Article Text
PDF
Abstract
Introduction National practice guidelines recommend early aspirin administration to reduce mortality in acute coronary syndrome (ACS). Although timely administration of aspirin has been shown to reduce mortality in ACS by 23%, prior regional Emergency Medical Service (EMS) data have shown inadequate prehospital administration of aspirin in patients with suspected cardiac ischaemia.
Objectives Using the National EMS Information System (NEMSIS) database, we sought to determine (1) the proportion of patients with suspected cardiac ischaemia who received aspirin and (2) patient and prehospital characteristics that independently predicted administration of aspirin.
Methods Analysis of the 2011 NEMSIS database targeted patients aged ≥40 years with a paramedic primary impression of ‘chest pain’. To identify patients with chest pain of suspected cardiac aetiology, we included those for whom an ECG or cardiac monitoring had been performed. Trauma-related chest pain and basic life support transports were excluded. The primary outcome was presence of aspirin administration. Patient (age, sex, race/ethnicity and insurance status) and regional characteristics where the EMS transport occurred were also obtained. Multivariate logistic regression was used to assess the independent association of patient and regional factors with aspirin administration for suspected cardiac ischaemia.
Results Of the total 14 371 941 EMS incidents in the 2011 database, 198 231 patients met our inclusion criteria (1.3%). Of those, 45.4% received aspirin from the EMS provider. When compared with non-Hispanic white patients, several groups had greater odds of aspirin administration by EMS: non-Hispanic black patients (OR 1.49, 95% CI 1.44 to 1.55), non-Hispanic Asians (OR 1.62, 95% CI 1.21 to 2.18), Hispanics (OR 1.71, 95% CI 1.54 to 1.91) and other non-Hispanics (OR 1.27, 95% CI 1.07 to 1.51). Patients living in the Southern region of the USA (OR 0.59, 95% CI 0.57 to 0.62) and patients with governmental (federally administered such as Veteran's Health Care, but not Medicare or Medicaid) insurance (OR 0.67, 95% CI 0.57 to 0.78) had the lowest odds of receiving aspirin. Age and sex (OR 1.00, 95% CI 1.00 to 1.00) were not associated with aspirin administration.
Conclusions It is likely that prehospital aspirin administration for patients with suspected cardiac ischaemia remains low nationally and could be improved. Reasons for disparities among the various groups should be explored.
- ECG
- cardiac care, acute coronary syndrome
- emergency ambulance systems
- paramedics
- prehospital care, clinical management
Statistics from Altmetric.com
- ECG
- cardiac care, acute coronary syndrome
- emergency ambulance systems
- paramedics
- prehospital care, clinical management
Key messages
What is already known on this subject
National practice guidelines recommend early aspirin administration to reduce mortality in acute coronary syndrome (ACS). Although timely administration of aspirin has been shown to reduce mortality in ACS by 23%, prior regional Emergency Medical System (EMS) data have shown inadequate prehospital administration of aspirin in patients with suspected cardiac ischaemia. Using the National EMS Information System (NEMSIS) database, we sought to determine (1) the proportion of patients with suspected cardiac ischaemia who received aspirin and (2) patient and prehospital characteristics that independently predicted administration of aspirin.
What this study adds
In a large EMS database review, prehospital aspirin administration for patients with suspected cardiac ischaemia remains low nationally and could be improved.
Introduction
Early aspirin administration is widely recommended for patients with acute coronary syndrome (ACS) to reduce adverse events and subsequent mortality.1–3 Aspirin reduces mortality by 23%1 by mechanisms that make its delivery time sensitive.4 The mortality benefit of aspirin administered early in the care of a patient with myocardial infarction makes the prehospital administration of aspirin standard of care in most Emergency Medical Service (EMS) systems.4–7
In addition to prehospital standards, other expert organisations and regulatory agencies have produced guidelines that include aspirin as part of the standard of care treatment of acute coronary syndromes. The 2012 practice guidelines by the American Heart Association (AHA)/American College of Cardiology Foundation advocate aspirin with a class I level of evidence8 (highest recommendation based on evidence and agreement on effectiveness). The Center for Medicare and Medicaid Services (CMS) and the Joint Commission on Accreditation of Health Care Organizations (JCAHO) use aspirin administration in acute myocardial infarction as a hospital-based quality indicator of care.9 ,10 Hospital CMS reimbursement is dependent on adherence to quality measures, although these guidelines do not currently extend to the prehospital arena.
Prior studies have revealed inadequate prehospital aspirin administration in cases of suspected cardiac ischaemia. Regional studies in the USA show aspirin administration ranging from 27% to 60% of patients with suspected cardiac ischaemia.11 ,12 Although the prevalence of prehospital aspirin administration in the cardiac ischaemia population has been studied, patient level and prehospital level factors associated with prehospital administration of aspirin have not been studied nationally. Our objectives were to determine the proportion of patients with suspected cardiac ischaemia who received aspirin and to identify patient characteristics that independently predict the administration of aspirin using the National EMS Information System (NEMSIS) database.
Methods
Study design
This cross-sectional descriptive study uses the 2011 Public-Release Data Set from the NEMSIS project. The study received an exemption from the University of California, San Francisco Institutional Review Board.
Data source
The NEMSIS database13 was established by the Office of Emergency Medical Services under the National Highway Traffic Safety Administration, after the 1996 EMS Agenda for the Future document called for a comprehensive EMS information system. The goal of NEMSIS was to create a nationwide database to collect EMS-related data from participating EMS agencies through state-level data submission. The NEMSIS project defined a set of common data elements to compile the ‘national dataset’ for EMS agencies (http://www.nemsis.org/index.html). The database is maintained by the NEMSIS Technical Assistance Center (University of Utah, School of Medicine, Salt Lake City, Utah, USA). To submit data to NEMSIS, states work with local EMS agencies to aggregate each component of the ‘national dataset’ for every EMS run. The most current version (V.3) contains over 578 elements in the data dictionary. NEMSIS reviews all state submitted data files and contacts individual provider agencies for any missing information. At the time of this study, the 2011 dataset was the most recent and complete version available. The 2011 dataset contains contributions from 35 US states and territories and a total of 14 371 941 EMS encounters (figure 1). All EMS incident data submitted by agencies are included in the dataset; there is no sampling.
National EMS Information System (NEMSIS) 2011 database. States submitting to NEMSIS are highlighted in grey (blue in the online version) and the NASEMSO regions are outlined.
Data collection
The target population for the study was defined as EMS events entered into the NEMSIS 2011 dataset representing adult patients with suspected cardiac ischaemia. The dataset includes patients entered from 1 January 2011 through 31 December 2011. To identify this group from the dataset, we selected patients with a primary complaint of chest pain. We defined ‘adult’ as a patient aged ≥40 years. We further narrowed our study group to patients with an EKG or cardiac monitor completed, as a proxy for suspected cardiac aetiology of chest pain. We excluded patients with trauma-related chest pain as well as basic life support (BLS) transports, as aspirin may not be available to all providers at that level of service. The primary outcome variable was aspirin administration for patients transported by EMS with suspected cardiac ischaemia. A meticulous evaluation of the study population data was performed not to miss any aspirin administration that was coded in a different format. For example, ‘ASA’, ‘Aspirin’, and ‘aspirin 81 mg’ were each listed separately and required a free text search for combination into the final analysis. The covariates used in our model include age, sex, race/ethnicity, insurance status and region served by the EMS agency. Race/ethnicity variables were based on predetermined NEMSIS groupings including non-Hispanic white patients, non-Hispanic black patients, Hispanics, Asians and Pacific Islanders. Insurance status was defined as private provider, Medicare, Medicaid, governmental (federally administered such as Veteran's Health Care, but non-Medicare and non-Medicaid), not billed, workman's compensation or self-pay. Regions were defined based on State EMS Officials (NASEMSO) classification and include North Central, East, South Central and West.14 This is illustrated in figure 1. Data were cleaned using SAS (V.9.3, Cary, North Carolina, USA) and analysed using Stata (V.12.1, College Station, Texas, USA).
Results
The algorithm for the final study population selection is presented in figure 2. The total number of patients in the database during the study period was 14 371 941. Of those, the number of patients with a ‘Provider Primary Impression’ coded as ‘chest pain’ was 651 615 observations (4.5%). Among those, patients aged ≥40 years accounted for 552 722 (3.8%) observations. We excluded 157 802 patients with trauma and BLS transports, which resulted in 394 920 patients (2.7%). Of these patients, 198 231 (1.4%) had an EKG or cardiac monitoring performed and this is the total number of transports available for analysis. This subgroup consisted of 30% of the total patients presenting with chest pain. This population was further stratified by aspirin use. Among the total number of transports for suspected cardiac ischaemia, 90 005 (45.4%) received aspirin, 108 226 (54.6%) patients did not receive aspirin.
Flow chart showing patient selection for the analysis. ALS, advanced life support; BLS, basic life support; EMS, Emergency Medical System; NEMSIS, National EMS Information System.
Demographics of patients who received aspirin and those who did not are presented in table 1.
Patient and regional characteristics of EMS-transported patients with suspected cardiac ischaemia from the National EMS Information System database 2011 subgroups (n=198 231)
Among patients who received aspirin, 45% (n=45 132) were men. There was no association between aspirin administration and age categorisation. In patients who received aspirin with a recorded race/ethnicity categorisation, non-Hispanic black patients (51%), Hispanics (56%) and non-Hispanic Asians (57%) had a higher proportion of patients receiving aspirin, whereas non-Hispanic white patients had a lower proportion receiving aspirin (43%). A third of total cases did not have race/ethnicity noted in the prehospital record. We observed that Medicare recipients (57%) or those on other forms of government-issued (federally administered such as Veteran's Health Care) insurance (67%) had a higher proportion of patients who did not receive aspirin by prehospital providers. There is an additional subset of patients without insurance status recorded in the prehospital record. Patients transported in the North Central (59%), South Central (63%) and West (52%) regions had a higher proportion of not receiving aspirin by prehospital providers than did those in the East (49%).
To assess the independent association between patient and community characteristics with administration of aspirin, we created a logistic regression model using patient variables from the database and the outcome of aspirin administration (table 2).
Independent predictors of aspirin administration by EMS providers for suspected cardiac ischaemia from National EMS Information System database 2011 (N=60 986)
We did not observe any association between age (OR 0.99, 95% CI 0.99 to 0.99), gender (OR 1.00, 95% CI 1.00 to 1.00) and aspirin administration. We found that compared with non-Hispanic white patients, the following groups had higher odds of receiving aspirin by prehospital providers: non-Hispanic black patients (OR 1.49, 95% CI 1.44 to 1.56), non-Hispanic Asians (OR 1.62, 95% CI 01.21 to 2.18), Hispanics (OR 1.71 95% CI 1.54 to 1.91) and non-Hispanic others (OR 1.27, 95% CI 1.07 to 1.51). In comparison with patients with private insurance, patients with government insurance (OR 0.67, 95% CI 0.57 to 0.78) had a lower odds of receiving aspirin and patients with Medicaid (OR 1.30, 95% CI 1.23 to 1.38) had a higher odds of receiving aspirin. Compared with the East NASEMSO region, patients in the North Central (OR 0.70, 95% CI 0.67 to 0.73), South (OR 0.59, 95% CI 0.57 to 0.62) and West (OR 0.76, 95% CI 0.68 to 0.86) regions had lower odds of receiving aspirin.
Discussion
This analysis of the very large NEMSIS dataset confirms that adult patients with suspected cardiac ischaemia were documented to have received aspirin only about half of the time. This low proportion is consistent with findings of prior published regional studies.11 ,12 There could be several reasons for lack of aspirin administration by prehospital providers.15 There may be a subset of patients directed to take aspirin by emergency medical dispatchers, given aspirin by a first responder unit, or patients may be taking aspirin as part of primary or secondary prevention efforts and self-administering aspirin prior to the arrival of EMS. While these estimates cannot be derived from our database, previous studies have shown that aspirin administration prior to EMS arrival in their region varies from 8% to 26%.12 ,15 However, the database does not record the reasons for lack of aspirin administration and hence the true proportion of aspirin use in the population could be underestimated in our study. Other possible reasons for lower prevalence of aspirin administration could include absolute contraindications to aspirin such as a history of an anaphylactic response to aspirin or relative contraindications such as gastrointestinal bleed or kidney disease. However, prehospital use of aspirin is safe and rarely associated with adverse events.16 True aspirin intolerance or ‘allergy’ is estimated at only 2% to 3% of the general population.17 Lack of documentation of aspirin administration could also be another reason for low prevalence of aspirin use in the prehospital setting. While we could argue that the stated prevalence is likely to be lower than actual prehospital aspirin use in communities for acute coronary syndromes, a major concern to the prehospital and cardiology communities would be failure of the provider to administer aspirin to a patient with suspected cardiac ischaemia due to lack of recognition of symptoms or if adequate prevention efforts are not in place leading to disparities in care.
Subgroup analysis of patient characteristics associated with aspirin administration as indicated for suspected cardiac ischaemia revealed a higher odds of aspirin administration in non-Hispanic black, non-Hispanic Asian and Hispanic patients when compared with non-Hispanic white patients. While higher aspirin use in this population seems to be contrary to the predicted results, we believe that some of these differences are due to lack of prevention efforts in these populations. This may represent populations with lower availability of primary care services and may not have aspirin available to take when prompted by the dispatcher or as part of their daily medications. Minority patients could be perceived as healthcare disadvantaged and in compensation are more regularly offered aspirin. Previous publications show that aspirin use for secondary prevention of heart disease is less in African–Americans and Asians compared with white patients.18 In addition, living in an area with lower neighbourhood household income and Medicaid insurance has been shown to be associated with longer prehospital delay times (symptom onset to hospital arrival), highlighting the importance of effective ACS treatment.19 National studies on prevention and treatment for ACS show disparities in care,20 which support our findings and interpretations of the differences in aspirin use among different groups. Other research has shown ethnic and racial disparities in treating time-sensitive conditions.21 Since we did not have zip code availability from this public database, it is difficult to focus on specific regional variability. Future steps could explore zip code differences and neighbourhood effects using geographic information systems (GIS) mapping to better understand community factors that play a role in the use of aspirin and other prevention treatments.22 While regional differences in EMS use has been shown for stroke and other time-sensitive conditions,23 disparities in secondary prevention has not been shown nationally.
Another important finding revealed lower odds of receiving aspirin for patients in the South region of the USA and those with government-issued insurance. Both residents of Florida (large NEMSIS contributor in the South region) and Veteran's healthcare patients are typically older and may have aspirin as part of a daily medication plan or have a contraindication to receiving it. There could also be increased awareness in this patient group regarding rapid administration of aspirin in the event of possible cardiac ischaemia, a ready availability of aspirin in the medicine cabinets or purses of friends or associates or an actual difference in EMS administration patterns.
Few treatments in medicine have been shown to be as effective as aspirin in decreasing mortality from acute coronary syndrome. Therefore, it is important that aspirin be administered quickly by EMS if the patient has not already taken it. Educational interventions targeting paramedics have been shown to improve aspirin administration to patients with suspected cardiac ischaemia.24 The proven mortality benefit of aspirin in ACS as part of standardised EMS protocols for chest pain should make aspirin administration and documentation in ACS a prehospital quality core measure. One way to evaluate compliance with this intervention in the prehospital setting is to track aspirin administration as a quality measure among EMS systems as is being done in California as part of its EMS Core Measure System.25
Currently, only hospital-based measures are used to assess the quality of care provided to patients with suspected myocardial infarction using evidence-based criteria from the JCAHO and the Centers for Medicare and Medicaid Services (CMS),9 ,10 but prehospital care may soon be recognised as an important component of the patient care spectrum. Prior study has measured the quality indexes and composite metrics (bundle score) in the prehospital care of patients with chest pain and found that aspirin was given in 62% of patient encounters and the composite score was met for only 39% of patients.7 These scores show that care could be improved and one approach would be to extend the metrics to prehospital care systems. In this changing scene of performance-based healthcare and bundle payments, EMS may face different reimbursement guidelines in the future and it is the responsibility of the EMS providers to consistently provide aspirin as an intervention in patients with suspected cardiac ischaemia.
Limitations
We recognise several limitations of our study. The NEMSIS database is a convenience sample of contributing EMS agencies and is not population based. Therefore, it may not be a representative sample of patients receiving prehospital care in the USA, but NEMSIS is the only multiregional prehospital database and provides an opportunity to look at national prehospital practices. Additionally, there were missing data for the race/ethnicity and insurance status variables in the regression model. Although the NEMSIS Technical Assistance Center reviews every data field to reduce missing data, if the individual EMS provider did not document data in the field provided, the variable is coded as ‘not reported’ or ‘not applicable.’ While statistical analysis such as multiple imputation could be used for a large number of missing data, we were limited by the availability of data variables in the dataset, also with the large sample size, regaining the missing data would not have made much of a quantitative difference. A third limitation is the lack of documentation of the reason that aspirin was not administered. For example, if aspirin was taken by the patient before EMS arrival or prescribed by a sending physician, it might not have been documented as such in the prehospital record. Although this was not the objective of the study, knowing if the disparities in aspirin administration were due to lack of adequate primary prevention efforts in particular racial/ethnic groups versus a clinical reason could contribute to targeted intervention efforts in this population.
Conclusions
In conclusion, prehospital aspirin administration for patients with suspected cardiac ischaemia is low nationally. Whether this is due to patients self-administering before EMS arrival, failure of paramedics to follow local protocols, protocols not in compliance with AHA recommendations or increasing use of aspirin as part of the public health measures must be better understood and is a question for future studies. Targeting interventions based on results from follow-up studies would help to improve the prehospital care of patients with ACS nationally.
Acknowledgments
We thank the Department of Statistics and Epidemiology at the University of California Berkeley for their assistance with the analysis.
References
Footnotes
This research was presented as an electronic poster at the ACEP Annual Conference 2013 in Seattle, Washington. It was also presented as a poster at the National Association of Emergency Medical Services Physicians (NAEMSP) Annual Conference 2014 in Tucson, Arizona, where it was awarded the Best Poster Abstract Presentation and the Best Student/Resident/Fellow Research Presentation.
Contributors KLT, MPM and PG planned the study. KLT and PG completed the data analysis. KLT, PG and MPM wrote the manuscript.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.