Racial disparities in endovascular thrombectomy: it’s more than just access

Ischemic stroke is a leading cause of long-term disability.¹ From 2010 to 2050, incident strokes are predicted to more than double.² This is a major public health issue in particular for Black and Hispanic patients, who will account for a disproportionate share of that increase.²

Ischemic stroke due to large vessel occlusion (LVO) are typically the most debilitating,³ but morbidity and mortality from LVO is substantially reduced with endovascular thrombectomy (EVT)⁴.

In this issue of Stroke, Kim and colleagues report findings from an analysis of statewide Texas data demonstrating lower rates of EVT among Black patients relative to White and Hispanic patients.⁵ The findings are confirmed in a national sample, validating this is more than a regional phenomenon in Texas, but rather a systemic, national issue.

These disparities are consistent with prior research in acute stroke regarding both EVT^6,7 and thrombolytic administration^8,9. Given the effectiveness of EVT, the stakes are high and thus begs the question – what is the cause of this disparity, and how can we address it?

Is it about access?

Often, we assume that disparities in treatment between groups are due to simple problems like differential access, leading to simple solutions such as redistribution of resources. Increasing attention has been focused on matters of spatial justice, recognizing that it matters where people live.¹⁰ This is true in terms of the built environment, patients’ social support at home, strategic locations of hospitals and outposts, and accessibility of care in the communities in which they reside.

Prior research on the role of hospital choice and its contribution to differential rates of EVT between groups has been mixed.^6,11 In the current study, access to EVT-capable hospitals and to comprehensive stroke centers (CSCs) was similar across groups. Yet despite similar presentation and access to EVT-capable sites, rates of EVT were different. This problem will not be solved by addressing access alone. Consistent with the framework on health disparities research developed by the National Institute on Minority and Health Disparities,¹² there are multiple important levels to consider to fully understand and address this problem.

The Patient Level

In studies using stroke registries or state-wide databases it is important to consider how the study population is identified. The lack of Hispanic/Latino patients in this study population is notable in context of the ethnic composition of Texas (21% versus 40% Hispanic/Latino in Texas).¹³ Disparities in treatment may be masked if there are concurrent disparities in initial stroke diagnosis. Here, assumptions about a patient’s baseline and the etiology of their presentation (e.g., stroke versus substance use) are critical. These assumptions are open to both explicit and implicit bias that may reduce the probability of stroke identification for Black or Hispanic patients.

If incidence of LVO is similar between groups¹⁴, the lack of data on eligibility for EVT is a limitation. We do not know whether differences in presentation time, differential use of emergency medical services (EMS), or differential EMS response times impacted eligibility.

The authors provided some additional analyses to address this limitation. In one analysis they limit comparison to patients who were treated with alteplase, where even among those earliest presenters Black patients were still less likely to receive EVT. Thus the lower rate of EVT among Black patients is not exclusively explained by delays in presentation. The authors subsequently narrowed analysis to patients with an LVO that presented to a CSC. The disparity persisted, despite the group being timely and having the necessary access.

Another issue is how we understand EVT is made available to patients. Utilization suggests a choice was made to receive a service and assumes that the service was offered. However, patients are often not in the position to request EVT if the treatment team is not already considering it. Rather than advocate for themselves by ‘shopping around’ patients are limited to the hyperacute treatment options offered where they are. Can we assume equal treatment offerings across racial and ethnic groups?

The Hospital Level

In other medical literature, disparities in access, quality of care, and patient outcomes have been attributed to hospital-level factors. One example is poor critical care outcomes for Black patients attributed to ‘minority-serving hospitals’ with less improvement in care and outcomes over time.¹⁵ Another from the surgical literature reports Black patients were more likely to live closer to high-quality hospitals and yet received surgery at lower-quality hospitals.¹⁶ Finally in obstetrics, Black and Hispanic women were more likely to deliver in hospitals with higher complication rates.¹⁷

This hospital-choice argument carries into the stroke literature as well. One study reported that disparities in thrombolytic administration were greater at hospitals treating more minority patients, whereas hospitals treating a higher proportion of White patients had less disparity between groups.¹⁸ If a similar effect is in play in this study, could there be within group differences between hospitals that may help explain the concerning findings reported? More plainly, are some CSCs behaving differently than other CSCs?

How would such hospital-level factors manifest? We might ask whether there are differences in the ‘activation energy’ required to call a stroke code. Is the culture around stroke care such that false positives are considered appropriate and necessary, or an error? We might also consider the relationships between disciplines and whether there is a collegial culture of discourse. Increasingly, we should also ask whether hospitals are performing self-assessments across racial and ethnic groups to ensure equitable care and outcomes.

The System Level and Larger Network Level

We also know that setting matters. Stroke care is part of a multi-disciplinary system. We must consider where patients live and initially present, transfer requirements and protocols, and what resources exist in each setting. In this study, disparities were exaggerated in rural settings, as has been described previously.^19,20 In rural areas there are more disparities in access to stroke centers (more due to socioeconomic status than ethnicity).²¹ Texas is a state that has some of the largest cities in the country as well as large rural areas such that both environmental extremes are contained in the data.

The findings reported by Kim and colleagues suggest a systemic issue, and both a statewide and national phenomenon. It cannot be simply chalked up to geography, nor attributed just to patients. We must critically examine the source of this disparity at every level. Who is allowed to be sick? What are we assuming about patients’ baselines? What is the hospital culture around stroke alerts? And what is the system doing to equitably match patients to resources?

As we recognize that the problem is not at one level alone but at all levels simultaneously, we are left with a picture of structural racism. Kim et al paint the picture of a systemic issue where there isn’t one simple problem to solve. This is a structural problem. Some components will be more straightforward to address, and others will take more creativity. We may change the culture, organization of medical teams, and the structure to optimize access and care delivery for all patients. But until we do all of these things and health systems consider their environment while holding themselves accountable to their communities, and advocate for change – we will continue to see studies like this one that suggest we need to improve our methods of caring for our patients.¹⁰ The hyperacute, time-sensitive nature of acute ischemic stroke care means that patients are limited to what is offered to them. Accepting the status quo does not work when we are talking about access to the most effective disability-reducing intervention there is. We can do better.

Non-standard Abbreviations and Acronyms

LVO

large vessel occlusion

EVT

endovascular thrombectomy

CSC

comprehensive stroke center

EMS

emergency medical services