Page 52 - Contributed Paper Session (CPS) - Volume 7

CPS2028 Ayon M.
1. Introduction
There is great interest in the possibility that clinical trials can be designed
with adaptive features that may make the studies more efficient, more likely
to demonstrate an effect of the drug if one exists, or more informative. An
adaptive clinical trial is a trial that evaluates a treatment by observing patient
responses on a prescribed schedule, and modifies parameters of the trial
protocol in accordance with those observations. The adaptation process
generally continues throughout the trial. Clinical trials are often designed with
adaptive features to force balance in sequential allocation of patients between
two or more competing treatments. In order to address the ethical criteria of
a clinical trial, it is also used to force imbalance by allocating a greater number
of study subjects to the better-performing treatment.
A clinical trial is a complex experiment on humans with multiple and often
competing experimental objectives. Here, several treatments for a disease are
compared with the purpose of obtaining information on the performance of
the treatments. Since it involves human patients, there is an ethical concern to
treat as many patients as possible with the best treatment. At the same time,
there must be some allocation of patients to the worse treatment arm for
making useful statistical inferences about treatment comparisons.
An increase in the number of patients receiving better treatments leads to
sequential experiments in which data are analysed and new allocations are
made in the light of the estimated parameters. However, the advocates of
traditional balanced randomization designs argue with the fact that having a
balanced allocation of patients across the treatment arms helps estimate
treatment effects efficiently. Since clinical trials involve human patients,
balanced allocation can be a serious problem as one would be more inclined
to be treated with the better treatment and such balanced allocation allocates
almost half of the patients to the worse treatment arm. To balance these
competing goals of ethics and statistical efficiency in a clinical trial, response-
adaptive designs have been developed and used. A response-adaptive design
uses the available response and treatment allocation histories to skew the
treatment allocation probabilities in favour of the treatment arm found best
thus far in the trial. However, human patients are heterogeneous and therefore
one needs to take into account such concomitant information when allocating
a particular patient to a treatment arm.
Covariate-adjusted response-adaptive designs balance the competing
goals of assigning a greater number of study subjects to the better treatment
and achieving high statistical efficiency in estimating treatment effects in the
presence of covariates, while maintaining randomness in treatment
assignments. Investigators are often aware of important baseline covariates
that may have a strong influence on patient responses and they may wish to
adjust the randomization procedure for these covariates. Rosenberger and

39 | I S I W S C 2 0 1 9