This course is an introduction to absolute risk, the probability of developing a specific outcome, over a specified time interval, in the presence of competing causes of mortality.  This course will define absolute risk and discusses methodological issues relevant to the development and evaluation of absolute risk models. We will present the cause-specific and cumulative incidence approaches to incorporating covariates, and discuss various study designs and data for model building, including cohort, nested case-control, and case-control data combined with registry data.  We will show how to evaluate the performance of risk prediction models and discuss the use of absolute risk in individual counseling for prevention strategies, including interventions that can have adverse effects.  We will address the impact of different distributions of model predictors and differences in verifying the disease status or outcome on measures of calibration, accuracy and discrimination of a model. We will present approaches for incorporating new information into existing risk prediction models, i.e. on “model updating”. We also discuss the potential use of such models for disease prevention in the population, including designing prevention trials, estimating the absolute risk reduction in the population from modifying risk factor distributions, the “high risk” preventive intervention strategy, risk-based disease screening, and resource allocation. 

In summary, attendees of the short course will learn what absolute (or “crude”) risk is, what it can be used for, how to estimate it from data obtained through various study designs, and how to assess the usefulness and validity of a model of absolute risk.

Much of the course material and additional details can be found in the book, “Absolute Risk: Methods and Applications in Clinical Management and Public Health”, by Ruth M. Pfeiffer and Mitchell H. Gail, CRC Press, Boca Raton, 2018.  Module registrants are not required to purchase the book, but it could be a useful and comprehensive resource.

Course prerequisites: The course is presented at a level that can be handled by statistics or biostatistics master’s students, or epidemiologists and medical researchers who have knowledge of basic statistics, epidemiologic designs, and a foundation in survival analysis. Those without a background in survival analysis could prepare for this module by taking SISCER Module 4, “Introduction to Survival Analysis,” on July 11-12.  SISCER Module 10, “The Evaluation of Biomarkers and Risk Models,” on July 22-23 would also provide useful preparation for this module. 

Ruth Pfeiffer and Mitchell Gail are Senior Investigators in the Division of Cancer Epidemiology and Genetics, National Cancer Institute, USA.  Both have published extensively on methods for developing, validating, and applying models for absolute risk.