Abstract A67: Cancer rate estimates for the metropolitan Detroit Arab American population

Objective: Arab Americans (ArA) are classified racially as white by many federal agencies and thus few statistics are kept on this group, but local and regional studies have shown this group is not only culturally different but may face unique health challenges. Because of racial classification as white, both numerator and denominator data are lacking. In addition, as with other minority/immigrant groups, ArA are subject to undercounting, thus rendering population estimates suspect. Therefore, little is known about cancer in this population because incidence and mortality rates cannot be calculated due to lack of numerator and denominator data. We assessed different population numbers for estimating Detroit-area ArA cancer incidence and mortality rates, while using a name list and algorithm for numerator data.

Methods: Age-adjusted rates were calculated following the National Cancer Institute's Surveillance, Epidemiology, and End Results (SEER) methodology. Numerator data (new cancers and cancer-related fatalities, 5-year average from 2000–2004) were obtained by linking the Detroit SEER database with an Arab/Chaldean surname database and algorithm previously developed and validated. For the denominator data, we used the Public-Use Microdata Sample (PUMS) from the U.S. 2000 decennial census to calculate two different population estimates. The first population estimate (ArA Pop1) used self-report of Arab or Chaldean ancestry. The second population estimate (ArA Pop2) was a combined value of Arab/Chaldean ancestry, birthplace in an Arab League country, or Arabic/Syriac language spoken at home. These responses are from the U.S. Census Bureau long form questionnaire. To assess the feasibility of calculated rates, we compared incidence and mortality rates to those published by SEER for Detroit black and white population groups.

Results: Both cancer incidence and mortality rates calculated using ArA Pop2 as the denominator were comparable to published rates for Detroit black and white subpopulations. We also calculated incidence rates for the 4 most common cancers (breast, prostate, lung, colorectal), as well as overall mortality, cancer mortality and heart disease mortality. ArA Pop2 again provided comparable estimates. Confidence intervals are fairly large due to small numbers. The published rates for Detroit black and white cancers are included within the confidence intervals for ArA rates; the rates for the 3 sub-populations are similar for all sites combined and the most common cancers.

Conclusions: Using a surname matching system for numerator estimates and several indicators of ArA ethnicity from a publicly available dataset to estimate the denominator provides a reasonable approximation of the cancer burden in the Detroit ArA population. This approach may work for other common chronic diseases and is useful for public health surveillance activities. It also will be helpful for health planning in geographic areas with large ArA sub-populations, such as metropolitan Detroit.