Stroke Disparities Among Non-Racial Minorities in the All of Us Research Program

INTRODUCTION

Numerous studies have demonstrated stroke disparities among racial and ethnic groups.¹ However, stroke disparities have not been well-described in non-racial minority groups, partly due to the underrepresentation of these groups in research. The National Insitutes of Health All of US study aims to provide health data on over 1 million Americans with a focus on groups that have been underrepresented in research.² We sought to describe the burden of stroke among non-racial minorities, including those defined by disability, income, education, and sexual orientation, in All of Us.

METHODS

We performed a cross sectional study in All of Us.² Data are available at www.allofus.nih.gov. We identified minority groups based on self-reported qualifiers of diversity, including race/ethnicity; age >=75 years; physical disability; education less than high school; income level <= $35k; and gender identity and sexual orientation.³ We identified stroke using survey and health record data. We used chi-squared tests and logistics regression to model the association between minority status and stroke. Multivariable models were adjusted for age, race/ethnicity, insurance status, smoking, hypertension, hyperlipidemia, diabetes, and atrial fibrillation.

Standard Protocol Approval

This study was deemed exempt from IRB review because all data are de-identified and publicly available.

RESULTS

We included 230,577 participants with available data. The prevalence of stroke was 2.7% (n=6,275). Disparities in stroke prevalence were present among non-racial minority groups, including disabled versus non-disabled persons (5.5% vs 2.4%; p<0.001) and persons with income <$35k versus >$35k (3.1% vs 2.5%; p<0.001) (Table 1). Minorities defined by sexual orientation and gender identity had a lower prevalence of stroke (2.4% vs 2.8%; p=0.001). In adjusted analyses, disability (OR 1.72; 95% CI 1.58–1.86; p<0.001) and income <$35k (OR 1.23; 95% CI 1.16–1.31; p<0.001) were associated with stroke.

Table 1.

Stroke prevalence across minority groups in All of Us.

Minority groups	Prevalence of minority groups among all study participants	Prevalence of stroke among minority groups	P
	All study participants, n =230,577	Participants with stroke, n =6,275
Race/ethnicity (n, %)
Whites	126,324 (55.8)	3,800 (3.0)	<0.001
Blacks	43,370 (19.2)	1,359 (3.1)
Hispanics	42,396 (18.7)	744 (1.8)
Asians	6,776 (3.0)	83 (1.2)
Other/More than one	7,526 (3.3)	163 (2.2)
Age (n, %)
<75 years	206,959 (89.8)	4,866 (2.4)	<0.001
> 75 years	23,618 (10.2)	1,409 (6.0)
Disability (n, %)
No disability	204,086 (89.9)	4,933 (2.4)	<0.001
Disability	22,893 (10.1)	1,258 (5.5)
Sexual Minority (n, %)
Non-LGBTQIA+	202,493 (87.8)	5,596 (2.8)	0.001
LGBTQIA+	28,084 (12.2)	679 (2.4)
Education (n, %)
High school completed	204,232 (90.4)	5,580 (2.7)	0.192
Less than high school	21,570 (11.2)	556 (2.6)
Income (n, %)
Household income >35K	108,836 (58.4)	2,738 (2.5)	<0.001
Household income <35K	77,431 (41.6)	2,384 (3.1)

LGBTQIA+ = lesbian, gay, bisexual, transgender, queer, intersex, asexual, other

DISCUSSION

In the All of Us study, we found that minority groups defined by age, disability, and income had a higher burden of stroke compared to individuals not in these groups.

We highlight a need for greater inclusion and retention of non-racial minorities in stroke research. Our data suggest that older persons and those with disability and low income have a high stroke burden. However, these groups have been excluded from stroke clinical trials through the use exclusions based on upper age limits, pre-existing disability, and indicators of nonadherence including poor attendance at clinic appointments and comorbid psychiatric or substance use disorders.⁴ Low-income populations also face several barriers to accessing care and participating in research including homelessness or housing insecurity, limited paid sick-leave, and inability to afford transportation to study sites.⁵ These factors lead to a viscious cycle of worsening disparities, as a lack of data on these populations is often invoked as a reason to withold treatment. There is likely a synergistic effect for intersectional identities; racial/ethnic minorities with disability or low income likely have a greater burden of stroke and greater barriers to participation in research.

This study has several limitations. First, the higher prevalence of stroke among disabled persons may reflect stroke-related disability. Second, while our models adjust for race and ethnicity, health differences among non-racial minority groups cannot be separated from the experience of being a racial minority.

In summary, we call attention to stroke disparities among non-racial minorities in All of Us. Further studies are needed to identify causes of stroke disparities in non-racial minorities and to promote the inclusion of these groups in research.

Non-Standard Abbreviations and Acronyms

LGBTQIA+

lesbian, gay, bisexual, transgender, queer, intersex, asexual, other