Most clinical trials are designed with no reference to the size of the population in which the research is conducted. Whilst this may be reasonable in a large population, in rare diseases or other small populations it can lead to designs that are inappropriate. If the population under investigation is small, a large proportion of the patient group may be recruited to a clinical trial. Recruitment to one trial may thus have an impact on the conduct of other trials or even reduce significantly the size of the population receiving usual care. The value of one clinical trial must therefore be compared with that of other trials if research is to proceed efficiently[1]. We will develop novel methods for the design and sequential monitoring of small population clinical trials based on a formal comparison of the gain in information from a clinical trial with the cost of the trial.

Decision-theoretic methods explicitly enable evaluation of the level of evidence required from a clinical trial to best inform clinical practice. This in turn can lead to efficient and appropriate clinical trial design. In a small population setting, the small available sample size, the fact that recruitment to one clinical trial may limit recruitment to concurrent trials and the small size of the target population may mean that standard clinical trial designs, proposed in the setting of large patient populations, may be inappropriate. In this project we will explore the use of decision-theoretic and value-of-information methods for the design and sequential monitoring of clinical trials in small populations. This will lead to smaller studies that are more able to lead to appropriate decision-making.

The gain in information will reflect the size of the population for which the therapy is being developed, while consideration of costs will be in both resource use and opportunity terms, the latter being particularly important when a series of clinical trials is considered since in a small population recruitment to many trials simultaneously is likely to be impossible. Conventional sample size calculations will be replaced with calculations that explicitly allow for the opportunity cost associated with undertaking a trial of a certain size and sequential monitoring of trials might be used to be able to terminate trials rapidly when early results do not appear sufficiently promising. We will also explore the use of multi-arm trials in the small population setting. Such designs allow a number of potential novel therapies to be compared with each other and with a control simultaneously in a single trial. This can prove an efficient way to rapidly evaluate a number of treatments, particularly if early trial results can be used to drop effective treatments part-way through the trial[2],[3]. The decision-theoretic approach will be used to formally assess the potential benefits from such an approach as well as developing optimal strategies for choosing which treatments should be included or dropped from the trial.

The methodology developed will also build on the health economic value of information approach[4], and enable an assessment of the appropriate level of information required for clinical decision-making in the small population setting. Value-of-information methods have been explored extensively in the health economic literature, where value is defined in terms of the likelihood of adoption decisions changing in light of trial results, and is therefore linked to decision-making concerning the reimbursement of drugs. Extending this approach to rare diseases and small populations would be particularly novel and challenging, since reimbursement decisions involve balancing cost-effectiveness against complex and subjective considerations of equity.

Tasks

We will perform a thorough review of literature in the area of decision-theoretic approaches to clinical trial design and monitoring.

• We will develop decision-theoretic methods for the design and sequential monitoring of clinical trials in small populations. Such methods are based on an explicit modelling of the consequences of decisions, for example that to stop or continue with a particular trial, or to recommend or abandon a particular therapy, and a comparison of these consequences with the costs of conducting the trial to obtain the information on which these decisions are based. We will start by developing models for the decision-making process. These will be based on previous work that we and others have published, identified in the review in 2.1. The aim is to have a representation of the trial and decision-making process that is both realistic enough to lead to appropriate designs whilst still being computationally tractable. The study design process can then be optimised with respect to the costs and consequences specified.
• Costs and consequences of different trial designs or drug development programs will be compared on the basis of some gain or utility function. This can be based on either monetary costs and rewards or the costs of suboptimal treatment and gains from effective treatment. We will develop alternative utility functions and explore the resulting optimal designs. The utility functions developed will reflect the small population size and include potential gains from conducting alternative studies, so that the impact of reallocation of resources between studies can be explored. We will also explore the use of multi-arm trials in the small population setting.
• When the utility function coincides with that of bodies making drug adoption / reimbursement decisions, the decision-making problem is similar to that considered in the health economics value of information (VoI) approach, when the cost of information is weighed against the likelihood of adoption decisions changing in light of the new information, and the expected gain from the changed decision. We will develop novel VoI methods specifically in the small population setting and apply these to obtain optimal trial designs.
• We will use decision-theoretic methods to determine when it is optimal to introduce a novel therapy to clinical practice based on trial results, and when it is optimal to continue with further evaluation. This will indicate the level of evidence required for making this decision. Consideration of the population size may mean that this is quite different to conventional levels used in larger population clinical trial evaluation. The methodology developed will be illustrated through retrospective application to setting of one or more concrete examples of rare diseases. This will illustrate the gains from use of the novel approach relative to more traditional study designs. Through dissemination of the results via publications and the project conference we will also attempt to start to influence policy-makers in this area; we see this as a key step towards application of the methods developed.

Deliverables

• Report of literature review of decision-theoretic methods in clinical trial design: Initial report presenting findings of a literature review of use of decision-theoretic methods in clinical trial design. [month 13]
• Report on decision-theoretic methods for clinical trials in small populations: Interim report presenting developed methodology including possible decision models and gain functions for clinical trials in small populations. [month 22]
• Report on value-of-information methods for clinical trials in small populations: Interim report presenting developed methodology for use of a value-of-information approach in small population clinical trial designs. [month 34]
• Report on decision-theoretic designs for clinical trials in small populations: Final report presenting developed methodology, including evaluation of levels of evidence required from small clinical trials to inform decision-making and use of a value-of-information approach along with an example of retrospective application based on a realistic setting. Report to form the basis of a submission for publication in an appropriate high-quality peer-reviewed journal. [month 40]Schedule of relevant Milestones