ST elevation myocardial infarction (STEMI) is a medical emergency typically caused by sudden thrombotic occlusion of a major coronary artery. Untreated, STEMI is associated with a high chance of death and major morbidity. In the acute phase outcomes are improved by:

1. Access to immediate defibrillation in the event of cardiac arrest. This is dependent on the availability of automated external defibrillators (AEDs) and how quickly a patient comes under either paramedic or hospital care after the onset of symptoms.1–3
2. Prompt diagnosis and acute reperfusion therapy to open the occluded artery. These ‘acute reperfusion therapies’ are pharmacological fibrinolysis and primary percutaneous coronary intervention (primary PCI) combined with evidence-based anti-thrombotic treatment.4–7 In adjusted analyses, every 30-minute delay in delivering reperfusion therapy is associated with a 7.5% increase in mortality.8 Several other studies have also demonstrated that treatment delays in STEMI are associated with a higher risk of adverse outcomes.7,9,10

In New Zealand, all patients receiving coronary angiography after a STEMI are captured in the nationwide All New Zealand Acute Coronary Syndrome Quality Improvement (ANZACS-QI) registry.11 This registry captures information regarding the pre-hospital and in-hospital phases of a STEMI, including key time points from symptom onset to the delivery of reperfusion, method of transport to hospital and reperfusion therapy strategy. The process of acute STEMI care can be divided operationally into two broad components.12,13 The first component is the time from symptom onset to first medical contact (FMC), which depends on patients’ decisions regarding if and when to call the ambulance service or self-present directly to hospital (‘patient delay’). This is the period during which patients are unlikely to be rapidly defibrillated in the event of a cardiac arrest unless there are community AEDs close by. The second component is from FMC until reperfusion therapy is delivered, either pharmacologically, using a fibrinolytic agent, or by primary PCI. This ’system delay’ depends on the response and processes of both ambulance and hospital services.

The National Cardiac Network and New Zealand ambulance services have recently developed a National Out-of-Hospital STEMI Pathway to improve STEMI care.14 Key elements of this pathway include:

1. Primary PCI, which is the reperfusion strategy of choice if estimated FMC to device time is less than 120 minutes (min).
2. Fibrinolysis, which is the recommended reperfusion strategy if estimated FMC to device time is greater than 120 min.
3. Pre-hospital fibrinolysis should be administered, where possible, in patients not undergoing primary PCI as a reperfusion strategy (estimated FMC to device time greater 120 min).
4. Rapid (and immediate if possible) transfer of patients receiving fibrinolysis to a hospital with the facilities to provide immediate (or rescue) PCI for failed re-perfusion and early, routine angiography for others.

The aim of this report is to describe patient and system delays, and their determinants, in the delivery of acute reperfusion therapy for STEMI in New Zealand. The purpose of this description is to identify areas for improvement.

Methods

Cohort

Between 1 January 2015 and 31 December 2017 the ANZACS-QI registry captured a mandatory set of data of 4,464 patients admitted to New Zealand public hospitals with a confirmed diagnosis of STEMI13,14 who received acute reperfusion therapy (either primary PCI or fibrinolysis and a subsequent coronary angiogram). Prior validation work has established that the ANZACS-QI registry captures at least 95% of eligible patients in New Zealand (Submitted to NZMJ). There were 607 patients (14%) excluded from the analysis. This included 197 patients who developed STEMI while in-hospital, and a further 410 patients who had inconsistent data regarding whether fibrinolysis was in-hospital or pre-hospital, or reperfusion temporal data that were judged inaccurate because of clinical likelihood. The study cohort therefore comprised 3,857 patients.

Rescue PCI for those receiving fibrinolysis

This data was available in a subset of patients, as time-frame data was only available from the registry from 26 November 2015 on. From that date, 872 patients received fibrinolysis. Of these, 58 had their records excluded due to invalid data (rescue PCI time same as or before needle time, or rescue PCI to needle time >12 hours). The cohort for this sub-analysis was therefore 814 patients.

ANZACS-QI programme

Individual patient demographic, risk factor, diagnostic, treatment and outcome data were collected prospectively as part of ANZACS-QI. ANZACS-QI is a comprehensive national cardiac registry funded by the Ministry of Health under the auspices of the New Zealand Branch of the Cardiac Society of Australia and New Zealand. The registry is a web-based electronic database that captures a mandatory ACS dataset including patient demographics, clinical presentation, cardiovascular risk factors, investigations, inpatient management, inpatient outcomes and discharge medications. The registry is subject to monthly auditing to ensure data completion in at least 95% of patients. Data quality in each centre is audited annually. Details on data collection, data items, definitions and audit have previously been reported.11

Data definitions

Ethnicity

Up to three ethnic groups can be recorded for each patient in the datasets. However, for this analysis, each patient was categorised into one of five main ethnic groups using a prioritisation process that prioritised groups in the following order: Māori, Pacific, Indian, Other Asian and New Zealand European/Other. The ‘Other’ group comprised only six percent of the latter group, and were composed of Other Europeans and people from the Middle East, Africa and Latin America.

Hospital of admission

Patients were dichotomised by the availability (or not) of a routine all-hours primary PCI service at the hospital of first presentation.

Clinical and risk factor data

The presence of heart failure at admission was defined as a Killip Class of II to IV.15 Cardiovascular disease (CVD) risk factor and comorbidity data included: smoking status, diabetes status, systolic blood pressure (SBP), low density lipoprotein (LDL) and high density lipoprotein (HDL), serum creatinine, body mass index (BMI), prior CVD, MI and congestive heart failure (CHF) diagnoses.

Investigations

The timing and findings at angiography were collected. The findings at angiography were grouped into one of the following: (i) no significant coronary artery disease, defined as the absence of any stenosis with ≥50% diameter loss in the epicardial vessels, (ii) significant (≥50% stenosis) single/double vessel coronary artery disease, (iii) significant three vessel disease and/or left main stem (LMS) disease ≥50%. Left ventricular ejection fraction (LVEF) assessment using echocardiography or left ventriculography was classified into: normal (EF ≥50%), or mildly (EF=40–49%), moderately or severely (EF<40%) impaired. In-hospital coronary intervention by either PCI or coronary artery bypass grafting (CABG) was recorded.

Transport and temporal data (Figure 1)

All patients were recorded as either transported to hospital by ambulance or ‘self-transported’ by other means. Date and time of onset of the most severe symptom(s) were recorded when available. This data was not recorded in 371 (9.6%) of patients.

Figure 1: Temporal components of the total ischaemic time: for those receiving acute reperfusion therapy in-hospital transported by ambulance (Figure 1a) or self-transported (Figure 1b), and those treated with pre-hospital fibrinolysis (Figure 1c).

FMC = First medical contact.

For ambulance-transported patients the date and time of ambulance despatch was taken from the ambulance record. For these patients this was defined as the FMC time. Prior work has established that FMC time is within a few minutes of when the patient calls an ambulance.16 For patients using the ambulance service, time of arrival at hospital was taken from the ambulance report. In all non-ambulance (‘self-transported’) cases, time of hospital arrival (‘door time’) was defined as the time when they were first seen in the emergency department. All patients were managed by a primary PCI or fibrinolysis strategy.

For primary PCI, ‘device time’ was defined as the time the first device was deployed, regardless of the type of device used (either balloon inflation, stent deployment or thrombectomy). If the lesion could not be crossed with a guidewire or device, the time of guidewire introduction was used.

For fibrinolysis ‘needle time’ was the time that fibrinolysis drug administration commenced. Fibrinolysis was recorded as occurring pre-hospital or in-hospital. Rescue PCI was recorded from November 2015 on, where rescue PCI was defined as an emergency PCI in a STEMI patient treated with fibrinolytic therapy (typically given ≤24 hours after onset of ischaemic symptoms) performed within 24 hours of fibrinolysis. Indications for rescue PCI included: failed reperfusion evidenced by <50% ST recovery at 60 minutes after fibrinolysis, recurrence of ST-segment elevation, ongoing ischaemic symptoms, haemodynamic instability or cardiogenic shock.

Patient delay

This was defined as the time from symptom onset to FMC, where FMC was defined as either ambulance despatch time for ambulance-transported patients, or hospital arrival time for self-transported patients (symptom to FMC time).

System delay

This was defined as the time from FMC to needle (FMC to needle time) for fibrinolysis or device (FMC to device time) for primary PCI.

Statistical analysis

Descriptive statistics for continuous variable were summarised as mean with standard deviation (SD) or median with interquartile range (IQR). Categorical data were reported by frequency and percentage.

The relative risks (RRs) with accompanying 95% confidence intervals (CIs) of independent predictors of demographic variables for ambulance transport, and patient and system delays were estimated using multivariable log-binomial regression models. Dichotomous cut-points for analysis were chosen based on clinical consensus (patient delay >1 hour), and the recommendation from the 2016 National STEMI Pathway (primary PCI system delay >120 min).14 For this analysis a fibrinolysis system delay cut-off of >90min was used to be consistent with the primary PCI delay.

All p-values reported were two tailed and a p-value <0.05 was considered significant. Data was analysed using SAS statistical package, version 9.4 (SAS Institute, Cary, NC).

Boxplots were done using boxplot function RStudio version 1.1.419 while maps showing the percentage arriving by ambulance were made using ArcGIS version 10.5.1.

Ethics approval

ANZACS-QI is part of the wider Vascular Informatics Using Epidemiology and the Web (VIEW) study. The VIEW study was approved by the Northern Region Ethics Committee Y in 2003 (AKY/03/12/314), with subsequent amendments to include the ANZACS-QI registries, and with annual approvals by the National Multi-region Ethics Committee since 2007 (MEC07/19/EXP).

Results

In the three-year study period, 3,857 STEMI patients who underwent angiography were treated with acute reperfusion therapy. Of these, 2,691 (70%) received primary PCI and 1,166 (30%) received fibrinolysis. Of those receiving fibrinolysis, 122 (10.5%) received pre-hospital fibrinolysis.

The characteristics and management of these subgroups are shown in Table 1. Despite nearly 70% of patients being younger than 70 years, the risk factor and clinical burden was high. One third of patients were current smokers, 17% had diabetes, 10% had a resuscitated cardiac arrest associated with their presentation and 9% had heart failure on arrival. Most patients (77%) were transported to hospital by ambulance. The majority of patients (95%) had coronary revascularisation therapy, predominantly PCI. From the total patient group, 21% had three vessel coronary artery and/or left main stem disease and 3% had no identifiable coronary disease, the relative proportion higher in patients receiving fibrinolysis.

Table 1: Characteristics of the STEMI cohort.

‍

SD=standard deviation; BMI=body mass index; CVD=cardiovascular disease; MI=myocardial infarction; CHF=congestive heart failure; IQR=interquartile range; SBP=systolic blood pressure; LDL=low density lipoprotein; PCI=percutaneous coronary intervention; LVEF=left ventricular ejection fraction; EF=ejection fraction; CABG=coronary artery bypass grafting; CAD=coronary artery disease.

Māori were more likely to receive fibrinolysis than primary PCI compared with New Zealand European/Others or Pacific people (44%, 22%, 11% respectively). Following fibrinolysis, 93% of patients had residual obstructive coronary artery disease and the coronary revascularisation rate (PCI & CABG) was high (86%), although lower than the 98% in patients treated with a primary PCI strategy.

Temporal components of the reperfusion process (Figure 2, Appendix Table)

Patient delay

Figure 2 shows patient and system delays according to reperfusion strategy and whether the patient was transported to hospital by ambulance. For ambulance-transported patients, the median symptom to FMC time was 45 min but for a quarter of patients, this delay was over two hours. Delays were similar for ambulance-transported patients subsequently managed with fibrinolysis. Delays were much longer for self-transported patients, with a median of 105 min for those who received primary PCI and a quarter experiencing delays of more than four hours before FMC.

Figure 2: Patient and system delay according to reperfusion strategy and whether the patient was transported to hospital by ambulance. Median, mean and interquartile range data are shown. The value of the top whisker for patient delay in the top right panel is beyond the y-axis range. Its value is 562 min.

System delay

Times were more closely clustered than for patient delay times. The median FMC to device time for the largest subgroup of patients—those transported by ambulance who received primary PCI—was two hours, with a quarter delayed beyond 2.5 hours. For those who self-transported, times were slightly shorter (median 87 min (IQR, 65–120 min)), consistent with the ambulance transport time not being included.

As expected, FMC to needle times were consistently shorter than for FMC to device due to the time delay in getting the patient into the catheterisation laboratory. For ambulance-transported patients, the FMC to needle times for fibrinolysis delivered in hospital were longer (86 min (63–117 min)) than for fibrinolysis delivered prior to hospital arrival (39.5 min (31–52min)).

Predictors of patient delay and ambulance utilisation

A. Patient delay (Figure 2, Table 3 and Appendix Table)

Almost half (47% of patients) experienced a delay between symptom onset and FMC of more than 60 min. After adjustment for covariates, delay was more common with increasing age, for Māori and Indian patients, and in those who did not call an ambulance. Sex and whether the local hospital provided routine primary PCI did not influence this delay.

B. Ambulance transport (Table 2 and Figure 3)

After adjustment for covariates, the likelihood of travelling to hospital by ambulance increased with age. Women were also slightly more likely than men to travel by ambulance. Although Māori were slightly less likely to travel by ambulance, after adjustment for covariates, ethnicity did not independently predict method of transport. Patients living at district health boards (DHBs) providing a routine primary PCI service were 30% more likely to be transported by ambulance. This geographical variation is seen in more detail in Figure 3.

Figure 3: Variation in the proportion of patients who travelled to hospital by ambulance according to district health board of domicile.

Table 2: Patient delay: multivariable model for symptom onset to FMC >60min.

‍

SX=symptoms; FMC=first medical contact; PCI=percutaneous coronary intervention; RR=relative risk.

Table 3: Multivariable model for transportation by ambulance.

‍

PCI=percutaneous coronary intervention; RR=relative risk.

Predictors of system delay (Figure 2, Table 4a and b, Appendix Table 1)

The predictors of system delay are analysed separately for primary PCI and fibrinolysis strategies. They are shown only for those transported by ambulance as these comprise the largest number.

Table 4a: Primary PCI system delay: multivariable model for FMC to Device >120 min.

‍

Multivariable regression entering age band, sex, ethnic group, routine primary PCI hospital.
FMC=first medical contact; PCI=percutaneous coronary intervention; RR=relative risk.

Table 4b: Fibrinolysis system delay: multivariable model for FMC to needle >90 min.

‍

Overall, 48% of primary PCI patients experienced FMC to device times of >120 min. Both increasing age and female sex were independently associated with a delay >120 min. Other Asian patients were less likely to be delayed than European/Other patients but other ethnic groups were similar. The small number of patients treated with primary PCI admitted after initially presenting to a hospital without routine primary PCI had a longer system delay than other patients.

Overall, 96%, 68% and 39% of fibrinolysis patients had FMC to needle times of over 30, 60 and 90 min, respectively. The median system delay for Māori was 19 min longer than for European/Other patients. After adjustment, Māori were 36% more likely to exceed the 90 min target. Patients treated with fibrinolysis in-hospital, rather than pre-hospital, were nearly five times as likely to exceed the 90 min target.

Pre-hospital versus in-hospital fibrinolysis (Figure 4, Appendix Table 1)

Only 122 (19%) of the 631 ambulance-transported patients treated with fibrinolysis received it pre-hospital. Of those who received pre-hospital fibrinolysis, the median system time was 46 min shorter than for those than those who received it in hospital (39.5 min (IQR 31 to 52 min) and 86 min (IQR 63 to 117min), respectively).

Figure 4: Distribution of First Medical Contact to needle time in pre-hospital fibrinolysis vs in-hospital fibrinolysis.

This is in contrast to the patient delay times, which were similar.

Rescue PCI

From November 2015 there were 814 fibrinolysis patients with data that included rescue PCI time-frames. Of these, 215 (26%) underwent rescue PCI. Of the remaining 599, 32% had routine coronary angiography on the day of admission, 35% the next day and the remainder beyond two days. Ninety-one percent of these 599 patients had coronary revascularisation (76% PCI, 15% CABG).

For those receiving rescue PCI, the median needle-to-rescue time was 237 min (IQR 185 to 332 min) (see Figure 5).

Figure 5: Distribution of Needle-to-Rescue PCI times.

Discussion

In this comprehensive three-year national NZ STEMI experience, nearly half of patients who received reperfusion therapy took more than an hour to call an ambulance or reach hospital, a period during which defibrillation is not immediately available unless there are community AEDs available. The importance of this is emphasised by the fact that 10% of our study cohort had a resuscitated cardiac arrest. Factors independently associated with a longer patient delay were not being transported by ambulance, Māori or Indian ethnicity and older age. The system delay for patients treated with primary PCI exceeded the National STEMI pathway target of 120 min in 48% of patients, and for those treated with fibrinolysis was greater than 30 min in 96% and greater than 90 min in 39%. Māori patients treated with fibrinolysis were more likely than European/Other patients to exceed the very conservative 90 min FMC to needle cut-off. The 10% of fibrinolysis patients treated pre-hospital started treatment 45 min earlier than those started in-hospital, suggesting that this is a useful strategy to improve performance. After fibrinolysis, only a quarter of patients received rescue PCI, which is lower than reported rates in trials of fibrinolysis therapy.17 Furthermore, the median delay from needle-to-rescue is very long at four hours, which is beyond the period of greatest benefit.

How do patients get to hospital?

Seventy-seven percent of STEMI patients were transported to hospital by ambulance. These patients have more timely access to defibrillation in the event of cardiac arrest than those who self-present to hospital. After adjustment, men and younger patients were less likely to be transported by ambulance. The most important independent determinant of not being transported by ambulance was admission to a regional hospital without routine primary PCI. This is a concern because these smaller hospitals typically serve rural populations, and patients living in the catchment of these hospitals are most likely to benefit from pre-hospital fibrinolysis, which can be provided by the ambulance team. We found that Māori were less likely than other ethnic groups to travel by ambulance. Compared with other ethnic groups, Māori are more likely to live in rural areas and this may account for the lack of ethnic differences in ambulance transport after adjustment for covariates including admitting hospital,17 which is a surrogate for rural residence.

Actions

Further investigation to understand the relative underutilisation of ambulances in non-metropolitan areas and initiatives to improve this are needed. Public education programmes should include a strong message that people should call an ambulance, rather than driving themselves to hospital or have other people take them. Other potential barriers to calling an ambulance which could be reduced are the cost of calling an ambulance for some people and even access to a phone.

What reperfusion therapy do patients receive?

Of NZ STEMI patients receiving acute reperfusion therapy, 70% were managed with a primary PCI strategy and 30% received fibrinolysis.

There is major geographical variation in reperfusion therapy with primary PCI, the dominant treatment in all centres with routine primary PCI capability—predominantly the larger metropolitan DHBs. Fibrinolysis predominates in the smaller non-metropolitan hospitals without such primary PCI access. Furthermore, there were low levels of rescue PCI performed with significant treatment delays. Of all ethnic groups, Māori were most likely to receive fibrinolysis, reflecting the relatively higher numbers of Māori living in rural New Zealand.

Actions

This reiterates the importance of optimising the fibrinolysis pathways to achieve more equitable outcomes. Routine transfer of all suitable STEMI patients to a facility capable of providing immediate rescue PCI in line with the National STEMI pathway, or development of primary PCI services in regional New Zealand are important to address the inequity of STEMI care for regional New Zealand.18

Timeliness of reperfusion therapy

Patient delay

This is the most dangerous period for patients in regards to risk of arrhythmias and cardiac arrest as they are not under paramedic or medical care. Patients took a median of 45 min to call an ambulance and around one and a half hours to self-present to hospital. Patients should not drive themselves or have someone else drive them to the hospital. They should first call an ambulance because of the ability to be monitored and treated if arrhythmias occur and to transmit ECGs so the PCI intervention team can be ready earlier to perform PCI. There are multiple possible reasons for delays in patients contacting medical services. These include patient factors such as recognition of symptom significance, family support and health beliefs as well as potential system barriers including access to a phone, language barriers and charges for ambulance services.

Of concern, on average, both Māori and Indian patients were more delayed than European/Other patients and this difference persisted after adjustment for covariates. This casts some light on prior work showing that Māori in particular are disproportionately represented in community cardiac arrest statistics, and both Māori and Pacific people are 50% more likely to die, both in and out of hospital, when they have an acute ischaemic heart disease event, than European people.19,20 These delays in seeking medical help are therefore an important target for improvement. Intervention studies to reduce pre-hospital delays have had mixed results. Two European community intervention studies in the 1980s and 90s reduced the median delay to hospitalisation in myocardial infarction.21,22 In contrast, a similar study from the US,23 which had shorter delays at baseline, found no significant reduction in delay after community intervention programmes. Another US study which randomised communities to community intervention versus no intervention also found no effect on delay, although they did increase the proportion of patients calling an ambulance.24

Actions to reduce patient delay

Further work to identify and minimise any of the system barriers for patients calling for help is needed. These findings also suggest that a community intervention program targeted at rural and higher risk ethnic groups which encourages earlier call for help directly to the ambulance service may be a useful part of an overall strategy to reduce disparity and improve cardiac outcomes. In New Zealand the National Heart Foundation have run an annual public heart attack awareness programme for the last two years and are continuing work in this area. The St John Ambulance service has been working with marae around New Zealand on a Marae Cardiac Arrest Programme to make AEDs available in marae, particularly in more rural areas. This initiative also includes training in CPR and how to use AEDs.19 Every patient who goes home from hospital after an ACS should be given a chest pain action plan prior to discharge, as half of all patients presenting with a new heart attack have had a prior ischaemic heart disease hospitalisation; this is an important strategy.25

System delay

For patients treated with primary PCI, increasing age and female sex were associated with longer system delay, a finding which requires further investigation. There was little difference in these times across ethnic groups. There remains room for significant improvement, with just under half of ambulance-transported primary PCI patients achieving an FMC-to-device time of less than 120 min, and only a small proportion receiving timely fibrinolysis.

In contrast, of those treated with fibrinolysis, the mean system time for Māori patients was nearly 20 min longer than for European/Other patients. After adjustment a delay beyond 90 min occurred 50% more often in Māori. Possible factors contributing to this difference in system delay include differences in ambulance travel times, patient complexity and regional utilisation of pre-hospital fibrinolysis.

Actions to improve system delays

The National STEMI Pathway recommends that if the FMC-to-device time is likely to exceed 120 min that an initial fibrinolysis strategy be considered. There are clear opportunities to improve outcomes using an initial fibrinolysis strategy. In our analyses, only one in five ambulance-transported fibrinolysis patients were treated pre-hospital, but those that did received fibrinolysis 46 min earlier on average than patients receiving in-hospital fibrinolysis. Greater national adoption of pre-hospital fibrinolysis as advocated by the National STEMI Pathway is likely to reduce times for many of the other patients currently treated in hospital, with associated improvements in outcomes. Furthermore, of the one in four fibrinolysis patients who received rescue PCI for incomplete reperfusion, there were very long delays. The National STEMI Pathway also recommends that all suitable patients receiving fibrinolysis be immediately transferred to a hospital with the facilities to provide rescue PCI, without waiting to determine whether there are signs of reperfusion, provided the patent is appropriate to receive PCI in the event of failure of reperfusion and routine angiography within 24 hours with PCI or CABG as appropriate. This would ensure timely treatment for those requiring rescue PCI and coronary angiography within a day for fibrinolysis patients. With the progressive implementation of the STEMI Pathway, improvements in system delay and outcomes are anticipated. ANZACS-QI registry will be used to monitor system improvement and assess outcomes.

International comparisons

Some caution needs to be exercised in making comparisons with international data due to differences in STEMI pathways and differences in reporting methodology. Most quality indicator data in the past has been for door-to-balloon and door-to-needle times. The reporting of patient and system delay is more recent and less data is available. The most reliable comparisons are for ambulance transferred patients receiving primary PCI. Our median patient delay of 45 min in these patients is similar to that reported in the US in 2016 (50 min) and shorter than in Denmark (74–106 min) in 2008.26,27 The New Zealand median system delay of 119 min is similar to the comparable US (108 min) and Danish patients (97–139 min). The New Zealand median door-to-balloon times of 49 min is similar to the US (63 min) but longer than that in Denmark (29–39 min).

Limitations

Nearly 10% of eligible patients were excluded due to incomplete timeframe data. It is unlikely that this would have led to any systematic bias in results. The ANZACS-QI in-form electronic validation rules have been updated based on this observation to reduce the ineligible number for future audits. This ANZACS-QI cohort includes only out-of-hospital STEMI patients who received acute reperfusion therapy and a subsequent coronary angiogram. While this includes all STEMI patients who received primary PCI, it does not include patients treated with fibrinolysis who either did not receive an angiogram because they died or were considered clinically inappropriate for angiography. The focus of this study was on management of patients receiving reperfusion therapy, so we did not include other STEMI patients who did not receive any form of acute reperfusion. There are also other relevant time points which are not collected in ANZACS-QI. These include ambulance arrival time, time of first ECG, when the ambulance leaves the scene, ‘door-in to door-out’ times when patients arrive initially at a small hospital and are then referred on to a tertiary hospital. St John and Wellington Free ambulance services now capture these data electronically and the ANZACS-QI team is collaborating with these organisations to augment the pre-hospital timeframe data we can report. Prior studies have established the relationship between delay to reperfusion and adverse outcomes.7–10 Further analyses are planned to investigate this relationship in the NZ STEMI cohort when sufficient numbers and follow-up time has accrued. The multivariable modelling performed assumed linear associations with age and did not investigate possible statistical interactions. For this study we needed to exclude temporal data which were clinically invalid. Subsequent to this study additional data validation rules have been implemented on the data entry templates to minimise this source of data inaccuracy.

Conclusion

Systematic implementation of the New Zealand National STEMI pathway is needed to reduce system delays in the delivery of reperfusion. Ongoing public and patient education initiatives are required to reduce delays in patient decision time. The ongoing auditing of patient and system delays with ANZACS-QI will be able to assess whether there are improvements.

Appendix

Appendix Table 1:

‍

Appendix Table 2: