RCT class # 13 Flashcards
Lecture 13 is on what
additional trial designs: simple; N of 1; adaptive,
SMART
Platform trials
what is a large simple trial
an RCT with a relatively large sample size, few variables and pheraps short follow up.
Pros to large simple design
low cost, suitable if intervention is expected to have modest consistent benefits and is easily administered in a large recruitable population.
Cons of large simple design
they are unsuitable if the interventions are complex.
- long term followup with adherence required
- subgroups analysis or supportive analysis are needed
one period of an N-of-1 trial
one pair of an N of 1 trial
one period is the experimental trial then the second period of the pair is the placebo. This is a pair.
pairs of treatment for an n of 1 trial
ie period one is experimental
and the other is alt period. the order of the pair is randomized.
how does the clinician monitor the therapy in n of 1 trials.
drug diaries from the pts
the condition of the n of .1 trial must be ___
chronic bc acute conditions will resolve too quickly.
effects of a therapy (experimental or alt )must be short or long in 1 trial
short (or will last too long)
to make a decision of efficacy, harm or lack of effect usually need how many treatment periods
3 periods.
who is the control of an n of 1 trial
the patient is there own control. (they are the best control, basically a counter factual
example he gave of n of 1
hiv med in a patient for whom SOC has not been effective.
main advantage of N of 1 trial
avoids long term expense.
How does n of 1 move a drig toward approval w FDA
it doesn’t.
adaptive design def
adapt some aspect of design during the trial
adoative design formal def
Adaptive Design for a clinical trial is a trial protocol that allows certain design features to change from an initial specification based on evolving trial information while maintaining statistical, scientific, and ethical integrity. (interm data assessment
example of adaptive designs
stop for futility/benefit./efficacy
-change of a dose to reach
example of adaptive designs
changing antidepression drugs based on adaptive designs
key for adaptive designs
must be prespecified
key best way to use adaptive designs
adjust sample size based on nuisance parameters. (ie not the observed treatment effect itself
3 examples of nusciance parameters of adaptive designs
i. event rate lower than expected
ii. variability in a contious outcome is gr than expected
iii. worse adherence than expected.
crossover design only done for __ dz
chronic (bc cant go away)
withdrawal design is only valid if ____
nonrestrictive
factorial designs problem
low power need large sample size.
why is it usually inappropriate to have an adaptive design as a phase III (other than interim analysis etc) (2)
exceptions (2)
- issues with type I errors control.
- The question changes.
exceptions (2)
- prospective
- type I error controlled and question is clear
adaptive designs must maintain these 3 forms of integrity
- statistical 2. scientific. 3. ethical integrity
statistical problem if adaptive designs
Statistically valid only if all adaptations which may occur are specified in advance, and control of overall type 1 error at 0.05 or less error is demonstrated.
4 expamples of newer adaptive designs he gives
- dose selection studies drop an inferior arm
- adaptive randomization based on relative treatment group responsive
- adapt of sample size based on interim effect size estimates not just nuciance parameters
(such as
sd)
platform trials
test multiple treatments within the same trial (ie genetic studies that base treatments based on gene signature ie personalized medicine).
problem with platform trials
- control of type i error is unresolved still need phase III trial.
**platform type trials are useful as at this level of trial
phase II
SMART designs stands for
Sequential multiple assignment randomized trial design –> basically decision analysis type designs
smart designs are a fancy type of __
adaptive design
problem with SMART designs
increased chance of alpha error.
again biggest problem with adaptive, SMART and platform designs
controlling the alpha error.
Dr. JT likes adaptive designs for phase __
II.
what is an enrichment design
only take patients with treatment effects or failures.
based on FDA rules for a phase III trial a ___ trial design should be used
simple
what is a program of successor trials
?
