Analysis of matched case-control data with multiple ordered disease states: possible choices and comparisons

In an individually matched case-control study, effects of potential risk factors are ascertained through conditional logistic regression (CLR). Extension of CLR to situations with multiple disease or reference categories has been made through polychotomous CLR and is shown to be more efficient than carrying out separate CLRs for each subgroup. In this paper, we consider matched case-control studies where there is one control group, but there are multiple disease states with a natural ordering among themselves. This scenario can be observed when the cases can be further classified in terms of the seriousness or progression of the disease, for example, according to different stages of cancer. We explore several popular models for ordered categorical data in this context. We first adopt a cumulative logit or equivalently, a proportional-odds model to account for the ordinal nature of the data. The important distinction of this model from a stratified dichotomous and polychotomous logistic regression model is that the stratum-specific nuisance parameters cannot be eliminated in this model via the conditional-likelihood approach. We discuss a Mantel-Haenszel approach for analysing such data. We point out possible difficulties with standard likelihood-based approaches with the cumulative logit model when applied to case-control data. We then consider an alternative conditional adjacent-category logit model. We illustrate the methods by analysing data from a matched case-control study on low birthweight in newborns where infants are classified according to low and very low birthweight and a child with normal birthweight serves as a control. A simulation study compares the different ordinal methods with methods ignoring sub-classification of the ordered disease states.     

非吸烟女性肺癌危险因素的多分类logistic模型分析

目的 利用多分类logistic模型分析和评价非吸烟女性不同组织学类型肺癌的危险因素。方法 资料来源于非吸烟女性肺癌病例504例及人群对照601例，拟合多分类logistic回归模型估计各因素的比数比。结果 影响女性腺癌发病的危险因素是厨房小环境污染、经常食用菜油、BMI较低、一级亲属肺癌家族史，而保护因素是活产次数较多和胡罗卜素。女性鳞癌的危险因素则是厨房小环境污染和一级亲属肺癌家族史，保护因素是饮茶。女性其他类型肺癌危险因素主要是油炸烹饪次数较多等，而类型不明肺癌发病的危险因素类似于腺癌。结论 不同组织学类型的非吸烟女性肺癌病因可能存在一定的差异。     

Methodology for the differential diagnosis of a complex data set. A case study using data from routine CT scan examinations

Data from routine CT scan examinations are employed to illustrate the use of the polychotomous logistic regression model as a statistical diagnostic tool. The assumptions of the model, the interpretation of its parameters, and its capabilities are described in detail. In carrying out the analysis on the CT data, a large, relatively sparse data set, many technical difficulties were encountered. Modifications to the methodology that were necessary to permit its implementation are described, and it is demonstrated that an unbiased analysis of T + 1 diagnostic categories can be implemented by separately performing T individual simple logistic analyses. The limitations of the methodology are discussed. It is hoped that this paper may serve as a basis for the practical implementation of the polychotomous logistic model in similar diagnostic settings.     

Regressive logistic models for ordered and unordered polychotomous traits: application to affective disorders

Regressive models are extended to disease phenotypes with two or more affection classes through the use of polychotomous logistic regression. The classes of affection may be ordered (ranked as on a liability continuum), or unordered. Data on affective disorders are used for illustration.     

Statistical methods for the evaluation of diagnostic measurements concerning paired organs

Diagnostic measurements involving paired organs may be modelled using marginal or conditional approaches. Relevant information could be ignored by using marginal models with a mean structure not including information from the fellow organ. The conditional Rosner model can be generalized to a symmetric polychotomous logistic regression model. The rather strong assumptions of the Rosner model are relaxed in this symmetric model. The use of the symmetric model is justified by its relation to discriminant analysis. Additionally, the information of the diagnostic measurements from the fellow organ is taken into account explicitly. Furthermore, in case-control studies this approach corresponds to the sampling scheme. When the status of the fellow organ and the diagnostic measurement of this organ both are available, a surprising effect may be observed; pathological values of the diagnostic measurement from the fellow organ decrease the disease probability of the selected organ. As an example, four different diagnostic procedures from the Erlangen Glaucoma Registry are investigated.     

Polychotomous logistic regression methods for matched case-control studies with multiple case or control groups

Two statistical methods, a polychotomous and pairwise approach, are presented to derive estimates of the relative odds in a matched case-control design when multiple case or control groups are used. Test statistics are derived to determine if the relative odds between groups are different. The polychotomous method is limited to case-control sets, i.e., where data are available on all members of a matched set. In contrast, the pairwise method makes use of data from both complete and incomplete sets. Nonetheless, efficiency calculations show that the polychotomous logistic regression model is more efficient even when 40 per cent of the case-control sets are incomplete. An example using a single dichotomous variable is provided.     

Statistical methods for relating several exposure factors to several diseases in case-heterogeneity studies

Statistical methods for characterizing risks are now well-developed, although little attention has focussed on the problem of risk identification, as in the surveillance of adverse drug reactions or occupational cancers. In this paper we consider the analysis of studies in which one ascertains and compares cases of several disease groups in terms of exposure histories. We address the problem of adjusting each risk for the confounding effects of all the other risks in the data. Data analysis consists of multidimensional contingency tables or polychotomous logistic regression. The latter approach focuses attention on the exposure-disease relations of primary interest rather than on those among the exposure factors and higher-order interactions, and applies easily to many exposure variables and to continuous exposure variables. We describe a stepwise approach to selecting effects for inclusion in the model. Application to preliminary data from a study aimed at identification of hitherto unsuspected occupational carcinogens illustrates the general approach.     

Finding a method for optimizing risk adjustment when comparing surgical-site infection rates.

OBJECTIVE: To investigate whether stratification of the risk of developing a surgical-site infection (SSI) is improved when a logistic regression model is used to weight the risk factors for each procedure category individually instead of the modified NNIS System risk index. DESIGN AND SETTING: The German Nosocomial Infection Surveillance System, based on NNIS System methodology, has 273 acute care surgical departments participating voluntarily. Data on 9 procedure categories were included (214,271 operations). METHODS: For each of the procedure categories, the significant risk factors from the available data (NNIS System risk index variables of ASA score, wound class, duration of operation, and endoscope use, as well as gender and age) were identified by multiple logistic regression analyses with stepwise variable selection. The area under the receiver operating characteristic (ROC) curve resulting from these analyses was used to evaluate the predictive power of logistic regression models. RESULTS: For most procedures, at least two of the three variables contributing to the NNIS System risk index were shown to be independent risk factors (appendectomy, knee arthroscopy, cholecystectomy, colon surgery, herniorrhaphy, hip prosthesis, knee prosthesis, and vascular surgery). The predictive power of logistic regression models (including age and gender, when appropriate) was low (between 0.55 and 0.71) and for most procedures only slightly better than that of the NNIS System risk index. CONCLUSION: Without the inclusion of additional procedure-specific variables, logistic regression models do not improve the comparison of SSI rates from various hospitals.     

四分类logistic回归模型对肺癌分组织类型的多因素研究

logistic回归模型已经广泛用于医学研究中,但多分类logistic回归模型应用较为少见。本文建立了四分类logistic回归模型,分析了云南锡矿公司126例病例-对照资料,提示了鳞癌与氡暴露有一定关系,其相对危险度为1.7671, P<0.01;鳞癌发病年龄较其他组织类型肺癌高;吸烟对腺癌的危险性较其他类型低。     

Predictors of atherosclerosis in the Honolulu Heart Program. II. Adjustment for autopsy bias

Although autopsy is considered the final word on many medical questions, there has long been concern over possible bias in inference about a living population from analysis of autopsy material. Focus of the present paper is on the relationship between results obtainable only at autopsy and risk factors recorded as part of a prospective study of the entire "target population." Since pathologies are sure to be overrepresented in an autopsy sample compared to the target population, the dependence of autopsy scores upon risk factors may be distorted in the autopsy sample. The present paper proposes a method of adjustment for this bias. When both autopsy sample and target population can be stratified by major disease categories, under certain assumptions of equal effect, adjustment similar to the direct method for age adjustment may be applied. If, in addition, dependence can be characterized accurately by linear regression of both autopsy score and disease category frequency onto risk factors, then a very convenient calculation produces adjusted regression coefficients. This "parametric" method usually provides the most convenient results, with the greatest statistical power.     

围产儿先天性心脏病危险度预测模型

通过对先天性心脏病(CHD)影响因素的分析,建立胎儿CHD危险度预测模型.采用单因素logistic回归分析筛选影响因素后用多因素非条件logistic回归和决策树法建立胎儿CHD危险度预测模型,分析比较两种预测方法 的优势与不足.实例分析表明,logistic回归模型和决策树模型对215例训练样本和55例测试样本的分类正确率分别为80.93%、82.79%和85.45%、89.09%.将logistic回归和决策树方法 联合应用,不仅能提高预测的准确率,还能克服因素间共线性的影响,从而保证分析的准确和完善.     

Correlates of Condom Use and Number of Sexual Partners Among High School Adolescents

ABSTRACT: Factors associated with condom use and number of sexual partners were examined in a statewide sample of public high school students in grades 9–12 (N = 3,893). Data were collected in spring 1990 using the 70-item, self-report Youth Risk Behavior Survey, developed and piloted by the Centers for Disease Control. Composite scores were constructed to measure aggression, physical recklessness, alcohol use, illegal drug use, cigarette use, lack of exercise, and academic self-image. Since simple polychotomous logistic regression models revealed a significant race by gender interaction, multivariate models were run separately for each race-gender group. Odds ratios and 95% confidence intervals were calculated from polychotomous logistic regression of lifetime sexual activity and condom use with their potential correlates. Risky sexual behavior appears to be correlated with a complex of other behaviors that place students at risk. A pattern of declining condom use with increasing number of partners was evident, especially for White students.     

高龄老年H型高血压患者靶器官损害的研究分析

H型高血压定义为伴有血浆同型半胱氨酸(Hcy)水平升高的原发性高血压.目前鲜有高龄老年人群H型高血压对靶器官损害的文献报道,为此本研究探讨高龄老年H型高血压与靶器官损害的关系. 1.对象与方法: (1)研究对象:选择2010年9月至2012年8月在滨州医学院附属医院老年医学科住院高血压患者1024例,其中年龄＞60岁者934例.     

Risk factors for monoinfections and coinfections with HIV,hepatitis B and hepatitis C viruses in northern Spanish prisoners.

A cross-sectional study was conducted in prisons of Cantabria (northern Spain) from June 1992 to December 1994. Inmates were asked to participate in a survey on prevalence and risk factors for monoinfections and coinfections with HIV, HBV and HCV. Crude and multiple odds ratios of risk factors were calculated (by polychotomous logistic regression). Prevalence of coinfections was higher than that of monoinfections. IDU risk factors were the main independent variables associated with monoinfections and coinfections with these agents. The strength of association increased with the degree of coinfection for IDU risk factors and penal status, e.g. duration of injecting drug use for more than 5 years yielded an adjusted OR ranging from 1.3 (95% CI: 0.4-5.1) for HBV monoinfection to 180 (95% CI: 61.0-540.0) for HIV-HBV-HCV coinfection. In comparison, sexual behaviours were less important than IDU risk factors.     

An insight on the use of multiple logistic regression analysis to estimate association between risk factor and disease occurrence

Multiple logistic regression is an accepted statistical method for assessing association between an anticedant characteristic (risk factor) and a quantal outcome (probability of disease occurrence), statistically adjusting for potential confounding effects of other covariates. Yet the method has potential drawbacks which are not generally recognized. This article considers one important drawback of logistic regression. Specifically the so-called main effect logistic model assumes that the probability of developing disease is linearly and additively related to the risk factors on the logistic scale. This assumption stipulates that for each risk factor, the odds ratio is constant over all reference exposure levels, and that the odds ratio exposed to two or more factors is equal to the product of individual risk factor odds ratios. If the observed odds ratios in the data follow this pattern, the model-predicted odds ratios will be accurate, and the meaning of the odds ratio for each risk factor will be straightforward. But if the observed odds ratios deviate from the model assumption, the model will not fit the data accurately, and the model-predicted odds ratios will not reflect those in the data. Although satisfactory fit can always be achieved by adding to the model polynomial and product terms derived from the original risk factors, the odds ratios estimated by such an interaction logistic model are difficult to interpret, viz., the odds ratio for each risk factor depends not only on the reference exposure levels of that factor, but also on the exposure level in other factors.(ABSTRACT TRUNCATED AT 250 WORDS)     

Multivariate methods for binary longitudinal data with heterogeneous correlation over time.

Clustered binary data occur frequently in biostatistical work. One particular application is in binary longitudinal data, where several visits are available for the same individual. Several approaches have been proposed for the analysis of clustered binary data. In Rosner, a polychotomous logistic regression model was proposed which is a generalization of the beta-binomial distribution and allows for person- and visit-specific covariates, while controlling for clustering effects. One assumption of this model is that all pairs of visits within an individual are equally correlated, which may be inappropriate if several visits are available over a long follow-up period. In this paper, this approach is extended to allow for heterogeneous correlation over time. The total time period is divided into subintervals and a beta-binomial mixture model is introduced to estimate odds ratios relating outcomes for pairs of visits both within a subinterval as well as in different subintervals. To include covariates, an extension of the polychotomous logistic regression model is proposed, which allows one to estimate effects of person-, subinterval-, and visit-specific covariates, while controlling for clustering using the beta-binomial mixture model. This model is applied to the analysis of respiratory symptom data in children collected over a 14-year period in East Boston, MA, in relation to maternal and child smoking, where the unit is the child and symptom history is divided into early-adolescent and late-adolescent symptom experience.     

二分类 Logistic 回归在冠心病危险因素研究中的应用

目的：应用二分类Logistic回归在可能引发冠心病影响因素中筛选危险因素，建立冠心病危险因素“最优”回归方程。方法采取系统抽样方法，对某几所医院心血管内科初诊为冠心病并进行冠脉造影病例，抽取400例30～65岁患者病例作为样本。通过二分类Logistic回归方法分析冠心病与危险因素的相关关系。结果以是否冠心病为因变量，各因素为自变量，筛选影响因素。经相关分析、共线性诊断，筛选出冠心病危险因素为年龄、合并疾病、吸烟、收缩压、血糖、尿酸、低密度脂蛋白、脂蛋白（ a）。结论用二分类Logistic回归找出危险因素，可以有效分析各因素的相对重要性。     

Alternative parameterization of polychotomous models: theory and application to matched case-control studies

A method is proposed for transforming a class of models having an outcome variable with more than two levels into an equivalent binary model. The polychotomous logistic model is used to demonstrate the method. The equivalency to a simple logistic regression model after some data transformation (augmentation) is shown. The method is applied to the data from two case-control studies each with two control groups, and further applications are indicated.     

Regression modeling of consumption or exposure variables classified by type

Consumption or exposure variables, as potential risk factors, are commonly measured and related to health effects. The measurements may be continuous or discrete, may be grouped into categories and may, in addition, be classified by type. Data analyses utilizing regression methods for the assessment of these risk factors present many problems of modeling and interpretation. Various models are proposed and evaluated, and recommendations are made. Use of the models is illustrated with Cox regression analyses of coronary heart disease mortality after 24 years of follow-up of subjects in the Framingham Study, with the focus being on alcohol consumption among these subjects.