What do forest fires, drug addicts, and diabetes patients have in common?

Jason Sutherland successfully defended his Ph.D. thesis entitled "Multi-List Methods in Closed Populations with Stratified or Incomplete Information" where new methods were developed to estimate prevalence of diseases, conditions or populations-at-risk.

In multi-list studies, lists of members of the population of interest are created. For example, to estimate the number of diabetes patients, lists from hospitals, physicians, and clinics may be used. Then, subjects are matched across lists and the pattern of how many patients appear on all lists, on all but one list, on all but two lists, etc is used to predict how many patients didn't appear on any list. Jason extended this methodology to where lists may be incomplete (e.g. some lists may only record males 18-40, while other lists record males and females from 18-65), or different matching methods are used for different lists (e.g. some lists have name and phone number, some lists use phone number and health care number). These methods have also been used in resource management problems to estimate the number of forest fires, count the number of salmon returning to spawn, or the number of pairs of ducks in a breeding area.

This type of interdisciplinary work is a hallmark of our program in Applied Statistics at Simon Fraser University. For more information, please contact Jason Sutherland (sutherli (at) stat.sfu.ca) or his supervisor Carl Schwarz (cschwarz (at) stat.sfu.ca).