Page 394 - Contributed Paper Session (CPS) - Volume 6
P. 394
CPS1995 Daniel B. et al.
A mixed models approach to extrapolation of
clinical data
Daniel Bonzo, Evelyn Wang, Jillian Prescod
Global Biometry, LFB, 175 Crossing Blvd, Framingham, MA 01702, USA
Abstract
Extrapolation of clinical trials data is being accepted increasingly by regulatory
agencies as a means of generating data in diverse situations during drug
development process. We consider this problem of extrapolation using the
concept of estimand [Akacha, M., et. al. (2017)] under a mixed models setting.
The concept of estimand captures population, endpoint, and a measure of
effect – in general, one can think about extrapolation of historical data from
one estimand to another closely related estimand. A likelihood procedure is
presented for estimating the parameters of interest under a generalized linear
models setting. Allowing the possibility of censored/grouped data transforms
the likelihood expression into a likelihood involving counts of interval data by
utilizing the latent variable concept. A relatively simple estimation and testing
construction is obtained when one assumes that the underlying distribution
comes from the family of exponential distribution. Using large sample
approximation, we show an approach for goodness-of fit testing and
estimation of parameters of interest. Finally, we demonstrate the utility of this
construction in a setting were we evaluate efficacy in a subgroup of a clinical
trial population using a marker of efficacy. In conclusion, the concept of
estimand allows an extrapolation approach that can cover a broad array of
applications and settings, including the case when censoring is allowed. Useful
expressions of estimators and tests are given for application purposes, though
they require sufficiently large sample to be efficient. These expressions have
an intrinsic weighting mechanism for the different sources of data.
Keywords
Estimand; latent variable; censored; likelihood; goodness-of fit
1. Introduction
In this paper we present a procedure for extrapolation that can be applied
in a general setting where the underlying distribution comes from the family
of exponential distributions. This procedure can be used to treat a variety of
problems in drug development that call for extrapolation of results, e.g., from
adult to pediatric population, from one or several geographic regions to
another such as in bridging studies from one indication to a related indication,
383 | I S I W S C 2 0 1 9