GRAPHICAL SENSITIVITY ANALYSIS WITH DIFFERENT METHODS OF IMPUTATION FOR A TRIAL WITH PROBABLE NON‐IGNORABLE MISSING DATA

Graphical sensitivity analyses have recently been recommended for clinical trials with non‐ignorable missing outcome. We demonstrate an adaptation of this methodology for a continuous outcome of a trial of three cognitive‐behavioural therapies for mild depression in primary care, in which one arm had unexpectedly high levels of missing data. Fixed‐value and multiple imputations from a normal distribution (assuming either varying mean and fixed standard deviation, or fixed mean and varying standard deviation) were used to obtain contour plots of the contrast estimates with their P‐values superimposed, their confidence intervals, and the root mean square errors. Imputation was based either on the outcome value alone, or on change from baseline. The plots showed fixed‐value imputation to be more sensitive than imputing from a normal distribution, but the normally distributed imputations were subject to sampling noise. The contours of the sensitivity plots were close to linear in appearance, with the slope approximately equal to the ratio of the proportions of subjects with missing data in each trial arm.     

Joint modeling tumor burden and time to event data in oncology trials

The tumor burden (TB) process is postulated to be the primary mechanism through which most anticancer treatments provide benefit. In phase II oncology trials, the biologic effects of a therapeutic agent are often analyzed using conventional endpoints for best response, such as objective response rate and progression‐free survival, both of which causes loss of information. On the other hand, graphical methods including spider plot and waterfall plot lack any statistical inference when there is more than one treatment arm. Therefore, longitudinal analysis of TB data is well recognized as a better approach for treatment evaluation. However, longitudinal TB process suffers from informative missingness because of progression or death. We propose to analyze the treatment effect on tumor growth kinetics using a joint modeling framework accounting for the informative missing mechanism. Our approach is illustrated by multisetting simulation studies and an application to a nonsmall‐cell lung cancer data set. The proposed analyses can be performed in early‐phase clinical trials to better characterize treatment effect and thereby inform decision‐making. Copyright © 2014 John Wiley & Sons, Ltd.     

Response‐adaptive designs for binary responses: How to offer patient benefit while being robust to time trends?

Response‐adaptive randomisation (RAR) can considerably improve the chances of a successful treatment outcome for patients in a clinical trial by skewing the allocation probability towards better performing treatments as data accumulates. There is considerable interest in using RAR designs in drug development for rare diseases, where traditional designs are not either feasible or ethically questionable. In this paper, we discuss and address a major criticism levelled at RAR: namely, type I error inflation due to an unknown time trend over the course of the trial. The most common cause of this phenomenon is changes in the characteristics of recruited patients—referred to as patient drift. This is a realistic concern for clinical trials in rare diseases due to their lengthly accrual rate. We compute the type I error inflation as a function of the time trend magnitude to determine in which contexts the problem is most exacerbated. We then assess the ability of different correction methods to preserve type I error in these contexts and their performance in terms of other operating characteristics, including patient benefit and power. We make recommendations as to which correction methods are most suitable in the rare disease context for several RAR rules, differentiating between the 2‐armed and the multi‐armed case. We further propose a RAR design for multi‐armed clinical trials, which is computationally efficient and robust to several time trends considered.     

The correlation structure of longitudinal measurements of vision in patients with macular degeneration

Background: In age‐related macular degeneration (ARMD) trials, the FDA‐approved endpoint is the loss (or gain) of at least three lines of vision as compared to baseline. The use of such a response endpoint entails a potentially severe loss of information. A more efficient strategy could be obtained by using longitudinal measures of the change in visual acuity. In this paper we investigate, by using data from two randomized clinical trials, the mean and variance–covariance structures of the longitudinal measurements of the change in visual acuity. Methods: Individual patient data were collected in 234 patients in a randomized trial comparing interferon‐ α with placebo and in 1181 patients in a randomized trial comparing three active doses of pegaptanib with sham. A linear model for longitudinal data was used to analyze the repeated measurements of the change in visual acuity. Results: For both trials, the data were adequately summarized by a model that assumed a quadratic trend for the mean change in visual acuity over time, a power variance function, and an antedependence correlation structure. The power variance function was remarkably similar for the two datasets and involved the square root of the measurement time. Conclusions: The similarity of the estimated variance functions and correlation structures for both datasets indicates that these aspects may be a genuine feature of the measurements of changes in visual acuity in patients with ARMD. The feature can be used in the planning and analysis of trials that use visual acuity as the clinical endpoint of interest. Copyright © 2010 John Wiley & Sons, Ltd.     

A flexible parametric survival model for fitting time to event data in clinical trials

Time‐to‐event data are common in clinical trials to evaluate survival benefit of a new drug, biological product, or device. The commonly used parametric models including exponential, Weibull, Gompertz, log‐logistic, log‐normal, are simply not flexible enough to capture complex survival curves observed in clinical and medical research studies. On the other hand, the nonparametric Kaplan Meier (KM) method is very flexible and successful on catching the various shapes in the survival curves but lacks ability in predicting the future events such as the time for certain number of events and the number of events at certain time and predicting the risk of events (eg, death) over time beyond the span of the available data from clinical trials. It is obvious that neither the nonparametric KM method nor the current parametric distributions can fulfill the needs in fitting survival curves with the useful characteristics for predicting. In this paper, a full parametric distribution constructed as a mixture of three components of Weibull distribution is explored and recommended to fit the survival data, which is as flexible as KM for the observed data but have the nice features beyond the trial time, such as predicting future events, survival probability, and hazard function.     

Restricted mean survival time as a summary measure of time‐to‐event outcome

Many clinical research studies evaluate a time‐to‐event outcome, illustrate survival functions, and conventionally report estimated hazard ratios to express the magnitude of the treatment effect when comparing between groups. However, it may not be straightforward to interpret the hazard ratio clinically and statistically when the proportional hazards assumption is invalid. In some recent papers published in clinical journals, the use of restricted mean survival time (RMST) or τ ‐year mean survival time is discussed as one of the alternative summary measures for the time‐to‐event outcome. The RMST is defined as the expected value of time to event limited to a specific time point corresponding to the area under the survival curve up to the specific time point. This article summarizes the necessary information to conduct statistical analysis using the RMST, including the definition and statistical properties of the RMST, adjusted analysis methods, sample size calculation, information fraction for the RMST difference, and clinical and statistical meaning and interpretation. Additionally, we discuss how to set the specific time point to define the RMST from two main points of view. We also provide developed SAS codes to determine the sample size required to detect an expected RMST difference with appropriate power and reconstruct individual survival data to estimate an RMST reference value from a reported survival curve.     

A recommended analysis for 2 × 2 crossover trials with baseline measurements

In many two‐period, two‐treatment (2 × 2) crossover trials, for each subject, a continuous response of interest is measured before and after administration of the assigned treatment within each period. The resulting data are typically used to test a null hypothesis involving the true difference in treatment response means. We show that the power achieved by different statistical approaches is greatly influenced by (i) the ‘structure’ of the variance–covariance matrix of the vector of within‐subject responses and (ii) how the baseline (i.e., pre‐treatment) responses are accounted for in the analysis. For (ii), we compare different approaches including ignoring one or both period baselines, using a common change from baseline analysis (which we advise against), using functions of one or both baselines as period‐specific or period‐invariant covariates, and doing joint modeling of the post‐baseline and baseline responses with corresponding mean constraints for the latter. Based on theoretical arguments and simulation‐based type I error rate and power properties, we recommend an analysis of covariance approach that uses the within‐subject difference in treatment responses as the dependent variable and the corresponding difference in baseline responses as a covariate. Data from three clinical trials are used to illustrate the main points. Copyright © 2014 John Wiley & Sons, Ltd.     

Design and inference for 3‐stage bioequivalence testing with serial sampling data

A bioequivalence test is to compare bioavailability parameters, such as the maximum observed concentration (C_max) or the area under the concentration‐time curve, for a test drug and a reference drug. During the planning of a bioequivalence test, it requires an assumption about the variance of C_max or area under the concentration‐time curve for the estimation of sample size. Since the variance is unknown, current 2‐stage designs use variance estimated from stage 1 data to determine the sample size for stage 2. However, the estimation of variance with the stage 1 data is unstable and may result in too large or too small sample size for stage 2. This problem is magnified in bioequivalence tests with a serial sampling schedule, by which only one sample is collected from each individual and thus the correct assumption of variance becomes even more difficult. To solve this problem, we propose 3‐stage designs. Our designs increase sample sizes over stages gradually, so that extremely large sample sizes will not happen. With one more stage of data, the power is increased. Moreover, the variance estimated using data from both stages 1 and 2 is more stable than that using data from stage 1 only in a 2‐stage design. These features of the proposed designs are demonstrated by simulations. Testing significance levels are adjusted to control the overall type I errors at the same level for all the multistage designs.     

Missing data in clinical trials: control‐based mean imputation and sensitivity analysis

In some randomized (drug versus placebo) clinical trials, the estimand of interest is the between‐treatment difference in population means of a clinical endpoint that is free from the confounding effects of “rescue” medication (e.g., HbA1c change from baseline at 24 weeks that would be observed without rescue medication regardless of whether or when the assigned treatment was discontinued). In such settings, a missing data problem arises if some patients prematurely discontinue from the trial or initiate rescue medication while in the trial, the latter necessitating the discarding of post‐rescue data. We caution that the commonly used mixed‐effects model repeated measures analysis with the embedded missing at random assumption can deliver an exaggerated estimate of the aforementioned estimand of interest. This happens, in part, due to implicit imputation of an overly optimistic mean for “dropouts” (i.e., patients with missing endpoint data of interest) in the drug arm. We propose an alternative approach in which the missing mean for the drug arm dropouts is explicitly replaced with either the estimated mean of the entire endpoint distribution under placebo (primary analysis) or a sequence of increasingly more conservative means within a tipping point framework (sensitivity analysis); patient‐level imputation is not required. A supplemental “dropout = failure” analysis is considered in which a common poor outcome is imputed for all dropouts followed by a between‐treatment comparison using quantile regression. All analyses address the same estimand and can adjust for baseline covariates. Three examples and simulation results are used to support our recommendations.     

A proposal for plotting positions in probability plots

Probability plots allow us to determine whether a set of sample observations is distributed according to a theoretical distribution. Plotting positions are fundamental elements in statistics and, in particular, for the construction of probability plots. In this paper, a new plotting position to construct different probability plots, such as Q–Q Plot, P–P Plot and S–P Plot, is proposed. The proposed definition is based on the median of the ith order statistic of the theoretical distribution considered. The main feature of this plotting position formula is that it is independent of the theoretical distribution selected. Moreover, the procedure developed is ‘almost’ exact, reaching, without a high cost of time, an accuracy as great as we want, which avoids using approximations (proposed by other authors).     

Determination of hazard ratio for progression‐free survival considering the tumor assessment schedule in sample size calculation

Progression‐free survival is recognized as an important endpoint in oncology clinical trials. In clinical trials aimed at new drug development, the target population often comprises patients that are refractory to standard therapy with a tumor that shows rapid progression. This situation would increase the bias of the hazard ratio calculated for progression‐free survival, resulting in decreased power for such patients. Therefore, new measures are needed to prevent decreasing the power in advance when estimating the sample size. Here, I propose a novel calculation procedure to assume the hazard ratio for progression‐free survival using the Cox proportional hazards model, which can be applied in sample size calculation. The hazard ratios derived by the proposed procedure were almost identical to those obtained by simulation. The hazard ratio calculated by the proposed procedure is applicable to sample size calculation and coincides with the nominal power. Methods that compensate for the lack of power due to biases in the hazard ratio are also discussed from a practical point of view.     

Joint probability of statistical success of multiple phase III trials

In drug development, after completion of phase II proof‐of‐concept trials, the sponsor needs to make a go/no‐go decision to start expensive phase III trials. The probability of statistical success (PoSS) of the phase III trials based on data from earlier studies is an important factor in that decision‐making process. Instead of statistical power, the predictive power of a phase III trial, which takes into account the uncertainty in the estimation of treatment effect from earlier studies, has been proposed to evaluate the PoSS of a single trial. However, regulatory authorities generally require statistical significance in two (or more) trials for marketing licensure. We show that the predictive statistics of two future trials are statistically correlated through use of the common observed data from earlier studies. Thus, the joint predictive power should not be evaluated as a simplistic product of the predictive powers of the individual trials. We develop the relevant formulae for the appropriate evaluation of the joint predictive power and provide numerical examples. Our methodology is further extended to the more complex phase III development scenario comprising more than two (K > 2) trials, that is, the evaluation of the PoSS of at least k₀ () trials from a program of K total trials. Copyright © 2013 John Wiley & Sons, Ltd.     

The size accuracy of combination tests

Adaptive clinical trials typically involve several independent stages. The P‐values from each stage are synthesized through a so‐called combination function which ensures that the overall test will be valid if the stagewise tests are valid. In practice however, approximate and possibly invalid stagewise tests are used. This paper studies how imperfections of the stagewise tests feed through into the combination test. Several general results are proven including some for discrete models. An approximation formula which directly links the combined size accuracy to the component size accuracy is given. In the wider context of adaptive clinical trials, the main conclusion is that the basic tests used should be size accurate at nominal sizes both much larger and also much smaller than nominal desired size. For binary outcomes, the implication is that the parametric bootstrap should be used.     

Improving interim decisions in randomized trials by exploiting information on short‐term endpoints and prognostic baseline covariates

Conditional power calculations are frequently used to guide the decision whether or not to stop a trial for futility or to modify planned sample size. These ignore the information in short‐term endpoints and baseline covariates, and thereby do not make fully efficient use of the information in the data. We therefore propose an interim decision procedure based on the conditional power approach which exploits the information contained in baseline covariates and short‐term endpoints. We will realize this by considering the estimation of the treatment effect at the interim analysis as a missing data problem. This problem is addressed by employing specific prediction models for the long‐term endpoint which enable the incorporation of baseline covariates and multiple short‐term endpoints. We show that the proposed procedure leads to an efficiency gain and a reduced sample size, without compromising the Type I error rate of the procedure, even when the adopted prediction models are misspecified. In particular, implementing our proposal in the conditional power approach enables earlier decisions relative to standard approaches, whilst controlling the probability of an incorrect decision. This time gain results in a lower expected number of recruited patients in case of stopping for futility, such that fewer patients receive the futile regimen. We explain how these methods can be used in adaptive designs with unblinded sample size re‐assessment based on the inverse normal P‐value combination method to control Type I error. We support the proposal by Monte Carlo simulations based on data from a real clinical trial.     

Handling of missing data in long‐term clinical trials: a case study

Missing data in clinical trials is a well‐known problem, and the classical statistical methods used can be overly simple. This case study shows how well‐established missing data theory can be applied to efficacy data collected in a long‐term open‐label trial with a discontinuation rate of almost 50%. Satisfaction with treatment in chronically constipated patients was the efficacy measure assessed at baseline and every 3 months postbaseline. The improvement in treatment satisfaction from baseline was originally analyzed with a paired t‐test ignoring missing data and discarding the correlation structure of the longitudinal data. As the original analysis started from missing completely at random assumptions regarding the missing data process, the satisfaction data were re‐examined, and several missing at random (MAR) and missing not at random (MNAR) techniques resulted in adjusted estimate for the improvement in satisfaction over 12 months. Throughout the different sensitivity analyses, the effect sizes remained significant and clinically relevant. Thus, even for an open‐label trial design, sensitivity analysis, with different assumptions for the nature of dropouts (MAR or MNAR) and with different classes of models (selection, pattern‐mixture, or multiple imputation models), has been found useful and provides evidence towards the robustness of the original analyses; additional sensitivity analyses could be undertaken to further qualify robustness. Copyright © 2012 John Wiley & Sons, Ltd.     

Study design of single‐arm phase II immunotherapy trials with long‐term survivors and random delayed treatment effect

In the traditional study design of a single‐arm phase II cancer clinical trial, the one‐sample log‐rank test has been frequently used. A common practice in sample size calculation is to assume that the event time in the new treatment follows exponential distribution. Such a study design may not be suitable for immunotherapy cancer trials, when both long‐term survivors (or even cured patients from the disease) and delayed treatment effect are present, because exponential distribution is not appropriate to describe such data and consequently could lead to severely underpowered trial. In this research, we proposed a piecewise proportional hazards cure rate model with random delayed treatment effect to design single‐arm phase II immunotherapy cancer trials. To improve test power, we proposed a new weighted one‐sample log‐rank test and provided a sample size calculation formula for designing trials. Our simulation study showed that the proposed log‐rank test performs well and is robust of misspecified weight and the sample size calculation formula also performs well.     

A statistical method to convert published response rates into marginal distributions with an example application in psoriasis

Assessment of severity is essential for the management of chronic diseases. Continuous variables like scores obtained from the Hamilton Rating Scale for Depression or the Psoriasis Area and Severity Index (PASI) are standard measures used in clinical trials of depression and psoriasis. In clinical trials of psoriasis, for example, the reduction of PASI from baseline in response to therapy, in particular the proportion of patients achieving at least 75%, 90%, or 100% improvement of disease (PASI 75, PASI 90, or PASI 100), is typically used to evaluate treatment efficacy. However, evaluation of the proportions of patients reaching absolute PASI values (eg, ≤1, ≤2, ≤3, or ≤5) has recently gained greater clinical interest and is increasingly being reported. When relative versus absolute scores are standard, as is the case with the PASI in psoriasis, it is difficult to compare absolute changes using existing published data. Thus, we developed a method to estimate absolute PASI levels from aggregated relative levels. This conversion method is based on a latent 2‐dimensional normal distribution for the absolute score at baseline and at a specific endpoint with a truncation to allow for baseline inclusion criterion. The model was fitted to aggregated results from simulations and from 3 phase III studies that had known absolute PASI proportions. The predictions represented the actual results quite precisely. This model might be applied to other conditions, such as depression, to estimate proportions of patients achieving an absolute low level of disease activity, given absolute values at baseline and proportions of patients achieving relative improvements at a subsequent time point.     

Repeated‐measures models in the analysis of QT interval

Because of the recent regulatory emphasis on issues related to drug‐induced cardiac repolarization that can potentially lead to sudden death, QT interval analysis has received much attention in the clinical trial literature. The analysis of QT data is complicated by the fact that the QT interval is correlated with heart rate and other prognostic factors. Several attempts have been made in the literature to derive an optimal method for correcting the QT interval for heart rate; however the QT correction formulae obtained are not universal because of substantial variability observed across different patient populations. It is demonstrated in this paper that the widely used fixed QT correction formulae do not provide an adequate fit to QT and RR data and bias estimates of treatment effect. It is also shown that QT correction formulae derived from baseline data in clinical trials are likely to lead to Type I error rate inflation. This paper develops a QT interval analysis framework based on repeated‐measures models accomodating the correlation between QT interval and heart rate and the correlation among QT measurements collected over time. The proposed method of QT analysis controls the Type I error rate and is at least as powerful as traditional QT correction methods with respect to detecting drug‐related QT interval prolongation. Copyright © 2003 John Wiley & Sons, Ltd.     

A simple test for the treatment effect in clinical trials with a sequential parallel comparison design and negative binomial outcomes

In placebo‐controlled, double‐blinded, randomized clinical trials, the presence of placebo responders reduces the effect size for comparison of the active drug group with the placebo group. An attempt to resolve this problem is to use the sequential parallel comparison design (SPCD). Although there are SPCDs with dichotomous or continuous outcomes, an SPCD with negative binomial outcomes—with which investigators deal eg, in clinical trials involving multiple sclerosis, where the investigators are still concerned about the presence of placebo responders—has not yet been discussed. In this article, we propose a simple test for the treatment effect in clinical trials with an SPCD and negative binomial outcomes. Through simulations, we show that the analysis method achieves the nominal type I error rate and power, whereas the sample size calculation provides the sample size with adequate power accuracy.