Christy Chuang-Stein’s research while affiliated with Kalamazoo College and other places

The Acute Stroke Therapy by Inhibition of Neutrophils (ASTIN) study, initiated in November of the year 2000, is now widely recognized as having been a landmark study in the history of clinical trials. We look at why this is the case by considering its key features and impact. These key features are: the use of Bayesian design and analysis; the use of the normal dynamic linear model; the response adaptive nature of the study; the use of real‐time dosing decisions; and the use of an integrated model to predict 90‐day response on the Scandinavian Stroke Scale. Our overall conclusion is that the ASTIN study's main impact came from showing the clinical trial community the feasibility of the novel design and analysis used when most of these key features were rarely used in industry trials, let alone used together in one trial in a disease area with a tremendous unmet medical need.

According to ICH-E4, “Elucidation of the dose–response function” is a key stage in drug development. Consequently, designing a dose–response study with the desired characteristics is an important activity in drug development. Inadequate dose–response knowledge has been known to lead to a delay or denial in regulatory approvals of initial drug applications. There have also been cases when the dose initially approved for a marketed product had to be reduced subsequently. In this chapter, we focus on using the Emax model to describe a dose–response relationship, but the discussion applies equally to other dose–response models or to a collection of models. We examine in detail the three metrics introduced in Chapter 6 for assessing a dose–response study design, for both continuous and binary endpoints. We include an example of dose–response studies for an investigational medicinal product to illustrate the importance of covering an adequate dose range in dose-ranging studies.

A diagnostic test is often performed to diagnose a disease or to sub-classify a disease regarding its severity and susceptibility to treatment. In recent years, companion diagnostics have been developed to pre-select patients for specific treatments based on their own biology. In this chapter, we review the metrics commonly used to characterize the performance of a diagnostic test. We then discuss what it means when a diagnostic test returns a positive or a negative result. We illustrate the discussion with the use of several decision rules.

Adaptive designs can potentially help with quantitative decision making by allowing designs to be amended so that better decisions can be made in the light of emerging data. The chapter starts with some principles for adaptive designs before reviewing the many types of adaptive design covered in the adaptive design guidance issued by the Food and Drug Administration in the United States. We discuss four particular types of adaptive design. They are a group sequential design with an interim analysis for futility, a Phase 2 dose-ranging study where doses can be dropped or added, a Phase 2/3 enrichment design where the trial could be enriched by focusing on the correct patient population and a confirmatory trial where the adaptation is to decide on two doses before accumulating confirmatory evidence on these doses. We offer some comments concerning each of these four adaptive designs.

A proof-of-concept (POC) study is where a developer finds out for the first time if a new drug has any effect on the endpoint of interest in targeted patients, and if the drug appears to have an effect, whether the size of the effect warrants investment in further development. A number of approaches have been proposed to make a decision for a POC study. In this chapter, we review five of them. They are the traditional hypothesis-testing approach, the Early Signal of Efficacy (ESoE) approach, the approach implemented in the Learn Phase Development Assessment Tool (LPDAT) and two approaches that can be considered as special cases of the LPDAT approach. We review the metrics introduced in Chap. 6 in the context of these five design types. We show how prior information on historical controls can be used to help plan a POC study and share an example on this application involving a rare disease.

Developing a new drug is a high-risk and high-reward enterprise. The high risk is reflected by the low success rate of turning a new molecular entity (NME) into an approved drug even after the NME has successfully met the preclinical testing requirements. In this chapter, we offer a high-level summary of the clinical testing which is composed of four distinct phases under a traditional development plan. We discuss deviations from the traditional development plan and new regulatory approval pathways. In particular, we highlight the speed with which COVID-19 vaccines were developed and authorized for emergency use through extensive public–private partnerships and the adoption of a high-risk business model. We offer some examples of recent advances in clinical trial designs and conclude the chapter with a short summary of the salient points covered in the chapter.

Cost is an important consideration when planning a clinical development program. Yet, many drug development programs do not formally incorporate cost into their decision process. In this chapter, we review two approaches that have been proposed to quantitatively incorporate cost into design considerations. The first one optimizes a benefit–cost efficiency score that measures the cost-effectiveness of a proof-of-concept trial design. The second approach combines cost and potential commercial return and chooses the approach that would optimize the expected net present value (eNPV) of the investment. We include in the chapter a detailed discussion on the calculation of eNPV which will be of interest to readers without much exposure to product valuation.

In this chapter, we compare successive trials designed and conducted to assess the efficacy of a new drug to a series of diagnostic tests. The condition to diagnose is whether the new drug has a clinically meaningful efficacious effect. This comparison offers us the opportunity to apply properties pertaining to diagnostic tests discussed in Chap. 3 to clinical trials. Building on the results in Chap. 3, we discuss why replication is such a critically important concept in drug development and show why replication is not as easy as some might have hoped. The difference between replicability and reproducibility is briefly discussed. We end the chapter by highlighting the difference between statistical power and the probability of a positive trial. This last point becomes more important as a new drug moves through the various development stages as will be illustrated in Chap. 9.

Drug development is a continuum. Information from completed trials and other sources may be available when we design a new trial. In this chapter, we consider the Bayesian approach to incorporate existing information into future trial planning. Under the Bayesian approach, prior distributions are used to describe the accumulated knowledge about unknown parameters such as responses to controls in previous trials and estimates pertaining to treatment benefit. Information on historical controls can be used to help design and analyze the future trials while prior distributions on the treatment effect can be used to calculate the probability of a successful trial or assurance probability. In this chapter, we look at closed-form expressions for assurance probabilities under some definitions of success. We can use simulations to estimate these probabilities when closed-form expressions do not exist. With the use of prior distributions for the parameters of interest, we show how to assess the positive and negative predictive values of a design with its companion decision rule.

By the time a new drug moves into the confirmatory stage, its developer should in theory have a reasonable amount of information on the effect of the drug on several efficacy endpoints. Based on this assumption, we begin this chapter by assuming that some treatment effect information on the primary endpoint and for the appropriate patient population is available for planning a Phase 3 trial. We review how this information can be used to assess the probability of success of the study (POSS) and discuss how POSS can in turn help a sponsor assess the adequacy of the sample size calculated from the hypothesis-testing perspective. In addition, we discuss factors that could affect POSS and how these factors should be incorporated into the planning of the confirmatory program. The chapter highlights the importance of a robust investment in Phase 2 development in order to achieve a desirable level of POSS at the Phase 3 stage.