Since 2002, Aetna Corporation has boldly walked where many insurers feared going. Led down this path by departing CEO and Chairman John Rowe, MD, the insurer has spent the past four years collecting data on the race and ethnicity of its members. The effort has been described as "courageous" by some and "racist" by others. Rowe, who is about to assume a job as health policy professor at Columbia University's Mailman School of Public Health, recently spoke to PHSI on why self-reporting information on race and ethnicity is an important step to narrow health care disparity gaps.

Q: What led Aetna to start collecting this data?

JR: I came to Aetna from academic medicine. I was president of Mt. Sinai Hospital and School of Medicine so I was very familiar with the issue of racial and ethnic disparities and the importance of cultural competence.

When I arrived, I realized that while we knew a lot about our members, we did not know their race or ethnicity. It turned out that no large insurance company was collecting information on this. I began to question how Aetna can make programs available to its members at high risk if we lack the data to know who is at high risk.

Q: How challenging was it for Aetna to seek out this kind of data?

JR: I was warned that this was the "third rail" of health insurance politics. There was a lot of initial resistance internally, with company staff issuing admonishments. If you started collecting this kind of data, staff warned that Aetna would be suspected of disadvantaging some of its members and feared it would be perceived as a racist company. My response to those concerns was that when a health plan knows that these groups are at special risk and does nothing about it, then it is perpetuating a racist policy.

A key experience that led to our decision was our involvement with Katie Couric's colorectal cancer screening program. It turned out that people of different races are at different risks for cancer. That tipped the scales and we began to collect data in 2002, coinciding with the release of the Institute of Medicine's landmark report on unequal treatment. That report really put a spotlight on this issue, and that was helpful to us when we went around and talked with customers.

Q: So what have you learned and what have you developed as a result?

JR: The programs we've developed include a pre-term labor program. African American women have consistently higher rates of premature birth rates than white women. When a health plan knows you are African American, it can provide additional educational materials and have nurses call you, monitor you, and enroll you in programs to avert pre-term labor. But a plan can't tell just by your name that you are African American. We need to have you report that information to us. We have other programs dealing with hypertension and cervical cancer. Vietnamese women are at high risk for cervical cancer and we have a program where we will pay for Pap smears for members at high risk. So if race or ethnic background constitutes a special risk, knowing this information permits us to identify those who are at special risk for specific conditions and target educational programs to them.

Q: Is Aetna the only health plan doing this?

JR: When we rolled this program out, there was a front page article about it in the Wall Street Journal and we got national attention. Afterwards, the federal government started a program called The National Health Plan Collaborative that brought together a dozen large health plans on how to collect data and how to use the data. Many health plans are now doing this, which I think is great. I feel that Aetna played a role in moving the field forward.

Q: This has been a voluntary strategy. Do you think self-identification should be mandatory?

JR: The literature is replete with evidence of racial and ethnic disparities. But it would be a mistake to mandate the reporting of this information. The insurance industry needs a longer track record of collecting this data without any untoward reactions. The people at Aetna were at no risk when we were collecting this data. Since it was primarily through large group insurance, there was no fear of medical underwriting.

Q: How do you collect the data?

JR: There are two ways to collect data. One is through individual reporting, and the other way is through geocoding. Geocoding uses Census data. It is not quite as good as individual reporting but if you have a large population of patients, it can help. If I was talking to a male whose last name is Wong living in Chinatown, I might assume that he is Chinese but if it's a woman, I can't assume anything. She could be married to a Chinese person but she may not be Chinese. Most health plans do both geocoding and collecting on individuals.

Q: How have you built public trust to get customers to report this information?

JR: Many of our customers are large companies/employers who are very interested in having their employees report this information. A lot of employees who are racial or ethnic minorities are not getting access to care they should be getting and paying the same premium. For business leaders that is a persuasive argument to urge employees to report this kind of information.

Q: Do you think you've made a difference? In what way?

JR: I think we have made a difference. This is a competitive business and to whatever extent people see this as a positive step, everyone will copy it. Some customers will criticize companies for not doing this.

Second, this isn't all we're doing. I required every clinician at Aetna to participate in cultural competency training. About 1,500 nurses and doctors were required to take a course on cultural competency, a skill they never learned in medical or nursing school. Massachusetts General Hospital under the leadership of Joe Betancourt has developed a course for our employees and other health plans are interested in such courses.

Q: If the data are proprietary, how much will it do for the greater public good?

JR: For proprietary reasons, we can't let others have access to our data but we can let them know how the collection is going, how much it costs to collect this data, and which groups respond less than others. Essentially, we can teach them about the perils and pitfalls of collecting data. If there was a nationalized effort we can inform how it might be structured.