Regulatory and scientific issues regarding use of foreign data in support of new drug applications in the United States: an FDA perspective.

Globalization of clinical research has led to an increase in clinical trials conducted outside of the United States that are submitted to the US Food and Drug Administration (FDA) in new drug applications. This article discusses the FDA's experience with these submissions in specific therapeutic areas, including the extent of this practice, differences between the effectiveness and safety outcomes of studies conducted inside and outside the United States, and the FDA's approach to acceptance of these trials.

The prevention and treatment of missing data in clinical trials: an FDA perspective on the importance of dealing with it.

At the request of the Food and Drug Administration (FDA) and with its funding, the Panel on the Handling of Missing Data in Clinical Trials was created by the National Research Council's Committee on National Statistics. This panel recently published a report(1) with recommendations that will be of use not only to the FDA but also to the entire clinical trial community so that the latter can take measures to improve the conduct and analysis of clinical trials.

FDA's critical path initiative: a perspective on contributions of biostatistics.

This article describes the motivation for, description of, and the objectives and plans for the FDA's initiative that was introduced in March of 2004 by way of a report titled 'Innovation or Stagnation?--Challenge and Opportunity on the Critical Path to New Medical Products'. The FDA initiative is very much an outreach effort and a wake-up call to many constituencies to contribute and partner to improve the product development process and thereby to contribute to the success rate of new products that will benefit the public. We discuss in general terms where some of the opportunities and challenges exist for the discipline of biostatistics to make contributions to this effort over the next few years. In particular, guidance development in five areas is considered as is the need to devote new energy and efforts to quantitative risk assessment and safety evaluation, an area that has lagged the attention received in the efficacy evaluation area.

Biostatistical considerations in pharmacovigilance and pharmacoepidemiology: linking quantitative risk assessment in pre-market licensure application safety data, post-market alert reports and formal epidemiological studies.

This paper deals with a conceptual discussion of a variety of statistical concepts, methods and strategies that are relevant to the quantitative assessment of risk derived from safety data collected during the pre- and post-marketing phase of a new drug's life cycle. A call is made for the use of more standard approaches to the analysis of safety data that are statistically and epidemiologically rigorous and for attempts to link the strategies for pre-market safety assessment with strategies for post-market safety evaluation. This link may be facilitated by recognizing the limitations and complementary roles played by pre- and post-market safety data collection schemes and by linking the quantitative analyses utilized for either exploratory or confirmatory purposes of risk assessment in each phase of safety data collection. Examples are provided of studies specifically designed to evaluate risk in a post approval setting and several available guidelines intended to improve the quality of these studies are discussed.

Secondary endpoints cannot be validly analyzed if the primary endpoint does not demonstrate clear statistical significance.

There is lack of consensus surrounding the interpretation of observed treatment effects for secondary clinical endpoints when the primary endpoint for which the clinical trial was initially designed does not meet the objective of a demonstrated effect. We provide some arguments to support caution in making inferences for secondary endpoints in this situation. We examine the definitions of primary and secondary endpoints within the context of a hypothesis-testing framework for multiple endpoints, and we address the relationship of the correlation structure of these endpoints and the statistical adjustments needed to preserve experiment-wise type I error for a valid inference. We also address the hypothesis-testing framework and the estimation framework for valid inference, focusing on the interpretation of p-values associated with differentially powered hypothesis tests for each endpoint to detect an important clinical effect. We point out the limitations on the strength of evidence (and quantification of uncertainty) for a secondary endpoint effect that can be derived from only one study and introduce the likelihood of replication of the finding in another study of identical size and design as a useful concept to guide this interpretation.

The behavior of the P-value when the alternative hypothesis is true.

The P-value is a random variable derived from the distribution of the test statistic used to analyze a data set and to test a null hypothesis. Under the null hypothesis, the P-value based on a continuous test statistic has a uniform distribution over the interval [0, 1], regardless of the sample size of the experiment. In contrast, the distribution of the P-value under the alternative hypothesis is a function of both sample size and the true value or range of true values of the tested parameter. The characteristics, such as mean and percentiles, of the P-value distribution can give valuable insight into how the P-value behaves for a variety of parameter values and sample sizes. Potential applications of the P-value distribution under the alternative hypothesis to the design, analysis, and interpretation of results of clinical trials are considered.

Efficacy evaluation for monotherapies in two-by-two factorial trials.

For factorial clinical trials in which two monotherapy treatments under study can interact only in the presence of treatment effects for each treatment, the always-pooled test statistic using data from all four groups has a correct size in detecting the simple effect of an individual treatment used alone. However, this test statistic may have an unbounded bias in estimation of the simple effect. The never-pooled test statistic that uses only data from the treatment group not receiving the other treatment has poor precision for estimating the simple effect. Two alternative test statistics under consideration are the two-stage statistic involving a preliminary test of treatment interaction and the maximum test statistic taking the larger of the always-pooled and the never-pooled statistics. The power, bias, and mean square error of all four tests are compared. When negative interactions exist, the two-stage and maximum statistics are generally superior to the always-pooled statistic and compare reasonably well with the never-pooled statistic; the maximum statistic seems slightly more favorable than the two-stage statistic. The two-stage statistic is the best choice when a treatment interaction can be large.

Statistical concepts in the planning and evaluation of drug safety from clinical trials in drug development: issues of international harmonization.

The assessment of the safety of new drugs during pre-marketing clinical studies is an important and integral part of the drug development and regulatory evaluation process. The International Conference on Harmonization of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) is a project that brings together the regulatory decision-makers of Europe, Japan and the United States of America and the experts from the pharmaceutical industry in the three regions to seek ways to eliminate redundant and duplicative technical requirements among the developed countries for registering new medicinal substances and products. The ICH is developing guidelines or position papers to achieve the goal of harmonizing technical standards in three broad areas, namely, drug efficacy, safety and quality. Within the area of drug safety, this paper will discuss three of the safety topics because of their relevant statistical framework, and because these topics have not, to date, received any attention by the statistical community. The three issues under consideration by the International Conference on Harmonization (ICH), are: 1. Dose-response information to support drug registration (especially dose/toxicity relationships). 2. Studies in support of special populations; Geriatrics, A Draft Guideline. 3. ICH Draft Guideline 3 on 'The extent of population exposure required to assess clinical safety for drugs intended for long-term-treatment of non-life-threatening conditions'. The ICH special population guideline concerning studies in geriatric patients is closely related to a recent Food and Drug Administration 'Guideline for the study and evaluation of gender differences in the clinical evaluation of drugs', which is another example of a 'subgroup' for whom specific interest exists to evaluate drug safety and efficacy.

Some FDA perspectives on data monitoring in clinical trials in drug development.

The FDA's interest in data monitoring of clinical trials derives from its public health responsibility to assure the safety and efficacy of new drugs based on evidence from adequate and well-controlled studies. Therefore the FDA is concerned that clinical trials of new drugs are designed and carried out in a manner that will insure the integrity and validity of study inferences. The FDA regulation and guidelines recognize the role of data monitoring and the variety and diversity of situations utilizing a data monitoring process in clinical studies. This paper describes relevant aspects of the regulations and guidelines, some concerns the FDA has with regard to monitoring of both government- and industry-sponsored trials and the consequences of early termination of trials of new drugs in the investigational and marketed stages. Comments include advice on communication between the FDA and data monitoring committees.

Competing risk analysis of life table data: application to lifetime risk computation.

In this pedagogic note we propose to assess the safety of treatment in a clinical trial, or the effect of risk exposure in a chronic animal study, in terms of two lifetime risks. These risks are computable from life table type data and take into account the effects of competing risks. We first describe their computational procedures in detail to demonstrate the need for their implementation in a computer program. We then illustrate their practical application through use of the data obtained from an actual clinical study.

On sample sizes to estimate the protective efficacy of a vaccine.

To estimate vaccine protective efficacy, defined as VE = 1 - ARV/ARU where ARV is the disease attack rate in the vaccinated group and ARU is the disease attack rate in the controls, investigators have used both cohort and case-control designs. For each design, we present a method for calculation of the sample size required to provide an approximate confidence interval for VE of predetermined width and probability of coverage. The required sample size is a function of the desired width of the confidence interval, the probability of coverage, the assumed VE, and, for cohort designs, the assumed disease attack rate in the controls, and for case-control designs, the assumed vaccine exposure prevalence for the controls.

Statistical analyses of adverse event data from clinical trials. Special emphasis on serious events.

The purpose of this paper is to describe some methods for analyzing and summarizing adverse event rates from clinical trials, emphasizing, in particular, serious adverse drug events and their time of occurrence, and the impact of differential subject exposure and pretreatment status on the estimation of rates.

Sample sizes for estimation of the odds ratio in unmatched case-control studies.

A method is presented to obtain sample sizes for cases and controls that are required to provide approximate confidence intervals on the log odds ratio of predetermined width 2d and probability of coverage as a function of assumed exposure rates in the control group, assumed odds ratio psi, required d, and ratio C:1 of controls to cases.

A method for post-marketing screening of adverse reactions to drugs: initial results.

The FDA is pilot-testing a methodology for signaling previously unsuspected relationships between drugs and important adverse events. This method uses data it receives through the FDA spontaneous reporting program. Reviewing drugs used primarily on an outpatient basis, this screening methodology focuses on "tracer" adverse events and the organization of these reactions into body/functional systems. This review process enables a clinical evaluator to perceive more easily the clinically important drug-adverse event patterns. The method can incorporate drug use data; this enables a drug's proportional share of specific adverse events, relative to its therapeutic class, to be compared to its respective proportional share of drug use. The assumption is that the adverse event distribution of drugs in a therapeutic class should be the same as the distribution of drug use in that class, if all drugs in the class were to carry the same risk. Actual examples of drug-adverse event associations signaled by the screening method are presented. The potential uses of this methodology in other settings, and under other data situations, are discussed.

Case-control studies: a sequential approach.

A sequential approach to the design of a matched pair case-control study is proposed as an alternative to fixed sample plans when information on case-control pairs is acquired sequentially in time. The method described is that of Wald using the Sequential Probability Ratio Test (SPRT) for comparing two binomial populations. The test is an open, non-truncated procudure. Several tables are presented for the average sample numbers needed under the sequential plan and fixed sample plan for selected Type 1 and 2 errors, alternative relative risks of interest and exposure probabilities in case and controls. A hypothetical example is presented to illustrate the use of the method and discussion is given as to the possible advantages and disadvantages of the sequential approach to case-control studies.

A statistical procedure useful in evaluating time of onset and termination of response in clinical trials.

A statistical model jointly characterizing the onset and termination of treatment response of a subject over a fixed observed time period is presented. The model requires that the observations for each subject are made at a set of pre-selected time points during the observed time period. A useful index characterizing the probability of being in response is developed along with maximum likelihood estimates and variances. A likelihood ratio test is developed to simultaneously compare two treatment groups with respect to this index for all times. The proposed procedure is applied to a set of data from a clinical trial of two bronchodilator drugs from which our procedure is motivated.