Doubly marginalized: the interplay of racism and disability in outcomes for minoritized people with Down syndrome

Abstract

Background:

Intersectionality, or the multidimensional influence of social identity and systems of power, may drive increased morbidity and mortality for adults of color with Down syndrome. We documented racial and ethnic differences in death and hospitalizations among Medicaid enrolled adults with Down syndrome and assessed interaction of racial–ethnic group and Down syndrome.

Methods:

Our sample consisted of 119,325 adults with Down syndrome and >3.2 million adults without intellectual disability enrolled in Medicare at any point from 2011–2019. We calculated age-adjusted mortality and hospitalization rates by racial–ethnic group among those with Down syndrome. We examined additive interaction between Down syndrome and racial and ethnic group on mortality and hospitalization rates.

Results:

Among those with Down syndrome, age-adjusted mortality rate did not differ between Black and White racial groups (rate ratio: 0.96, 95%CI: 0.92, 1.01) while mortality rate was lower for Pacific Islander (0.80), Asian (0.71), Native (0.77), and Mixed-race groups (0.50). Hospitalization rates were higher for all marginalized groups compared to the White group. When assessing the interaction between racial–ethnic group and Down syndrome, Black, Native Americans, and Mixed-race groups exhibited a negative additive interaction for mortality rate and all groups except Native Americans exhibited positive additive interaction for hospitalization.

Conclusions:

Increased hospitalization rates for adults with Down syndrome from marginalized racial and ethnic groups suggest worse health and healthcare. Similar mortality rates across racial and ethnic groups may result from increased infant mortality rate in marginalized groups with Down syndrome leading to reduced mortality among those surviving to adulthood.

Introduction

Down Syndrome is the trisomy of chromosome 21 and is the most common genetic cause of intellectual disability.¹ While the average lifespan has greatly increased in the past 70 years, people with Down syndrome still have average life expectancy <65 years and experience burdensome co-occurring conditions compared to peers without Down syndrome.^2–4 Individuals with Down syndrome face increased risk for congenital heart defects; autoimmune diseases; obstructive sleep apnea; dementia; vision and hearing problems; and recurrent infections⁵. Given these health conditions, people with Down syndrome are at an increased risk of hospitalization.^4,6.

There is some evidence to suggest people with Down syndrome from marginalized racial or ethnic groups have higher risks for poor health outcomes compared to White peers. Black children with Down syndrome have poorer survival in infancy compared to White children with Down syndrome⁷. Among adults with Down syndrome, those from marginalized racial or ethnic groups still live shorter lives compared to White non-Hispanic people.^3,8 These disparities exist in the wider context of racial and ethnic health disparities. In the U.S., widespread health disparities in disease burden and prevention disproportionately affect marginalized racial and ethnic groups.⁹ Racism is a central cause of these health inequities.^10–12.

In all adults with Down syndrome, health outcomes are shaped by systems of power and oppression interlocking with marginalized identities, specifically their disability identity and the societal response to disability¹³. Historically this social marginalization, i.e. ableism, stems from attitudes and beliefs that disabled people are inherently inferior because of physical or mental impairment.¹⁴ Ableism is structural, for example, people with Down syndrome need to maintain an income of less than $750/month to receive social security benefits, which keeps them impoverished and limits opportunity.¹⁵ Historically public health and epidemiology have contributed to the ableist approach to disability, aiming to ‘prevent’ and ‘cure’ disability.^16,17 Therefore, less effort has been placed on understanding social mechanisms and life course outcomes for disabled people¹⁸.

Individuals who identify as members of more than one marginalized social group (e.g. marginalized racial or ethnic group, disability status) experience a disproportionate burden of disease due to compounding structural inequity ¹⁹. Intersectionality is a framework that identifies the multidimensional influence of social identity and systems of power and privilege that drive health inequities.^20–22 Using an intersectional lens frames research to consider how systemic factors, whether intentional or unintentional, influence individual health outcomes. At the analytical level, health sciences often use cursory analysis of race (e.g. not reflecting that race is a downstream effect of racism), rarely analyze disability, and seldom ever address how these multifaceted identities interact with systems and each other.^23,24 The exclusion of disability (which affects one in four US adults ²⁵) in analysis leads to challenges in distinguishing the impact of ableist health outcomes or the intersectional effect of racism and ableism. Critically examining health disparities via intersecting social and structural factors may highlight the relationality between disability as a social identity and power that shape health outcomes.

One way to examine the intersection of ableism, racism, and identity in driving systemic disparity is through leveraging the US public insurance system. Medicaid is a public insurance program for low-income and disabled children and adults in the US. Medicaid serves as a primary insurance provider for individuals with Down syndrome, as nearly all people with Down syndrome are automatically eligible due to meeting disability eligibilty.²⁶ Additionally, adults with Down syndrome can also be covered by Medicare, a federal health insurance program for adults ≥65 years and certain younger people with disabilities²⁷. If an individual meets the eligibility criteria for both Medicaid and Medicare, they can be enrolled in both programs where Medicare would serve as the primary payer, and Medicaid covering any additional costs not reimbursed by Medicare (e.g., copays)²⁷.

Given the need for analyses at the intersection of racial and ethnic group and disability and the utility of Medicaid data, our objective was to document racial and ethnic differences in mortality and hospitalization rates among Medicaid enrolled adults with Down syndrome. We then assessed whether mortality and hospitalization rates were modified due to the intersection of racial and ethnic group and having Down syndrome.

Methods:

Data Source

Data are from the Down Syndrome Toward Optimal Trajectories and Health Equity using Medicaid Analytic eXtract project (DS-TO-THE-MAX). DS-TO-THE-MAX is a longitudinal claims cohort of adults ≥18 years enrolled in Medicaid at any point between 2011–2019, with any one inpatient claim or ≥2 outpatient claims for Down syndrome and a random sample of >3 million enrollees without developmental disability. The non-Down syndrome sample was derived by the Centers for Medicare and Medicaid Systems at baseline and represents a roughly 1% sample of all adult Medicaid enrollees. Individuals with intellectual disability, inclusive of Down syndrome, were excluded from the random sample. Data include demographic files; inpatient, other service, and long-term care claims and encounters; and pharmacy prescription claims. Data were purchased from the Centers for Medicare and Medicaid Services after application and approved data use agreement. More information about the data acquisition process can be found online.²⁸ This project was deemed exempt and participant consent was waived by the Boston University Medical Campus institutional review board. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.²⁹

Aligning Data Collection Systems

DS-TO-THE-MAX spans two data collection systems: Medicaid Analytic eXtract (MAX; 2011–2015) and the Transformed Medicaid Statistical Information System Analytic Files (TAF; 2014–2019). There was a transitionary period in 2014 and 2015 with some states using MAX, whereas other states had already transitioned to TAF. A unique beneficiary identifier linked the two data systems. Many demographic variables had the same format and we checked them for consistency. We reparametrized other variables to align in both systems.²⁶ Further information on MAX and TAF data alignment and processing is provided in Supplement eAppendix 1.

Demographic Data

We assessed demographic characteristics from the person file and demographic enrollment file. Age at each year of enrollment was determined by date of birth. Racial and ethnic group were self-reported at Medicaid enrollment. Each state determined how their data were collected (e.g., number of racial / ethnic categories to choose from). Results were then harmonized by the Centers for Medicare and Medicaid Services, who categorized racial groups as: Asian, Black, Native American, Pacific Islander, and White; and ethnic groups as: Hispanic or non-Hispanic. If racial or ethnic group were reported in any year, we considered that to be the individual’s racial or ethnic group, even if other years were missing racial or ethnic group data. If an individual had two or more different racial groups reported over time, we classified them as multiple races. If racial or ethnic group was missing in all years (24% of Down syndrome group was missing racial group, 12% missing ethnic group; 30% of the non-Down syndrome sample was missing racial group, 13% were missing ethnic group) we used multiple imputation to probabilistically account for missingness. We used the American Community Survey 2016–2020 summary data file ³⁰ to calculate the percentage of each racial and ethnic group at the zip code level. These proportions and demographic data were used for imputing the missing race and ethnicity. We did 30 imputations of a model that included age, sex, disability eligibility, dual enrollment, and zip code–level racial or ethnic group distribution (pre-imputation data are presented in Supplement eTable 1).

Enrollment Data

We linked data across years by beneficiary identifier. We examined the number of individuals enrolled and person–time. An individual’s person–time would count toward the Down syndrome in any year they were enrolled if they had a Down syndrome claim in any year (e.g., if the first Down syndrome claim was in 2012 and they were enrolled in 2011, their 2011 person–time would be in the Down syndrome cohort). We used this approach because Down syndrome is a lifelong condition, and a lack of claim likely indicates limited health care use in the year or clinician coding practice. If an individual turned 18 years of age after 2011, their data for years that they were younger than 18 years were excluded. We calculated enrollment by year. We identified loss to follow-up (i.e., disenrollment with no death), and continuous enrollment (i.e., no breaks in enrollment). We examined ever dual enrollment (individuals who were ever enrolled in both Medicare and Medicaid at the same time) and person–time dual enrollment through yearly indicators of dual enrollment.

Outcome measurement

Our main outcomes of interest were mortality and hospitalization. Date of death is a variable provided by the Centers of Medicare and Medicaid Services as part of their demographic data sets. Date of death is only provided if they are enrolled at time of death. We calculated time from study entry until death as our person–time unit for incidence analyses. We defined hospitalization as unique claims in the inpatient data set with nonoverlapping admission dates. These dates are of low concern for data quality issues.³¹

Statistical Analysis

We presented descriptive data for demographic characteristics and aggregate data by cohort (Down syndrome or non-Down syndrome group). Then, we assessed mortality and hospitalization rates among racial and ethnic group within the Down syndrome cohort. We calculated these statistics stratified by age category, then calculated rate ratios using the white and non-Hispanic groups as referents for race and ethnicity, respectively. Rate ratios were age standardized to account for different age distributions between racial or ethnic groups. We first calculated age-specific death rates for each racial and age group by dividing the number of deaths by person–years for each age bracket. Then we standardized the rate to the reference groups by multiplying the death rates of each marginalized age group to the corresponding death rates by age group in the reference group. We also ran sensitivity analysis assessing race and ethnicity as one construct (e.g., White non-Hispanic, White Hispanic, Black non-Hispanic and so on).

Our goal was to run Poisson and negative binomial regressions to quantify mortality and hospitalization rate differences between the Down syndrome groups and our non-Down syndrome group. Mortality rate met the Poisson regression assumptions (mean=0.16, variance=0.14) but hospitalization did not (mean=1.14, variance=8.20). Therefore, we used Poisson models to estimate mortality and negative binomial regressions for hospitalization. However, our models for differences measures did not converge and we calculated mortality rate ratio and hospitalization rate ratio. To account for differing person–time, we used person time as an offset. We adjusted for sex, age, and region (see DAG in Supplement eAppendix 2). As sensitivity analyses, we explored adjustment for enrollment factors- dual enrollment and continuous enrollment (i.e. no gaps of >1 month without enrollment). These policy related factors could be confounders (eAppendix 2), as an individual with Down syndrome may be more likely to qualify and these policies may decrease risk of poor outcomes.

Using our mortality and hospitalization rate ratios, we calculated additive interaction on mortality and hospitalization rate by calculating Relative Excess Risk due to Interaction (RERI). RERI allows for assessment of departure from additivity on a relative risk scale. Assessment on the additive scale is of more relevant public health because it reflects the departure from expectation when evaluating two variables in combination and can be more useful in allotting treatments.^32,33 For each non-White racial group (or non-Hispanic ethnicity) we restricted to the group and the white (or non-Hispanic) referent group and created a four-level variable that accounted for all combinations of Down syndrome and race / ethnicity. We calculated RERI as ${\beta }_{\mathit{Down\; Syndrome}+\mathit{Racial\; group}}–{\beta }_{\mathit{Down\; syndrome}}–{\beta }_{\mathit{Racial\; group}}–1$

95% confidence intervals were calculated around the RERI using standard errors for the corresponding RERI coefficient. RERI provides information about the direction of the interaction but we cannot make statements on the underlying magnitude of interaction. We also calculated the observed compared to expected rate ratios and corresponding P value when assessing multiplicative interaction which does have a magnitude and should be presented with additive interaction analyses.³⁴

Statistical analyses were performed in R version 4.3.0.

Results

Demographics

Our sample of Medicaid-enrolled adults with Down syndrome from 2011 to 2019 included 119,325 unique individuals (796,984 person–years) of which 14% were self-identified as Black, 3% Asian, 1% Pacific Islander, 1% Native American, 6% multiple races, and 75% White (Table 1). For ethnicity, 17% self-identified as Hispanic. When looking at race and ethnicity as one construct, we found 9% self-identified as White-Hispanic, 12% Black non-Hispanic, 2% Black-Hispanic, and 6% as other-Hispanic. Our non-Down syndrome sample consisted of 3,055,494 enrollees, of which 61% were White, 24% Black, 2% Pacific Islander, 6% Asian, 1% Native American, and 5% were multiple races. Individuals in the Down syndrome cohort were more likely to be male and living in the Northeast or South compared to the non-Down syndrome sample. Among those with Down syndrome, 60% were ever dual enrolled in Medicare (46% in the non-Down syndrome sample). A full list of demographic characteristics by racial and ethnic group can be found in Supplement eTable 2.

Table 1.

Characteristics of adults with claims for Down Syndrome in Medicaid, 2011–2019

		Down Syndrome	Random Sample
		N= 119,325 Person–years: 796,984	N= 3,055,494 Person–years: 11,995,827
Deaths		N= 19,335	N=144,015
Hospitalizations		Mean=1.14 per year SD=2.9, Median=0, IQR=1	Mean=0.65 per year SD=2.6, Median=0, IQR=1
Age at First Claim (years), n(%)	18–25 26–34 35–44 45–54 55–64 65–89	42,643 (36) 20,586 (17) 19,096 (16) 22,573 (19) 12,065 (10) 2,362 (2)	1,047,252 (34) 566,367 (19) 416,024 (14) 366,986 (12) 331,776 (11) 327,089 (11)
Race, n(%) a	White Black Pacific Islander Asian Native American Mixed	89,081 (75 16,835 (14) 1,256 (1) 4,032 (3) 1,035 (1) 7,086 (6)	1,873,446 (61) 723,538 (24) 67,045 (2) 187,571 (6) 39,711 (1) 164,183 (5)
Ethnicity, n(%) b	Non-Hispanic/Latino Hispanic/Latino	99,444 (83) 19,881 (17)	2,424,101 (79) 631,393 (21)
Sex, n(%)	Male Female	61,473 (52) 57,852 (49)	1,338,581 (44) 1,716,913 (56)
Region, n(%)	Northeastern Midwest South West Other/US territory	26,711 (22) 27,615 (23) 38,175 (32) 26,200 (22) 622 (1)	612,522 (20) 615,027 (20) 1,019,754 (33) 764,493 (25) 43,691 (1)
Eligibility Type (Ever), n(%)	Disability Income	96,879 (81) 56,175 (47)	745,606 (24) 1,644,528 (53)
Ever dual enrollment with Medicare, n(%) Person years	Yes No Mean, (SD) Median, (IQR)	71,086 (60) 48,239 (40) 3.80, (4) 6.0 (0, 8)	680,356(22) 2,375,138 (78) 3.92 (3) 3(0, 4)

^aDS missing race pre imputation: 23,041 (24%); Random sample missing race: 71,356 (30%)

^bMissing ethnicity pre imputation: 14,438 (12%); Random sample missing ethnicity 31,979 (13%)

Mortality and hospitalization rates in adults with Down syndrome by race and ethnicity

Mortality rates were qualitatively similar across racial groups and ethnic groups until age 45. (Figure 1a and 1b; Supplement eTable 3). At older ages mortality was greater for White and non-Hispanic adults compared to non-White and Hispanic adults. Black enrollees with Down syndrome had the highest rate of hospitalizations among all racial groups in all age groups except for 65+, where Pacific islanders and Native Americans had the highest rate (Figure 2a). There was a higher rate of hospitalizations among Hispanics compared to non-Hispanics in all age groups (Figure 2b).

Figure 1. Mortality rate by racial or ethnic group for Medicaid enrollees with Down syndrome, 2011-2019.

1a. Mortality rate by racial group

1b. Mortality rate by ethnic group

Figure 2. Hospitalization rates by racial or ethnic group for Medicaid enrollees with Down syndrome, 2011-2019.

2a. Hospitalization rate by racial group

2b. Hospitalization rate by ethnic group

Mortality and Hospitalization rates by race and ethnicity in the Down syndrome group

Age-adjusted mortality rates were lower for Black 0.96 [95% CI: 0.92–1.01], Pacific Islander 0.80 [0.66–0.96], Asian 0.71 [0.63–0.79], Native American 0.77 [0.62–0.94], and Mixed racial groups 0.50 [0.45–0.56] compared to the White group. All marginalized groups had higher rates of hospitalization compared to the White group; Black 1.38 [95% CI: 1.36–1.40], Pacific Islander 1.05 [1.00–1.11], Asian 1.07 [1.01–1.14], Native American 1.06 [1.02–1.09], and Mixed racial groups 1.15 [1.12–1.17] (Figure 3a; eSupplement Table 4). Hispanic people with Down syndrome had 0.71 times the mortality rate compared to non-Hispanic people with Down syndrome [95% CI: 0.68–0.76] and 1.09 times the rate of hospitalizations [95% CI: 1.07–1.11], (Figure 3b). In our sensitivity analysis where we looked at race and ethnicity as one construct (Supplement eTable 5), we saw mortality rate ratios for all marginalized groups were less and hospitalizations were greater than the White referent group.

Figure 3. Age-standardized mortality and hospitalization incident rate ratios comparing racial or ethnic group of Medicaid enrollees with Down syndrome, 2011-2019^a.

A. Age-standardized mortality rate ratio by racial group

B. Age-standardized mortality rate ratio by ethnic group

C. Age-standardized hospitalization rate ratio by racial group

D. Age-standardized hospitalization rate ratio by ethnic group

^a. White and non-Hispanic as reference group

Main Effect and Interaction Analysis on Mortality and Hospitalization rates across marginalized Down syndrome groups

In our Poisson model examining mortality rates among those with and without Down syndrome, we found an individual with Down syndrome had 3.69 [95% CI: 3.62–3.75] times the mortality rate compared to an individual without any intellectual disabilities (adjusted for age, sex, and region). When looking at ethnicity, we found Hispanics with Down syndrome had 3.83 [95% CI: 3.77–3.89] (CI) times the mortality rate compared to non-Hispanics without any intellectual disabilities. Individuals with Down syndrome had 1.17 [95% CI: 1.16–1.19] (CI) times the hospitalization rate compared to those without disability, and Hispanics had 1.21 [95% CI: 1.19–1.23] (CI) times the rate compared to non-Hispanics.

Black (−0.20; 95% CI: −0.24, −0.15), Native Americans (−0.50; 95% CI: −0.71, −0.29), and mixed-race (−0.33; 95% CI: −0.44, −0.21)) groups exhibited a negative RERI based on an additive model, suggesting a protective effect against mortality compared to our reference group (Table 2). Pacific Islander (0.65; 95% CI: 0.46, 0.83) and Asian (0.10; 95% CI: −0.02, 0.22) had a positive RERI. Hispanic enrollees with Down syndrome had a negative RERI as well (−0.1, 95% CI: −0.15, −0.04). When examining the RERI for hospitalizations, we found Black (0.22; 95% CI: 0.19, 0.25), Pacific Islanders (0.49; 95% CI: 0.37, 0.61), Asian (0.52; 95% CI: 0.46, 0.59), and mixed (0.14 95% CI: 0.08, 0.19) racial groups had a super additive interaction effect. RERI was also super additive for hospitalization rate among Hispanic people with Down syndrome (0.10; 95% CI: 0.07, 0.13). When examining interaction terms in multiplicative models, there was super-multiplicative interaction for mortality and hospitalization for all groups except the Native American group (eSupplement Appendix 3; eSupplement Appendix 4).

Table 2.

Interaction Analysis on Mortality and Hospitalization rates across marginalized Down syndrome (DS) groups

Mortality Rate looking at Race and Down syndrome as an interaction
Racea	Mortality Rate (IRR)	RERIb	95 % CI
White no DS	1.00
White DS	1.30
Black no DS	0.80
Black DS	3.27	-0.20	(−0.24, −0.15)
White no DS	1.00
White DS	3.73
Pacific Islander no DS	0.65
Pacific Islander DS	4.02	0.65	(0.46, 0.83)
White no DS	1.00
White DS	3.75
Asian no DS	0.41
Asian DS	3.25	0.10	(−0.02, 0.22)
White no DS	1.00
White DS	3.73
Native no DS	1.05
Native DS	3.28	-0.50	(−0.71, −0.29)
White no DS	1.00
White DS	3.74
Mixed no DS	0.47
Mixed DS	2.88	-0.33	(−0.44, −0.21)
Ethnicitya	Mortality Rate (IRR)	RERIb	95 % CI
Non-Hispanic no DS	1.00
Non-Hispanic DS	3.83
Hispanic no DS	0.58
Hispanic DS	3.31	-0.10	(−0.15, −0.04)
Hospitalization Rate looking at Ethnicity and Down syndrome as an interaction
Racea	Hospitalization Rate (IRR)	RERIb	95 % CI
White no DS	1.00
White DS	1.16
Black no DS	1.14
Black DS	1.53	0.22	(0.19, 0.25)
White no DS	1.00
White DS	1.18
Pacific Islander no DS	0.94
Pacific Islander DS	1.61	0.49	(0.37, 0.61)
White no DS	1.00
White DS	1.19
Asian no DS	0.67
Asian DS	1.38	0.52	(0.46, 0.59)
White no DS	1.00
White DS	1.18
Native no DS	1.28
Native DS	1.46	0.01	(−0.13, 0.14)
White no DS	1.00
White DS	1.18
Mixed no DS	1.00
Mixed DS	1.32	0.14	(0.08, 0.19)
Hospitalization Rate looking at Ethnicity and Down syndrome as an interaction
Ethnicitya	Hospitalization Rate	RERIb
Non-Hispanic no DS	1.00
Non-Hispanic DS	1.21
Hispanic no DS	1.08
Hispanic DS	1.39	0.10	(0.07, 0.13)

^aAdjusted for sex, age, and region

^bRelative Excess Risk due to Interaction

When we included service use variables in the model, we saw the observed mortality rate and hospitalization rates moved toward the null compared to our primary model (eSupplement Appendix 5). We observed similar movement toward the null when we restricted to those who were continuously enrolled in specific racial and ethnic groups. We saw our estimate move away from the null for hospitalization.

Discussion

We examined the complex intersectionality between having Down syndrome and being a member of a marginalized racial or ethnic group in the Medicaid population. In our cohort of 119,325 adults with Down syndrome in the Medicaid system from 2011 to 2019, those from marginalized racial or ethnic groups had higher hospitalization rates compared to White peers with Down syndrome and greater hospitalization rates than expected based on an additive model. Exact mechanisms on how racial or ethnic group affects health outcomes in the Down syndrome population need to be determined, but the stark differences in mortality and hospitalization rates by marginalized groups are clear.

We found that after age adjustment, all marginalized groups with Down syndrome had the same or lower mortality rates compared to White and non-Hispanic peers with Down syndrome. In interaction analyses, Black, mixed race, Native American, and Hispanic groups had negative relative excess risk due to interaction. For Black and Hispanics with Down syndrome, results mirror survival paradoxes in the general population, where Hispanic and Black groups exhibit lower mortality rates in adulthood relative to non-marginalized groups, despite having a disadvantaged risk factor profile ^35–39. In our data, this survival advantage is likely driven by conditioning on survival to age 18. When looking at Black children with Down syndrome, the literature suggests that they are at increased risk of death compared to White peers with Down syndrome ^8,39 because of increased risk of death in infancy for Black children with Down syndrome. We hypothesize that pattern in mortality rates observed in our data is partially due to sicker Black and Hispanic children with Down syndrome not surviving until adulthood, and the ones that do survive are generally healthier than their white peers. Given the historic increased infant mortality rates in Black children with Down syndrome, it is not surprising we see the decreased rates in adulthood across all age-strata we assessed ⁸. Continued work is needed to explore cause of death factors that privilege health among some and have also created systems perpetuating inferior treatment and access to healthcare among others.

When assessing rate of claims for hospitalizations, which are not contingent on longevity, we saw increased hospitalization rates comparing marginalized racial or ethnic groups compared to White groups. Increased hospitalization in Black beneficiaries compared to White beneficiaries is consistent in non-Down Syndrome populations, highlighting the roles of racism and structural disparity.⁴⁰ These findings are also in line with the general paradox noted above, where racialized groups have higher hospitalization rates than White peers.

We found that adults with Down syndrome from marginalized racial or ethnic backgrounds experienced higher hospitalization rates than what was expected than the additive effects of having Down syndrome and being from a marginalized group. Higher rates may be related to access to care, especially access to Down syndrome specialty clinics. Most people with Down syndrome do not have access to specialty clinicians and given geographic distribution it may be even less likely for those from marginalized racial and ethnic groups to gain access ^41,42. A 2020 study that looked at Down syndrome specialty clinic accessibility in the US, found that 1 in 5 individuals with Down syndrome must travel >2 hours to reach their nearest clinic, and more than 33% of these patients lived in the South.⁴² One-third of our study population were geographically located in the South, highlighting geographic disparities in our study population. Preventative care is often lacking among people with developmental disabilities ⁴³, like Down syndrome, which may be exacerbated by dismissive care and racist medical algorithms that leads to more hospitalization and death ⁴⁴. For instance, caretakers of Black individuals with Down syndrome report worry about their loved one with Down syndrome being treated with respect and that dismissive healthcare providers were a barrier to quality care ⁴⁵.

The intersectionality of disability and race extends beyond those with Down syndrome. In our sensitivity analyses with adjustments that made the non-Down syndrome group more like the Down syndrome group (e.g. more likely to be consistently enrolled, adjusting for disability status), our estimates were moved toward the null, suggesting that disability more generally might be a driving factor as compared to Down syndrome specifically. Pervasive ableist attitudes and practices in the medical profession may also contribute to the increased mortality rate and hospitalization rate than expected ⁴⁶. Disparities in access may signify larger systemic barriers reinforced by social inequalities and overlayed power structures. A recent study found individuals in marginalized racial and ethnic groups with intellectual and developmental disability who lived in areas with more racism and ableism had poorer quality of life, regardless of demographics, compared to peers ⁴⁷. Our results in the context of systemic racism and ableism underscore the importance of culturally sensitive, person-centered care and interventions for marginalized populations and communities.

We were able to identify mortality and hospitalization disparities among marginalized groups in adults with Down syndrome using a full Medicaid dataset spanning 9 years. Our study benefited from a large sample size and the identification of Down syndrome cases through Medicaid claims. Our analysis relies on the use of racial and ethnic group and Down syndrome status as downstream proxies of racism and ableism. We did not have individual or community level metrics of experienced discrimination which could have been useful to find more specific sub-groups that experience these inequities differently. While approximately 90% of adults with Down syndrome are enrolled in Medicaid ²⁶ our cohort does not capture those who are only covered by Medicare or are privately insured. We did not have more refined ethnicity categories than Hispanic / non-Hispanic. We were limited by the nature of claims data, and thus did not have more detailed phenotypic information or important social factors, such as employment status or living arrangement. Incidence rate difference models did not converge, which prevented us from being to directly assess risk due to interaction.

Conclusions

Adults with Down syndrome from marginalized racial and ethnic groups are at increased risk of hospitalization but have similar mortality rates before age 45 compared to White peers with Down syndrome. Hospitalization rates were higher in adults with Down syndrome in marginalized racial ethnic groups compared to expectations from an additive interaction model. There is still work to be done to ensure that all adults with Down syndrome achieve a healthier and longer lifespan, which involves understanding the mechanisms linking racism and ableism to health outcomes to address these disparities.

Data access:

The data were purchased and used under a data use agreement with the Center for Medicare and Medicaid statistics. The data are not allowed to be shared. Code are proprietary products of the Boston University Biostatistics and Epidemiology Data Analytic Center and may be available upon request.