			Home About us Contact

Correlated Data (correlated + data)

Distribution by Scientific Domains

Medical Sciences	28%

Selected Abstracts

Joint Regression Analysis of Correlated Data Using Gaussian Copulas

BIOMETRICS, Issue 1 2009
Peter X.-K.
Summary This article concerns a new joint modeling approach for correlated data analysis. Utilizing Gaussian copulas, we present a unified and flexible machinery to integrate separate one-dimensional generalized linear models (GLMs) into a joint regression analysis of continuous, discrete, and mixed correlated outcomes. This essentially leads to a multivariate analogue of the univariate GLM theory and hence an efficiency gain in the estimation of regression coefficients. The availability of joint probability models enables us to develop a full maximum likelihood inference. Numerical illustrations are focused on regression models for discrete correlated data, including multidimensional logistic regression models and a joint model for mixed normal and binary outcomes. In the simulation studies, the proposed copula-based joint model is compared to the popular generalized estimating equations, which is a moment-based estimating equation method to join univariate GLMs. Two real-world data examples are used in the illustration. [source]

Low-Dose Topiramate Versus Lamotrigine in Migraine Prophylaxis (The Lotolamp Study)

HEADACHE, Issue 3 2007
Praveen Gupta MD
Objective.,To assess the efficacy and safety of topiramate and lamotrigine for prophylaxis in patients with frequent migraine as compared to each other and to placebo. Methods.,Sixty patients with frequent migraine (more than 4 attacks per month) from the headache clinic at a tertiary referral centre in India were randomized to receive 50 mg topiramate/lamotrigine or matching placebo for 1 month each in 2 divided doses in 4 phases in a crossover manner with a washout period of 7 days in between. Primary efficacy measure was responder rate (50% decrease in mean migraine frequency/intensity). Secondary efficacy measures included reduction in mean monthly frequency, intensity, duration, rescue medication use, migraine associated symptoms, and adverse events. Statistical analysis.,Analysis was on intention to treat basis. Data were analyzed as correlated data. Generalized estimation equation was used to compute overall mean standard deviation and 95% confidence intervals for each of the outcome variables. Bonferroni's correction done for multiple comparisons. P value of <.017 was taken as significant. Results.,Fifty-seven patients comprised the intent-to-treat population. Four patients withdrew from the study at various phases, none because of the side effects. Responder rate for frequency was significantly higher for topiramate versus placebo (63% vs 30%, P < .001), and versus lamotrigine (63% vs 46 %, P= .02). For intensity of headache also a responder rate of topiramate versus placebo (50% vs 10%, P < .001), and versus lamotrigine (50% vs 41%, P= .01) was observed. Topiramate showed statistically significant benefits (P < .017) in most of the secondary efficacy measures while lamotrigine was beneficial for reduction in headache frequency, and migraine associated symptoms. Adverse events were similar. Conclusion.,Low-dose topiramate is efficacious in migraine prophylaxis as compared to both placebo and lamotrigine. Lamotrigine in low doses might be beneficial for headache frequency; however, longer trials are required to establish its efficacy on the intensity and frequency of migraine. [source]

Regression modelling of correlated data in ecology: subject-specific and population averaged response patterns

JOURNAL OF APPLIED ECOLOGY, Issue 5 2009
John Fieberg
Summary 1.,Statistical methods that assume independence among observations result in optimistic estimates of uncertainty when applied to correlated data, which are ubiquitous in applied ecological research. Mixed effects models offer a potential solution and rely on the assumption that latent or unobserved characteristics of individuals (i.e. random effects) induce correlation among repeated measurements. However, careful consideration must be given to the interpretation of parameters when using a nonlinear link function (e.g. logit). Mixed model regression parameters reflect the change in the expected response within an individual associated with a change in that individual's covariates [i.e. a subject-specific (SS) interpretation], which may not address a relevant scientific question. In particular, a SS interpretation is not natural for covariates that do not vary within individuals (e.g. gender). 2.,An alternative approach combines the solution to an unbiased estimating equation with robust measures of uncertainty to make inferences regarding predictor,outcome relationships. Regression parameters describe changes in the average response among groups of individuals differing in their covariates [i.e. a population-averaged (PA) interpretation]. 3.,We compare these two approaches [mixed models and generalized estimating equations (GEE)] with illustrative examples from a 3-year study of mallard (Anas platyrhynchos) nest structures. We observe that PA and SS responses differ when modelling binary data, with PA parameters behaving like attenuated versions of SS parameters. Differences between SS and PA parameters increase with the size of among-subject heterogeneity captured by the random effects variance component. Lastly, we illustrate how PA inferences can be derived (post hoc) from fitted generalized and nonlinear-mixed models. 4.,Synthesis and applications. Mixed effects models and GEE offer two viable approaches to modelling correlated data. The preferred method should depend primarily on the research question (i.e. desired parameter interpretation), although operating characteristics of the associated estimation procedures should also be considered. Many applied questions in ecology, wildlife management and conservation biology (including the current illustrative examples) focus on population performance measures (e.g. mean survival or nest success rates) as a function of general landscape features, for which the PA model interpretation, not the more commonly used SS model interpretation may be more natural. [source]

Optimal Design of VSI ,X Control Charts for Monitoring Correlated Samples

QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL, Issue 8 2005
Yan-Kwang Chen
Abstract This paper develops an economic design of variable sampling interval (VSI),X control charts in which the next sample is taken sooner than usual if there is an indication that the process is off-target. When designing VSI,X control charts, the underlying assumption is that the measurements within a sample are independent. However, there are many practical situations that violate this hypothesis. Accordingly, a cost model combining the multivariate normal distribution model given by Yang and Hancock with Bai and Lee's cost model is proposed to develop the design of VSI charts for correlated data. An evolutionary search method to find the optimal design parameters for this model is presented. Also, we compare VSI and traditional ,X charts with respect to expected cost per unit time, utilizing hypothetical cost and process parameters as well as various correlation coefficients. The results indicate that VSI control charts outperform the traditional control charts for larger mean shift when correlation is present. In addition, there is a difference between the design parameters of VSI charts when correlation is present or absent. Copyright © 2005 John Wiley & Sons, Ltd. [source]

Course of the modified Rodnan skin thickness score in systemic sclerosis clinical trials: Analysis of three large multicenter, double-blind, randomized controlled trials

ARTHRITIS & RHEUMATISM, Issue 8 2009
Sogol Amjadi
Objective To assess the course of the modified Rodnan skin thickness score (MRSS) in 3 large, multicenter, double-blind, randomized controlled trials (RCTs) of patients with diffuse cutaneous systemic sclerosis (dcSSc) with different baseline disease durations, as defined from the date of onset of the first dcSSc symptom (excluding Raynaud's phenomenon) or from the date of onset of the first dcSSc-related symptom (including Raynaud's phenomenon). Methods Data from 3 RCTs examining high-dose versus low-dose D-penicillamine (D-Pen Trial), recombinant human relaxin versus placebo (Relaxin Trial), and oral bovine type I collagen versus placebo (Collagen Trial) treatment in patients with dcSSc were pooled and analyzed. Patients were divided into 5 groups according to their disease duration at baseline. The linear mixed model for correlated data was used to model the 2 predictors of MRSS: time in study (expressed in months after baseline) and baseline disease duration (expressed in months, calculated from the date of onset of the first symptom characteristic of dcSSc with and without Raynaud's phenomenon). Results At study entry, the mean MRSS value was 21.0 in the D-Pen Trial cohort, 27.3 in the Relaxin Trial cohort, and 26.1 in the Collagen Trial cohort. Time in study was a significant predictor of improvement in MRSS regardless of the disease duration at baseline (P < 0.0001). Patients with a disease duration of ,24 months showed a greater rate of decline as compared with patients with a disease duration of <24 months (P < 0.05). Similar results were obtained when disease duration was reclassified by including the time of the first Raynaud's phenomenon symptom in the definition. Conclusion Our study confirms recent findings that in patients entered into these 3 RCTs, skin thickening did not follow the same trend in natural history as that seen in the dcSSc populations entered into early, open longitudinal studies previously reported. These findings have important implications for study design, in which "prevention of worsening" is the main objective. [source]

Joint Regression Analysis of Correlated Data Using Gaussian Copulas

Residual-Based Diagnostics for Structural Equation Models

BIOMETRICS, Issue 1 2009
B. N. Sánchez
Summary Classical diagnostics for structural equation models are based on aggregate forms of the data and are ill suited for checking distributional or linearity assumptions. We extend recently developed goodness-of-fit tests for correlated data based on subject-specific residuals to structural equation models with latent variables. The proposed tests lend themselves to graphical displays and are designed to detect misspecified distributional or linearity assumptions. To complement graphical displays, test statistics are defined; the null distributions of the test statistics are approximated using computationally efficient simulation techniques. The properties of the proposed tests are examined via simulation studies. We illustrate the methods using data from a study of in utero lead exposure. [source]