What cardiac arrest registries can tell us about health differences and disparities

Abstract

Cardiac arrest is common and associated with significant mortality and morbidity among survivors. To better understand the epidemiology and outcomes of cardiac arrest, many countries and regions have developed cardiac arrest registries. In the United States, with the diversity of its population, these registries have illuminated differences and disparities in the care and outcomes of cardiac arrest patients based on their race, ethnicity, and sex. These findings raise concerns as traditionally vulnerable patient groups have lower survival rates for cardiac arrest—a condition for which overall survival is already low. Although leveraging registries to raise awareness of disparities in cardiac arrest outcomes is an important first step, further research is needed to understand the sources of these differences, narrow observed disparities and improve overall outcomes.

Introduction

Cardiac arrest is the sudden cessation of mechanical function of the heart and, in the absence of prompt resuscitative efforts, leads to death within minutes. Given the differences in circumstances, etiology, responder types, immediate resources, and outcomes, cardiac arrest is pragmatically categorized as two separate medical conditions—based on whether it occurs in the hospital (in-hospital cardiac arrest [IHCA]) or in community settings (out-of-hospital cardiac arrest [OHCA]). About 292,000 IHCA occur each year in the United States and survival to hospital discharge is ∼ 20% to 24%.1, 2, 3 In contrast, approximately 350,000 OHCA occur each year in the United States and survival to hospital discharge is ∼ 10%.⁴

The systematic capture, analysis and dissemination of cardiac arrest data and outcomes informs contemporary practice and highlights the presence of potential differences and disparities in care. Although overall survival rates for cardiac arrest are low, not all patients have the same odds of survival. For OHCA, there can be variation in bystander response and outcomes based on an individual’s sex, race, ethnicity, and socioeconomic status, as well as by neighborhood characteristics. For IHCA, while there is no evidence to suggest differential treatment of patients based on their demographic variables, differences and disparities in care and outcomes are seen and appear to be explained by the hospital at which a given patient receives care. Thus, efforts to understand (and improve) cardiac arrest care and outcomes for patient populations with historically worse outcomes (e.g., Black and Hispanic patients; women) rely on systematic collection of patient demographic data, arrest characteristics, treatment practices, and pertinent outcomes. To date, the primary strategy for this level of data collation is through cardiac arrest registries.

Cardiac arrest registries

Many countries maintain cardiac arrest registries, with several of them covering the entire population. These include the likes of the Swedish Cardiac Arrest Register (SCAR), a nationwide registry that has been collecting data on OHCA since 1990 and on IHCA since 2004.⁵ They recently provided insights into the merits of mobile application-based volunteer activation for OHCA.⁶ The All-Japan Utstein Registry of the Fire and Disaster Management Agency has been in existence since at least the mid-2000s and facilitated studying the efficacy of a nationwide public access defibrillation program.⁷ The Denmark Cardiac Arrest Registry has prospectively captured data for all OHCAs since 2001,⁸ while the Danish registry for IHCA (DANARREST) was established in 2013, and has data submitted from 95% of its hospitals.⁹ As seen above, these registries have been critical in advancing our understanding of the epidemiology and outcomes of cardiac arrest.5, 6, 7 However, due to a lack of data on key socio-demographic variables or the relative racial homogeneity of populations in these countries, these registries have been limited in providing insights on how cardiac arrest care and outcomes might differ based on the race and ethnicity of its patients. In fact many of the countries with leading cardiac arrest registries, like the three listed above, intentionally do not collect race or ethnicity data.10, 11 Therefore, the current review will focus on the cardiac arrest registries in the United States, given its diverse racial-ethnic makeup.

In the United States, registries for OHCA and IHCA are generally voluntary site-based registries. The foremost contemporary OHCA registry in the United States is the Cardiac Arrest Registry to Enhance Survival (CARES).⁴ Created as a joint collaboration between the Centers of Disease Control and Prevention and the Emory University School of Medicine’s Department of Emergency Medicine in 2004, CARES is an emergency medical services (EMS) agency-based registry for OHCA. CARES now spans 33 state-based registries, and community sites within 14 additional states with a catchment area of approximately 175 million people (or ∼ 53% of the U.S. population). CARES data is collected by participating EMS agencies and hospital sites using a data collection form which captures typical Utstein elements, and a handful of in-hospital and post discharge measures. The data is audited by CARES staff to ensure completeness.

The Resuscitation Outcomes Consortium (ROC) Epistry-Cardiac Arrest is a North American prospective population-based registry for OHCA at eleven geographical locations (British Columbia, Toronto, Ottawa, Seattle, San Diego, Pittsburgh, Portland, Milwaukee, Dallas and the states of Alabama and Iowa). With emergency medical service agencies submitting OHCA data, ROC had a catchment area of about 23.7 million persons.¹² After their launch, ROC also served as a multicenter, research network for conducting clinical trials in OHCA.¹³ However, ROC has not been actively collecting OHCA data since 2015.

Get With The Guidelines (GWTG)-Resuscitation is the IHCA registry in the United States. Originally known as the American Heart Association’s National Registry of Cardiopulmonary Resuscitation registry when it was launched in 2000, participating hospitals submit data on all IHCAs and receive reports from an interactive online Patient Management Tool^TM. In any given year, there are ∼ 350 to 400 participating hospitals (representing approximately 6% of hospitals in the United States) with 25,000 to 30,000 IHCA events each year.

Differences, disparities, bias

A seminal report, “Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care” commissioned by the Institute of Medicine¹⁴ as well as insights introduced by Rathore et al,¹⁵ provide a practical conceptual framework for determining what constitutes a difference, disparity or bias. Differences in care between groups, refers to any variation in care or outcomes that may be due to clinical appropriateness (indications and contraindications), patient preferences, comorbidities, or structural issues, such as healthcare access or insurance and bias. Thus, differences encompass all sources of variation between groups, including disparities and bias. Disparities refer to the residual differences in care or outcomes that persist after accounting for clinical appropriateness, patient preferences, and comorbidities. It refers to the ways in which people and systems knowingly and unknowingly contribute to poor care and outcomes for a subset of patients. Bias refers to disparities that are directly a result of an individual’s decisions (e.g., care providers), that disadvantage certain patients based on their attributes (e.g, race, ethnicity or sex). Biases can be explicit (racism) or subconscious (implicit bias). Since it is often difficult to measure individual provider motivations, identifying bias can be challenging. As a result, registries primarily allow investigators to characterize differences and disparities in care and outcomes by patient characteristics, such as by race/ethnicity or sex.

Out of hospital cardiac arrest

Racial and ethnic differences

Three decades ago, a field defining study reported significantly higher incidence and mortality of OHCA in Black residents of Chicago, compared to White residents.¹⁶ Poor survival in Black individuals was accompanied by lower rates of witnessed arrest and bystander interventions; however, there was no difference in EMS response times by race. Other studies have also observed lower rates of bystander cardiopulmonary resuscitation (CPR) in Black¹⁷ and Hispanic persons with OHCA.¹⁸ However, these earlier studies were conducted in limited geographic regions, and it remained unclear whether these findings were representative of other regions in the United States.

The inception of the CARES registry in the United States has provided opportunities to examine patient and EMS site level differences and disparities in OHCA care and outcomes across broad geographic regions in the country. Within CARES, one study was among the first to describe disparities in rates of bystander CPR based on the racial composition of the neighborhood in which the arrest occurred. They observed that individuals with OHCA were 51% less likely to receive bystander CPR in low-income Black neighborhoods, and that these neighborhood factors of race and income were additive.¹⁹ Subsequently, this work was extended using data from the ROC registry and observed that persons with an OHCA in a predominantly Black neighborhood had lower odds of hospital admission (OR 0.40; 95% CI, 0.39–0.55), survival to hospital discharge (OR 0.65; 95% CI, 0.51–0.82), and favorable neurologic survival (OR: 0.56; 95% CI, 0.38–0.82), as compared to persons with OHCA in a predominantly White neighborhood.²⁰ However, these racial differences were attenuated after adjusting for ROC site. In contrast, a subsequent CARES study observed that persons with OHCA in Black and low-income neighborhoods had lower rates of survival to hospital admission and survival to discharge than those with OHCA in White or high-income neighborhoods.²¹ In that study, differences in survival by neighborhood persisted even after controlling for differences in bystander CPR and neighborhood site, suggesting that other factors may have accounted for the observed survival differences. Similar findings were noted when neighborhoods were examined by Hispanic ethnicity. Within the ROC registry, persons with OHCA in neighborhoods with > 75% Hispanic residents were less likely to receive bystander CPR (OR 0.72; 95% CI, 0.55–0.96) or survive to discharge (OR 0.56; 95% CI, 0.34–0.93).²²

Lower rates of bystander CPR in Black and Hispanic neighborhoods may be due to lower rates of CPR training among residents of these communities,²³ or lack of resources to support dispatcher-assisted CPR programs. However, it remained unclear whether there were patient-level differences in rates of bystander CPR even in low-income and predominantly Black or Hispanic neighborhoods. A recent study observed that, compared to White persons, the odds of receiving bystander CPR were much lower in Black and Hispanic persons with OHCA, and these treatment differences explained, in part, the lower survival rate for OHCA in Black and Hispanic persons.²⁴ A concerning finding in the study was that the magnitude of disparities in bystander CPR were larger when the arrest occurred at a public location (OR 0.63; 95% CI, 60–0.66) compared to a home location (OR 0.74; 95% CI, 0.72–0.76), and this pattern was consistent regardless of the racial/ethnic composition or income of the neighborhood in which the arrest occurred. These findings raise additional concerns about implicit and explicit bias in layperson response to OHCA given the larger gap in bystander CPR rates in public locations and highlight that strategies to reduce racial/ethnic disparities in OHCA survival may need to go beyond improving CPR training rates.

Prior studies on racial/ethnic disparities in OHCA have largely focused on geographic regions (e.g., census tracts or counties), which are rarely amenable to interventions. In contrast, EMS agencies play a fundamental role in OHCA care. EMS agencies have a well-defined administrative structure, provide training to paramedics and emergency medical technicians, and interface with 9–1-1 dispatch centers, first responders as well as the community in organizing and delivering pre-hospital care to OHCA victims. Therefore, understanding how care patterns differ between EMS agencies can provide important insights into addressing racial/ethnic disparities in OHCA, particularly since we know there is large variation in OHCA survival at the EMS agency level (median odds ratio 1.35 (95% CI, 1.32–1.39).²⁵ This variation in EMS agency performance in OHCA has now been described by the race/ethnicity of the community in which the arrest occurs. A recent study observed that EMS agencies in Black and Hispanic communities had lower survival rates to hospital admission than EMS agencies working in White communities.²⁶ Moreover, existing measures on EMS practices (e.g., response time and rates of termination of resuscitation) collected in a national registry such as CARES did not explain the survival differences between the two EMS agency groups. Similar findings were also reported in secondary analyses of the PART trial which found no difference in measures of EMS CPR efficacy and intubation success rates.27, 28 Thus, a key research priority is to identify best practices at EMS agencies with the highest survival rates, especially high-performing EMS agencies which work in predominantly Black and Hispanic communities.

Sex differences

Both the CARES and ROC registries have also provided insights into potential sex differences and disparities in OHCA practices and outcomes for women. A study from the ROC registry showed that women with an OHCA in a public setting were 27% less likely to receive bystander CPR than men with OHCA in a public setting, although there were no clear sex differences in rates of bystander CPR for OHCAs occurring at home.²⁹ Also within the CARES registry, overall rates of bystander CPR were similar between men and women (40% vs. 38%), and rates of survival to hospital admission, survival to hospital discharge and favorable neurologic survival were higher in women than in men.³⁰

To better understand factors that explain lower rates of bystander CPR in women with OHCA in public, a national survey of 548 lay persons was conducted and observed that women might be less likely to receive CPR than men because of the sexualization of their bodies, fear that women were more likely to be injured by CPR, and failure to identify cardiac arrests in women.³¹ Such an application of registry data to design further research and spur implementation is critical to addressing disparities.

In hospital cardiac arrest

Racial and ethnic differences

In 2009, a National Registry of Cardiopulmonary Resuscitation (the precursor of GWTG-Resuscitation) study demonstrated differences in IHCA survival to hospital discharge in Black vs White patients.³² Black patients had an unadjusted survival to hospital discharge rate of 25.2%, while White patients had a 37.4% rate (relative risk 0.73 (95% CI, 0.67–0.79)). After adjusting for differences in patient characteristics the gap narrowed (adjusted relative risk 0.81 (95% CI, 0.75–0.88)), and further adjusting for admitting hospital further attenuated the disparity (adjusted relative risk 0.89 (95% CI, 0.82–0.96)). Thus about 2/3 of the racial differences in IHCA arrest survival were related to patient-case mix and the hospital in which care was administered. Of note, these differences persist even after hospital discharge.³³

Another study from the National Registry of Cardiopulmonary Resuscitation observed that among IHCA due to an initial rhythm of ventricular fibrillation or pulseless ventricular tachycardia, delays in defibrillation (>2 minutes) were common, and occurred more frequently in Black individuals (OR 1.23; 95% CI, 1.05–1.43), compared to White individuals.³⁴ Because patients with delays in defibrillation had a survival odds that was 52% lower as compared to those who were promptly defibrillated (OR 0.48; 95% CI, 0.42–0.54), these findings highlighted a potential opportunity to target interventions to narrow previously noted racial disparities in survival.

The link between racial disparity and hospital quality was further highlighted in a subsequent study that observed that Black-White difference in IHCA survival had narrowed markedly between 2000 to 2014, largely due to the elimination of racial difference in acute resuscitation survival, and substantial improvement in racial differences in post-resuscitation survival. Importantly, reduction in racial disparities in that study were driven by a larger improvement in survival at hospitals that disproportionately treated Black patients.³⁵

Sex difference

Within GWTG-Resuscitation, sex differences in rates of do-not-attempt resuscitation orders after successful resuscitation from an IHCA have been described, with women more likely to have a do-not-attempt resuscitation order placed (adjusted relative risk 1.15%; 95% C1, 1.10–1.20).³⁶ Despite this, there were no differences in rates of survival to discharge (34.5% in women vs. 36.7% in men, [relative risk 1.00; 95% CI, 0.99–1.02]). Two other studies have observed that in-hospital processes and outcomes may be comparable, if not better, for women. In patients that died during a resuscitation, female sex was associated with longer resuscitation times,³⁷ while childbearing women aged 14 to 44 were observed to have higher odds of return of spontaneous circulation, survival to hospital discharge, and favorable neurologic survival in a different study.³⁸

Next steps for cardiac arrest registries and researchers

Over the past few decades cardiac arrest registries have advanced our understanding of potential differences and disparities in the care and outcomes for patients with cardiac arrest. However, they have been unable to discern the underlying mechanisms behind observed differences and disparities.

Race is a social not a biologic construct. Thus, differences observed by race ultimately originate from the ways in which systems interact with race and not race itself. For centuries, skin pigmentation dictated socioeconomic status, opportunity for upward mobility, housing, education, and access to health care. The persistent and downstream effects of these factors predispose communities to comorbid conditions, and other detrimental physiologic responses that may underlie some of the differences in cardiac arrest outcomes by race. Although capturing comorbidities in cardiac arrest populations is challenging, understanding whether increased comorbidity burden or other facets of our systems are driving differences in outcomes, is critical to improve survival. Associations between sex and outcomes are even more complex to untangle as there are true biologic differences in addition to the differential societal impacts. Enhancing our registries to capture more data elements to uncover the underpinning mechanisms of differences and disparities is a critical next step for our registries.

Literature on racial and ethnic disparities outside the U.S. remains limited largely because many countries such as Germany and France intentionally do not collect race/ethnicity data in their registries.10, 11 Collecting these data are of paramount importance to document the existence of disparities and for developing strategies to eliminate them.

Another shortcoming of current US cardiac arrest registries is that not all hospitals or EMS systems contribute data. Thus, although CARES’ coverage area now collects OHCA data on ∼ 53% of the United States population, its data may not be generalizable to non-participating areas, especially rural areas of the country. Hospitals participating in GWTG-Resuscitation represent < 10% of all hospitals in the United States. Moreover, it remains unknown if and how EMS agencies in CARES and hospitals in GWTG-Resuscitation use registry data to track their performance and monitor the implementation of best practices, especially for traditionally vulnerable populations with worse survival outcomes.

Registries in the United States are also limited by the lack of longitudinal data among survivors. Routine linkage of CARES and GWTG-Resuscitation data with data from the Centers for Medicare and Medicaid Services would provide rich information on long term survival and patterns of post-discharge care in survivors of OHCA and IHCA. If realized, this would provide important information regarding the long-term trajectory of health outcomes of cardiac arrest and its associated economic impact. Moreover, it would inform potential strategies to enhance the health status of cardiac arrest survivors.

Another critical intervention that could improve our understanding of cardiac arrest and facilitate efforts to improve outcomes is making cardiac arrest a reportable condition. If cardiac arrest is reportable, this would provide incentives for EMS agencies and hospitals to innovate their care of patients with OHCA and IHCA, respectively, as there would be national benchmarking of cardiac arrest outcomes. It would also ensure that OHCA and IHCA data is systematically collected throughout the country, reducing potential participation bias in current registries. Finally, mandating cardiac arrest reporting would allow for more systematic evaluation of care differences and survival disparities by patients’ race/ethnicity and sex, as well as create greater urgency in developing interventions to narrow differences that exist.

There also remains a paucity of studies to understand why some EMS agencies or hospitals are achieving much higher survival rates for OHCA and IHCA than others. If future research is able to identify best practices used at EMS agencies or hospitals with much higher rates of cardiac arrest survival, this may provide a roadmap for all sites to improve cardiac arrest outcomes. Such an approach has been previously used for IHCA in the NIH-funded Hospital Enhancement of Resuscitation Outcomes for In-hospital Cardiac arrest (HEROIC). These investigators identified a range of best practices through a quantitative survey followed by qualitative interviews during on-site visits.39, 40 For instance, they observed that top performing hospitals had regular IHCA case review, monitored for chest compression interruptions, defined clear team roles and conducted in-depth mock codes. For reducing disparities in OHCA, ongoing initiatives such as the Reducing Ethnic-racial Disparities in Cardiac Arrest Survival Outcomes (RED-CASO, NIH R01HL160734) hold promise. Like HEROIC, the RED-CASO is a mixed methods study focused on identifying best practices at EMS agencies with the highest survival rates. A unique aspect of RED-CASO is that > 50% of included EMS agencies are those working in predominantly Black and Hispanic communities. It is the hope that identified strategies and best practices could provide a roadmap for other EMS agencies working in Black and Hispanic communities, with the goal of narrowing existing racial disparities in OHCA survival. Although such mixed-methods studies are labor, time and resource intensive, these studies will move beyond describing the existence of disparities and provide insights for improving care quality and reducing disparities. In many ways, registries hold unique potential if we are able to identify and disseminate best practices within the registries. We can then study their implementation and registries can serve as laboratories for quasi-experimental implementation of best practices in the future.

Conclusion

Racial, ethnic and sex-based differences and disparities in cardiac arrest outcomes have been described for the better part of three decades in the United States with little progress. We must extend the current bounds of our registries and registry facilitated research to ensure these disparate outcomes do not endure. The registries have spoken, now we must act.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Dr. Girotra (R01HL160734 and R01HL166305) and Dr. Chan (R01HL160734) are supported by grants from the National Institutes of Health.

CRediT authorship contribution statement

Anezi Uzendu: Writing – review & editing, Writing – original draft, Methodology, Investigation, Data curation, Conceptualization. Saket Girotra: Writing – review & editing, Supervision, Methodology, Data curation, Conceptualization. Paul S. Chan: Writing – review & editing, Supervision, Methodology, Investigation, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.