Anemia is a common complication of chronic kidney disease (CKD) and end-stage renal disease. A high hemoglobin level targeted in the treatment of anemia has been controversial because recent overseas studies have reported that it did not affect renal survival or increased the risk of cardiovascular events. In the motivation study, patients with CKD were randomly assigned to high or low hemoglobin target group (11.0-.13.0 or 9.0-11.0 g/dL). The comparison of groups for the composite of renal events as the primary endpoint revealed no significant differences (p=0.111). In these studies, ad hoc analyses suggested that a high hemoglobin level may potentially reduce cardiovascular events. However, those results could not precisely estimate the effect of treatment with high hemoglobin because of post-treatment selection bias. To address this problem, we used the method based on principal stratification approach to estimate the causal effect of Partial Responders by which the treatment effect of high hemoglobin can be evaluated. The results suggested that not only Partial but also Always Responders may benefit more from high hemoglobin treatment than Never Responders. These data suggest that patients with CKD can receive benefit from high hemoglobin treatment, who can respond to that treatment.
In clinical investigations designed to demonstrate the efficacy of a diagnostic procedure, the procedure is usually evaluated by multiple independent raters. Although the sensitivity and specificity may be estimated by considering consensus evaluations to treat results from multiple raters as if there were a single rater, raters are not considered independent in consensus evaluations. Typically, estimation methods are based on an “average rater” or a “majority rater” to account for multiple raters. In this paper, we propose a method for summarizing sensitivities and specificities evaluated from multiple independent raters based on a bivariate random effects model (BVRM) to account between-rater variance and correlation between sensitivity and specificity. In addition, we propose methods to draw joint confidence regions of sensitivity and specificity based on the BVRM. Simulation results show that the differences in the biases between the proposed method and the average rater method are small and that the empirical coverage probabilities of the proposed joint confidence regions are close to the nominal level. The proposed methods are illustrated using data from florbetapir F 18 positron emission tomographic imaging to predict the presence of β-amyloid in the brains of subjects with Alzheimer’s disease.
In this article, we propose a strategy to show the combinability of multiple animal datasets in a parallel-line assay to estimate the relative potency. The following three assumptions are made in the linear fixed-effect modeling, and we examine if any of them result in nonconformance:
a) Intrasubject parallelism (parallel dose-response for each subject),
b) Intersubject homogeneity of the slopes of the mean response (averaged across study substances),
c) Intersubject homogeneity of the differences between intercepts.
For inferences about relative potency, a) is essential, and we derived a new metrics, intrasubject parallelism criterion (ISP), via the translation of aggregated individual bioequivalence criterion stated in the regulatory guidance (Food and Drug Administration, 2001). For b) and c), we used the 95% confidence interval of the I2 criterion, which is commonly used to evaluate the interstudy homogeneity in a meta-analysis (Higgins and Thompson, 2002). For choosing the thresholds, we applied the conventions used in the guideline.
The proposed procedure is demonstrated in an example analysis, and its properties are evaluated through a Monte Carlo simulation. The power of our proposed intrasubject parallelism criterion was shown to be high for designs of moderate size, but the demonstration of homogeneity via I2 was rather conservative.