Record Details
Abstract
In December 2003, the NAACCR Call for 1995-2001 Data required that all submitting registries run the NAACCR Hispanic Identification Algorithm (NHIA) on the file and the 5 years, 1997-2001, were used to evaluate the outcomes of the algorithm in registries throughout the United States. This evaluation was conducted in anticipation of NAACCR publishing rates for Latinos annually, beginning with the December 2004 NAACCR Call for Data submission for the 2005 Cancer in North America. Fifty-one registries submitted a data file and only four did not run NHIA. For 26 registries, this was the first time that NHIA had been run. Of the 47 registries that ran NHIA, four registries encountered execution problems. Among the 43 registries that correctly ran NHIA, seven registries had rates that were inconsistent with rates produced by other registries. Among these seven, data from two registries produced rates that were too low. Neither registry had attained the NAACCR data quality standard for completeness (i.e., at least 90 percent), which may contribute to the lower incidence rates. Data from five registries produced rates that appeared to be too high, three of the states with extremely small numbers of Latino cases, which apparently were not able to produce a stable estimate, even over a 5-year period. NHIA is a useful and generally reliable method to enhance the ethnic identification of the U.S. Latino population. NHIA will continue to be evaluated to improve the validity of indirect assignment of a Latino ethnicity.
Keywords
Keyword |
evaluation
|
nhia
|
submission
|