Ross Prentice (l) speaks with Marvin Zelen.

The award recognizes outstanding leadership that has greatly impacted the theory and practice of statistical science, especially the awardee's contribution to the creation of an environment in which statistical science and its applications have flourished. The lecture was the concluding event of a two-day Harvard/Schering-Plough biostatistics workshop on statistical issues in drug safety monitoring.

"There are great leads that haven't been exploited," said Prentice, a professor of biostatistics at the University of Washington School of Public Health and Community Medicine in Seattle. For example, evidence suggests that Asian women who move to the United States end up with a 60 percent jump in breast cancer rates as a group within a decade of migration, and no one knows why.

Prentice aims to improve the design and analysis of disease prevention trials and observational studies.

"It's worth pondering how the research agenda and research portfolio can be developed for large-scale clinical testing of chronic disease research prevention efforts," he said. "An enhanced primary prevention research program has great potential to identify practical means of reducing the risk of prominent cancers and other chronic diseases."

For example, testing the effectiveness of lifestyle interventions, while seldom done, may reap benefits. The Diabetes Prevention Program, a major clinical trial involving people who don't process glucose properly, incorporated both physical activity and improved diets.

The resulting 60 percent reduction in the incidence of type 2 diabetes proved so effective that researchers stopped the trial early to enable people in other arms of the trial to take advantage of the benefits. The Diabetes Prevention Program was primarily funded though NIH and Department of Health and Human Service agencies.

Unfortunately, randomized control trials with disease outcomes often may be prohibitively expensive and impractically long, Prentice said. Researchers need to develop new methods and study designs that can improve reliability and reduce cost. Until then, the public health community must often rely on association and cohort studies to develop and test hypotheses, which can be affected by random or systematic biases, he said.

To help identify prevention interventions that are worth testing, Prentice called for a standing multidisciplinary body that includes basic, clinical, and population research and that deals with a wide range of health outcomes.

--CCM