Rare Diseases Challenge Modelers to Deliver Equity

In many countries, economic modeling plays a pivotal role in demonstrating the value and overall impact of new treatments to Health Technology Assessment (HTA) agencies. That role is significantly more challenging when economic modelers find themselves supporting HTA submissions in very rare diseases. In this article, we explore these challenges and highlight the adaptations to the usual modeling approaches that are needed to bring new treatments to patients.

Introduction

Economic modelers face multiple challenges when supporting HTA submissions in very rare diseases. Given the small patient populations, modelers must reexamine established modeling techniques and adapt or fine tune them to capture how a rare disease may progress during treatment. A rare disease can be defined as a condition that typically affects less than 200,000 people in the United States,¹ or less than 5 per 10,000 people in the European Union.² Estimates suggest that 80% of rare diseases are genetic and 70% of them begin in childhood.

Uncertainty in clinical efficacy estimates caused by small numbers

For modelers, one of the consequences of such small patient populations in very rare diseases is the lack of data to allow robust estimates of clinical efficacy. The small numbers of patients that can be enrolled in trials and the variation in response to treatment seen in all conditions, whether rare or common, combine to create large confidence intervals (CI) in estimates of efficacy. This, in turn, may lead to models which, at one end of the CI, yield an acceptable estimate of cost-effectiveness for the treatment and, at the other, an estimate that is completely unacceptable, providing little guidance to a decision maker on whether the treatment should be funded. Modelers must be prepared to understand and use a broader set of information than that which comes from clinical trials alone, leveraging patient and clinician experience using systematic techniques such as expert elicitation or Delphi panels – both of which collect insights and judgements from specialists or experts in the disease.

Bridging clinical effects to hard outcomes

Beyond efficacy estimates, modelers have to establish plausible numerical links between observed clinical effects and the hard outcomes that healthcare payers are interested in – notably survival, health-related quality of life (HRQoL) and costs. The use of real-world data and, again, elicitation techniques, may well be required. This is likely to be particularly important in linking treatment effects to quality of life. Rare diseases that are the subject of drug development are often devastating in their effects on patients and their families. Important impacts of treatment on HRQoL may fall not just on the patient but also on their carers or family members. A particular problem is that these aspects of rare diseases are often not much investigated, if at all, until the possibility of treatment is real. A novel treatment means there is a good chance that the model being constructed will be the first to try to incorporate these hard outcomes. The absence of data may be a massive source of uncertainty, and so imagination and close collaboration with the affected patient community are usually necessary to address it. Being able to conceive and possibly execute secondary elicitation studies is a capability for the modeler which is much more rarely needed in more conventional modeling.

HTA agencies evolving approach to rare diseases

HTA agencies have responded to these challenges by being more prepared to consider conditional approval for reimbursement of a promising treatment. In such circumstances there has to be belief, founded in the modeling and its supporting argumentation, that the treatment has a real likelihood of being demonstrated to be cost-effective by the future evidence collected. For example, in Batten disease the National Institute for Health and Care Excellence (NICE) approved funding for cerliponase alfa in treating neuronal ceroid lipofuscinosis type 2, subject to an agreement for collecting long-term data. A similar agreement was made for migalastat in Fabry disease.^3,4 In these circumstances, the modeler has to utilize every piece of available evidence they can put together to climb the first hurdle and address what feasible future data collection could confirm the results.

Practical demands on economic modelers and their collaborators

Economic modeling in rare diseases requires a multidisciplinary approach. Beyond technical proficiency, modelers must have a real understanding of the condition being modeled, the health service framework in which patients are currently managed (in particular, how they are diagnosed, with late diagnosis often severely limiting the possibility of effective treatment in practice) and how the condition and its treatment impact patients and their carers. Modelers need to:

• Creatively supplement limited trial data
• Employ advanced elicitation methods to better understand both efficacy and its links to outcomes of interest to payers
• Ensure that the data creating the most uncertainty at the time of conditional approval can be collected practically in future

Conclusion

In the rare disease space modelers must embrace a wide variety of new challenges and be part of a wider multidisciplinary team than is typical in more common, well-studied diseases. If we want to bring new treatments to address these often devastating, rare diseases, we must continue to innovate and collaborate. Doing so will help us to navigate the challenges to deliver more equitable access to potentially lifechanging solutions for people with rare disease.