Patient Preference Elicitation and Incorporation into HTA/Access
Medical stakeholders and decision-makers have frequently held that patient preference information (PPI) (defined as the relative acceptability of a choice or alternative among outcomes that differ among alternative health interventions) and the methods for its elicitation are key components of a robust market access strategy; after all, who better to consider what is best for the patient than patients and those closest to them? Speakers at this year’s conference delved into the theoretical and practical challenges of PPI while offering guidance on how best to incorporate it into drug development strategies.
From a theoretical perspective, there are substantial concerns around PPI elicitation. Juan Marcos Gonzales (Associate Professor in Population Health Sciences, Duke University) posed the question of whether a caregiver had the right to provide preference–based information on a therapy that could help the patient.1 Controversially, he suggests that a caregiver, or any non-patient agent, should not have preferences measured, no matter how useful, because the mix between the caregiver as an agent and beneficiary of the therapy is too difficult to disentangle. The theoretical challenges to collecting data are compounded when considering how an individual’s own preference changes over time; Dr Eric Finklestein (Professor of Health Services and Systems Research, Duke-National University of Singapore Medical School) noted the concept of “short-term time-inconsistent behaviour”, the common phenomenon where individuals would make decisions that contrast with their own decisions at a separate point in time. On the other hand, some decisions do not change over time, and in yet other circumstances a patient may not be able to adequately express their own preferences. With all of these challenges, can PPI ever be collected in a robust manner?
Despite the challenges, PPI is still collected. However, for decision-makers to consider the findings as valid, the methods must be well-designed, validated and should not lead to burnout in patients. Another key consideration is “questionnaire burnout”, where individuals can be inundated with tests and questionnaires, particularly in clinical trials. A poorly-designed trial could mean that patients are asked preference-based questions multiple times, at the wrong time, or not at all. The FDA and EMA have published position papers on PPI, and the FDA has suggested that there are several points in a product’s lifecycle where PPI could be integrated and collected, from the prototype stage all the way up to the regulatory decision stage (Figure 1).2-4
ISPOR also convened a task force on “Using Patient Preferences to Inform Decision Making Good Practices” with the aim of providing a framework on how best to use PPI.5 With existing research focused heavily on methodological advancement to address some of the concerns mentioned above, the task force presented results on how to ensure usefulness and impact of PPI, regardless of the method used. The draft framework presented five key considerations when assessing the likely impact of a patient preference study: context, purpose, population, method, and impact (Figure 2). While all of these points are important themselves, the key takeaway is that the relevance and usefulness of a preference study requires interaction and alignment with the decision maker. This may be more difficult in countries where decisions are made through strict HTA processes where PPI often does not factor directly into a decision, but in the US, where value can be assessed in a more holistic manner, the discussions could lead to more targeted and better planned preference studies. Ultimately, this could allow for wider access and let American researchers lead the way in a field that will continue to grow in importance over the coming years.