ABSTRACT: Population-based cancer registries rely on various methods to assign Hispanic ethnic identifiers to patients in the registry. The methods may result in misclassification of patient ethnic identities. Such misclassification may obscure the real incidence of cervical cancer among Hispanic women. This review summarizes previous literature on the accuracy of methods used to ascertain Hispanic ethnicity in numerator and denominator data for the calculation of cancer incidence. In addition, cancer registry ethnicity ascertainment methods were examined for six United States states (California, Florida, Illinois, New Mexico, New York, and Texas) that have a high proportion of Hispanics. The percentage of persons classified as Hispanic who self-identified as Hispanic (predictive value positive) in various reported studies ranged from 54 to 76% for women. The accuracy of ethnicity assignments based on either the United States census list or the Generally Useful Ethnic Search System (GUESS) program show slight differences in percentages of self-identified Hispanics who were classified as Hispanic (sensitivity among women: 62-80% for 1980 United States census list, 63-82% for GUESS program). Higher sensitivity and lower predictive value positive is achieved with a greater number of sources used. In conclusion, decisions about collecting racial and ethnicity information are influenced by demographic changes, immigration trends, changes in ethnic and racial identity, legislative needs, and public policies. The rapidly growing Hispanic population and the excess incidence of cervical cancer in this population requires improving the accuracy of ethnicity information.
Cancer Epidemiology Biomarkers & Prevention 11/2002; 11(10 Pt 1):979-84. · 4.12 Impact Factor