Background While adaptive trials tend to improve efficiency, they are also subject to
some unique biases.
Purpose We address a bias that arises from adaptive randomization in the setting of a
time trend in disease incidence.
Methods We use a potential-outcome model and directed acyclic graphs to illustrate
the bias that arises from a changing subject allocation ratio with a concurrent
change in background risk.
Results In a trial that uses adaptive randomization, time trends in risk can bias the
crude effect estimate obtained by naively combining the data from the different
stages of the trial. We illustrate how the bias arises from an interplay of
departures from exchangeability among groups and the changing randomization
proportions.
Limitations We focus on risk-ratio and risk-difference analysis.
Conclusions Analysis of trials using adaptive randomization should involve attention to
or adjustment for possible trends in background risk. Numerous modeling strategies
are available for that purpose, including stratification, trend modeling,
inverse-probability-of-treatment weighting, and hierarchical regression.
GalloPChuang-SteinCDragalinV. Adaptive designs in clinical drug development—an
executive summary of the PhRMA working group. J Biopharm
Stat2006; 16:
275–83.
GottliebS. Deputy Commissioner for Medical and Scientific
Affairs, Food and Drug Administration—2006 Conference on Adaptive Trial
Design—July 10, 2006. http://www.fda.gov/NewsEvents/Speeches/ucm051901.htm (accessed 20
March 2011).
5.
DragalinV. Adaptive designs: terminology and
classification. Drug Inf J2006; 40:
425–35.
6.
BerryDA. Bayesian statistics and the efficiency and ethics of
clinical trials. Stat Sci2004; 19:
175–87.
7.
ChowS-CChangM. Adaptive Design Methods in Clinical TrialsChapman & Hall / CRC, Boca
Raton, 2006.
8.
ChowS-CChangMPongA. Statistical consideration of adaptive methods in
clinical development. J Biopharm Stat2005; 15:
575–91.
9.
FengHShaoJChowS-C. Adaptive group sequential test for clinical trials
with changing patient population. J Biopharm Stat2007; 17:
1227–38.
10.
ArmitageP. Sequential Medical Trials19752nd. BlackwellOxford.
HuFRosenbergerWF. Analysis of time trends in adaptive designs with
application to a neurophysiology experiment. Stat
Med2000; 19:
2067–75.
15.
RobinsJMHernánMABrumbackBA. Marginal structural models and causal inference in
epidemiology. Epidemiology2000; 11:
550–60.
16.
KangJDYShaferJL. Demystifying double robustness: a comparison of
alternative strategies for estimating a population mean from incomplete data (with
discussion). Stat Sci2007; 22:
523–80.
17.
GreenlandS. Principles of multilevel modelling.
Int J Epidemiol2000; 29:
158–67.
18.
BerryDA. Bayesian clinical trials. Nat
Rev Drug Discov2006; 5:
27–36.
19.
GreenlandSMaldonadoG. The interpretation of multiplicative-model parameters
as standardized parameters. Stat Med1994; 13:
989–99.
20.
GreenlandSRothmanKJLashTL. Measures of effect and measures of
association. In: RothmanKJGreenlandSLashTL. (eds). Modern Epidemiology20083rd. Lippincott Williams &
Winkins51–70.
21.
GreenlandSRobinsJMPearlJ. Confounding and collapsibility in causal
inference. Stat Sci1999; 14:
29–46.
22.
GreenlandSPearlJRobinsJM. Causal diagrams for epidemiologic
research. Epidemiology1999; 10:
37–48.
23.
GlymourMMGreenlandS. Causal diagrams. In: RothmanKJGreenlandSLashTL. (eds). Modern Epidemiology3rd. Lippincott Williams &
Winkins, Philadelphia,
2008; 183–209.
