Martin Kulldorff

I develop new statistical and epidemiological methods for disease surveillance, many of which are relevant for cancer research.

For geographical surveillance, cancer cluster evaluations and the early detection of disease outbreaks, I have developed spatial and space-time scan statistics to detect and determine if observed disease clusters are likely to be due to chance or not, adjusting for the multiple comparison inherent in the many potential cluster locations and sizes possible. These methods are available for different probability distribution to handle caner incidence, mortality, stage, histology, treatment, screening or survival.

For pharmacovigilance, I have developed tree-based scan statistic data mining methods to detect and evaluate drug and vaccine adverse events with minimal prior specificantions.

In addition to my methodological work, I have collaborated on various cancer research projects as a biostatistician. This work includes health services research on screening, vaccines and other forms of cancer prevention, nutritional cancer epidemiology studies and cancer surveillance efforts.