Absolute Risk Modeling
DCEG investigators in the Biostatistics Branch have developed models (such as the Gail model) for projecting the individualized absolute risk of certain types of cancer. These models have been used to counsel individual patients on their disease risk; to make more formal management recommendations, such as whether or not to take tamoxifen to prevent breast cancer; to design cancer prevention trials; and to assess the potential reductions in population absolute risk from preventive activities.
For clinical and public health applications, the absolute risk of disease is typically more useful than relative risk. Methodologic research is used to define the best kinds of data for estimating absolute risk, to develop strategies for selecting and combining risk factors to project risk, to develop criteria to evaluate the performance of risk models in various applications, and to conduct validation studies using independent data to determine the performance of a risk model.
NCI Risk Assessment Tools
DCEG investigators and their colleagues have employed absolute risk modeling to develop tools that clinicians and their patients can use to calculate risk of developing the following cancers:
The breast cancer risk assessment tool has been monitored to check its calibration over time and has been enhanced to improve projections for African-American and Asian-American women. Studies to improve projections for Hispanic women are ongoing. Computer programs for projections for these models are available in the Tools and Resources section of this web site under Risk Assessment Macros, Software Programs, and Tools.
For more information, contact Mitchell Gail and Ruth Pfeiffer.