Solutions for Quantitative Problems in Cancer Surveillance and Control
The Statistical Research and Applications Branch is involved in a wide variety of projects which use statistical and mathematical modeling to improve the presentation and analysis of national cancer statistics, and to provide a vital nexus between data and their utility for assessing progress in cancer control.
Examples of past and ongoing projects include:
- the estimation of the variability of vital rates: an improved confidence interval for directly standardized rates has been developed, and a method has been developed to estimate the degree of extra-Poisson variation of vital rates;
- methods to estimate the representativeness of the SEER data system and to project national incidence from SEER incidence;
- methods to estimate medical costs for cancer patients when follow-up is incomplete;
- models of relative and cause-specific survival including possible leveling off of the survival curve (i.e., cure);
- Joinpoint regression methods to estimate statistically significant changes in cancer trends.
