稀有事件logistic回归在医学研究中的应用

目的 探讨稀有事件logistic回归模型的参数估计及概率估计方法.方法 在普通logistic回归结果的基础上进行先验校正、加权校正和MCN校正,并根据Vuong检验原理编程实现非嵌套模型间的对比分析.结果 logistic 回归MCN加权校正对实例数据拟合较好.结论 针对稀有事件数据的分析,在模型的参数估计和预测预报方面,稀有事件logistic回归结果优于普通logistic回归,是一种值得推广的应用统计模型.     

重复观测数据的半参数回归分析

目的 研究重复预测数据的回归分析技术。方法 利用半参数回归分析的原理与方法，结合重复观测数据的特点，建立重复观测数据的半参数回归模型，并进一步讨论模型参数的估计方法及假设检验公式。结果 讨论了重复观测数据的半参数回归模型的模型误差，分析了重复因素的效应及参数的影响，给出了其模型的方差分析表。结论 通过实例分析，表明对重复观测数据的处理，半参数回归分析的效果优于普通的最小二乘法和广义最小二乘法。     

基于Logistic回归模型分析影响产科住院分娩费用的因素

目的探索基于Logistic回归模型对产科住院分娩费用的分析运用方法。方法通过选取2007～2010年产科分娩病例作为研究对象,观察产科分娩费用的影响因素。以Logistic回归模型(多水平模型)对产科住院分娩费用进行分析处理,第一层以住院患者个体,第二层以医疗付款方式,拟合两水平模型对费用影响因素分析。结果 Logistic回归模型适用于分析评价住院费用类的数据。结论产科分娩费用的分析评价基于Logistic回归模型的研究结果会更科学。     

省级中医机构财政拨款预测模型应用分析

[目的]探讨省级中医机构财政拨款预测建模办法,为中医类医院预算管理提供参考依据。[方法]基于全国中医药统计摘编1987-2016年中医机构财政拨款中全国及湖北省有关数据,探讨回归、ARIMA及指数平滑等3类多个模型,通过比较3类最优模型预测值与观测值的组内相关系数(intraclass correlation co-efficient,ICC),选择ICC值最大者为最终最优模型。[结果]最终最优回归预测模型表示为:湖北省中医机构财政拨款=-138.279+0.028×全国中医机构财政拨款。R 2为0.982,预测值与观测值之间的双向混合效应模型单个测量绝对一致ICC值为0.991。[结论]模型数据与实际数据拟合度高,选用的线性回归模型对湖北省中医机构财政拨款预测具有可行性,对该省中医类医院预算管理以及对其他省份的预测参考具有重要意义。     

ICU患者医院感染logistic回归模型预测

目的构建ICU患者医院感染logistic回归预测模型,并对模型进行评价。方法以入住ICU>48 h的患者为研究对象,构建医院感染logistic回归模型,对模型进行拟和优度检验、ROC曲线下面积分析。结果入住ICU天数、气管插管、前列腺肥大、动静脉插管、基础疾病(肿瘤)等变量进入logistic回归方程,模型ROC曲线下面积为0.856。结论 logistic回归模型对ICU患者医院感染预测拟合度较好。     

三种时间序列模型探讨1989～2012广西肺结核发病趋势

目的 探讨不同时间序列预测模型分析肺结核发病趋势的可行性,评价模型的拟合度,为制定肺结核预防控制策略提供科学依据.方法 收集广西壮族自治区1989 ～ 2009年肺结核疫情报告数据,分别采用曲线回归法、指数平滑法和ARIMA模型模拟肺结核疫情的动态轨迹,比较三种方法的拟合精度,评价拟合及外推效果.结果 1989～ 2007年广西肺结核发病率逐年增高,2007年后逐渐降低.其变动轨迹,以ARIMA模型拟合效果较好(拟合度R2=0.84),三种模型均具有一定的外展预测能力.结论 ARIMA模型适用于拟合类似肺结核发病率的动态趋势,在传染病疫情预测领域具有重要的现实意义.     

湖南省手足口病普通病例EV71阳性率与重症率的曲线回归分析

目的 探索湖南省2011年1月~2015年12月手足口病普通病例EV71阳性率与重症率之间的关联。方法 应用SPSS 18.0对2011-2014年的肠道病毒71型（enterovirus,EV71）阳性率与重症率建立直线回归模型和曲线回归模型,并用2015年的数据验证模型,应用组内相关系数（intraclass correlation co-efficient,ICC）评价观测值与预测值的一致性。结果 湖南省手足口病普通病例EV71阳性率和重症率的关联,用三次曲线回归模型的拟合优度最高（校正的R2=0.687）,二次曲线回归模型次之（校正的R2=0.594）,直线回归模型最差（校正的R2=0.420）。三次曲线回归模型对重症率的预测值与观测值之间的两因素混合效应模型单个测量绝对一致ICC值为0.497。结论 利用三次曲线回归模型,用普通病例EV71阳性率预测重症率所得预测值与实际值的一致性一般。     

偏最小二乘回归的原理及应用

[目的]探讨偏最小二乘回归的理论及其应用。[方法]应用医学实例计算,对偏最小二乘回归与主成分回归及一般最小二乘回归进行比较。[结果]偏最小二乘回归对数据的拟合度优于主成分回归和一般最小二乘回归法。[结论]偏最小二乘回归适用于处理有多重共线性的资料,当解释变量个数多、样本量少时尤为有效,是稳健的数据“软”建模的统计方法。     

多种数据模型在手足口病发病预测的应用探讨

目的 应用多种数学模型拟合出手足口病发病最优模型,预测手足口病发病趋势,为疾病防控提供科学依据。 方法 采用SPSS 18.0软件建立差分自回归移动平均模型(autoregressive integrated moving average,ARIMA模型)、差分自回归移动平均模型与多层感知神经网络组合模型(autoregressive integrated moving average and multilayer perceptron, ARIMA-MLP模型)和差分自回归移动平均模型与径向基函数神经网络组合模型(autoregressive integrated moving average and radial basis function , ARIMA-RBF模型),分别对手足口病发病情况进行拟合,通过对三种模型比较,找到预测最优模型。 结果 ARIMA模型的拟合度R²和平均绝对误差MAE值分别为0.725、2.672,ARIMA-MLP模型为0.724、2.672,ARIMA-RBF模型为0.801、2.206。 结论 ARIMA-RBF模型的拟合度最大、平均绝对误差最小模型,拟合效果优于其他两种模型。     

多水平分析在调查问卷结果分析中的应用北大核心CSCD

目的探讨应用两水平分层logistic回归模型分析调查问卷二分类结局变量资料的影响因素,对影响因素做出正确的评价和解释。方法以学生的调查问卷作为分析对象,运用两水平分层logistic回归模型,拟合一系列零模型、随机截距模型,识别学生水平和学校水平因素的影响大小。结果学生对所研究问题的看法受到学生个体特征和学校特征的影响,学生层面中学生成绩影响较大。结论多水平分析方法是处理分层嵌套数据的有用工具,利用两水平分层logistic回归模型可以同时探讨学生层面和学校层面解释变量对结局变量的效应,对于分析嵌套数据结构资料结局变量的影响因素有优越性。     

Multivariate multilevel spline models for parallel growth processes: application to weight and mean arterial pressure in pregnancy

Growth models are commonly used in life course epidemiology to describe growth trajectories and their determinants or to relate particular patterns of change to later health outcomes. However, methods to analyse relationships between two or more change processes occurring in parallel, in particular to assess evidence for causal influences of change in one variable on subsequent changes in another, are less developed. We discuss linear spline multilevel models with a multivariate response and show how these can be used to relate rates of change in a particular time period in one variable to later rates of change in another variable by using the variances and covariances of individual‐level random effects for each of the splines. We describe how regression coefficients can be calculated for these associations and how these can be adjusted for other parameters such as random effect variables relating to baseline values or rates of change in earlier time periods, and compare different methods for calculating the standard errors of these regression coefficients. We also show that these models can equivalently be fitted in the structural equation modelling framework and apply each method to weight and mean arterial pressure changes during pregnancy, obtaining similar results for multilevel and structural equation models. This method improves on the multivariate linear growth models, which have been used previously to model parallel processes because it enables nonlinear patterns of change to be modelled and the temporal sequence of multivariate changes to be determined, with adjustment for change in earlier time periods. Copyright © 2012 John Wiley & Sons, Ltd.     

Misuse of the linear mixed model when evaluating risk factors of cognitive decline

The linear mixed model (LMM), which is routinely used to describe change in outcomes over time and its association with risk factors, assumes that a unit change in any predictor is associated with a constant change in the outcome. When it is used on psychometric tests, this assumption may not hold. Indeed, psychometric tests usually suffer from ceiling and/or floor effects and curvilinearity (i.e., varying sensitivity to change). The authors aimed to determine the consequences of such misspecification when evaluating predictors of cognitive decline. As an alternative to the LMM, they considered 2 mixed models based on latent processes that handle discrete and bounded outcomes. Model differences are illustrated here using data on 4 psychometric tests from the Personnes Agées QUID (PAQUID) Study (1989-2004). The type I error of the Wald test for risk-factor regression parameters was formally assessed in a simulation study. It demonstrated that type I errors in the LMM could be dramatically inflated for some tests, such that spurious associations with risk factors were found. In particular, confusion between effects on mean level and effects on change over time was highlighted. The authors recommend use of the alternative mixed models when studying psychometric tests and more generally quantitative scales (quality of life, activities of daily living).     

Peut-on changer les comportements alimentaires ?

The naïve model of feeding behaviour determination, stating it stems from individual conscious will alone, is widespread but fallacious and roots many a failure in diets and public health campaigns.Feeding behaviour is determined at three levels: technico-economics, psychology, and sociology. Individual capacity of change, facing a whole system of affordances and habits, is limited. Changes are possible, but not when relying upon individual will alone. We list some realistic paths to foster and support change.     

Correlates of fat intake behaviors in participants in the eating for a healthy life study

OBJECTIVE: This study identified socioeconomic correlates of total dietary fat intake in the Eating for a Healthy Life Study, a community-based dietary intervention study involving religious organizations. SUBJECTS/DESIGN: To create our sampling frame for the Eating for a Healthy Life Study, a pool of religious organizations was identified from a list of religious faith organizations provided by the Church Council of Greater Seattle. Individual members were randomly selected from recruited religious organizations to complete a telephone-administered, baseline, cross-sectional survey. There were 2,507 respondents who were eligible, consenting participants. Dietary behavior was assessed using a modified version of the Fat- and Fiber-Related Diet Behavior Questionnaire. Socioeconomic status was assessed using individual demographic variables. MAIN OUTCOME MEASURES: Fat-related behaviors indicated by fat summary score and stage of change. STATISTICAL ANALYSES: A multivariable linear regression model was used to examine the association of individual demographic variables with the Fat- and Fiber-Related Diet Behavior Questionnaire-derived measure of dietary fat intake. The same individual variables were used in a multivariable logistic regression model of dietary stage of change. RESULTS: Age, race, sex, education, and self-assessed health status were found to be statistically significant correlates of fat intake. Variables associated with stage of dietary fat change included sex, education, and religious organization cohesion. CONCLUSIONS: Demographic variance is an important factor in understanding dietary fat intake.     

Patterns over time in quality of life, coping and psychological adjustment in late stage melanoma patients: An application of multilevel models

Background: Quality of life (QOL) assessment is an important component within cancer research. There is often variability in QOL scores both between patients and across time. Understanding this variability in terms of personal characteristics and psychosocial factors would be useful but is often obstructed by the types of analyses that are applied to longitudinal data sets. Improved understanding can be gained with the application of multi-level or hierarchical models that allow for greater flexibility for modelling individual patterns of change over time. Methods and patients: Questionnaires were sent to a cohort of stage IV melanoma patients seen at the Sydney Melanoma Unit between 1991 and 1996, approximately every 3 months for up to 2 years. The data reported here are from a sub-sample of 44 patients who each completed between 3 and 8 questionnaires. Three aspects of QOL were measured (effort to cope, mood and physical well-being), each with a single LASA line. Multilevel techniques were used to model the patterns of QOL over time. Covariates were added to explain variation between patients in their average QOL and change in QOL over time. Results: The scores of each of the three QOL measures showed marked fluctuations over time. However, there was little systematic change during the study in either effort to cope ( p = 0.32) or mood ( p = 0.06). In contrast, the physical well-being scores of some patients improved while others deteriorated ( p < 0.001). On average, physical well-being deteriorated ( p < 0.001). Variability between patients accounted for 60% (effort to cope), 45% (mood) and 44% (physical well-being) of the total variance of each scale. A range of psycho-social factors including active and avoidant coping styles and psychological adjustment accounted for significant amounts of the variability between patients in each QOL measure. Conclusion: Individual coping and psychological adjustment are related to individual changes in QOL and to differences among patients' QOL. The study illustrates the use of multi-level techniques to further our understanding of differences between patients in their QOL and how it changes over time.     

苏州市中小学女生超重肥胖与月经初潮年龄提前的相关性研究

目的:分析苏州市中小学女生超重肥胖与月经初潮年龄提前的相关性,为预防中小学女生月经初潮年龄提前提供依据.方法:采用分层整群随机抽样法抽取苏州市2个行政区内12所学校的1472名中小学女生为调查对象,收集调查对象的相关资料,评估其营养状况.采用多因素logistic回归模型分析中小学女生超重、肥胖与月经初潮年龄提前的相关...     

Motivierende Gespr?chsf��hrung in der Adipositas-Therapie von Jugendlichen

Group and individual counseling sessions play a key role in adolescent obesity treatment. The requirements of these training sessions are defined in national guidelines, e.g., the German "Adipositas-Konsensuspapier" (Obesity Consensus Paper). Literature research between 2005 and 2010 showed that the long-term effects of these sessions were low, based on five reviews considering obesity treatment and six reviews dealing with motivational interviewing (MI). The reviews underline high attrition and dropout rates associated with decreasing compliance. Individual requirements for resilient motivation and relapse prevention are described with the help of the transtheoretical model of behavioral change and the self-control model. By integrating MI into group and individual counseling sessions, behavioral change can be encouraged over the long term. The stages of change can be ascertained and coping strategies of three different risk situations can be established. MI seems to be helpful especially at the beginning and end of treatment, during the state of maintenance.     

Stages of change for fruit and vegetable consumption in deprived neighborhoods.

This article describes the association of external and psychosocial factors on the stages of change for fruit and vegetable consumption, among 2,781 inhabitants, aged 18 to 65 years, in deprived neighborhoods (response rate 60%). To identify correlates of forward stage transition, an ordinal logistic regression model, the Threshold of Change Model (TCM), was used to analyze the data. The results indicate that men, those from Dutch origin, those with (rather) low health locus of control, those displaying risk behaviors, and those without knowledge of dietary guidelines were less likely to move from one stage to another and therefore were more likely to be in a lower stage-of-change category. The results make it possible to distinguish target groups, which should receive more attention in future health promotion campaigns, and to identify factors to be addressed in those programs.     

Comparison of case-deletion diagnostic methods for Cox regression

Case-deletion diagnostics are a routine component of regression analysis since they identify unusual observations that substantially affect parameter estimates. The exact approach is to compute the change in each regression parameter by dropping that individual and refitting the model. Repeating a Cox regression for the removal of each individual is very time consuming and therefore not done in practice. The two methods commonly used to approximate the exact case-deletion change for Cox regression are the empirical influence function approach and the covariate-vector augmentation approach. This paper reports the results of a simulation study on how well these methods estimate the exact change in a parameter estimate when deleting a known outlier or a known non-outlier. Additionally, we investigate how well these methods correctly identify outliers and non-outliers. The covariate augmentation approach clearly outperformed the influence function approach in these simulations.