Comments

1
Comments

• Kay Dickersin, MA, PhD
• University of Pennsylvania 4th Annual Emerging
  Statistical Issues in the Design and Monitoring
  of Clinical Trials
• April 13, 2011

2
Challenges to trials in 2011

• Maintaining a balance in initiating group
• Commercial interests (opportunities for
  blockbuster drugs)
• Public interests clinicians, payers, patients,
  those interested in evidence-based healthcare
  (conditions lacking adequate treatments, varying
  costs)

3
(No Transcript)
4
Challenges to trials in 2011 (contd)

• Costs of trials an impediment to implementation
  for industry and the public

5
Are design issues that reduce costs differently
important to industry and the public? If yes,
does it matter? (yes, if good science threatened)

• Adaptive designs
• Sample size can be modified mid-trial if
  pre-planned decision rules (theres the rub)
• Consumers (at least in cancer) are on board
• Smaller sample sizes
• Choice of primary outcome
• Desire for positive primary outcome and a smaller
  trial may lead to choosing cause-specific death
  as primary endpoint (not all cause mortality)
• Surrogate outcomes
• Can meta-analysis help?
• Non-inferiority trials
• Use of multiple treatments (network)
  meta-analysis - direct and indirect comparisons
• More people doing (and monitoring) trials without
  formal training or full partnership with
  methods folks(CTSAs,CER are factors).
• Is training necessary? What are consequences of
  lack of training, if any?

6
April 11, 2011
7
Requirements of the NHLBI RFA

• Successful applicants to this FOA will propose a
  design that overcomes the following challenges
• Minimize specialized infrastructure developed
  solely for conduct of the clinical trial.
• Minimize visits designed solely for the purpose
  of retrieving measurements for the trial.
• Explore novel methods of obtaining patient
  consent that minimize the need for specialized
  clinical trial staff and participant visits while
  still meeting ethical and legal requirements.
• Employ low-cost methods of monitoring study
  conduct such as adherence.
• The proposal should test the feasibility of the
  design in a clinical trial.

8
Examples of new design features

• Development of a method of randomization that is
  embedded within the electronic health record.
• Development of a method of participant consent
  that is brief and embedded within the electronic
  medical record while still meeting legal and
  ethical requirements.
• Development of a method to consent participants
  in a group format while still meeting legal and
  ethical requirements.
• Development of a method to utilize electronic
  health records to obtain study outcomes that is
  valid, complete, and easily generalized across a
  range of electronic health record systems.
• Development of a method to utilize and compensate
  health care providers to identify, recruit, and
  randomize eligible participants during routine
  clinic visits.

9
Role of the consumer