Modified Goldilocks Design with strict type I error control in confirmatory clinical trials

If used only when a preliminary F test yields significance, the usual multiple range procedures can be modified to increase the probability of detecting differences without changing the control of Type I error. The modification consists of a reduction in the critical value when comparing the largest and smallest means. Equivalence of modified and unmodified procedures in error control is demonstrated. The modified procedure is also compared with the alternative of using the unmodified range test without a preliminary F test, and it is shown that each has advantages over the other under some circumstances.

Download Full-text

Efficient Noninferiority Testing Procedures for Simultaneously Assessing Sensitivity and Specificity of Two Diagnostic Tests

Computational and Mathematical Methods in Medicine ◽

10.1155/2015/128930 ◽

2015 ◽

Vol 2015 ◽

pp. 1-7 ◽

Cited By ~ 1

Author(s):

Guogen Shan ◽

Amei Amei ◽

Daniel Young

Keyword(s):

Sensitivity And Specificity ◽

Diagnostic Tests ◽

Error Control ◽

Type I Error ◽

Binary Outcomes ◽

Type I ◽

Test Statistics ◽

Asymptotic Approach ◽

Testing Procedures ◽

Assessing Sensitivity

Sensitivity and specificity are often used to assess the performance of a diagnostic test with binary outcomes. Wald-type test statistics have been proposed for testing sensitivity and specificity individually. In the presence of a gold standard, simultaneous comparison between two diagnostic tests for noninferiority of sensitivity and specificity based on an asymptotic approach has been studied by Chen et al. (2003). However, the asymptotic approach may suffer from unsatisfactory type I error control as observed from many studies, especially in small to medium sample settings. In this paper, we compare three unconditional approaches for simultaneously testing sensitivity and specificity. They are approaches based on estimation, maximization, and a combination of estimation and maximization. Although the estimation approach does not guarantee type I error, it has satisfactory performance with regard to type I error control. The other two unconditional approaches are exact. The approach based on estimation and maximization is generally more powerful than the approach based on maximization.

Download Full-text

Statistical Monitoring of Clinical Trials Using Brownian Bridges

International Journal of Statistics and Probability ◽

10.5539/ijsp.v6n1p142 ◽

2016 ◽

Vol 6 (1) ◽

pp. 142

Author(s):

Qiang Zhang ◽

Michael R. Kosorok

Keyword(s):

Clinical Trials ◽

Type I Error ◽

Brownian Bridge ◽

Type I ◽

Group Sequential Design ◽

Group Sequential ◽

Cancer Trial ◽

Statistical Monitoring ◽

Brownian Bridges ◽

Group Sequential Methods

The Brownian bridge is not yet used widely in the statistical monitoring of clinical trials. In this paper, we investigate properties of the Brownian bridge and formally derive monitoring rules from these results. We will present four related main methods: (1). derivation of group sequential boundaries; (2). calculation of conditional power; (3). a new alpha spending function and (4). repeated confidence intervals, all under a Brownian bridge framework. Simulation results show that the type I error rate is well controlled and power is satisfactory for the group sequential design. We apply the proposed methods to monitor the interim results from the Beta Blocker Heart Attack Trial (BHAT) and a Head and Neck cancer trial with comparisons to the commonly used monitoring tools. Overall, the proposed methods when used together as one framework are more powerful and sensitive to interim positive and negative trends that are clinically meaningful and lead to timely early stopping with potentially more savings on sample sizes, time and costs. These tools are valuable additions to the existing group sequential methods which can be utilized in trial design, routine monitoring, and to answer important questions from data monitoring committees.

Download Full-text

Statistical inference of genetic pathway analysis in high dimensions

Biometrika ◽

10.1093/biomet/asz033 ◽

2019 ◽

Vol 106 (3) ◽

pp. 651-651

Author(s):

Yang Liu ◽

Wei Sun ◽

Alexander P Reiner ◽

Charles Kooperberg ◽

Qianchuan He

Keyword(s):

Statistical Inference ◽

Pathway Analysis ◽

Genetic Variants ◽

Error Control ◽

Genome Wide Association Study ◽

Type I Error ◽

High Density Lipoproteins ◽

Type I ◽

Genetic Pathway ◽

A Genome

Summary Genetic pathway analysis has become an important tool for investigating the association between a group of genetic variants and traits. With dense genotyping and extensive imputation, the number of genetic variants in biological pathways has increased considerably and sometimes exceeds the sample size $n$. Conducting genetic pathway analysis and statistical inference in such settings is challenging. We introduce an approach that can handle pathways whose dimension $p$ could be greater than $n$. Our method can be used to detect pathways that have nonsparse weak signals, as well as pathways that have sparse but stronger signals. We establish the asymptotic distribution for the proposed statistic and conduct theoretical analysis on its power. Simulation studies show that our test has correct Type I error control and is more powerful than existing approaches. An application to a genome-wide association study of high-density lipoproteins demonstrates the proposed approach.

Download Full-text

No counts, no variance: allowing for loss of degrees of freedom when assessing biological variability from RNA-seq data

Statistical Applications in Genetics and Molecular Biology ◽

10.1515/sagmb-2017-0010 ◽

2017 ◽

Vol 16 (2) ◽

Cited By ~ 1

Author(s):

Aaron T. L. Lun ◽

Gordon K. Smyth

Keyword(s):

Software Package ◽

Error Control ◽

Degrees Of Freedom ◽

Linear Models ◽

Type I Error ◽

Real Data ◽

Type I ◽

Rna Seq ◽

Study Gene Expression ◽

Complex Models

AbstractRNA sequencing (RNA-seq) is widely used to study gene expression changes associated with treatments or biological conditions. Many popular methods for detecting differential expression (DE) from RNA-seq data use generalized linear models (GLMs) fitted to the read counts across independent replicate samples for each gene. This article shows that the standard formula for the residual degrees of freedom (d.f.) in a linear model is overstated when the model contains fitted values that are exactly zero. Such fitted values occur whenever all the counts in a treatment group are zero as well as in more complex models such as those involving paired comparisons. This misspecification results in underestimation of the genewise variances and loss of type I error control. This article proposes a formula for the reduced residual d.f. that restores error control in simulated RNA-seq data and improves detection of DE genes in a real data analysis. The new approach is implemented in the quasi-likelihood framework of the edgeR software package. The results of this article also apply to RNA-seq analyses that apply linear models to log-transformed counts, such as those in the limma software package, and more generally to any count-based GLM where exactly zero fitted values are possible.

Download Full-text

Association of human endogenous retrovirus (hERV) expression with clinical efficacy of PD-1 blockade in metastatic clear cell renal cell carcinoma (mccRCC).

Journal of Clinical Oncology ◽

10.1200/jco.2019.37.15_suppl.4568 ◽

2019 ◽

Vol 37 (15_suppl) ◽

pp. 4568-4568 ◽

Cited By ~ 1

Author(s):

Jean-Christophe Pignon ◽

Opeyemi Jegede ◽

Sachet A Shukla ◽

David A. Braun ◽

Christine Horak ◽

...

Keyword(s):

Error Control ◽

Type I Error ◽

Endogenous Retrovirus ◽

Phase Iii ◽

Human Endogenous Retrovirus ◽

Rank Test ◽

Type I ◽

Limit Of Quantification ◽

Cell Renal Cell Carcinoma ◽

Exact Test

4568 Background: hERV levels positively correlate with tumor immune infiltrate and were recently shown to be associated with clinical benefit to PD-1/PD-L1 blockade in two small cohorts of patients (pts) with mccRCC (Smith C.C. et al and Panda A. et al; 2018). We tested whether hERV levels correlate with efficacy of nivolumab in a prospective phase II study of pts with mccRCC (Checkmate 010). Methods: Reverse transcribed RNA extracted from 99 FFPE pretreatment tumors were analyzed by RT-qPCR to assess levels of pan- ERVE4, pan- ERV3.2, hERV4700 GAG or ENV, and the reference genes 18S and HPRT1. Normalized hERV levels were transformed as categorical value (high or low) using population quartiles as cutoffs. For each cutoff, samples with non-quantifiable hERV levels for which the limit of quantification was above the tested cutoff could not be categorized and were excluded from analysis. Log rank test was used to test the association of hERV levels with PFS/irPFS (RECISTv1.1/irRECIST) at each cutoff using Holm-Bonferroni correction for Type I error control; adjusted P-values are reported. Fisher’s exact test was then used to explore the association with ORR/irORR (RECISTv1.1/irRECIST). Results: Among the hERV studied, only hERV4700 ENV was significantly associated with PFS/irPFS. At the 25th percentile cutoff, 45 pts had high levels of hERV4700 ENV and 24 pts had low levels of hERV4700 ENV. Median PFS and irPFS were significantly longer in the high- hERV4700 ENV group [7.0 (95% CI: 2.2 - 10.2) and 8.5 (95% CI: 4.2 - 14.1) months, respectively] versus the low- hERV4700 ENV group [2.6 (95% CI: 1.4 - 5.4) and 2.9 (95% CI: 1.4 - 5.7) months, respectively] ( P = 0.010 for PFS and P = 0.028 for irPFS). At the same cutoff, ORR and irORR rates were significantly higher in the high- hERV4700 ENV group [35.6 (95% CI: 21.9 - 51.2) % for both ORR/irORR] versus the low- hERV4700 ENV group [12.5 (95% CI: 2.7 - 32.4) and 8.3 (95% CI: 1.0 - 27.0) %, respectively] ( P = 0.036 for ORR and P = 0.012 for irORR). Conclusions: hERV4700 ENV levels may predict outcome on nivolumab in mccRCC. Validation of our results and correlation of hERV levels with immune markers in a controlled phase III trial (CheckMate 025) is ongoing.

Download Full-text

Use of interval estimations in design and evaluation of multiregional clinical trials with continuous outcomes

Statistical Methods in Medical Research ◽

10.1177/0962280217751277 ◽

2018 ◽

Vol 28 (7) ◽

pp. 2179-2195 ◽

Cited By ~ 1

Author(s):

Chieh Chiang ◽

Chin-Fu Hsiao

Keyword(s):

Clinical Trials ◽

Sample Size ◽

Type I Error ◽

Interval Estimation ◽

Error Rates ◽

New Drugs ◽

Sample Size Determination ◽

Type I ◽

Size Determination ◽

Interval Estimators

Multiregional clinical trials have been accepted in recent years as a useful means of accelerating the development of new drugs and abridging their approval time. The statistical properties of multiregional clinical trials are being widely discussed. In practice, variance of a continuous response may be different from region to region, but it leads to the assessment of the efficacy response falling into a Behrens–Fisher problem—there is no exact testing or interval estimator for mean difference with unequal variances. As a solution, this study applies interval estimations of the efficacy response based on Howe’s, Cochran–Cox’s, and Satterthwaite’s approximations, which have been shown to have well-controlled type I error rates. However, the traditional sample size determination cannot be applied to the interval estimators. The sample size determination to achieve a desired power based on these interval estimators is then presented. Moreover, the consistency criteria suggested by the Japanese Ministry of Health, Labour and Welfare guidance to decide whether the overall results from the multiregional clinical trial obtained via the proposed interval estimation were also applied. A real example is used to illustrate the proposed method. The results of simulation studies indicate that the proposed method can correctly determine the required sample size and evaluate the assurance probability of the consistency criteria.

Download Full-text