Pharmaceutical Statistics最新文献

Flexible machine learning algorithms are increasingly utilized in real-world data analyses. When integrated within double robust methods, such as the Targeted Maximum Likelihood Estimator (TMLE), complex estimators can result in significant undercoverage-an issue that is even more pronounced in singly robust methods. The Double Cross-Fitting (DCF) procedure complements these methods by enabling the use of diverse machine learning estimators, yet optimal guidelines for the number of data splits and repetitions remain unclear. This study aims to explore the effects of varying the number of splits and repetitions in DCF on TMLE estimators through statistical simulations and a data analysis. We discuss two generalizations of DCF beyond the conventional three splits and apply a range of splits to fit the TMLE estimator, incorporating a super learner without transforming covariates. The statistical properties of these configurations are compared across two sample sizes (3000 and 5000) and two DCF generalizations (equal splits and full data use). Additionally, we conduct a real-world analysis using data from the National Health and Nutrition Examination Survey (NHANES) 2017-18 cycle to illustrate the practical implications of varying DCF splits, focusing on the association between obesity and the risk of developing diabetes. Our simulation study reveals that five splits in DCF yield satisfactory bias, variance, and coverage across scenarios. In the real-world application, the DCF TMLE method showed consistent risk difference estimates over a range of splits, though standard errors increased with more splits in one generalization, suggesting potential drawbacks to excessive splitting. This research underscores the importance of judicious selection of the number of splits and repetitions in DCF TMLE methods to achieve a balance between computational efficiency and accurate statistical inference. Optimal performance seems attainable with three to five splits. Among the generalizations considered, using full data for nuisance estimation offered more consistent variance estimation and is preferable for applied use. Additionally, increasing the repetitions beyond 25 did not enhance performance, providing crucial guidance for researchers employing complex machine learning algorithms in causal studies and advocating for cautious split management in DCF procedures.

The creation of the ICH E9 (R1) estimands framework has led to more precise specification of the treatment effects of interest in the design and statistical analysis of clinical trials. However, it is unclear how the new framework relates to causal inference, as both approaches appear to define what is being estimated and have a quantity labeled an estimand. Using illustrative examples, we show that both approaches can be used to define a population-based summary of an effect on an outcome for a specified population and highlight the similarities and differences between these approaches. We demonstrate that the ICH E9 (R1) estimand framework offers a descriptive, structured approach that is more accessible to non-mathematicians, facilitating clearer communication of trial objectives and results. We then contrast this with the causal inference framework, which provides a mathematically precise definition of an estimand and allows the explicit articulation of assumptions through tools such as causal graphs. Despite these differences, the two paradigms should be viewed as complementary rather than competing. The combined use of both approaches enhances the ability to communicate what is being estimated. We encourage those familiar with one framework to appreciate the concepts of the other to strengthen the robustness and clarity of clinical trial design, analysis, and interpretation.

Major depressive disorder (MDD) is one of the leading causes of disability globally. Despite its prevalence, approximately one-third of patients do not benefit sufficiently from available treatments, and few new drugs have been developed recently. Consequently, more efficient methods are needed to evaluate a broader range of treatment options quickly. Platform trials offer a promising solution, as they allow for the assessment of multiple investigational treatments simultaneously by sharing control groups and by reducing both trial activation and patient recruitment times. The objective of this simulation study was to support the design and optimisation of a phase II superiority platform trial for MDD, considering the disease-specific characteristics. In particular, we assessed the efficiency of platform trials compared to traditional two-arm trials by investigating key design elements, including allocation and randomisation strategies, as well as per-treatment arm sample sizes and interim futility analyses. Through extensive simulations, we refined these design components and evaluated their impact on trial performance. The results demonstrated that platform trials not only enhance efficiency but also achieve higher statistical power in evaluating individual treatments compared to conventional trials. The efficiency of platform trials is particularly prominent when interim futility analyses are performed to eliminate treatments that have either no or a negligible treatment effect early. Overall, this work provides valuable insights into the design of platform trials in the superiority setting and underscores their potential to accelerate therapy development in MDD and other therapeutic areas, providing a flexible and powerful alternative to traditional trial designs.

For toxicology studies, the validation of the concurrent control group by historical control data (HCD) has become requirements. This validation is usually done by historical control limits (HCL), which should cover the observations of the concurrent control with a predefined level of confidence. In many applications, HCL are applied to dichotomous data, for example, the number of rats with a tumor versus the number of rats without a tumor (carcinogenicity studies) or the number of cells with a micronucleus out of a total number of cells. Dichotomous HCD may be overdispersed and can be heavily right- (or left-) skewed, which is usually not taken into account in the practical applications of HCL. To overcome this problem, four different prediction intervals (two frequentist, two Bayesian), that can be applied to such data, are proposed. Based on comprehensive Monte-Carlo simulations, the coverage probabilities of the proposed prediction intervals were compared to heuristical HCL typically used in daily toxicological routine (historical range, limits of the np-chart, mean $\pm$ 2 SD). Our simulations reveal, that frequentist bootstrap calibrated prediction intervals control the type-1-error best, but, also prediction intervals calculated based on Bayesian generalized linear mixed models appear to be practically applicable. Contrary, all heuristics fail to control the type-1-error. The application of HCL is demonstrated based on a real life data set containing historical controls from long-term carcinogenicity studies run on behalf of the U.S. National Toxicology Program. The proposed frequentist prediction intervals are publicly available from the R package predint, whereas R code for the computation of the two Bayesian prediction intervals is provided via GitHub.

A major emphasis in personalized medicine is to optimally treat subgroups of patients who may benefit from certain therapeutic agents. One relevant study design is the targeted design, in which patients have consented for their specimens to be obtained at baseline and the specimens are sent to a laboratory for assessing the biomarker status prior to randomization. Here, only biomarker-positive patients will be randomized to either an experimental or the standard of care arms. Many biomarkers, however, are derived from patient tissue specimens, which are heterogeneous leading to variability in the biomarker levels and status. This heterogeneity would have an adverse impact on the power of an enriched biomarker clinical trial. In this article, we show the adverse effect of using the uncorrected sample size and overcome this challenge by presenting an approach to adjust for misclassification for the targeted design. Specifically, we propose a sample size formula that adjusts for misclassification and apply it in the design of two phase III clinical trials in renal and prostate cancer.

Sample size determination in Bayesian randomized phase II trial design often relies on computationally intensive search methods, presenting challenges in terms of feasibility and efficiency. We propose a novel approach that greatly reduces the computing time of sample size calculations for Bayesian trial designs. Our approach innovatively connects group sequential design with Bayesian trial design and leverages the proportional relationship between sample size and the squared drift parameter. This results in a faster algorithm. By employing regression analysis, our method can accurately pinpoint the required sample size with significantly reduced computational burden. Through theoretical justification and extensive numerical evaluations, we validate our approach and illustrate its efficiency across a wide range of common trial scenarios, including binary endpoint with Beta-Binomial model, normal endpoint, binary/ordinal endpoint under Bayesian generalized linear model, and survival endpoints under Bayesian piecewise exponential models. To facilitate the use of our methods, we create an R package named "BayesSize" on GitHub.

In this paper, we carried out extensive simulation studies to compare the performances of partial and maximum likelihood based methods for estimating the area under the bi-Weibull ROC curve. Further, real data sets from HIV/AIDS research were analyzed for illustrative purposes. Simulation results suggest that both methods perform well and yield similar results for Weibull data. However, for non-Weibull data, both methods perform poorly. The bi-Weibull model yields smooth estimates of ROC curves and a closed-form expression for the area under the ROC curve. Moreover, by adjusting its shape parameter, the bi-Weibull model can represent a variety of distributions, such as exponential, Rayleigh, normal, and extreme value distributions. Its compatibility with Cox's proportional hazards model facilitates the derivation of covariate-adjusted ROC curves and supports analyses involving correlated and longitudinal biomarkers. These properties make the model very useful in the ROC curve analyses. Thus, the bi-Weibull model should be considered as an alternative when the restrictive distributional assumptions of the commonly used parametric models (e.g., binormal model) are not met.

With the ongoing advancements in cancer drug development, a subset of patients can live quite long, or are even considered cured in certain cancer types. Additionally, nonproportional hazards, such as delayed treatment effects and crossing hazards, are commonly observed in cancer clinical trials with immunotherapy. To address these challenges, various cure models have been proposed to integrate the cure rate into trial designs and accommodate delayed treatment effects. In this article, we introduce a unified approach for calculating sample sizes, taking into account different cure rate models and nonproportional hazards. Our approach supports both the traditional weighted logrank test and the Maxcombo test, which demonstrates robust performance under nonproportional hazards. Furthermore, we assess the accuracy of our sample size estimation through Monte Carlo simulations across various scenarios and compare our method with existing approaches. Several illustrative examples are provided to demonstrate the proposed method.