Department of Biostatistics Seminar/Workshop Series

Statistical Methods and Statistical Pitfalls in Biomarker Research

Frank E. Harrell, Jr., PhD

Professor & Chair, Department of Biostatistics, Vanderbilt University School of Medicine

Wednesday, June 4, 2008, 1:30-2:30pm, MRBIII Conference Room 1220

Intended Audience: Persons interested in applied statistics, statistical theory, epidemiology, health services research, clinical trials methodology, statistical computing, statistical graphics, R users or potential users.

Biomarkers may be useful for diagnosis, prognosis, or monitoring the effects of treatments when the actual clinical outcome is difficult to measure or occurs too far in the future. Discovery of biomarkers and determining how best to use them are important areas for research. Like some other high technology arenas, biomarker research often involves cutting corners in the rush for success. This talk will catalog some of the common mistakes in biomarker research, with special attention to the following issues:

1. Failing to demonstrate that the biomarker information is new
2. Limitations of sample information for picking the "winner" from among hundreds of potential markers
3. Categorizing continuous biomarkers
4. Using improper accuracy scores such as classification accuracy
5. Failure to use reliable methods for validating prediction or classification rules