Date

Book/Article summary

Presented/ led by

23 Sep 2015

The meeting was based on the FDA draft guidance on Rare Diseases in Drug Development.

The FDA draft guidance is intended to provide guidance for sponsors of drugs and biological products to treat rare diseases. A rare disease for the purpose of this document is defined as a condition that affects less than 200,000 persons in the United States. The group discussed several aspects of the document. Some key points were:

- Several participants noted that they were surprised that this document provides guidance on the whole drug development programme and very little information specifically on conducting clinical trials in this setting.
- The remit of the guidance could be clearer. Would a targeted therapy for a very specific subgroup of a common disease be covered by the guidance?
- As this is a guidance document it would be beneficial if it were more descriptive than prescriptive, i.e. suggest appropriate approaches rather than describing what is currently used.
- There is no discussion in the guidance of extrapolation between studies in adults and paediatrics, or between similar populations or drugs, which might be an important consideration in some rare disease settings to make full use of all available information.
- Similarly, use of emerging large electronic databases and disease registries seems like an omission.

Overall the group found that this is a much needed guidance document which requires clarification in some areas. On behalf of the InSPiRe team attendees of the meeting sent comments on the draft guidance to the FDA.

- Thomas Hamborg

20 Oct 2015

24 Nov 2015

Meeting was based upon articles on EVI by Ades, AE, Lu, G, Claxton, K (doi:10.1177/0272989X04263162) and Eckermann, S and Willan, AR (doi: 10.1002/hec.1161).

The group discussed the paper by Ades et al on calculating the Expected Value of Sample of Information (EVSI) for Value of Information (VoI) analysis. The paper gave a short overview of the objectives of VoI analysis, and a summary of the distinction between EVSI, EVPI and EVPPI (the Expected value of Perfect Information and Partial Perfect Information respectively). The authors present a framework motivated by model-based cost-effectiveness analysis, which involve multiple parameters. Optimal study design involves identifying parameters for which data should be collected, as well as sample size and duration of follow-up. The paper provides algorithms for EVSI estimation in a range of situations, and show how estimation can be simplified when there are no correlations between focal and non-focal parameters. The group discussed the challenges involved in VoI analysis, including the difficulties involved in correctly specifying current uncertainty and capturing the relevant decision framework.

- Jason Madan

26 Jan 2016

.

19 Apr 2016

Meeting was based upon articles on Medical Statistics and ethics by Hutton, JL (2000) (doi:10.1177/0272989X04263162) and Hutton, JL (2001) (doi: 10.1002/hec.1161).

I found the speaker, Jane Hutton, really interesting to listen to. She shared with us ideas about the role of ethics in statistics, and statistics in ethics, that I hadn't really engaged with before. The discussion refreshed my memory about ethical issues I don't usually engage with, say, religious virtues of justice/faith/ hope and combined them with ones I think about a little more often, like informed consent and issues around involvement in RCTs. I found our discussion of drug trial recruitment the most interesting topic of the session, mainly because it provoked thought about the role of the scientist and study designer as an ethicist whether they are aware of it or not - and the potential consequences of less-than-fully-ethically-considered trial (and other study) designs.

- Rebecca Johnson