October 25, 2019

Mr. David S. Wichmann Chief Executive Officer
UnitedHealth Group Incorporated
9900 Bren Road East
Minnetonka, Minnesota 55305

Dear Mr. Wichmann:

We write in response to reports that Optum's data analytics program, Impact Pro, significantly underestimates health needs for black patients. Specifically, a recent study published in the journal Science showed that Impact Pro's flawed algorithm ranked healthier white patients as equally at risk for future health problems-and therefore in need of more intensive healthcare intervention-as black patients who suffered from far more chronic illnesses. ¹ Based on these results, which Optum markets to health insurers and healthcare providers, black patients' health concerns are deemed "less than" those of white patients. These discriminatory results, whether intentional or not, are unacceptable and are unlawful in New York. We call on you to immediately investigate these reports and demonstrate that this algorithm is not racially discriminatory or to cease using Impact Pro (or any other data analytics program) if you cannot demonstrate that it does not rely on racial biases or perpetuate racially disparate impacts.

As you are aware, the N.Y. Insurance Law, N.Y. Human Rights Law, N.Y. General Business Law, and federal Civil Rights Act, protect against discrimination for certain classes of individuals, including disparate treatment and disparate impact. These laws govern the activities of insurers who are responsible for complying with these anti-discrimination laws irrespective of whether they themselves are collecting data and directly underwriting consumers, or using and developing algorithms or predictive models that are intended to be partial or full substitutes for direct underwriting. In short, neither you nor any other healthcare or insurance entity may produce, rely on, or promote an algorithm that has a discriminatory effect. In addition, we are troubled by the potential for conflicts of interest to the extent that the entity that controls the algorithm also is affiliated with both providers and insurers, and the algorithm considers costs.

This country has a long and troubled history of racism with respect to healthcare, both conscious and unconscious. Stemming from the earliest days of segregation and exclusion, black communities today still face barriers to healthcare and unconscious bias in the care that they do receive. It is well documented that black patients endure longer Wait times than white patients when seeking treatment,² and that their claims of pain are taken less seriously than white patients'.3 Their medical histories, therefore, are less likely to reflect their true medical needs than white patients who have historically been given greater medical attention. This is a well­ documented problem. For example, in 2018, New York Governor Andrew Cuomo created a comprehensive initiative to combat maternal mortality and reduce racially disparate outcomes. Earlier this year, Governor Cuomo signed legislation to create a Maternal Mortality Review Board charged with reviewing the cause of each maternal death in New York State and making recommendations to the Department of Health on strategies for preventing future deaths and improving overall health outcomes.

This is why reports of the discriminatory effects from your algorithm's reliance on past health costs are so troubling: a black patient's actual medical needs may not be accurately captured by his or her prior health costs. By relying on historic spending to triage and diagnose current patients, your algorithm appears to inherently prioritize white patients who have had greater access to healthcare than black patients. This. compounds the already-unacceptable racial biases that black patients experience, and reliance on such algorithms appears to effectively codify racial discrimination as health providers' and insurers' policy. If true, this outcome has no place in New York or elsewhere. Patients should receive care based on their current needs, not their history of inclusion or exclusion, particularly where that history runs along racial lines.

New York will not allow racial bias, especially where it results in discriminatory effects that could mean the difference between life and death for an individual patient and the overall health of an already-underserved community.

Sincerely,

Linda A. Lacewell, Superintendent
New York State Department of Financial Services

Howard A.Zucker, M.D., J.D.
Commissioner New York State Department of Health

¹ Z. Obermeyer, 8. Powers, C. Vogeli & S. Mullainathan, "Dissecting racial bias in an algorithm used to manage the health of populations," Science, vol. 366, iss. 6464 (Oct. 25, 2019), available at https://science.sciencemag.org/content/366/6464/447.

² W. P. Qiao et al., "Relationship between racial disparities in ED wait times and illness severity," The American Journal of Emergency Medicine, vol. 34, iss. I, IO - 15 (Jan. 2016), available at https://www.ajemjoumal.com/article/S0735-67571500744-5/fuIItext.

³ K. M. Hoffman, S. Trawalter, J. R. Axt & M. N. Oliver, "Racial bias in pain assessment arid treatment recommendations, and false beliefs about biological differences between blacks and whites," Proceedings of the National Academy of Sciences of the United States of America, 201516047; DOI: 10.1073/pnas.1516047113 (Apr.2016), available at https://www.pnas.org/content/early/2016/03/30/1516047113/tab-article-info.