Statistics in medicine
-
Statistics in medicine · Jul 2001
A generalized concordance correlation coefficient for continuous and categorical data.
This paper discusses a generalized version of the concordance correlation coefficient for agreement data. The concordance correlation coefficient evaluates the accuracy and precision between two measures, and is based on the expected value of the squared function of distance. ⋯ We also introduce a stratified concordance correlation coefficient which adjusts for explanatory factors, and an extended concordance correlation coefficient which measures agreement among more than two responses. With these extensions, the generalized concordance correlation coefficient provides a unifying approach to assessing agreement among two or more measures that are either continuous or categorical in scale.
-
Statistics in medicine · Jul 2001
Group sequential test strategies for superiority and non-inferiority hypotheses in active controlled clinical trials.
In a group sequential active controlled clinical trial, the study hypothesis may be a superiority hypothesis that an experimental treatment is more effective than the active control therapy or a non-inferiority hypothesis that the treatment is no worse than the active control within some non-inferiority range. When it is necessary to plan for testing the superiority and the non-inferiority hypotheses, we propose an adaptive group sequential closed test strategy by which the sample size is planned for testing superiority and is to be increased for showing non-inferiority given that it is deemed more plausible than superiority based on the observed sample path during the course of the trial. The proposed adaptive test strategy is valid in terms of having the type I error probability maintained at the targeted alpha level for both superiority and non-inferiority. It has power advantage or sample size saving over the traditional group sequential test designed for testing either superiority only or non-inferiority only.
-
There are three approaches to health economic evaluation for comparing two therapies. These are (i) cost minimization, in which one assumes or observes no difference in effectiveness, (ii) incremental cost-effectiveness, and (iii) incremental net benefit. ⋯ Furthermore, if costs and effectiveness are not censored, this can be achieved using common two-sample statistical procedures. The above will be illustrated using two examples, one with censoring and one without.
-
Statistics in medicine · May 2001
Comparative StudyLogistic regression when binary predictor variables are highly correlated.
Standard logistic regression can produce estimates having large mean square error when predictor variables are multicollinear. Ridge regression and principal components regression can reduce the impact of multicollinearity in ordinary least squares regression. ⋯ Recommendations for choosing among standard, ridge and principal components logistic regression are developed. Published in 2001 by John Wiley & Sons, Ltd.
-
Missing data in public health research is a major problem. Mean or median imputation is frequently used because it is easy to implement. Although multiple imputation has good statistical properties, it is not yet used extensively. ⋯ The simulation showed large differences among various multiple imputation methods with a different number of variables for creating the matching metric for multiple imputation. Multiple imputation using only a few covariates in the matching model produced more biased coefficient estimates than using all available covariates in the matching model. The simulation also showed better standard deviation estimates for multiple imputation than for single mean imputation.