Article, Emergency Medicine

Racial and insurance disparities among patients presenting with chest pain in the US: 2009-2015

a b s t r a c t

Background: Nationally representative studies have shown significant racial and socioeconomic disparities in the triage and diagnostic evaluation of patients presenting to the emergency department (ED) with chest pain. How- ever, these studies were conducted over a decade ago and have not been updated amidst growing awareness of healthcare disparities.

Objective: We aimed to reevaluate the effect of race and Insurance type on triage acuity and diagnostic testing to assess if these disparities persist.

Methods: We identified ED visits for adults presenting with chest pain in the 2009-2015 National Hospital Am- bulatory Health Care Surveys. Using weighted logistic regression, we examined associations between race and payment type with triage acuity and likelihood of ordering Electrocardiography or cardiac enzymes.

Results: A total of 10,441 patients met inclusion criteria, corresponding to an estimated 51.4 million patients na- tionwide. When compared with white patients, black patients presenting with chest pain were less likely to have an ECG ordered (adjusted odds ratio [OR] = 0.82, 95% confidence interval [CI] = 0.69-0.99). Patients with Medi- care, Medicaid, and no insurance were also less likely to have an ECG ordered compared to patients with private insurance (Medicare: OR = 0.79, CI = 0.63-0.99; Medicaid: OR = 0.67, CI = 0.53-0.84; no insurance: OR = 0.68, CI = 0.55-0.84). Those with Medicare and Medicaid were less likely to be triaged emergently (Medicare: OR = 0.84, CI = 0.71-0.99; Medicaid: OR = 0.76, CI = 0.64-0.91) and those with Medicare were less likely to have car- diac enzymes ordered (OR = 0.84, CI = 0.72-0.98).

Conclusions: Persistent racial and insurance disparities exist in the evaluation of chest pain in the ED. Compared to earlier studies, disparities in triage acuity and cardiac enzymes appear to have diminished, but disparities in ECG ordering have not. Given current Class I recommendations for ECGs on all patients presenting with chest pain emergently, our findings highlight the need for improvement in this area.

(C) 2019

  1. Introduction

Racial and socioeconomic disparities exist in the diagnosis, treat- ment, and outcomes of patients with acute myocardial infarction, in- cluding higher rates of mortality and readmission, and lower rates of percutaneous coronary intervention, coronary artery bypass surgery, Aspirin administration, and beta-blocker use among low income and African American patients [1-6].

* Corresponding author at: 550 1st Avenue, New York, NY 10016, United States.

E-mail address: [email protected] (A. Mukhopadhyay).

Disparities in the initial evaluation of chest pain in the emergency department (ED) may in part be responsible for these differences. Na- tionally representative survey data have shown differences in triage cat- egorization (i.e., the assigned urgency of the visit) and cardiac-specific testing for African American patients and for patients with public insur- ance [7-10]. However, these studies were conducted between 1995 and 2006 and did not adjust for clinical confounders, such as vital signs on presentation, Pain severity, or comorbidities that could reasonably in- fluence the likelihood of cardiac ischemia.

Given rising awareness and efforts to improve health disparities

[11], we hypothesized that previously described inequities may now be reduced. In this context, our study aimed to reevaluate the associa- tions of race and insurance status with the triage and evaluation of pa- tients presenting to the ED with chest pain nationwide.

https://doi.org/10.1016/j.ajem.2019.11.018

0735-6757/(C) 2019

1374 A. Mukhopadhyay et al. / American Journal of Emergency Medicine 38 (2020) 13731376

  1. Methods
    1. Data source

We analyzed ED data from the National Hospital Ambulatory Medi- cal Care Survey (NHAMCS-ED), a cross-sectional survey conducted an- nually by the Center for Disease Control and Prevention National Center for Health Statistics from 2009-2015. This survey includes a na- tional probability sample of ambulatory visits made to non-federal, gen- eral, and short-stay hospitals in the U.S., from which we analyzed visits to hospital emergency departments only. NHAMCS uses a multi-staged sample design for emergency departments, with 3 stages: (1) primary geographic sampling units, (2) hospitals within sampling units, and

(3) patient visits within emergency service areas of selected hospitals. A detailed description of the history, design, data collection, and publicly accessible data for NHAMCS-ED is available at the website of the Na- tional Center for Health Statistics [12].

Field representatives of the U.S. Census Bureau inducted hospitals into the NHAMCS survey. Hospital staff or Census Bureau officials uti- lized the medical record to obtain information on patient demographics, reasons for visit, vital signs, cause(s) of injury, diagnoses rendered, diag- nostic tests ordered, procedures provided, medications prescribed, pro- viders consulted, and disposition. The total unweighted response rate ranged from 54.6% to 87.5% for years 2009-2015. The NHAMCS is annu- ally approved by the Ethics Review Board of NCHS, and our analyses of publicly available data were considered non-human-subjects research by the Beth Israel Deaconess Committee on Clinical Investigations.

Sample

Reason for visit was classified by NHAMCS-ED using a standard clas- sification system [13]. In our analysis, we included visit data from partic- ipants who were over 18 years old who presented with any of the following primary reasons: chest pain; discomfort, pressure, or tight- ness in the chest; heart pain; and burning in the chest. We excluded par- ticipants who were transferred from another facility (given potential work-up at an outside facility), arrived deceased, or presented for in- jury, trauma, adverse event, overdose, or poisoning as work-up for car- diac ischemia would not necessarily be indicated in those scenarios.

Measures

  1. Primary outcomes

Our primary outcomes of interest were triage acuity and evaluation with an electrocardiogram (ECG) and cardiac enzymes. Triage acuity was treated as an ordinal variable based on the following NHAMCS-ED categories: emergent (should wait less than 15 min), urgent (can wait 15-60 min), semi-urgent (can wait 1- 2 h), and non-urgent (can wait 2-24 h).

Independent variables and other demographics

Based on NHAMCS-ED predefined categories, we examined race and ethnicity as Non-Hispanic black, Non-Hispanic white, Hispanic, and other and insurance status as private, Medicare, Medicaid, and unin- sured. We defined uninsured as those with payment type listed as either “No charge” or “Self-pay.” Additional demographics included age, gen- der, and geographic region. Unlike prior studies [7,9,10], we could not analyze hospital ownership or urban setting, as these data are no longer available for NHAMCS-ED.

Clinical factors

Among clinical factors, we included vital signs on presentation, pain severity, and co-morbidities. We dichotomized data for vital signs on presentation using prior validated ED rapid response criteria [14], defin- ing normal ranges as: heart rate 40-130 beats per minute, systolic blood pressure greater than 90 mm Hg, oxygen saturation greater than 90%,

and respirations 8-30 per minute. self-reported pain severity was re- corded by NHAMCS-ED on a 10-point scale, and this was included in our analysis as an ordinal variable. Additionally, we included diabetes and heart failure as comorbidities, as these were the only co- morbidities affecting cardiac risk that were available for all years of NHAMCS-ED data analyzed.

Statistical analysis

We used ordinal logistic regression to assess the effect of race and in- surance status on the odds of having higher triage acuity. We used stan- dard logistic regression models to assess the effect of race and insurance status on the odds of having an ECG or cardiac enzymes ordered. The fol- lowing covariates were used for all models: age, gender, race, insurance status, region, co-morbidities, and survey year. Given the substantial amount of missing data for vital signs and pain severity, separate models were created both with and without these covariates to test the sensitivity of our results to their inclusion. We analyzed all data using the sampled visit weight provided by NHAMCS, a product of the corresponding sampling fractions at each stage in the sampling design adjusted for survey nonresponse within time of year, geographic region, and urban/rural ownership designations. We used STATA 14.1 for all analyses (StataCorp LLC, College Station, Tx).

  1. Results
    1. Sample Characteristics

Fig. 1 depicts the patient flow for this study. A total of 10,441 patients met inclusion criteria, corresponding to an estimated 51.4 million pa- tients nationwide. As seen in Table 1, approximately a quarter of pa- tients were over 65, with a little over half being female. The majority of patients were white and had private insurance.

Fig. 1. Patient Flow Diagram.

A. Mukhopadhyay et al. / American Journal of Emergency Medicine 38 (2020) 13731376 1375

Table 1

Sample Characteristics.

Table 3 Likelihood of higher triage acuity and evaluation with EKG and cardiac enzymes by insur- ance status.

Variable Percent or Mean*

Age–Mean (s.d.) years 50.4 (18.0)

Age–Over 65 22%

Sex–Female

Race/Ethnicity

Non-hispanic white

55%

63%

Private Medicare

Medicaid

1.00

1.22 (1.06-1.41)**

0.67 (0.56-0.80)***

1.00

1.06 (0.90-1.26)

0.59 (0.48-0.73)***

1.00

1.23 (1.07-1.41)**

0.81 (0.67-0.98)*

Non-hispanic black

23%

No insurance

0.69 (0.57-0.85)**

0.61 (0.49-0.75)***

0.74 (0.61-0.91)**

Hispanic 12% Adjusted+ Odds Ratio (95% Confidence Interval)

Non-hispanic/other

3%

Private

1.00

1.00

1.00

Insurance status

Medicare

0.84 (0.71-0.99)*

0.79 (0.63-0.99)*

0.84 (0.72-0.98)*

Private insurance

33%

Medicaid

0.76 (0.64-0.91)**

0.67 (0.53-0.84)***

0.99 (0.81-1.20)

Medicaid or CHIP

18%

No insurance

0.88 (0.72-1.06)

0.68 (0.55-0.84)***

0.83 (0.67-1.02)

Triage Acuity ECG Cardiac Enzymes

Unadjusted Odds Ratio (95% Confidence Interval)

* Percentages were weighted to reflect national population estimates.

Medicare

28%

Worker’s compensation

9%

No insurance

13%

+ Adjusted for age, sex, region, payment type, comorbidities, and survey year. Esti- mates were similar after additional adjustment for Abnormal vital signs and pain severity, although sample size was smaller, confidence intervals were widened, and some statistical significance was lost (See Supplemental Table).

Effect of race

Table 2 summarizes the effects of race on triage acuity and whether patients were evaluated with ECG and cardiac enzymes. Before adjust- ment, black patients were significantly less likely to be triaged emergently, evaluated with an ECG, and evaluated with cardiac en- zymes than white patients. After adjustment for age, sex, region, pay- ment type, comorbidities and survey year, these differences were attenuated aside from a lower rate of evaluation with an ECG, which remained statistically significant. Additional adjustment for vital signs and pain severity did not substantially change our estimates, but be- cause the sample size with both additional measures was smaller, con- fidence intervals were necessarily wider (see Supplemental Table). For Hispanic patients, significant differences were seen in triage acuity and evaluation with cardiac enzymes before adjustment, but these were attenuated and no longer statistically significant after adjustment.

Effect of insurance status

Table 3 summarizes the effects of insurance on the three outcomes. Prior to adjustment, compared to patients with private insurance, those with Medicaid and no insurance were less likely to be triaged emergently or to be evaluated with ECG and cardiac enzymes. In con- trast, Medicare patients (who were older) were significantly more likely to be triaged emergently and to have cardiac enzymes. After adjustment for age, sex, region, race, comorbidities, and survey year, patients with Medicaid remained significantly less likely to have an ECG or cardiac en- zymes; and patients with no insurance remained significantly less likely to have an ECG. Patients with Medicare were significantly less likely to be triaged emergently or to have an ECG or cardiac enzymes after ad- justment, reversing the previous direction. Estimates were similar

Table 2

Likelihood of higher triage acuity and evaluation with EKG and cardiac enzymes by race before and after adjustment.

Triage Acuity ECG Cardiac Enzymes

Unadjusted Odds Ratio (95% Cl)

Non-Hispanic White 1.00 1.00 1.00

Non-Hispanic Black 0.71 (0.59-0.85)** 0.72 (0.60-0.88)* 0.80 (0.66-0.96)*

Hispanic 0.78 (0.63-0.97)* 0.87 (0.69-1.09) 0.71 (0.57-0.89)**

Adjusted+ Odds Ratio (95% CI)

Non-Hispanic White 1.00 1.00 1.00

Non-Hispanic Black 0.84 (0.69-1.02) 0.82 (0.69-0.99)* 0.92 (0.76-1.11)

Hispanic 0.94 (0.78-1.14) 1.16 (0.91-1.48) 0.86 (0.70-1.07)

+ Adjusted for age, sex, region, payment type, comorbidities, and survey year. Esti- mates were similar after additional adjustment for abnormal vital signs and pain severity, although because the sample size with both additional measures was smaller, confidence intervals were necessarily wider (See Supplemental Table).

after adjustment for vital signs and pain severity in the subset of pa- tients with this information, although confidence intervals were again wider, and some statistical significance was lost (See Supplemental Table).

  1. Discussion

In this nationally Representative sample of patients presenting to the ED with chest pain between 2009 and 2015, black patients and patients with public or no insurance were significantly less likely to be evaluated with an ECG even after adjusting for potential confounders. Addition- ally, patients with Medicare and Medicaid were less likely to be triaged emergently and those with Medicare were less likely to have cardiac en- zymes ordered.

Our study adds to prior literature [7-10] by utilizing more recent, nationally representative ED data (2009-2015), and by accounting for additional patient factors that could confound the triage and manage- ment of chest pain, such as vital signs, pain severity, and comorbidities. Given increased awareness, research, and training in health disparities, we hypothesized that previously described inequities in the emergent evaluation of chest pain [7-10] would now be undetectable. Unfortu- nately, this is not what we found. Compared to prior work using data from 1997 to 2006 [9], our study found a similar degree of racial and in- surance disparities in the evaluation with ECG. Encouragingly, differ- ences in triage acuity and the ordering of cardiac enzymes were reduced, suggesting partial but incomplete improvement.

Currently, Class I recommendations advise that all patients with

chest pain have an ECG performed within 10 min of arrival to an emer- gency facility [15]. The reasons for the deviation from this standard ob- served in our study, specifically for black patients and patients with Public insurance, remain unclear. One hypothesis could be related to the well-established racial and insurance disparities in access to pri- mary care, and the increased use of the ED as a primary care substitute [16]. This could result in a higher proportion of ED visits for non-cardiac chest pain in this population. However, per current guidelines, all pa- tients with chest pain should be evaluated with ECG [15]. Another hy- pothesis is that EDs serving black and Lower socioeconomic status patients could have fewer ECG machines and technicians, or may not have formal chest pain protocols to facilitate rapid and systematic or- dering of ECGs. One would expect provider-level bias to play more of a role in triage acuity and the ordering of cardiac enzymes than the or- dering of ECGs, as these decision points require direct evaluation by a provider (triage nurse and ED physician, respectively). Therefore, our data suggest that improvements may have been made on the provider-level, but that systemic inequalities persist.

Our results should be interpreted in the context of several limita- tions. Our study was a cross-sectional, retrospective analysis, and was therefore limited in its ability to draw causal conclusions. Additionally,

1376 A. Mukhopadhyay et al. / American Journal of Emergency Medicine 38 (2020) 13731376

NHAMCS data did not include data on chest pain descriptors and associ- ated symptoms, additional clinical co-morbidities, hospital ownership, and urbanicity. Finally, our data could not be linked to adverse clinical outcomes. Strengths of this study include its large sample size and a na- tionally representative sample, allowing for maximal generalizability.

In summary, racial and insurance disparities persist in the evaluation of chest pain in the ED, specifically in the ordering of ECGs. Given cur- rent Class I recommendations for ECGs on all patients presenting with chest pain emergently, our findings highlight the need for targeted systems-level quality improvement efforts to reduce this disparity. Fur- ther research should be directed towards granular, hospital-level com- parisons to determine appropriate interventions, such as assessing the presence of standard chest pain protocols and their effect on disparate ECG ordering. Additionally, future work should attempt to quantify the effects of these disparities on patient outcomes.

Author contributions

All authors were involved in study design and conception. Amrita Mukhopadhyay, Robert D’Angelo, Ethan Senser, Christina C. Wee, and Kenneth J. Mukamal were involved in statistical planning and data anal- ysis. Amrita Mukhopadhyay drafted the manuscript and all authors con- tributed substantially to its revision.

Declaration of Competing Interest

The authors declared that there is no Conflict of Interest.

Appendix A. Supplementary material

Supplementary data to this article can be found online at https://doi. org/10.1016/j.ajem.2019.11.018.

References

  1. Popescu I, Vaughan-Sarrazin MS, Rosenthal GE. Differences in mortality and use of revascularization in black and white patients with acute MI admitted to hospitals with and without revascularization services. JAMA 2007;297(22):2489-95.
  2. Barnato AE, Lucas FL, Staiger D, Wennberg DE, Chandra A. Hospital-level racial dis- parities in acute myocardial infarction treatment and outcomes. Med Care 2005;43 (4):308.
  3. Ferdinand KC. Cardiovascular risk reduction in African Americans: Current concepts and controversies. Glob Cardiol Sci Pr [Internet]. 2016 Aug 21 [cited 2019 Jan 14]; 2016 (1). Available from: https://globalcardiologyscienceandpractice.com/index. php/gcsp/article/view/3.
  4. Bernheim SM, Spertus JA, Reid KJ, Bradley EH, Desai RA, Peterson ED, et al. Socioeco- nomic disparities in outcomes after acute myocardial infarction. Am Heart J 2007 Feb;153(2):313-9.
  5. Vaccarino V, Rathore SS, Wenger NK, Frederick PD, Abramson JL, Barron HV, et al. Sex and racial differences in the management of acute myocardial infarction, 1994 through 2002. N Engl J Med 2005;353(7):671-82.
  6. Wong MD, Shapiro MF, Boscardin WJ, Ettner SL. Contribution of major diseases to disparities in mortality. N Engl J Med 2002;347(20):1585-92.
  7. Pezzin LE, Keyl PM, Green GB. Disparities in the Emergency Department Evaluation of Chest Pain Patients. Acad Emerg Med 2007 Feb;14(2):149-56.
  8. Arnold AL, Milner KA, Vaccarino V. Sex and race differences in electrocardiogram use (the National Hospital Ambulatory Medical Care Survey). Am J Cardiol 2001;88(9): 1037-40.
  9. Lopez L, Wilper AP, Cervantes MC, Betancourt JR, Green AR. Racial and sex differ- ences in emergency department triage assessment and test ordering for chest pain, 1997-2006: racial and sex differences in ed triage for chest pain. Acad Emerg Med 2010;17(8):801-8.
  10. Wilper AP, Woolhandler S, Lasser KE, McCormick D, Cutrona SL, Bor DH, et al. Waits to see an emergency department physician: US trends and predictors, 1997-2004. Health Aff (Millwood) 2008;27(2):w84-95.
  11. National Committee for Quality Assurance. National Scan of CLAS and Disparities Re- duction Activities. National Committee for Quality Assurance; 2018 Jun.
  12. National Center for Health Statistics. NHAMCS Scope and Sample Design [Internet]. 2016 [cited 2018 Jan 15]. Available from: http://www.cdc.gov/nchs/ahcd/ahcd_ scope.htm.
  13. Schneider D, Appelton L, McLemore T. A Reasonfor Visit Classification for AmbulatoryCare. Natl Cent Ofr Heal Stat Vital Heal Stat Ser 2 Data Eval Methods Res. 78.
  14. McGillicuddy DC, O’Connell FJ, Shapiro NI, Calder SA, Mottley LJ, Roberts JC, et al. Emergency department abnormal vital sign “Triggers” program improves time to therapy: emergency medicine triggers. Acad Emerg Med 2011;18(5):483-7.
  15. Amsterdam EA, Wenger NK, Brindis RG, Casey DE, Ganiats TG, Holmes DR, et al. 2014 AHA/ACC Guideline for the Management of Patients With non-ST-elevation acute coronary syndromes: A Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation [Internet]. 2014 Dec [cited 2019 Jan 28];130(25). Available from: https://www.ahajournals.org/doi/ 10.1161/CIR.0000000000000134.
  16. Pamela Jo Johnson, Neha Ghildayal, Andrew C. Ward, Bjorn C. Westgard, Lori L. Boland, Jon S. Hokanson. Disparities in potentially avoidable emergency department (ED) care: ED visits for ambulatory care sensitive conditions. Med Care. 2012; 50 (12):1020-8.

Leave a Reply

Your email address will not be published. Required fields are marked *