Medicare's Unconvincing Racial Identification System

Normally, I would just have passed it by. The title, "More Accurate Racial and Ethnic Codes for Medicare Administrative Data" in the Spring 2008 Health Care Financing Review isn't exactly compelling.

Eighteen months ago I attended a symposium conducted by experts from Family Tree DNA, a service which matches and tracks mitochondrial and Y chromosome markers for those interested in genetic genealogy. One of the experts spoke about serious flaws in customary forensic identification of individuals' race assignments in criminal proceedings and the availability of technology advanced and convincing new methods.

As I read the Health Care Financing Review article, I noted no proof of the correctness of its claim that the authors' ethnic recoding and reidentification of beneficiaries improved the accuracy of coding for Hispanic and Asian or Pacific Islander beneficiaries. There was no independent technical verification of the validity of the authors' approach using available genetic confirmation. Basically, the authors seem to have claimed victory because their conclusions matched the biases they brought to the project.

This is no trivial issue because the statistical approach adopted by the authors, to "identify health care disparities between Medicare beneficiaries who are White, Black, Hospanic or Asian/Pacific Islander can safely proceed . . .it means it is possible to monitor efforts being made to reduce or eoiminate health care disparities among these groups." In other words, lots of dollars will be allocated differently if the authors' premises and conclusions are accepted.

Is this science or politics? Disease appears to match genetics more closely than loosely identified titles such as "race" or "ethnicity". Let's spend our money on science rather than biased conjecture which may lead to political decisions.