Record Details
Abstract
There is considerable interest in routinely reporting cancer incidence data for racial/ethnic groups, in addition to white or black, as part of general surveillance summaries and site-specific investigations. Such reports can detect disparities in magnitude, severity, or trends of disease, and can help shape group-specific prevention and control strategies. There are concerns, however, with the quality of rate estimates for racial/ethnic groups using multi-registry aggregated data, including geographic variation in the accuracy of race-specific population counts and in the classification of incident cancer cases by race. When selecting registries to include in group-specific rates, a fundamental tension exists between interests of accuracy and interests of precision and comparability. A preferred approach has not been identified, and different strategies have been employed in recent analyses of Cancer in North America (CINA) data. Based on descriptive studies of ovarian and breast cancer, this project evaluates the influence of alternative geographic selection strategies on race/ethnicity-specific incidence rates, confidence intervals, and interracial comparisons. Changes in age-specific and age-adjusted rates, as well as rates stratified by tumor behavior and summary stage, are quantified. Rates and rate ratios are compared statistically. The working hypothesis is that there will not be meaningful differences under the various approaches, and that study inferences will remain unchanged.
Keywords
| Keyword |
|
cancer
|
|
combined
|
|
data
|
|
evaluation
|
|
incidence
|