Table 1 Studies Examining Medicaid Expansion Impact on African American-white Health Care Coverage, Access, and Treatment, 2014-2021

Menon, Patel, Karmakar, & Tipirneni (2021) [14]

Assessed differential impacts of ME on racial/ethnic and racial/ethnic-sex disparities among HIV testing

Difference-in-Difference (DD) and Triple Difference-in-Difference (DDD)

Pre- & post-ACA

ME & non-ME

AA-white disparity

Behavioral Risk Factor Surveillance System (BRFSS), 2011-2018

Adults (ages 19-64), low-income (less than or equal to 138% FPL), non-pregnant, non-disabled

Age, sex, race, race-sex, percent FPL, education level, employment status

Absolute

N/A

ME associated with significant increase in reports of ever having HIV test; no significant changes in reporting HIV test in last year; no significant changes in AA-white disparities associated with ME in ever having HIV or having an HIV test in previous year

N/A

Johnson, Choi, & Herrera (2021) [15]

Compared changes in medication for opioid use disorder (MOUD) post-ACA to determine whether implementation was associated with increased MOUD for African American clients relative to white clients.

Descriptive and logistic regression analyses; associative analysis with interactions

Pre- & post-ACA

AA-white disparity

Substance Abuse and Mental Health Services Administration’s (SAMHSA) Treatment Episode Dataset-Admissions (TEDS-A), 2007-2018

White, African American, and Hispanic clients for opioid use disorder treatment, first annual episodes

Age groups, sex, education status, homelessness; client admission setting, heroin use, polysubstance use; geography (rural)

Absolute

N/A

No AA-white disparity changes associated with ME states; disparities increased in criminal justice-referred population in ME states; for Medicaid-covered population, disparities decreased; AA-white disparities in MOUD significantly reduced overall

Significance of disparity in change difference between ME and non-ME states not tested

N/A

Ji, Castellino, Mertens, Zhao, Nogueira, Jemal, Yabroff, & Han (2021) [16]

Examined association of ME with insurance coverage at diagnosis among young adults newly diagnosed with cancer

Difference-in-Difference (DD); used linear probability models

Pre- & post-ACA

ME & non-ME

Examined both public and private coverage

National Cancer Database (NCDB),

2011-2016

Young adults (ages 18-39) diagnosed with a first primary cancer between January 1, 2011 and December 31, 2016 in U.S.

Sex, age, group, self-reported race/ethnicity, urban or rural residence, median household income

Absolute

Not examined by race/ethnicity

Stage 1 diagnosis increased more in ME states than non-ME states for white patients and AA patients; increase larger but less significant for AA patients

AA-white disparity or changes in disparities associated with ME not tested for significance

N/A

Le Blanc, Heller, Friedrich, Lannin, & Park (2020) [17]

Reviewed association of ME with breast cancer state at diagnosis and disparities associated with insurance status, age, and race/ethnicity

Retrospective cohort analysis (XW analysis used, with 1-sided p < 0.05; Mann-Whitney U test used to assess significance of state income and employment data with 1-sided p < 0.05)

Pre- & post-ACA

Distinguished between Medicaid coverage and un-insurance

National Cancer Database (NCDB) public benchmark reports via the American College of Surgeons, 2007-2016

Patients with primary breast cancer diagnosed between 2007 and 2016 who were uninsured or had Medicaid, private insurance, or Medicare, and whose race/ethnicity, age, state of residence, and American Commission Joint Cancer summary were recorded

Race, insurance status (uninsured and Medicaid), breast cancer stage (early or late)

Absolute

ACA implementation associated with reduced number of uninsured white and AA patients in ME and non-ME states; reductions greater for all racial groups

Significance of AA-white disparities not tested; differences between ME and non-ME states not tested for significance

ACA associated with increased rates of cancer diagnosis at earlier stages for AA and white patients in ME states, but not in non-ME states; incidence of advanced disease in AA patients decreased in ME states, and remained approximately the same in non-ME states

Significance of AA-white disparities not tested; difference between ME and non-ME states not tested for significance

N/A

Baumgartner, Collins, Radley, & Hayes (2020) [18]

Examined degree to which racial/ethnic disparities have narrowed post-ACA

Difference-in-Difference (DD)

Pre- & post-ACA

AA-white disparity

Did not distinguish between private and public insurance

American Community Survey Public Use Microdata Sample (ACS PUMS), 2013-2018

Behavioral Risk Factor Surveillance System (BRFSS), 2013-2018

Adults (ages 18-64); white, African American, Hispanic

None reported

Absolute

AA adults living in ME states now less likely to be uninsured than white adults in non-ME states

No statistics of significance reported

Unmet need due to cost decreased for AA and white; AA-white differences in cost-related access problems have narrowed in ME and non-ME states

No statistics of significance reported

Having usual source of care increased for white and for AA; AA adults in ME states now almost as likely as white to have usual source of care

No statistics of significance reported

N/A

Buchmueller & Levy (2020) [19]

Considered how the ACA’s insurance coverage expansions have affected racial/ethnic disparities related to access to care

Estimated difference between groups; presented unadjusted mean outcomes and results that controlled for individual characteristics

Pre- & post-ACA

AA-white disparity

Did not distinguish between private and public insurance

Behavioral Risk Factor Surveillance System (BRFSS), 2008-2017

Adults (ages 19-64); ~ 400,000 adults each year

Age, education, employment status, sex, marital status, and interaction between sex and marital status

Absolute

National AA-white un-insurance disparity decreased after ACA implementation

Difference in AA-white disparity reductions between ME and non-ME states not tested for statistical significance

National AA-white foregone care due to cost disparity decreased after ACA implementation

Difference in AA-white disparity reductions between ME and non-ME states not tested for statistical significance

N/A

Artiga, Orgera, & Damico (2020) [20]

Examined how health coverage by race/ethnicity has changed post-ACA

Difference-in-Difference (DD), stratified by race/ethnicity

Pre- & post-ACA

ME & non-ME

Did not distinguish between private and public insurance

American Community Survey (ACS), 2010-2018

Non-elderly population of whites, African Americans, Hispanics, & Alaska Natives (ages 0-64)

National coverage time trends

Absolute

Un-insurance rates for whites and AA decreased in non-ME and ME states, 2010-2018

DDD effect not tested for significance for AA-white disparity change

N/A

N/A

Breslau, Han, Lai, & Yu (2020) [21]

Examined impact of ME on use of 4 types of mental health services in nationally representative samples of low-income individuals during first 2 years following implementation

Difference-in-Difference (DD); used survey design-adjusted linear regression models; binary modeled by logistic DD model; continuous modeled by linear DD regression on subsample of service users; impacts of ME within racial/ethnic groups examined by extending original DD models to DDD models

Pre- & post-ACA

ME & non-ME

Medical Expenditure Panel Survey (MEPS), 2007-2015

Adults (ages 18 and older) with incomes at or below Medicaid eligibility under expansion rules (138% FPL)

Age, sex, marital status, race/ethnicity, K-6 score category, education level

Absolute

N/A

Change in use of outpatient mental health visits significant among whites and Hispanics, but not AA; no significant changes observed in number of mental health-related hospital stays, emergency department visits, or prescription refills

Significance in changes in AA-white disparities due to ME implementation not tested for any outcomes

N/A

Wiggins et al. (2020) [22]

Determined association between ME and infant mortality rates (IMR) in U.S.

Difference-in-Difference (DD); multiple linear regression models using DD estimation and Huber-White robust standard errors

Pre- & post-ACA

ME & non-ME

CDC’s Wide-ranging Online Data for Epidemiologic Research (WONDER), 2019-2017

State-level aggregate data on U.S. IMR and population count

Sex, race/ethnicity

Absolute

N/A

N/A

No association between ME states and change in national IMR, 2010-2017; ME associated with reduction in IMR among Hispanics; ME not associated with IMR reduction in ME states relative to non-ME states for whites and AA

Significance of AA-white disparity changes related to ME or ACA not tested

Barrington, Simmot, Calo, Cohn, Cosgrove, & Felix (2020) [23]

Determined associations between ME adoption and changes in insurance status, early-stage diagnosis, and cancer survival among women with endometrial carcinoma

Difference-in-Difference (DD); overall survival was fit with Cox proportional hazards models; logistical regression compared epidemiological, hospital, tumor, and treatment characteristics of women according to dichotomized ME status

Pre- & post-ACA

ME & non-ME

Examined a variety of insurance sources

National Cancer Database, Participant User Files (PUF),

2004-2015

Patients diagnosed with invasive endometrial carcinoma (ages 40-64)

Facility location, ME category of patient ZIP code; year (grouped into pairs); age, race, comorbities, hospital type, hospital location, rurality, educational attainment, household income, year of diagnosis

Absolute

Statistically significant improvement of changes in percent insured associated with ME implementation in ME states among white, but not African American.

DDD effect not tested for significance for AA-white disparity change

No significant changes observed in early-stage diagnosis associated with ME for any race.

DDD effect not tested for significance for AA-white disparity change

Significant increases in overall survival rates observed to be associated with ME for whites; no significant changes in survival rates observed for AA women.

DDD effect not tested for significance for AA-white disparity change

Eliason, 2020 [24]

Examined the effect of Medicaid expansion under the Affordable Care Act on state-level maternal mortality

ratios in the United States.

Difference-in-Difference (DD)

Pre- & post-ACA

ME & non-ME

Stratified by race

Underlying Cause of

Death 2006–2017 data files from the National Center for Health

Statistics;

Centers for Disease Control and Prevention, National Center for Health Statistics Natality data files, CDC WONDER Online Database

50 states and DC, from 2006 to 2017, for a total of 612 state-year observations

State-wide pregnancy checkbox adoption and state-level women’s unemployment ratio

Absolute

N/A

N/A

ME was significantly associated with lower maternal mortality;

ME effects were concentrated among non-

Hispanic Black mothers.

AA-white disparity or changes in disparities associated with ME not tested for significance

Brown, Moore, Felix, Stewart, & Tilford (2020) [25]

Identified association of ME with changes in county-level geographic variation in rates of low birthweight and preterm births, overall stratified by race/ethnicity

Compared changes in coefficient of variation and ratio of 80^th to 20^th percentiles using bootstrap samples (n = 1,000) of counties drawn for all births and for white, African American, and Hispanic births, separately

Pre- & post-ACA

ME & non-ME

AA-white disparity

National Center for Health Statistics (NCHS) Vital Statistics Birth Data Files, 2011-2016

County-level rates of low birthweight and preterm birth outcomes; 3,145 counties in contiguous U.S., excluding counties in U.S. territories; sample counties (n = 372) in 6 contiguous states that expanded Medicaid after January 1, 2014, and 9 independent cities, leaving 2,728 counties; county-level rates in included counties among 19,454,243 singleton births to women (ages 19 and older at time of birth)

None reported

Absolute

N/A

N/A

County-level variation for low birthweight and preterm births among all racial/ethnic categories declined in ME states; in non-ME states, geographic variation reduced for both outcomes among Hispanic births and low birthweight white births, but increased in both outcomes among AA births

Significance in changes in AA-white disparities not tested for either outcome

Glance, Thirukumaran, Shippey, Lustik, & Dick (2020) [26]

Determined whether ME was associated with reduction in revascularization disparities in patients with acute myocardial infarction

Retrospective analysis study; comparative interrupted time series analysis

Pre- & post-ACA

ME & non-ME

AA-white disparity

Did not distinguish between private and public insurance

Vizient Clinical Database/Resource (CDB/RM), 2010-2018

White and African American patients (ages 18-64) hospitalized with ST-segment elevation (STEMI) or non-ST-segment elevation acute myocardial infarction (NSTEMI) after ME

Patient characteristics, pre-ACA temporal trends, hospital effects

Absolute

Among patients with STEMI and NSTEMI, AA-white un-insurance rate disparity reductions in ME vs. non-ME states before and after ACA (DDD) is significant at p < 0.001.

N/A

Differences in AA-white revascularization rates for patients with STEMI decreased by 2.09 percentage points per year in ME vs. non-ME states; 7.24 percentage point increase for AA patients hospitalized with STEMI in non-ME states

Semprini & Olopade (2020) [27]

Evaluated impact of ME on disparity between African American-white breast cancer mortality rates

Difference-in-Difference (DD) fixed effects regression model with AA-white mortality ratio as outcome

Pre- & post-ACA

ME & non-ME

AA-white disparity

CDC All-Cause Mortality Database, 2012-2016

State-level breast cancer mortality data; no additional inclusion or exclusion criteria reported

Age

Absolute

N/A

N/A

ME did not lower disparity in breast cancer mortality; AA-white mortality ratio increased in ME states for all Medicaid-eligible age groups with significant effects in younger age groups

Chaudry, Jackson, & Glied (2019) [28]

Determined the extent to which the ACA has reduced disparities in insurance coverage among different racial/ethnic groups

Difference-in-Difference (DD)

Pre- & post-ACA; estimates adjusted

Examined public, private, and no coverage separately

American Community Survey (ACS), 2013-2017

Adults (ages 19-64); white, African American, Hispanic (any race); grouped by income relative to federal poverty guidelines;

Stratified by income level

Absolute

Among all income levels, white un-insurance rates declined in non-ME and ME states; AA and non-Hispanic un-insurance rates decreased in non-ME and ME states

Rates unadjusted and no significance testing conducted

N/A

N/A

Singh & Wilk (2019) [29]

Examined changes in access to primary care, measured by insurance status, having usual source of care, and delaying care due to cost, following ACA ME

Difference-in-Difference (DD); logistic regression models

Pre- & post-ACA

ME & non-ME

AA-white disparity

Did not distinguish between private and public insurance

Behavioral Risk Factor Surveillance System (BRFSS), 2011-2016

Adults (ages 25-64); white, African American, and Hispanic; other non-Hispanic adults (ages 25-64) with incomes below 100% FPL

Age, education status, marital status, gender, self-rated health status; race/ethnicity, income

Absolute

No significant AA-white disparity changes in insurance rates due to ME

No significant AA-white disparity changes due to ME status in unmet need due to cost or having a usual source of care

N/A

Lipton, Decker, & Sommers (2019) [30]

Examined changes related to racial/ethnic disparities in health insurance coverage and access to care after implementation of dependent coverage provision and full ACA limitation in 2014, respectively, separate from preexisting trends

Interrupted time series approach with 2 distinct intervention periods: October 2010 to December 2013 and January 2014 to December 2014

Pre- & post-ACA

ME & non-ME

AA-white disparity

Examined public, private, and any coverage

National Health Interview Survey (NHIS), 2000-2014

48,358 young adults (ages 19-25)

Age, sex, marital status, education, employment status, family income, region of residence; models included linear quarterly trend to control for trends in each outcome prior to ACA implementation

Absolute

ME associated with significantly greater increases in rates of having health coverage and having Medicaid coverage compared to gains for whites; rates of reporting any type of health insurance increased at significantly greater rates than for whites in ME states

ME associated with significantly greater increases in reporting a usual source of health care for AA compared to gains for whites; cumulative changes for ACA associated with significantly greater increases in reporting at least one doctor’s visit for AA compared to gains for whites

N/A

Crocker, Zeymo, McDermott, Xiao, Watson, DeLeire, Shara, Chan, & A-Refaie (2019) [31]

Examined impact of ME on utilization of cancer surgery for uninsured overall, low-income persons, and racial minorities

Poisson interrupted time series (ITS) analysis

Pre- & post-ACA

ME & non-ME

AA-white disparity

Examined private and public insurance separately

Merged data from State Inpatient Database, American Hospital Association, and Area Resource File, 2012-2015

81,000 patients (ages 18-64) who underwent cancer surgery

Adjusting for age, sex, comorbidity score, population-level and provider-level characteristics; quarter of discharge; year of admission; payer type

Absolute

N/A

Medicaid and uninsured population in ME states substantially increased utilization relative to non-ME states in 2014

DDD effect not significant for AA-white change and cancer surgery utilization

N/A

Wehby & Lyu (2018) [32]

Examined ACA ME effects on Medicaid take-up and private coverage and coverage disparities by age, race/ethnicity, and gender

Stratified Difference-in-Difference (DD) regression with state fixed effects; excluded 14 states that had partial or full expansions prior to 2014

Pre- & post-ACA

ME & non-ME

Examined Medicaid coverage, uninsured, individually purchased, employer-sponsored coverage, any private coverage

American Community Survey, 2011-2015

3,137,989 low-educated (high school or less) adults (ages 19-64); did not select sample based on household income or poverty level because income is potentially endogenous to insurance

National coverage time trends stratified by age group, stratified by race/ethnicity and gender

Absolute

Slight change in coverage disparities by race/ethnicity

DDD effect not tested for AA-white disparity change

N/A

N/A

Han, Yabroff, War, Brawley, Jemal (2018) [33]

Examined changes in percent uninsured and percent reporting care unaffordability, pre- & post-ACA ME

Difference-in-Difference (DD)

Pre- & post-ACA

ME & non-ME

Did not distinguish between private and public insurance

Behavioral Risk Factor Surveillance System (BRFSS), 2011-2017

118,631 cancer survivors (ages 18-64) with no known sex

Gender, age, race/ethnicity, household income, education, employment status, marital status, number of comorbid conditions

Absolute

Percent uninsured and care affordability decreased in all racial groups; disparities between white and Hispanic survivors persisted; greater reductions in un-insurance rates for AA in ME states (not statistically significant)

DDD effect not tested for significance for AA-white disparity change

Greater reductions in rates of care affordability reports for AA in ME states (not statistically significant); no significant findings for ME impact on AA-white disparities in unmet need due to cost

DDD effect not tested for significance for AA-white disparity change

N/A

Lee & Porell (2018) [34]

Estimated impacts of ACA ME on racial/ethnic disparities in insurance coverage, access to care, and health status

Difference-in-Difference (DDD) model specification with treatment and comparison groups; linear probability and regression models

Pre- & post-ACA

ME & non-ME

AA-white disparity

Did not distinguish between private and public insurance

Behavioral Risk Factor Surveillance System, 2011-2016

Non-pregnant childless adults (ages 19-64) residing in U.S. state or D.C. with incomes less than 100% FPL

Age, gender, race, marital status, education, employment, chronic disease status, tobacco use; state-year variables including number of hospital beds and physicians per 1,000 population, unemployment rate, per capita income, racial/ethnic composition, Senate voting records

Absolute

DDD effect not significant for AA-white disparity change in un-insurance

No significant findings for ME impact on AA-white disparities in unmet needs due to cost, having a usual source of care, or having annual wellness exam

No significant findings for ME impact on AA-white disparities in reported fair or poor physical health days, number of poor mental health days, and days with health-related activity limitation

Yue, Rasmussen, & Ponce (2018) [35]

Examined impacts of ME on health insurance coverage, having personal doctor(s), being unable to see doctors because of cost, and receiving a flu shot; tested racial/ethnic differential impacts

Quasi-experimental design with Difference-in-Difference (DDD) analyses; multiple imputations and survey weights used; excluded 14 states that had partial or full ME prior to 2014

Pre- & post-ACA

ME & non-ME

AA-white disparity

Did not distinguish between private and public insurance

Behavioral Risk Factor Surveillance System (BRFSS); State Physicans Workforce Data Book; Bureau of Labor Statistics, 2013-2015

Adults (ages 18 or older), low-income, non-elderly based on household income and family size; 18,408 observations in non-ME group and 16,964 in ME group

Age, general health, annual household income, race, education level, employment status, language, number of children in household, number of adults in household

Absolute

DDD effect not significant for AA-white disparity changes in rates of any-health coverage

No significant findings for ME impact on AA-white disparities due to cost, having usual source of care or personal doctor, or having flu shot

N/A

Hayes, Riley, Radley, & McCarthy (2017) [36]

Investigated effects of ME on access to health care across three racial/ethnic groups

Did not distinguish between private and public insurance

Stratified analysis; calculated and compared national averages for each indicator pre- & post-ACA, comparing ME and non-ME states

No statistical tests performed

American Community Survey (ACS), 2013 & 2015

Behavioral Risk Factor Surveillance System (BRFSS), 2013 & 2015

Uninsured adults (ages 19-64); adults (ages and older) who identified unmet heath care need do to cost and lacks usual source of care

None reported

Absolute

AA-white absolute disparity in un-insurance rates decreased in non-ME and ME states

No statistics of significance reported

AA-white absolute disparity in rates of unmet health care need due to costs decreased in non-ME and ME states

No statistics of significance reported

AA-white absolute disparity in rates of lacking a usual source of care decreased in non-ME and ME states

No statistics of significance reported

N/A

Flores & Vargas (2017) [37]

Tested whether ME predicted change in ethnoracial disparities with health insurance coverage at the county level

Fixed-effect regression models

Pre- & post-ACA

ME & non-ME

AA-white disparity

Did not distinguish between private and public insurance

American Community Survey, 2012-2014

U.S. counties with large enough minority population to conduct meaningful cross-race analyses; not a nationally representative study of U.S. counties

County-level immigration-related policy (20112-2014 from National Conference of State Legislatures); state-level racial prejudice (Google); county obesity rates (CDC); annual county median age and ethnoracial composition; baseline insurance coverage levels; annual county unemployment rate

Absolute

Gaps in county any-insurance coverage rates between AA and whites decreased, 2012-2014

DDD effect not significant for AA-white disparity change

N/A

N/A

Buchmueller, Levinson, Levy, & Wolfe (2016) [38]

Examined ACA and ME effect on rates of un-insurance, public health coverage, and private health coverage by racial/ethnic groups

Difference-in-Difference (DD) stratified by income group and state ME status

Pre- & post-ACA

ME & non-ME

AA-white disparity

Distinguished between private and public insurance rates

American Community Survey, 2008-2014

Adults (ages 19-64), white, African Americans, and Hispanics (any race)

Stratified analysis by income group and state ME status; did not control for sociodemographic or health status factors

Absolute

AA-white coverage gap decreased for both public and private insurance; greater gains for AA adults in non-ME states; greater gains for whites in ME states; AA without health insurance decreased in ME and non-ME states

DDD effect not significant for AA-white disparity change in un-insurance rate

N/A

N/A

McMorrow, Long, Kenney, & Anderson (2015) [39]

Examined ME impacts on absolute & relative disparity changes in un-insurance rates for African Americans and whites

Difference-in-Difference (DD)

Pre- & post-ACA

AA-white disparity

Did not distinguish between private and public insurance

National Health Interview Survey (NHIS), 2012-2014

Adults (ages 18-64); white, AA, Hispanic

Age; sex; did not control for sociodemographic or health status factors

Absolute & Relative

Absolute disparity for AA uninsured adults in ME and non-ME states decreased, 2013-2014

DDD effect not tested for significant differences in AA-white disparity in ME vs. non-ME states

N/A

N/A