National Cancer Institute - NCI Division of Cancer Treatment and ...

O T H E R B I O S T A T I S T I C A L R E S E A R C H
Smoothing-Based Approaches
for Estimating the Risk of a
Disease by Quantile-Categories
of a Predictor Variable
When one collects data on a prospective
cohort, the standard method is simply
to categorize the key predictor variable
by the empirical quartiles. One may then
include indicator variables for these
empirical quartile-categories as predictors,
along with other covariates, in a
generalized linear model (GLM), with the
observed health status of each subject as
the response. The standard GLM method,
however, is relatively inefficient because it
treats all observations that fall in the same
quartile-category of the predictor variable
identically, regardless of whether they lie
in the center or near the boundaries of
that category.
Alternatively, one may include the key
predictor variable, along with other
covariates, in a generalized additive
model (GAM), again with the observed
health status of each subject as the
response. The alternative GAM method
non-parametrically estimates the functional
relationship between the key
predictor variable and the response. One
may then compute statistics of interest,
such as proportions and odds ratios, from
the fitted GAM equation using the empirical
quartile-categories. Simulations show
that both the GLM and GAM methods
are nearly unbiased but that the latter
method produces smaller variances and
narrower bootstrap confidence intervals.
This work by BRB’s Dr. Albert was motivated
by collaborative work on NCI’s
Polyp Prevention Trial.
24 ■ P R O G R A M A C C O M P L I S H M E N T S 2 0 0 6
Borkowf CB, Albert PS. Efficient estimation of
risk of a disease by quantile-categories of a
predictor variable using generalized additive
models. Stat Med 2005:24;623–45.
In case-control studies of genetic epidemiology,
participating subjects (probands)
are often interviewed to collect detailed
data about disease history and age-atonset
information in their family members.
Genotype data are typically collected for
the probands. In this article, Dr. Shih and
collaborators consider an approach that
utilizes family history data of the relatives.
They used the methods for estimation
of risk of breast cancer from BRCA1/2
mutations using data from the Washington
Ashkenazi Study.
Chatterjee N, Zeynep K, Shih JH, Gail M. Casecontrol
and case-only designs with genotype
and family history data: estimating relative-risk,
familial aggregation and absolute risk. Biometrics
[Epub Oct 20 2005].
Genomic Control for Association
Studies under Various Genetic
Models
Case-control studies are commonly used
to study whether a candidate allele and
a disease are associated. However, spurious
association can arise due to population
substructure or cryptic relatedness,
which cause the variance of the trend
test to increase. A novel genomic control
approach had been developed to estimate
the “variance inflation factor” using an
additive genetic model. Dr. Freidlin and
collaborators expand this approach to
recessive and dominant genetic models.
They also discuss appropriate uses for
their method and the one derived for
the additive model.