Abstract
I
It is very interesting to notice that the proportion of black people in the SEER registry 1973–2011 is just 10.2%; however, in the 2010 U.S. census, the African American population represents 13.6%. 3 This difference, which can be explained by an increase in the black population over the past four decades, could affect the interpretation of the study results. Underestimating the proportion of the population with higher mortality rates could produce an underestimation of mortality rates in the whole population and, therefore, in some of the measures of association like population attributable risk. In addition to demographic changes over time, the discrepancy between proportions could be due to factors that affect access to cancer healthcare as lower proportion of insurance coverage and lack of trust in the health system among not Hispanic black people. 2 We could not find this subject contemplated in the article discussion.
Another interesting point is that Hispanic ethnicity was not assessed in this study. This is important because 16.3% of U.S. population and 11.9% of all whites considered themselves Hispanic as per the last census. 4 Conversely, 53% of all Hispanics considered themselves whites. Categorizing most of the Hispanic population among white people could produce some bias in the results because they could have lower access to healthcare, lower socioeconomic status, different healthcare outcomes, and higher CV risk factors than whites. This is especially important since the authors are studying CV causes of death and those are higher among Hispanics. 5 It is possible that mortality rates among Hispanics are closer to those exhibited by blacks, or that the mortality effect shown in this article affects similarly the Hispanic population. We believe that a more accurate analysis could have employed four demographic groups, or at least should have excluded Hispanics from both study groups (as they did with other races) to obtain more rigorous results.