Use of treatment information from a state central cancer registry in prostate cancer research.

A method was developed to categorize prostate cancer treatments for epidemiologic and outcomes studies. A total of 60,497 prostate cancer cases from the Florida Cancer Data System diagnosed between 2001 and 2007 were used. The classification has the following properties. First, the treatments classified in the same group are clinically comparable and capture all prostate cancer treatments or combinations of treatments (exhaustive classification). Second, the grouping was set up in a way that each patient is captured in only 1 treatment category without leaving out any patient due to treatment type (mutually exclusive categories). The prostate cancer cases were initially categorized into 14 combinations of treatment, which were then collapsed into 8 broader groups in order to maintain a large sample size for all groups, with treatments remaining clinically comparable within a group.