Statistical Methods for Case-Control Studies

Table of Contents Previous Next

"The sophisticated use and understanding of case-control studies is the most outstanding methodologic development of modern epidemiology."footnote 1

Norman E. Breslow, UW biostatistics professor in the School of Public Health and Community Medicine, was honored for career achievements with two major awards in statistics in 1995: the R. A. Fisher Award and the George W. Snedecor Award. Breslow is known for his work with case-control methods, which are the very cornerstone of epidemiological studies of health risk factors and causes of medical and public health problems.

Furthermore, Breslow was instrumental in founding the National Wilms' Tumor Study (NWTS) program in 1969. He now directs its statistical center at the Fred Hutchinson Cancer Research Center in Seattle.footnote 2 Over the years, through a combination of laboratory, clinical, and statistical findings, the NWTS has helped to raise the survival rate for this rare childhood kidney cancer from less than 60% to about 90% today, while at the same time reducing the risk of long-term complications such as scoliosis (curvature of the spine) and secondary tumors.

The basic strategy of a case-control study is to compare a group which has a particular outcome of interest ("case") with a control group ("control") that does not have that outcome but often matches the first group with respect to selected characteristics. Case-control methods have been used for many years to study cancer. For example, they were used to uncover the connections between lip cancer and pipe smoking, and between endometrial cancer and estrogen replacement therapy. Case-control methods are used to study rare diseases because the methods are quite efficient with respect to sample size in comparison with other approaches.

Breslow and colleagues pioneered much of the statistical foundation for sound and rigorous case-control methods and data analysis procedures, including methods to accommodate pair matching of controls to cases by means of conditional logistic regression. Breslow's recent work has emphasized the use of complex stratified sampling procedures in order to further enhance the efficiency of the case-control design.

  1. K. Rothman, Modern Epidemiology, Little Brown and Company, Boston, 1986, quoted in "Statistics in Epidemiology: The Case-Control Study," N. E. Breslow, manuscript based on the 1995 R. A. Fisher Lecture, to be published in the March 1996 issue of the Journal of the American Statistical Association.
  2. Quest (Newsletter of the Fred Hutchinson Cancer Research Center) 18, Fall 1994, p. 8-9.

Table of Contents Previous Next