Variation in emergency medical service use for acute coronary syndromes by ethnicity: an Aotearoa New Zealand observational study

Māori (the Indigenous people of Aotearoa New Zealand), Pacific peoples (people predominantly from South Pacific Islands) and South Asian peoples living in Aotearoa New Zealand are disproportionately affected by cardiovascular disease (CVD).¹ Māori, Pacific peoples and Indians (who are the only South Asian group that can currently be identified in routinely collected Aotearoa New Zealand health data) experience the onset of CVD at a younger age,^2,3 and have a significantly higher prevalence of diabetes than Europeans.⁴ Compared with their European counterparts, Māori and Pacific peoples also have a much higher prevalence of other risk factors for CVD such as smoking and atrial fibrillation,^5–7 and are more likely to die from CVD.^4,8 Of all ethnic groups in Aotearoa New Zealand, Indians have the highest prevalence of coronary heart disease. Yet, compared with Europeans, Māori and Pacific peoples are the most likely to die from coronary heart disease before reaching hospital.⁹ Indeed, coronary heart disease is the most important cause of potentially avoidable deaths for Māori and Pacific peoples, contributing up to 1 year of the approximately 6–7 year reduction in life expectancy for Māori and Pacific peoples compared with other New Zealanders.¹⁰

Acute coronary syndrome (ACS), comprised of ST-elevation myocardial infarction (STEMI), non-STEMI (NSTEMI) and unstable angina (UA), is a major contributor to the burden of CVD. Highly effective treatments reduce the morbidity and mortality of ACS, but the effectiveness of reperfusion (required for STEMI) is critically time-dependent.^11,12 Prehospital triage and investigations like electrocardiograph by emergency medical services (EMS) to promptly identify and manage such patients has led to significant advances in their management, and is a critical factor in reducing times to definitive reperfusion treatment for patients with STEMI.¹³ A study from 2012 found that among patients admitted to the coronary care unit of an Aotearoa New Zealand hospital with ACS, the delay between symptom onset and defibrillator availability was 7 hours longer for patients arriving by self-transport compared with EMS.¹⁴ That study also found Māori, Pacific and Indian peoples, as well as those from areas of higher deprivation, were less likely to arrive by EMS.¹⁴ Another study found that communities with the highest levels of socio-economic deprviation had the least availablity of public access defibrillatiors, despite also having the highest incidence of out-of-hospital cardiac arrest.¹⁵ In addition to determining early interventions, ACS type is also associated with prognosis.¹⁶

The Aotearoa New Zealand health system is required under the Pae Ora (Healthy Futures) legislation to ensure equitable access to health services.¹⁷The aim of this study was to use contemporary national data to determine if there are ethnic variations in EMS use by ACS type.

Methods

Study design and inclusion and exclusion criteria

A retrospective cross-sectional study of all adult patients (aged 18+ years) hospitalised for an ACS event in Aotearoa New Zealand was undertaken between 1 July 2019 and 30 June 2021. Patients were identified from the national hospitalisation data collection (National Minimum Dataset [NMDS]). This 2-year period was chosen because it was the earliest date where there were reliable electronic EMS data, and, at the time of the analysis, the latest date that hospitalisation data were available. An ACS event was defined as a primary diagnosis (at discharge) of International Classification of Diseases (ICD) 10 codes for STEMI (I21.0–I123, I220, I221, I228, I229), NSTEMI (I214, I222) or UA (I200). For patients with multiple ACS events during the study, only the first was included in the analysis. Patients were excluded if they had duplicate EMS records for the same date.

Variables

EMS use for each included ACS event was determined by linkage (using an encrypted National Health Index [NHI], a unique patient identifier) to the national EMS registry, the Aotearoa New Zealand Paramedic Care Collection (ANZPaCC). The ANZPaCC contains all routinely collected electronic clinical record data for all patients in Aotearoa New Zealand attended by EMS. We determined for each patient if their ACS event could be matched with an EMS encounter on the day of or the day prior to admission. If a match was made, they were classified as having been transported by EMS. Otherwise, they were classified as having been transported by other means.

Age, sex, ethnicity, socio-economic deprivation and rurality were obtained from the NHI data collection. Ethnicity was prioritised using a slight modification of the Aotearoa New Zealand Health and Disability sector standard¹⁸ in the following order: Māori > Pacific > Indian (including Fijian Indian) > non-Indian Asians (including Chinese) > Middle Eastern, Latin American and African (MELAA) > European > Other. MELAA and Other ethnic groups were combined given small numbers. Indians comprise approximately 90% of the total South Asian group but due to limitations in the ethnicity categorisations currently available in routinely collected Aotearoa New Zealand health data, non-Indian South Asians cannot be disaggregated from the rest of the non-Indian Asian group. Socio-economic deprivation and rurality measures were based on patients’ domicile code, which reflected their residential address. Socio-economic deprivation decile or quintile was classified according to the New Zealand Index of Deprivation (NZDep) 2013, based on a range of 2013 Census data from people living in the same small area (of at least 100 people).¹⁹ The Geographic Classification for Health (GCH) was used to classify each patient into one of five levels of (reducing) rurality (two urban categories [U1, U2] and three rural categories [R1, R2, R3]).²⁰

Statistical analysis

Patients were described using categorical variables (summarised as counts and percentages) and continuous variables (summarised as means with standard deviations [SD] and medians with interquartile ranges [IQR]) according to ACS type (STEMI, NSTEMI, UA) and within each ACS type by ethnicity. Differences in continuous variables (using the Kruskal–Wallis rank sum test) and categorical variables (using the Chi-squared test, or Fisher’s exact test where any cells had a value <5) were assessed between ACS types and, within each ACS type, by ethnicity (using Episheet). The association between variables (ethnicity as well as age, sex, NZDep and rurality, either individually for univariable or in combination [ethnicity, age, sex, NZDep, rurality] for multivariable analyses) and arrival via EMS was assessed using logistic regression separately for each ACS type (using R Studio). Data are presented as (adjusted) odds ratios (OR or aOR) with 95% confidence intervals (CI). A p-value of <0.05 was considered statistically significant. Analyses were undertaken using Episheet or R Studio.

Ethics

This study was conducted under a long-term CVD research programme, which undergoes annual ethics approval, most recently approved by the Northern Region B Health and Disability Ethics Committee on 21 October 2022 (2022 EXP 13442). This approval includes a waiver for individual participant consent. Ethical approval was also obtained from the Auckland University of Technology Ethics Committee (21/369) as the study was undertaken as part of the master’s thesis of Haydn Drake.

Results

Between 1 July 2019 and 30 June 2021, 19,283 patients were admitted with an ACS event and included in this analysis (STEMI n=4,827, 25%; NSTEMI n=10,627, 55%; UA n=3,829, 19.9%) (Table 1). Their median age was 69 years (IQR 60–78 years) and nearly two-thirds were male (65.2%, n=12,567). Most were European (75%), followed by Māori (11%), Pacific (5%), Indian (4%), non-Indian Asians (3%) and Other (1%). Most lived in urban areas (U1 51%, U2 23%), and the proportion living in areas with the greatest deprivation quintile was 23%. While there were statistically significant differences in each of these characteristics by ACS type, most differences were relatively minor, apart from those by age (median age STEMI 66 vs NSTEMI 71 and UA 70 years) and sex (women STEMI 31% vs 37% NSTEMI 37% and UA 36%). Each ACS type is also described by ethnicity in Appendix Table 1–3.

There were substantial (as well as statistically significant) differences in EMS use by ACS type, ranging from 75% for STEMI to 61% for NSTEMI and 45% for UA. The association between patient characteristics and EMS use is described for each ACS type in Table 2, both by proportions arriving by EMS (with numerators and denominators provided in Appendix Table 4–6) as well as by logistic regression.

For STEMI, EMS use ranged from 63.7% for non-Indian Asians to 77.9% for Europeans. All ethnic groups assessed were less likely to arrive by EMS for STEMI than Europeans, both for unadjusted as well as for adjusted analyses, although the effect was not statistically significant for Other patients (Figure 1). A similar pattern was evident for NSTEMI, although the reduced odds of EMS use compared with Europeans was not statistically significant in adjusted analyses for Indian or Other patients. For UA, while there were reduced odds of EMS use for all ethnic groups compared with Europeans, the magnitude of the reduction was attenuated, and odds were only statistically significant in adjusted analyses for non-Indian Asian patients.

EMS use also varied according to the other characteristics assessed. Men were consistently less likely than women to arrive via EMS across all ACS types, for unadjusted as well as adjusted analysis. Similarly, EMS was strongly associated with increasing age. The observed association between EMS use and deprivation was inconsistent, with some indication of lower access with increasing deprivation for STEMI and NSTEMI, but the inverse evident for UA. There was some indication of reduced EMS access with increasing rurality, although the pattern was mixed. EMS access was lower for U2 (second most urban) compared with U1 for all ACS types in adjusted analyses. EMS access was similarly lower for R2 and R3 (most rural) compared with U1, although the effect wasn’t consistently statistically significant for STEMI and NSTEMI, and no statistically significant effects were observed for UA. In adjusted analyses, EMS access for R1 (mid-level on the five-level rurality scale) was slightly reduced (but not statistically different) to that for U1 for STEMI, and increased (though again not statistically significantly) to that for U1 for NSTEMI and UA.

Discussion

Between 1 July 2019 and 30 June 2021, 19,283 ACS events were identified (STEMI 25%, NSTEMI 55%, UA 20%). Consistent with the findings of an earlier national study,¹⁶ STEMI patients were younger and less likely to be women than NSTEMI or UA patients (median age 66 vs 71 and 70 years, respectively; women 31% vs 37% and 36%, respectively). EMS use ranged from 45% for UA to 61% for NSTEMI and 75% for STEMI. EMS use for STEMI was similar to that found in an earlier study from Aotearoa New Zealand (73%).²¹

EMS use was highest in Europeans (48–78%) compared with other ethnic groups (Māori 36–66%, Pacific 41–66%, Indian 37–68%, non-Indian Asians 26–64%, Other 32–71%). For STEMI, EMS use was lower in Māori (aOR 0.72, 95% CI 0.58–0.90), Pacific (aOR 0.64, 95% CI 0.48–0.87), Indian (aOR 0.63, 95% CI 0.43–0.86) and non-Indian Asian (aOR 0.52, 95% CI 0.37–0.74) but not Other patients (aOR 0.79, 95% CI 0.43–1.52) compared with Europeans. Similar findings by ethnicity were found for NSTEMI. Although odds of EMS use were also lower for UA in all ethnic groups compared with Europeans, the magnitude of the reduction was attenuated, and the effect was not statistically significant, apart from for non-Indian Asian patients.

Other studies have similarly found lower levels of EMS use for ACS among minority ethnic populations. One study focussed on STEMI found that Māori and Pacific patients were less likely than other ethnic groups to be transported by EMS than to self-transport.²¹ A systematic review²² identified three studies that found Latino,²³ Māori, Pacific and Indian peoples¹⁴ and South Asians²⁴ were less likely to have been transported to hospital by EMS than non-minoritised populations.

A systematic review of ethnic differences of the care pathway following an out-of-hospital cardiac event found that disparities in access were attributed to a range of factors (comorbidities, insurance status, socio-economic variables, transportation barriers) and that addressing these disparities required targeted context- and population-specific interventions, informed by qualitative research.²² Responding to this gap, a recent study of Māori patients and families accessing care for an out-of-hospital cardiac event found that EMS initiation was influenced by knowledge of symptoms and a desire to maintain personal dignity, alongside systemic barriers including racism, discrimination and inadequate resourcing.²⁵ The following factors were found to be critical to optimising healthcare journeys: relationships with health professionals, availability of good-quality information and family support and the use of cultural practices.²⁵ Similar factors are also relevant for Pacific and Asian (including South Asian) populations, who may also experience other barriers related to language and lack of knowledge of the Aotearoa New Zealand health system (particularly among migrants within Asian communities).^26–28 Key recommendations for achieving equity from that study, which are likely to be relevant to improving EMS access for Māori, as well as other groups with reduced access, are to increase awareness of the concepts and implications of cultural safety among health professionals and to integrate cultural practices into the healthcare journey of patients and their families.²⁵ Such approaches are likely to be important—alongside addressing other barriers such as lack of good information and cost—to supporting equitable EMS access. Strategies that make non-Europeans as comfortable using the EMS service as Europeans might include diversification of the EMS workforce as well as collaboration with ethnic groups to develop ethnic-appropriate education, communication and outreach. Further research is needed to determine the effectiveness of such strategies in ensuring that the Aotearoa New Zealand health system meets its responsibility for ensuring equitable access to pre-hospital care for ACS patients.

The strengths of this study were that it was based on recent and near complete national hospitalisation and EMS data, and analyses by ethnicity were undertaken separately by ACS type and were adjusted for age, sex, deprivation and rurality. On the other hand, EMS access may have been over-estimated because patients were matched with an EMS encounter on the day of or the day prior to admission. Some of the findings may be spurious because no adjustment was made for multiple statistical testing, and it is possible that COVID-19 and associated national lockdowns may have influenced the results observed given the study period. This study only assessed patients who were admitted to hospital with ACS, and therefore excluded patients who died with ACS prior to hospitalisation. This is an important consideration for future research given that as many as three-quarters of deaths from ACS occur outside the hospital, there is a strong relationship between delay in paramedic care and mortality¹⁴ and Māori and Pacific peoples are significantly more likely than Europeans to have an unwitnessed cardiac arrest²⁹ and to die from coronary heart disease before reaching hospital.⁹

Conclusions

EMS use for ACS (particularly STEMI, and similarly for NSTEMI) varies by ethnicity, with most ethnic groups (including Māori and Pacific peoples) less likely to have access to EMS than Europeans. This is despite Māori and Pacific peoples experiencing the greatest burden of CVD, and legislative requirements in Aotearoa New Zealand for equitable access to health services, including EMS. Previously published research from Aotearoa New Zealand on heart healthcare has identified enablers of access, including good-quality information, reduced cost and health professional cultural safety. Strengthening these enablers could increase access to EMS use for ACS by Māori, Pacific peoples and other non-Europeans.

View Table 1–2, Figure 1.

View Appendix.