Christina Kendziorski Talks About UWCCC Biostatistical Analysis
September 5, 2006

MADISON—Christina Kendziorski, PhD, member of the Chemoprevention research program, is intereviewed about biostatistics in cancer research.

What does a biostatistician do?
Biostatisticians help design experiments and analyze the results. Experimental design questions include specifying what, when, how many, and to what extent measurements should be obtained. For example, when searching for genetic factors involved in complex diseases such as cancer, the answers to questions such as—“What phenotypes should be measured?” “How many families should be included?” “For how long should the families be studied?”—are not always clear. What is clear is that some sampling designs are more likely than others to yield useful results. Biostatisticians can identify the data collection strategy most likely to accomplish the objectives of a given study at a fixed cost. Of course, measurements provide information only when appropriately analyzed. A biostatistical analysis often involves identifying differences between patients, or their illnesses, and determining which differences are important.

Can you tell us more about the UW’s Department of Biostatistics and Medical Informatics?
The Department of Biostatistics and Medical Informatics is part of the University of Wisconsin School of Medicine and Public Health. It was founded by Dr. David L. DeMets and is now a nationally recognized center for research and collaborative activity in biostatistics, bioinformatics, clinical informatics, biomedical computing, and data management. It grew from a division in the Cancer Center supported by a collaboration between the Cancer Center and the Department of Statistics in 1982. A Biostatistics Center was formed in 1984 and departmental status was attained in 1992.

Can you explain your research?
My research involves developing and applying statistical methods to address questions arising in genetics and genomics based studies of complex diseases. In the past 10 years, advances in high throughput technologies have made it possible to obtain thousands of measurements from patients, as opposed to just a few. This advance has provided the opportunity to address increasingly complex questions, but doing so requires the development of new statistical methods. That is where my research comes in. A good example of one of these high throughput technologies is the microarray. Microarrays allow us to follow the action of thousands of genes simultaneously. They are used to identify the genetic basis of complex diseases, to identify high risk populations, and to define disease subtypes that may respond differently to treatment. For each of these purposes, an investigator must be able to distinguish between spurious and significant patterns in the data. My group works on methods that allow this type of distinction to be made.

Can you describe your decision to work at the University of Wisconsin?
Many biostatisticians develop methods and then seek out a problem to which the methods can be applied. I find the problem first. Every problem that I am currently working on arose in the lab of one or more of my collaborators. When you work in this way, being near top scientists is critical. The University of Wisconsin is consistently ranked among the top research institutions in the world, and as a result we can attract the top scientists. There are, of course, comparable research institutions; but the UW has a history and personality unlike any other—one that seemed to fit me quite well.

What challenges lie ahead for you and individuals involved in your research?
As I mentioned earlier, recent advances in technology have allowed us to pose questions that could not have been addressed, even 10 years ago. With the increasingly complex questions comes the need for insights not only in statistics, mathematics and computer science, but also in the biological sciences. I cannot simply take a biological problem to my office, address it using some mathematical approach, and then move on to the next. Today’s problems require persistent iteration between methodological developments and biological reality checks. Most statisticians do not have labs of their own where these checks can be done; indeed most of us have very little training in biology. Changing that is the most significant challenge for our field.

In your opinion, what difference are you making in the fight against cancer?
Biostatisticians play a critical role in the fight against cancer. We don’t see patients, invent drugs, or build radiotherapy machines; we engage in the fight by turning data into information. Our experimental designs and analytic methods enable health scientists and professionals to better detect and treat disease.