Better Measure for Racial Disparities in Causes of Death
A recent statistic shows that a black woman is 20 percent more likely to die from cancer than a white female. A public health planner armed with this information would probably want to devote resources to eliminating the cancer risk for women, with a particular focus on African American women to address the obvious health inequity. But what if the statistic was wrong?
“It might mean that a planner would devote time and resources to a disease that actually exacerbates health inequality instead of eliminating it,” said IPR social demographer Quincy Thomas Stewart.
Currently, causal decomposition is the most common method used to estimate and compare death rates between groups of people. This method, however, harbors a major flaw, according to Stewart. “Causal decomposition does not account for the fact that underprivileged groups are more likely to die from nearly all other causes of death, skewing the results,” he said.
To correct this, Stewart developed the “cause-deleted index.” In it, he deletes each individual cause and then evaluates the impact of the missing data on the survival rates of blacks relative to whites. It tries to answer the question, “How much would the relative survival rates of blacks improve if deaths from cause X were eliminated?”
Stewart tested the method by compiling 2,000 unique pairs of hypothetical mortality profiles from vital statistics collected between 1940 and 2000 and then comparing the estimates both methods generated.
Quincy Thomas Stewart
In certain cases, results from both the causal decomposition method and cause-deleted index (CDI) agreed. So, for example, both identified heart disease as the leading cause in death disparities for women in the year 2000. However, they diverged on secondary causes. The CDI suggests that diabetes, AIDS, and strokes are the next leading contributors to disparities, while causal decomposition fingered cancer and diabetes.
Running the CDI for cancer, Stewart found a 35 percent increase in disparities between black and white women when it was eliminated as a cause of death. This means that in the real world, where cancer does exist, it operates to reduce the overall racial difference in female death rates.
In cases like cancer, which have a high prevalence, higher overall death rates among blacks are driving the causal decomposition estimates, which leads to inaccurate results about the contributions of these causes to racial differences in health, Stewart said. The cause-deleted index provides additional information that, when used with causal decomposition, can pinpoint the major causes of death behind health disparities.
“Using both methods would give us a better understanding of the causes driving the racial gap in death rates,” Stewart said. “And it would also provide a more accurate map for future research and policies to reduce these differences. “
Stewart, Q. T. 2011. The cause-deleted index: Estimating cause of death contributions to mortality. Mathematical Population Studies. 18(4):234-57.
Photo courtesy of flickr.