Adaptive Design Discussions
April 18, 2011 Meeting
What do we know?
Jim spoke about new challenges. Are we talking about dose selection? Are we talking about outcomes based research? Sue Jane at the FDA has produced very cautious guidance parameters.
Chris wanted to know where is the discussion of Phase I and Phase II?
Jim said that we do not want to exclude sequential trials.
Jerry pointed out that these are all operational issues. He thinks of them as one. The DIA has already produced lots of information on these subjects.
We realized that we did not know about all the information that the DIA has produced.
Chris suggested that we identify all the various groups with information about adaptive clinical trial design and bring the information together in one place. We need to start with Don Berry and the DIA information.
Frank suggested that we take a more radical approach. How about a final document on adaptive clinical trial design produced by industry, FDA, and academia? We would present a discussion of all that is currently known about adaptive clinical trial design. We could start with a survey and then a position paper.
Laurel liked the idea of a position paper that was open to discussion. She would like the state of current affairs laid out with concerns about how this type of methodology would work in practice.
What Is Really Needed
Frank suggested that we step back and look at what do biostatisticians need to know about adaptive clinical trial design to use it in practice. Let's narrow the focus. What is the single biggest obstacle in adaptive clinical trial design? What are the budget challenges?
Jerry said that the DIA is tackling these ideas in the Drug Information Journal. He said that Gaydos has published material on the operational issues.
Chris would like to see the issues compiled in one place. There are so many areas to take into consideration: regulatory, data management, budgetary, etc
Laurel would like to see case examples discussing the political and statistical issues.
Frank would like to see reviews discussing why we care about multiplicity issues and what is the Bayesian debate.
Jim would like to rethink how we can move the CTSpedia forward in this endeavor. Can we put together a compendium of links and information?
Chris wanted to know what would be useful. There is lots of junk out there. We need to know what is the best in the area.
Jerry said that this would be very valuable. We need a resource that we trust to link us to all the information.
Chris wanted a discussion of strengths and weaknesses. Frank said that we need templates and examples with scripts that others can use.
Mary said that she will post any information that the group would like to send in. We will also announce this activity at the BERD Face-to-Face Meeting on April 27th.
March 28, 2011 Report
DIA Computational Science Meeting
Presentation by Partnership in Applied Comparative Effectiveness Science (PACES) discussed plans for the studies that will be used at Johns Hopkins to look at comparative effectiveness research. A number of meeting participants were upset that only Johns Hopkins investigators and the FDA will have access to the FDA. The group will publish their methodology and findings but no data will be available to the public.
Steve Willson's Suggestions for Our Adaptive Design Group
Steve felt that an educational link is missing in all of the FDA work. He suggested that Jerry contact Sue Jane Wang and Bob O'Neill to talk about how we can work together. Steve was particularly interested in Chris' educational materials.
March 4, 2011 Meeting
Attendees: Peter Bacchetti, Mary Banach, Laurel Beckett, Frank Harrell, Chris Lindsell, Jerry Schindler, Sally Thurston, Jim Ware
Introductions - Jim Ware
Today we are bringing together many threads in clinical research. We have Chris Lindsell (BERD University of Cincinnati), who has just submitted a CTSA supplemental application with the University of Texas to study adaptive clinical trial design in cardiovascular research. We have Jerry Schindler (Merck), who studies adaptive clinical trial design in drug development. Jerry is one of the planners for the two day Harvard spring Merck-Schering-Plough Symposium on Adaptive Clinical Trial Design. We have Peter Bacchetti (BERD UCSF) and Sally Thurston (BERD University of Rochester), who are in charge of online resources and educational materials for BERD. We have Laurel Beckett and Mary Banach (BERD UC Davis), who are responsible for CTSpedia administration and posting online materials. And we have Frank Harrell (BERD Vanderbilt), who is taking the lead in developing materials for reproducible research in clinical research.
We are here to discuss sharing of activities and what role BERD and the CTSpedia should play in this endeavor.
Jerry asked for a brief history of BERD. Jim explained that the GCRCs were replaced by the clinical and translational science awards (CTSA) to academic medical centers to expand the scope of clinical research activities at these sites. One of the key function committees (KFC) within CTSA is the Biostatistics, Epidemiology, and Research Design (BERD) KFC.
Chris Lindsell's BERD Adaptive Design Supplement
The University of Texas BERD site contacted Chris about his work in the ER and acute heart failure. The Texas site, specifically MD Anderson Cancer Center, was working with Don Berry on adaptive clinical trial design. Last August they had a joint conference on adaptive clinical trial design. They realized that they knew maximum sample size but how do you do plan for drug changes and all the various groups, e.g., the several different phenotypes in heart failure and how they respond to therapy? They felt that they needed to design an educational portfolio for everyone involved in these trials.
Chris has case-based materials that they will be posting to the CTSpedia. They will be creating templates and dynamic documents for everyone to share. They will discuss regulatory, funding, protocol, administration, etc issues.
Jerry Schindler's Adaptive Clinical Trial Design Work
Jerry is very active in setting up DIA working groups in adaptive design and beyond. The impetus for these groups is scientific issues that arise at the pharmaceutical companies. In the adaptive design group they are looking at pharmaceutical and regulatory issues, such as operational issues, Type I Error, estimation when a change needs to be made in the trial design.
Some of the other DIA groups are looking at comparative effectiveness research, clinical trials using participants from different regions of the world, biosimilar products, etc. The committees are composed of FDA, regulatory, pharma, academic and other end users. The DIA steering committees each have a chair and generate materials. They have summaries, worked examples, case studies and are thinking about what is the best repository for these materials.
Where Do We Go from Here?
What are the opportunities? What can the BERD CTSpedia do?
Jim suggested that we step back and talk to each other and then meet again. We need to look at what we can and cannot do. CTSpedia has had a rocky history. The Statistical Graphics collaboration with the FDA, the pharmaceutical companies, and BERD is looking very good. Does CTSpedia now have a foothold? Is it appropriate for the CTSpedia to take on a collaboration with the DIA? From Frank's perspective is this the best way to use our resources?
Jerry suggested that we look at Robert Gentleman's Bioconductor (http://www.bioconductor.org/) for ideas on site organization.
We will email and then schedule a meeting when Mary gets back from the DIA Meeting. We can also use the comment fields below to discuss our questions and concerns. Please add a comment and then click on "Add comment". For new questions click on Edit on the top of this page, go down to "ADD Question," add a question for discussion, and click on "Save" on the bottom of this page.
Jerry suggested that Mary contact Steve Wilson at the FDA for FDA thoughts on this work. Mary will see Steve on March 14th and will talk to him about our discussions.