Introduction

Two stories set the background for this article. The first is a personal experience. The second appeared as a news report in the New York Times.

In the fall of 2021, I sat in a law school classroom as a talented health insurance enrollment counselor taught our law students how to help people fill out Medicaid applications. Our students were part of a community-led effort to find and enroll 275,000 people newly eligible for Medicaid expansion in Missouri. The enrollment counselor got to the questions on the Medicaid application asking about race and ethnicity. She looked up and said, “You can skip these questions. They are voluntary. People can leave them blank. They aren’t needed to process the application.”

In the fall 2022, Max Li chose not to declare his race when it came time to fill out his college application form. Even though Max knew his last name sounded Chinese, he selected “prefer not to say.” A few months before, the Supreme Court had heard oral arguments in a case accusing Harvard and North Carolina of systematically discriminating against Asian American applicants. Max feared his race would be used against him in the college admissions process.^[1] In June 2022, the Supreme Court ruled that the colleges’ admissions systems had disadvantaged and stereotyped Asian-Americans students.^[2]

This article is about Medicaid and the intertwined issues of obtaining and stratifying health care quality data by race and ethnicity for Medicaid patients. For over twenty years, the Institute of Medicine and others have recommended that state Medicaid agencies, health plans, and providers stratify health care quality measures—like prenatal care, child immunization rates, medication for diabetes—by race and ethnicity to be able to identify and remedy racial and ethnic disparities in care.^[3] For just as long, state agencies, plans, and providers have struggled with how to obtain data from Medicaid enrollees about their race and ethnicity to be able to stratify their quality measures.^[4]

The preferred and best source of data about an individual’s race and ethnicity is the information they provide. Race is not a biological construct, it is a social identity.^[5] Race is not what we think someone else looks like, but how they self-identify. While zip code data and surnames can be helpful proxies for self-reported information about race and ethnicity when doing population level analysis, they are less accurate than self-identified data and should be used with caution in informing assessments and patient-level interventions.^[6]

But why should people trust Medicaid agencies with sensitive information about their race and ethnicity? For far too long, government agencies, health care providers, and others have used race and ethnicity to exclude, harm, and abuse patients.^[7] During the early years of this country, doctors performed horrific “experiments” on African American enslaved women.^[8] During the Tuskegee Syphilis Project, doctors and public health nurses withheld life-saving treatment for African American men.^[9] Still today, researchers continue to report that minority patients typically receive less treatment and poorer quality treatment than white patients do.^[10]

This article offers lessons from one motivated state Medicaid program—Michigan—that has embraced a health equity approach to quality improvement in managed care that puts reducing racial and ethnic disparities front and center. For twenty years, Michigan state law and policy has been at the forefront of using Medicaid managed care quality data, stratified by race and ethnicity, to report, track, and tackle health care disparities.^[11]

Part I explains the importance of Medicaid to minority health and the studies documenting long-standing, pervasive disparities in the quality of care that Medicaid provides to minority patients. It also presents an overview of two decades of collaborations showing how state Medicaid agencies and managed care plans can use Medicaid enrollee data stratified by race to drive equity-focused quality improvement efforts.

Part II describes how the lack of federal Medicaid policy has contributed to the failure of many states to collect and use data stratified by race and ethnicity for quality improvement purposes. It also discusses recent rulemaking that will finally require states to begin reporting Medicaid quality measures stratified by race and ethnicity effective 2027.

Part III shows how Michigan has used state law and policies and managed care contracts to put health equity and healthcare disparities reduction at the center of its Medicaid managed care quality efforts. For almost two decades, the state has partnered with minority communities to develop policies and strategies to reduce racial and ethnic disparities.^[12] Public reporting and tracking of quality data stratified by race and ethnicity identifies where disparities exist and whether interventions are successful.^[13] State-led performance improvement projects and financial incentives to reduce disparities assure that managed care plans prioritize reducing racial and ethnic disparities.^[14]

Part IV offers lessons from Michigan for other states struggling to collect enough data about the race and ethnicity of their Medicaid enrollees to be able to stratify quality data by race and ethnicity. It concludes that the key to success may be connected to how states use the race and ethnicity they collect. States, like Michigan, that lead the way in collecting race and ethnicity data on their Medicaid applications are typically states that have a clearly articulated public policy and rely on such data to identify and address disparities in health care and health outcomes, seeking to help rather than harm minority patients.^[15] They are also states that engage with communities of color to enlist them as allies and partners in designing health equity quality improvement initiatives.^[16]

I. Medicaid, health care disparities, and equity-focused quality improvement

Medicaid is a critical source of health care for people of color. Because of historic discrimination and structural racism, people of color are more likely to be low-income and Medicaid-eligible than white Americans.^[17] More than 60% of the Medicaid’s 83 million enrollees identify as minorities or as multiracial.^[18] Medicaid and CHIP provide coverage for more than half of Black, Hispanic, and American Indian/Alaska Native children.^[19] The Affordable Care Act’s Medicaid expansion for low-income adults has reduced racial and ethnic disparities in insurance coverage and increased access to care for people of color.^[20]

Yet studies have long documented that Medicaid all too often provides less health care, different health care, and lower quality health care to the people of color it covers when compared to white people.^[21] For example, one recent 2022 study in the Journal of the American Medical Association (JAMA) analyzed Medicaid claims data stratified by race and ethnicity for almost 2 million adults and children enrolled in Medicaid managed care in three states.^[22] Researchers found that Black enrollees received fewer services, including primary care, and generated lower spending than White enrollees, but were more likely to use the emergency room for avoidable reasons. ^[23] These differences persisted even among enrollees living in the same zip codes who were treated by the same health care professionals.^[24] In another 2022 published in Health Affairs involving almost 250,000 non-elderly Medicaid managed care enrollees in thirty-seven states, minority enrollees across the board—Black, Hispanic, Asian American, Native Hawaiian or other Pacific Islanders—reported significantly worse care experiences than white enrollees in the same plan.^[25]

Health care disparities in Medicaid and elsewhere in the health care system are the products of a complex interplay of factors that operate at the structural, institutional, provider and patient level, including systemic inequity and racism.^[26] Many state Medicaid programs are attempting to address the impact that the social determinants of health on health by increasing access to housing, an important effort.^[27] However, reducing health care disparities requires interventions to ameliorate institutional racism that arises when race neutral policies—of health plans and providers—disproportionately impact people of color.^[28] It also requires acknowledging and addressing the explicit and implicit bias that operate at the interpersonal level when medical care is delivered.^[29]

Since the IOM’s groundbreaking 2002 report Unequal Treatment: Confronting Racial and Ethnic Disparities in Healthcare shone a light on such health care disparities, health care thought leaders have urged state Medicaid agencies, plans, and providers to make disparities reduction an integral part of their health care quality efforts, linking quality and equity.^[30] With more than 90% of Medicaid beneficiaries enrolled in managed care, many efforts have focused on managed care quality reporting and quality improvement.^[31]

The Robert Wood Johnson Foundation and others have established a research base documenting how state Medicaid agencies, health plans, and providers can get and use patient data stratified by race to design and implement interventions to reduce racial and ethnic disparities in care.^[32] The Center for Health Care Strategies has partnered with state Medicaid agencies and health plans to enhance the capacity of states to collect and use data stratified by race and ethnicity to drive quality improvement projects and help Medicaid managed care plans develop clinical and administrative quality improvement strategies aimed at reducing RE disparities in care.^[33] The California Foundation for Health, the National Commission for Quality Assurance, and others have highlighted the efforts that various state Medicaid agencies and Medicaid health plans are using to collect and stratify Medicaid enrollee data by race and ethnicity to use that data to drive health equity focused quality improvement projects.^[34]

But many state Medicaid programs have had a chicken and egg problem. They cannot engage in equity focused quality improvement projects because they lack good data about the race and ethnicity of Medicaid enrollees. As Dr. Marcella Nunez-Smith, Chair of the Biden Administration COVID-19 Health Equity Task Force, famously said about the lack of public health data stratified by race and ethnicity, “We can’t fix what we can’t see.” We need to be able to stratify quality data by race and ethnicity to be able to see the disparities. The next section explains how a lack of federal guidance has contributed to states’ chicken and egg problem.

II. HHS drops the ball: The lack of federal guidance on collecting and reporting race and ethnicity data to support quality improvement efforts

HHS has ample statutory authority to require states to collect and report quality measures stratified by race and ethnicity. Such data is needed to enforce civil rights statutes like Title VI of the Civil Rights Act and Section 1557 of the Affordable Care Act. Section 1902(a)(6) of the Social Security Act requires state Medicaid agencies to report data “in such form and containing such information” as HHS may require.^[35] Section 1902(a)(4) requires states to cooperate with “methods of administration,” including reporting requirements, found by the Secretary to be “necessary for the proper and efficient administration” of Medicaid."^[36] Section 1903 requires that states submit such claims, encounter, and enrollee demographic data that HHS determines is necessary to the electronic Medicaid Statistical Information System (T-MSIS).^[37]

But until recently, HHS guidance has barely addressed the issue of health disparities as part of its quality of care initiatives.^[38] Since April 2016 and the waning days of the Obama Administration, federal Medicaid managed care regulations have required states to include a health disparities plan as part of their written state quality strategy for assessing and improving the quality of services offered by managed care plans. ^[39] The regulations specify that the disparities plan should “identify, evaluate, and reduce, to the extent practicable” health disparities based on race and ethnicity, as well as age, sex, primary language, and disability.^[40] With nearly 90% of Medicaid enrollees in some form of managed care, this disparities plan requirement should have broad impact.^[41] However, CMS was slow to implement the regulation and does not require that state disparities plan stratify managed care quality data by race and ethnicity (or other demographics).^[42]

The Trump Administration never updated the CMS Managed Care Quality Strategy Toolkit for states to include the new requirement for a health disparities plan.^[43] Finally, in June 2021, five years after the regulation was issued and six months into the Biden Administration, CMS updated the Toolkit, listing the disparities plan as a required element of the state managed care quality strategy.^[44] The 2021 Toolkit, which is recommended, albeit voluntary, for states, offers considerations for states writing a disparities plan, and includes links to a CMS webinar on collecting and using stratified data and a CMS webpage with resources about quality of care health disparities.^[45] It recommends that a state disparities plan explain how the state intends to identify and evaluate racial and ethnicity disparities and suggests stratifying quality measures as one method a state might use, but stops short of requiring states to stratify managed care quality measures by race and ethnicity, or other demographic factors.^[46]

CMS does not now require states to stratify core quality measures by race and ethnicity, but newly promulgated regulations will oblige states to begin reporting core measure stratified by race and ethnicity by September 2027.^[47] Since 2010, Congress has required HHS to identify a Core Set of health quality measures for children enrolled in Medicaid and CHIP.^[48] The statute specifically provides that the measures shall be evidence-based, “designed to identify and eliminate racial and ethnic disparities in child health and the provision of health care,” and “in a standard format that permits comparison of quality data on a state, plan, and provider level.”^[49] In 2013, Congress adopted similar requirements for a Core Set of measures for adults enrolled in Medicaid.^[50]

The 2023 and 2024 Core Quality Measures include twenty-seven for children and thirty-four for adults.^[51] Seventeen of the children’s Core measures and twenty-one of the adult Core measures are Healthcare Effectiveness Data and Information Set (HEDIS) measures developed by the National Committee for Quality Assurance (NCQA). HEDIS measures are used by more than 90% of health plans to measure performance on important dimensions of care and service.^[52] Other Core measures were developed by various divisions of HHS or the Pharmacy Quality Alliance.^[53]

Congress originally provided that state reporting of core measures would be voluntary, but beginning in 2024, states will be required to report all children’s core measures and adult core behavioral health measures.^[54] In August 2023, as part of the move from voluntary to mandatory state reporting, CMS issued new regulations that, for the first time, will require states to stratify a sub-set of their mandatory Core Set quality measures by race and ethnicity, and other demographics including sex, age, rural/urban status, disability, language, or other factors as specified by the Secretary beginning in 2027.^[55] The rule will also require states that cover optional Medicaid health home benefits to stratify a sub-set of the Health Home Core Quality Measures by race, ethnicity, and other demographics.^[56] States that fall short could lose part or all of the federal Medicaid matching dollars, giving CMS leverage to enforce quality measure reporting stratified by race and ethnicity.^[57]

In May 2023, CMS issued notice of proposed rulemaking that would create a new Medicaid and CHIP Managed Care Quality Rating System to help new and returning enrollees select managed care plans that best suit their needs.^[58] States would be required to collect and display a mandatory group of quality measures for each Medicaid managed care plan that aligns with adult and child Core Set measures, with CMS initially proposing eighteen such measures.^[59] CMS also proposed designating a subset of these Quality Rating System managed care plan measures for which states would have to collect and report stratified by race and ethnicity (as well as sex and dual Medicaid/Medicare status).^[60]

In comments on the proposed new rules for stratified reporting of quality measures, many organizations praised the move to require quality reporting and stratification by race and ethnicity (and other factors) of core quality measures.^[61] However, many were also concerned that states do not have good data on race and ethnicity to be able to stratify their quality measures.^[62] Not only has CMS been slow to require states to stratify quality measures by race and ethnicity, the agency has also failed to provide minimum standards to guide states in collecting data about the race and ethnicity of Medicaid enrollees.

Most state Medicaid programs have long collected data about enrollees’ race and ethnicity,^[63] but the ACA, for the first time, called for the creation of uniform data collection and reporting standards to allow for comparisons across states, programs, and tracking of disparities over time. Section 4302(a) of the ACA amended Section 3101 of the Public Health Service Act to require the Secretary of HHS Section to develop data-collection standards for five demographic categories of race, ethnicity, sex, primary language, and disability status.^[64] The Secretary was charged with ensuring, to the extent practicable, that federally conducted or supported health care or public health programs, activities, and surveys collect and report these categories.^[65] Section 4302(b) specifically required that state Medicaid and CHIP programs collect and report race and ethnicity data that complies with the HHS developed standards under Section 4302(a).^[66]

However, Section 4302’s uniform data collection standards for state Medicaid programs never took effect.^[67] Implementation of Section 4302(a) was conditioned upon Congress specifically appropriating funds to support its data collection efforts.^[68] The ACA appropriated funds for Section 4302 for the first four years, 2010-2014. But as Congressional Republicans became increasingly hostile to the ACA, no further appropriations occurred.^[69] HHS has taken the position that other minimum standards for race and ethnicity data collection—the Office of Management and Budget (OMB) minimum standards for collecting race and ethnicity data in federally sponsored data collection efforts and the HHS 2011 minimum standards for collecting and reporting race and ethnicity in population health surveys—do not apply directly to state Medicaid programs.^[70]

All states now collect voluntary self-reported data about race and ethnicity on their Medicaid application forms.^[71] Federal regulations specify that states may not require Medicaid applicants to provide information about their race and ethnicity as part of the application process.^[72] Thus, states ask applicants about their race and ethnicity during the Medicaid application process, but the questions are labeled as optional.^[73] Since 2003, state Medicaid programs have been required to provide data they collect on enrollees’ race, ethnicity, and primary language obtained during the enrollment process to managed care plans.^[74]

But states design their own Medicaid applications.^[75] Not surprisingly, with so little federal guidance, state Medicaid applications vary widely in what they ask about race and ethnicity and how they ask it. In 2022, state Medicaid programs asked about race in sixty-four different ways, with response options ranging from five to fifty-seven for online applications.^[76] While a majority of states use response options consistent with either the OMB or HHS’s 2011 minimum standards, many states include additional categories for both race and ethnicity.^[77] Some states allow applicants to select multiple races and ethnicities, while other states allow only one selection, and a few have a combined race and ethnicity questions.^[78] Most states use different racial and ethnic response categories for their paper and online applications.^[79]

Regrettably, states also vary widely in how well they do in collecting and reporting data about the race and ethnicity of Medicaid enrollees.^[80] While CMS does not explicitly require states to collect data on the race and ethnicity of Medicaid enrollees, it does require states to submit any data they do collect to the Transformed Medicaid Statistical Information System (T-MSIS), the national electronic data repository that supports Medicaid program management, CMS oversight activities, and research.^[81] According to an analysis by the Medicaid and CHIP Payment and Access Commission (MACPAC) of state race and ethnicity data submitted in 2019 to T-MSIS, thirty states had data that meets the minimum data quality standards necessary to conduct analyses stratified by race and ethnicity, while twenty-one states did not.^[82] Nine states had data of low concern, twenty-one states had data of medium concern, seventeen states had data of high concern, and four states had data that was unusable because race and ethnicity data was missing for more than half of enrollees.^[83] CMS’s own analysis of state race and ethnicity data submitted to T-MSIS from 2014 to 2020 reports similar results and similar quality problems across the years.^[84]

Many have called on HHS to provide more and better policy guidance and technical assistance to help states collect more and better data about their Medicaid enrollees’ race and ethnicity.^[85] Yes, states need “more strenuous” guidance about how to collect racial and ethnicity data about their Medicaid enrollees, but ultimately, data about race and ethnicity is, and should be, self-reported. Why should people trust Medicaid agencies with this sensitive data? What can other states learn from states like Michigan that already succeed in collecting race and ethnicity data for almost all of the Medicaid enrollees?

III. Michigan’s Long Experience

Michigan has a long history of collecting and using Medicaid enrollee data stratified by race and ethnicity to identify and address disparities in the quality of care provided by Medicaid managed care plans.^[86] The state consistently reports race and ethnicity for more than 90% of Medicaid enrollees to CMS, with people willingly providing the information as part of the application process.^[87] In 2011-2013, Michigan was one of twelve pilot states that worked with CMS to refine the T-MSIS data set, which includes demographic data, and to modernize the submission and quality review process for the data set.^[88]

From 2002 to 2004, Michigan was part of an early six state demonstration project, led by the Centers for Health Care Strategies and funded by the Commonwealth Fund and HRSA, to explore how states and managed care plans could obtain and use data stratified by race and ethnicity to identify and reduce health care disparities.^[89] The Michigan state Medicaid agency partnered with three managed care plans in the Detroit area with high percentages of Black members in an effort to develop quality improvement initiatives to reduce disparities in diabetes care for Black enrollees.^[90] All three plans focused on quality measures for comprehensive diabetes care, testing, and control of HbA1c and LDL levels.^[91]

The state agency transmitted data on enrollee race that people voluntarily self-reported on their Medicaid applications to the health plans, and the plans used that data to stratify their HEDIS diabetes quality measure by Black and White. All three plans identified at least one racial disparity in diabetes testing and control for their Black enrollees, and each plan designed its own intervention and improvement goals aimed at reducing or eliminating a racial disparity in diabetes care.^[92] The range of interventions included a disease management program and registry, physician profiles to track practice patterns, culturally sensitive education materials for patients, diabetes case management, and partnering with a home health agency.^[93] When the state agency and the plans evaluated how well the interventions worked, only some of the plans met their disparities-reduction goals.^[94] However, all the plans and the state Medicaid agency learned useful lessons about how to identify and target racial and ethnic disparities in quality of care for Medicaid patients.^[95]

As the demonstration project drew to a close, the Medicaid state agency sponsored a series of workshops to share information learned with all nineteen managed care plans in the state to encourage them to organize similar racial and ethnic disparities data analysis and quality improvement efforts.^[96] In 2005, the state agency formed a state level Multi-Disciplinary Working Group on Health Disparities Reduction to continue efforts to address racial disparities.^[97] In 2005, the state agency also participated in another Center for Health Care Strategies pilot project which identified a number of racial and ethnic disparities by both manage care plan and individual provider.^[98] From 2008 to 2010, the state required Medicaid health plans to conduct an annual performance improvement project aimed at reducing the disparity in one of their quality measures.^[99]

Meanwhile, in 2006, Michigan passed Public Act 653, often referred to as the Michigan Minority Health Act, which tasks the Michigan Medicaid state agency with using a health equity framework to tackle health and health care disparities.^[100] Public Act 653 directs the state agency to establish a departmental structure to address racial and ethnic minority health disparities, monitor minority health, promote workforce diversity, and develop policy and strategies to reduce racial and ethnic disparities.^[101] It charges the agency with seeking input from minorities on state health policies and programs, and assisting and promoting local minority health coalitions.^[102] The law requires the agency to compile and share via a webpage race and ethnicity data including, but not limited to, morbidity and mortality statistics.^[103] It also instructs the agency to submit an annual report on the agency’s efforts to address racial and ethnic health disparities in Michigan.^[104] Implementation of Public Act 653 has focused on five racial, ethnic, and tribal populations in Michigan: African/American, Hispanic/Latino, Native American, Asian American/Pacific Islander, and Arab/Chaldean American.^[105]

Public Act 653 has been described as a catalyst for the state agency to rethink its approach to racial and minority health in Michigan.^[106] Since 2008, the agency’s annual Health Equity Report: Moving Health Equity Forward documents the state’s efforts to address racial and ethnic disparities and to promote health equity in public health, Medicaid, and other programs.^[107] Over the years, the reports present data on minority health status and disparities, tracking changes in health disparities over time and highlighting key challenges like COVID-19.^[108] They highlight evidence-based and promising practices to reduce disparities in health and health care, including those aimed at the social determinants of health.^[109]

The annual Moving Health Equity Forward reports also document how the agency is engaging with communities of color, creating authentic relationships, and building trust.^[110] The agency acknowledges that community engagement is an essential component to advance health equity, and that effective community engagement requires investing in long term relationships with community residents and trusted community leaders and organizations.^[111] It seeks to place the voices of community members and leaders, especially those experiencing health and social inequities, at the center of efforts to identify and shape culturally appropriate interventions and strategies.^[112] The reports document examples of community engagement, including community representation of advisory councils and community input through focus groups, listening sessions, and surveys.^[113] It also concedes the gaps and limitations in its community engagement that occur because of lack of staff, funding shortfalls, and lack of knowledge. Among its primary recommendations are that the agency assures that community engagement is a two-way process and involves communities earlier in processes, particularly in the strategic planning, priority setting, and decision-making stages which impact them.^[114]

The act also prompted the Medicaid Managed Care Division’s Quality Improvement and Program Development Section to develop a Medicaid Health Equity Project aimed at identifying and eliminating disparities in managed care.^[115] In early 2010, the Managed Care Division worked with the Medicaid managed care plans to frame the problem of health care quality disparities and plan a project to report Medicaid managed care quality data stratified by race and ethnicity.^[116] In 2011, all Medicaid managed care plans began reporting data stratified by race and ethnicity for eight HEDIS measures.^[117] Since 2012, plans have submitted a uniform set of 13 HEDIS quality measures stratified by race and ethnicity, including three related to diabetes care, that continue the disparity reduction efforts the state agency and managed care plans began in 2002-2004.^[118]

Most of the race and ethnicity data plans used to stratify health plans’ HEDIS data comes from applicants during the Medicaid enrollment process, although health plans may supplement this data using electronic health records and others sources.^[119] As required by federal law, the Michigan state agency transmits self-reported data about enrollee race and ethnicity it collects on Medicaid application forms to managed care plans at the time of enrollment.^[120] Plans stratify measures by five racial categories used on the application: Asian American/ Native Hawaiian/Pacific Islander, African American, White, American Indian/Alaska Native, and one ethnicity, Hispanic.^[121] African American, Hispanic, and White enrollees make up about 85% of Michigan’s managed care enrollees.^[122]

Since 2012, the agency has published an annual Medicaid Health Equity Project Report that analyzes and reports on disparities for the Medicaid managed care program as a whole for the thirteen Health Equity HEDIS measures.^[123] Disparities are measured by comparing the rates of care provided to minority groups to that provided to whites, recognizing the role racism plays in creating disparities and thus treating care for whites as the baseline.^[124] These reports are posted online and widely available to the public, stakeholders, and advocates.^[125] They serve as a report card on racial and ethnic disparities in managed care. Regrettably, the state does not report on disparities by health plan to be able to compare how plans compare to each other or to the statewide average.^[126]

A decade of disparities data collection and reporting has allowed Michigan Medicaid to track changes in disparities over time, target disparities efforts, identify successful interventions and failures, and leverage Medicaid managed care contracts to address disparities. From 2015 through 2020, the Index of Disparity, which describes the extent of disparities across all racial and ethnicity groups, got smaller.^[127] By 2020, only three Health Equity HEDIS measures showed a high Index of Disparity level, compared to six in 2015.^[128] The three measures that continue to show a high Index of Disparity—postpartum care, childhood immunizations, and chlamydia screening for women—have all been the focus of quality improvement initiatives. All showed improvement from 2019 to 2020.^[129]

Despite overall improvements, the Medicaid Health Equity Reports also document that African Americans managed care enrollees suffer from persistent harmful disparities in quality of care they receive: African Americans have had consistently lower rates of care compared to whites for nine of the thirteen HEDIS measures reported each year.^[130] African Americans are the only group to have disparities of ten percentage points or more, with childhood immunizations being 12.87 percentage points lower for Black children than for White children, and child and adolescent access to primary care being 10.55% lower for Blacks.^[131] The annual reports note that studies of health disparities in Michigan have consistently determined that social determinants of health alone cannot account for all the racial and ethnic disparities in health and health care for African Americans, requiring that efforts also focus on the role of race, discrimination, and structural racism in creating and sustaining health care disparities.^[132]

The 2020 Medicaid Health Equity Report also describes a notable success in reducing disparities in postpartum care, while acknowledging that it still falls far short of parity.^[133] With managed care data showing large and dogged disparities for postpartum care and low overall rates for postpartum care, the state agency has emphasized this quality measure for many years incorporating it into several performance monitoring and incentive efforts.^[134] Those efforts seem to be bearing fruit. From 2019 to 2020, post-partum care rates improved significantly for all groups with the disparities gap for African Americans and Hispanics growing smaller and disparities gap for African American women shrinking by almost four percentage points.^[135] Regrettably, though, the disparity in postpartum care for African American women in 2020 was still one of the largest disparities reported, 9.91 percentage points lower than for white women compared to a whopping 13.79 in 2019.^[136] The annual report both noted the good news and acknowledged that “continued efforts are needed to reduce this disparity.”^[137]

Data from the Medicaid Health Equity Project informs and undergirds the state’s quality improvement strategy.^[138] As for the Michigan 2020-2023 Comprehensive Quality Strategy, the Medicaid Health Equity Project creates a system to monitor racial and ethnic disparities in the managed care population that allows the agency to identify priority areas for quality improvement initiatives related to health disparities.^[139] Managed care plans must participate in the Health Equity Project and report the thirteen Health Equity HEDIS measures stratified by race and ethnicity.^[140]

Reducing racial and ethnic disparities is one of five goals for the state’s 2022-2023 Comprehensive Quality Strategy for managed care, which was drafted prior to CMS updating the guidance in its Toolkit for states.^[141] The state agency and the health plans have committed to using a data-drive approach to identify root causes of racial and ethnicity disparities and to address them at their source.^[142] This process includes gathering input for beneficiaries, communities, and other stakeholders to ensure that people of color are engaged in the quality improvement intervention design and implementation process.^[143] The Michigan Office of Equity and Minority Health was actively involved in the development of the state’s managed care Comprehensive Quality Strategy.^[144] Given the connections between prenatal care and low weight births, in 2019-2020, health plans were required to participate in a performance improvement project to address disparities in the timeliness of prenatal care.^[145] In 2022, health plans were required to initiate a second performance improvement project focused on disparities in the timeliness of prenatal care.^[146]

Michigan’s Medicaid managed care contract also embeds health equity and reducing racial and ethnic health disparities in law. The managed care contract creates the legal requirement that plans must participate in the Medicaid Health Equity Project and associated initiatives including annual reporting of the thirteen Health Equity HEDIS measures stratified by race and ethnicity.^[147] It also requires plans to implement a community-led initiative to reduce disparities and improve health equity.^[148]

The managed care contract also creates financial incentives for plans to reduce disparities. Michigan offers a performance bonus for quality improvement in five areas, three of which focus on minority health and disparities reduction: health equity scoring, low birth weight, and sickle cell.^[149] Plans can earn the health equity bonus by achieving statistically significant reductions in disparities for Black and Hispanic enrollees on ten HEDIS measures, six of which are part of the thirteen Health Equity HEDIS measures reported annually.^[150] Three of the HEDIS measures are ones that have showed long standing Black/White disparities in the annual Medicaid Health Equity Report: postpartum care, childhood immunization, and chlamydia screening for women.^[151] The low weight birth measure bonus is tied primarily to reducing low weight births for Black women and reducing disparities between African American and White women.^[152]

The state also offers a performance bonus for integration of behavioral and physical health services focusing on follow-up care after hospitalization for mental illness and emergency room visits for alcohol and drug dependency.^[153] The state provides plans with plan-specific data stratified by race and ethnicity.^[154] Plans can earn points toward the bonus by reducing a disparity for at least one minority group.^[155]

In 2022, the contract also provided that plans had to participate in a state-led pay for performance (P4P) initiative to reduce racial and ethnic disparities in low-birth rates in which health plans, home visiting programs, and community health workers collaborate to design and implement a three-year project.^[156]

In Michigan, the state has made health equity and the reduction and elimination of racial and ethnic disparities in the quality of care delivered by Medicaid managed care plans a central focus of quality improvement efforts. Michigan has embraced a three-pronged approach to health equity in managed care. First, public reporting and tracking of quality data stratified by race and ethnicity identifies where disparities exist and whether interventions are successful. Second, financial incentives to reduce disparities assure that plans prioritize reducing racial and ethnic disparities. Third, and most importantly, for almost two decades the state has worked to engage with minority communities in the development of policies and strategies to reduce racial and ethnic disparities.

IV. Lessons from Michigan

Michigan’s experience with disparities reporting and performance improvement initiatives shows what policy makers, state agencies, and managed care plans can learn when they have a decade of data about Medicaid quality of care stratified by race. The state has leveraged state law and policy—managed care contracts and their federally required state managed care quality strategy—to require Medicaid managed care plans to identify, report, and address health care disparities. The state is using performance improvement projects and pay for performance incentives to assure that plans prioritize reducing racial disparities in care.

What can states struggling to collect data about the race and ethnicity of their Medicaid enrollees learn from Michigan? The lesson is that the key to success in collecting race and ethnicity date may be connected to how states use the race and ethnicity data they collect: states, like Michigan, that lead the way in collecting race and ethnicity data for Medicaid enrollees are typically states that have a clearly articulated public policy and legal agenda that relies on enrollee data stratified by race and ethnicity to identify and address disparities in health care and health outcomes. Michigan’s annual Medicaid Health Equity Project Report, Comprehensive Quality Strategy, and managed care contracts communicate to the general public and minority communities why racial and ethnic data is being collected and why it is needed. Linking data stratified by race and ethnicity to performance improvement projects and pay for performance incentives signals that the state is serious about reducing disparities and will hold managed care plans accountable. Leading states, like Michigan, engage with minority members, organizations, and advocates to enlist them as partners and allies in efforts that support a legitimate need for and use of racial and ethnic data.

Providing race and ethnicity data on Medicaid applications is, and should be, voluntary, and some people are unwilling to disclose this information because of distrust, lack of understanding why the information is important, and mistrust about how it will be used.^[157] There is valid mistrust by those who have been historically disenfranchised about how race and ethnicity data is collected and used which affects their willingness to answer such questions.^[158] Community engagement is critical in assuring that race and data is collected and shared by the states and their partners in a way that respects enrollee’s privacy towards the goal of using race and ethnicity data effectively to advance health equity and reduce disparities in the Medicaid program.^[159]

Michigan is an example of a state that has adopted public policy requiring the Medicaid agency to engage with minority communities, organizations, and leaders to understand their concerns and to collaborate with them about how race and ethnicity data will be used to advance health equity. P.L. 653, Michigan’s Minority Health Act, requires the state agency to seek input from minorities on state health policies, including Medicaid.^[160] The state agency’s annual Moving Health Equity Forward reports document that the state agency takes its statutory obligation seriously, seeking to create collaborative processes and bidirectional exchange of expertise and information—including the lived experiences of those impact by policies—into the development of policies and interventions.^[161]

Analysis by other researchers seems to confirm that the key to collecting highly robust race and ethnicity data is probably not found merely by examining leading states’ Medicaid applications.^[162] Little research has been done examining best practices for structuring race and ethnicity data collection in the Medicaid application context. ^[163] We don’t know much about how question structure, response options, the development of instructional language, and assister training materials impact Medicaid applications specifically, although researchers have developed best practices for obtaining self-identified race and ethnicity data in health care and social services settings more generally.^[164]

However, we do know that engaging minority patients and minority community members in the design of questions, racial and ethnic categories, and outreach materials about data collection can increase response rates and help ensure the response options reflect the diversity of the community.^[165] For example, a 2022 study at Mt. Sinai Hospital in New York City reduced missing race and ethnicity data by 76%, resulting in the hospital obtaining race and ethnicity data for 90% of patients by year five.^[166] The hospital developed training materials for the frontline staff that included information on how race and racism can impact care and why and how such information can improve care for minority patients.^[167] At the beginning of the project, patients’ opinions were solicited and addressed, and a feedback loop created to get more feedback from patients.^[168] Technology and infrastructure were modified to support better data collection. Lead data collection personnel were identified, and the data collection protocol was periodically reassessed and adjusted.^[169] Overall, the study emphasized the importance of consultation with and buy-in from patients and frontline staff, among others.

A 2021 study in New York focused on the Medicaid application and enrollment process increased response rates by 20% for race and by 8% for ethnicity by, among other things, training enrollment counselors and providing better information to applicants about why the information was being collected.^[170] The state provided training to enrollment counselors on how to ask questions about race and ethnicity, including creating a standardized script that explained the importance and purpose of the data.^[171] The standardized script provided additional detail on how race and ethnicity would be used, specifying “to improve services to the community and enhance outreach efforts.”^[172] The state also added the following two new response options for race and ethnicity: “don’t know” and “choose not to answer.”^[173]

States, like Michigan, that have a clearly articulated public policy and legal agenda that relies on enrollee data stratified by race and ethnicity to identify and address disparities have a credible story to share with enrollment counselors and applicants about why race and ethnic data is needed and how they will use the data. Michigan’s annual Health Equity Report and Medicaid Health Equity Project Report help tell that story. Managed care contracts and quality strategies that link quality measures stratified by race and ethnicity data with performance measures, quality improvement initiatives, and financial incentives to reduce disparities show that the state is serious about health equity and reducing disparities.

Conclusion

Michigan has used a three-prong approach to put disparities reduction at the center of its Medicaid managed care quality efforts. For almost two decades, the state has built partnership with minority communities so that their voices and lived experiences inform the development of policies and quality improvement efforts. Public reporting and tacking of quality data stratified by race and ethnicity identifies where disparities exist and whether interventions are successful. State-led performance improvement projects and financial incentives to reduce disparities assure that managed care plans prioritize reducing racial and ethnic disparities in care.

Michigan is not alone. Other states have adopted all or part of Michigan’s three-prong approach to health equity in managed care. According to an analysis commissioned by Princeton’s State Health & Value Strategies, at least seventeen states and the District of Columbia have provisions in their Medicaid managed care contracts requiring that certain quality measures be stratified by race and ethnicity.^[174]At least fourteen require managed care plans to create a disparities’ plan or report.^[175] At least nine states provide financial incentives to incentivize plans to reduce racial and ethnic disparities.^[176] At least seven states are leveraging managed care contracts to ensure that plans are listening to, understanding, and reflecting the priorities and experiences of their enrollees through meaningful enrollee involvement in policy development, implementation, evaluations, and improvements efforts.^[177] North Carolina, the most recent state to adopt Medicaid managed care, seems to be closely following Michigan’s three-prong play book.^[178]

At the federal level, momentum towards race and ethnicity disparities reporting has picked up momentum. HHS has finally issued regulations requiring states to report a set of core quality measures stratified by race and ethnicity and the agency seems poised to issue regulations requiring states to stratify Medicaid managed care plan quality measures by race and ethnicity. Together, these regulations will create minimum standards for state collection and reporting of quality disparities for both managed care and fee for service Medicaid. Since 2022, NCQA is requiring all health plans it accredits (commercial, Medicare, and Medicaid) to stratify select HEDIS measures by race and ethnicity with the number of stratified measures growing to twenty-two by 2024.^[179]

Yes, embracing racial and ethnicity disparities reduction and a health equity framework requires that states have good quality, voluntarily provided data about the race and ethnicity data of Medicaid enrollees. Michigan’s experience shows how state law, agency policy, managed care contracts, and managed care quality strategies can create the legal framework needed to see and fix racial and ethnic disparities in health care. All states should be partnering with minority communities to understand their concerns and to collaborate with them to develop and disseminate resources about why Medicaid agencies ask about race and ethnicity and how it will be used to advance health equity. Community input is critical in assuring that race and ethnicity data is collected and shared by the states and in a way that respects enrollees’ privacy, while furthering the goal of using race and ethnicity data effectively to advance health equity and reduce disparities in the Medicaid program.