Renal Cancer Patients with Unknown Ethnicity in Cancer Registry Data: Comparisons to Patients with Known Ethnicity.

Background: Information on ethnicity is important for health disparity research and health service planning. However, information on ethnicity is often incomplete in large routine databases such as cancer registries. This study aimed to compare survival status and other characteristics between cancer patients with and without information on Hispanic ethnicity in cancer registry data.

Methods: The study included 2,426 patients with clear cell renal cell carcinoma (RCC) diagnosed between 1988 and 2004 and identified from the US Department of Defense (DoD)'s Automated Central Tumor Registry (ACTUR) database. There were 1,353 non-Hispanic patients, 134 Hispanic patients, and 939 patients with unknown ethnicity. Patients were followed through death, date of last contact, or censored on December 31, 2007.

Results: Patients with unknown ethnicity exhibited significantly shorter survival than non-Hispanic or Hispanic patients (Log Rank P < .0001). Further analysis showed that compared to patients with known ethnicity, patients with unknown ethnicity were more likely to have advanced tumor stage at diagnosis and more likely to have missing information on tumor grade, size, and some demographic characteristics. After adjustment for demographic, tumor and treatment variables, patients with unknown ethnicity still exhibited significantly higher mortality than non-Hispanic patients (hazard ratio [HR], 1.69; 95% CI, 1.48-1.92), while Hispanic patients were not different from non-Hispanic patients (HR, 0.95; 95% CI, 0.71-1.28). The shorter survival in the unknown ethnicity group was consistently observed in subgroups by age, race, stage, grade, and surgical treatment, suggesting factors other than those investigated in the current study may play a role in the survival differences between patients with and without information on Hispanic ethnicity.

Implications: The poor survival of patients with unknown ethnicity in ACTUR warrants further research to elucidate missing mechanisms. Improvement in collection of data by reaching out for more engagement of patients, clinicians and registrars and increasing patient-provider interactions is needed.