Rare disease drug development: Time for a new approach?

A limited patient population is a key challenge in rare disease drug development. Professor Shein-Chung Chow of Duke University School of Medicine, USA and his research partners explore innovative approaches to overcome this challenge. These include using external control, selecting appropriate study endpoints, and justifying sample size based on probability statements. They propose the use of a complex innovative two-stage adaptive design that demonstrates not-ineffectiveness and not-unsafeness of the test treatment in a randomised clinical trial, supported by data from real-world studies. Finally, an individual benefit-risk assessment would provide a complete clinical picture.

A rare disease is one that affects a small percentage of the population. However, there is no universally agreed definition. For example, it is defined as a disorder affecting fewer than 1 in 2,000 people in the European Union but fewer than 200,000 persons in the United States.

The US Food and Drug Administration (FDA) requires the same standards for providing substantial evidence for regulatory approval of treatments for rare diseases as it does for more common conditions. The evidence of efficacy and safety must be obtained from well-controlled randomised clinical trials (RCTs). However, the limited patient population available for clinical trials in rare diseases makes it difficult to meet these requirements. As a result, funding or research on the treatment of these diseases is often lacking and only possible with support or incentives from governments or other agencies. In such a situation, a rare disease is referred to as an ‘orphan disease’ and a drug that treats it as an ‘orphan drug’.

A major problem of rare disease drug studies with small patient numbers is that the results may not truly reflect the effects of the drug. Furthermore, other practical challenges such as the lack of universally accepted endpoints or suitable biomarkers (characteristics of body processes that can be measured to show normal or abnormal function, or response to an external stimulus) make it difficult to gather adequate efficacy and safety data.

To encourage drug development for rare diseases, the FDA introduced incentive programmes that include fast-track designation, breakthrough therapy designation, priority review designation, and accelerated approval. Nevertheless, these incentives do not address the issue of small patient numbers in clinical trials for rare diseases and are unable to support drug evaluation to the same standard as that for drugs for more common conditions.

Professor Shein-Chung Chow of Duke University, USA, and his colleagues explore these challenges and propose novel designs and analyses for assessing the safety and efficacy of drugs for rare diseases.

Rare disease clinical trial challenges

Before a clinical trial can begin, a power analysis is usually performed to calculate the minimum number of patients required. A key concern in rare disease clinical trials is that the study may not achieve the desired power (the probability of detecting a true effect) to confirm the safety and efficacy of the drug at the 5% level of significance required for these trials. Therefore, the usual power calculation for sample size cannot be used for rare disease clinical trials, and alternative statistical calculations must be performed to justify the smaller sample size.

Heterogeneity in patient demographics and characteristics will also impact power calculations for sample size. In clinical trials of more common conditions, a larger patient sample size and stratified randomisation would be applied to account for this and prevent bias. These approaches are not realistic in trials for rare diseases.

Due to ethical considerations and the lack of sufficient patients with the condition, rare disease clinical trials often omit a control arm. Alternative approaches, such as the use of historical information from real-world data (RWD) as a control arm, may be considered.

Another problem with rare disease clinical trials is the lack of a universally accepted study endpoint or biomarker that can be used to assess efficacy and safety. While the search for suitable endpoints continues, researchers need to select endpoints based on the objectives of each trial.

Study designs used for common diseases with a large pool of potential trial subjects are inflexible or inefficient for researching drugs for rare diseases. Due to small patient numbers, the usual parallel-group design is unsuitable. Alternative complex innovative designs may be more appropriate.

Innovative solutions

To overcome some of the difficulties in rare disease drug development, Chow and his colleagues propose some out-of-the-box innovative solutions described below.

External controls
Where a control arm is not possible or unethical, they propose the use of an external control. However, external controls may introduce selection bias. Therefore, Chow and colleagues propose the use of statistical methods such as propensity score-matching (PSM) techniques, which mimic randomisation to overcome selection bias. An artificial control group is created by matching each treated person with a non-treated person with similar characteristics (eg, demographics and patient characteristics). PSM calculates the probability that a person will enrol in a trial based on observed characteristics.

Endpoint selection
Since most rare diseases have no universally accepted endpoint or biomarker, Chow and colleagues propose the development of a composite therapeutic index that combines clinical endpoints and biomarkers. This index can then be used to assess the overall safety and efficacy of the drug.

Sample size calculation
Chow and colleagues propose sample size determination for rare disease trials based on probability statements. This requires the selection of an appropriate sample size for controlling the probability of crossing safety and efficacy boundaries.

Demonstration of not-ineffectiveness and/or not-unsafeness
To gather effectiveness and safety data, Chow and colleagues propose trials designed to first demonstrate not-ineffectiveness and not-unsafeness, then collect additional data to confirm effectiveness and safeness. They highlight that there is a gap between not-ineffectiveness and effectiveness and between not-unsafeness and safeness, which they refer to as ‘the area of inconclusiveness’. They suggest that this gap may be filled by RWD.

Complex innovative design
Chow and colleagues discuss complex innovative trial designs, focusing on two that are commonly used in rare disease drug development. The first, an adaptive trial design, has the flexibility to modify the study protocol after a review of interim data. The aspect of the study to be modified is preplanned and the modification is based on analysis of data at a specified point in the study. Any trends in the drug’s efficacy and safety can be identified and the protocol modified to account for these patterns. This approach shortens the development process and increases the likelihood of success without compromising the integrity of the data or the validity of the conclusions.

The second type of study design is known as the ‘complete n-of-1 trial design’. An n-of-1 trial is one in which different treatments are assessed in an individual patient over time. The letter n represents the number of treatments and number 1 is the single patient. An n-of-1 trial typically involves multiple crossovers (multiple treatments at different times) and random allocation may be used to decide the order in which the patient receives the treatment under study or the control. This type of trial design is efficient and provides accurate and reliable information from a small number of patients.

Use of RWD and real-world evidence (RWE)
Due to concerns about the robustness of RWD and RWE, the FDA will only allow the submission of RWE in addition to substantial efficacy and safety data from RCT. The FDA defines RWD as ‘data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources’ and RWE as ‘the clinical evidence about the usage and potential benefits or risks of a medical product derived from analysis of RWD’.

Some of the concerns around using RWE to support drug development for rare diseases include whether the data are representative of these patients, heterogeneity due to differences across studies, confounding and interaction due to differences in patient demographics and characteristics, missing or incomplete data, reproducibility and generalisability, and data quality and validity due to selection bias.

Chow and colleagues argue that once these concerns have been addressed, RWD and RWE can be used to support regulatory submission using a 4-step approach that includes gap analysis between RWE and RCT data, assessment of data relevancy, quality and reliability, as well as ensuring data are fit-for-regulatory purpose.

Performing individual benefit-risk assessment

Finally, Chow and colleagues address the difficulty in balancing the benefits and risks of treatment in rare disease drug development. In 2023, the FDA published guidance on conducting risk-benefit assessments to support applications for drug approvals. Chow and colleagues propose two individual benefit-risk assessment frameworks they believe can provide a complete clinical picture of drugs under investigation for rare diseases.

The first, Bayesian multi-criteria decision analysis accounts for uncertainty in assessing benefit-risk balance, and the second, stochastic multi-criteria acceptability analysis accounts for statistical uncertainty and provides a consensus weight for describing the relative importance of the different criteria.

Proposed approach for regulatory consideration

Weaving together the current knowledge and potential innovations, Chow and colleagues propose a two-stage complex innovative design for rare disease drug development that includes a small RCT and a larger real-world study (Figure 1). The first stage aims to demonstrate not-ineffectiveness and/or not-unsafeness with a small sample size, while the second stage combines data from RCT and real-world studies to assess treatment effectiveness and safety. In addition, individual benefit-risk assessments would provide a complete clinical picture of the drug’s benefits and risks.

Chow and colleagues believe that these approaches would lead to ‘a more efficient, accurate and reliable assessment of the safety and efficacy of the rare disease test treatment under study in a more efficient way, even with only a limited number of patients available’.