analysis for clinical diagnostic test.Conclusion Based on the bayesian estimation of regression model,the ROC curves can be effectively used to solve the test accuracy evaluation problem of the clinical diagnosis trial impacted by the covariates.

目的 阐明基于贝叶斯估计的ROC曲线回归模型.方法 通过实例对比分析,介绍WinBUGS软件ROC曲线回归模型参数估计与应用.结果 基于贝叶斯估计的ROC曲线回归模型不仅可考虑(平衡)协变量对诊断试验结果准确性评价的影响,而且可计算不同协变量取值条件下的ROC曲线下面积;不同先验分布的选取在一定范围内模型参数估计结果较稳定,可作为临床诊断试验结果分析的依据.结论 基于贝叶斯估计的ROC曲线回归模型,可有效地解决受协变量影响的临床诊断试验准确度评价问题.

重复测量资料(repeated measurement data)是指对同一观察单位进行重复观察或测量所得到的资料,它以节省样本含量、资料容易收集、检验效能高等优点受到医学界科研人员的青睐.当反应变量是二分类变量时,为二分类重复测量资料,其在临床研究中非常多见,如在乳腺增生患者疗效研究中,定期记录患者治疗期间的变化,检测指标为是否有改善的二分类变量;呼吸道疾病疗效记录为是否好转的二分类变量等.