Thesis Submitted in Conformity with the Requirements for the Degree of Master’S Of

Identifying causes of delay in interfacility transports of injured patients

transported by air ambulance in Ontario

Brodie Nolan

A thesis submitted in conformity with the requirements for the degree of Master’s of

Science

Institute of Health Policy, Management & Evaluation University of Toronto

Identifying causes of delay in interfacility transport of injured patients transported by air ambulance in Ontario

Brodie Nolan

Master of Science

Institute for Health Policy, Management & Evaluation University of Toronto

2019

ABSTRACT:

INTRODUCTION: The purpose of this thesis was to examine patient, paramedic, and institutional-related risk factors for delay and identify specific causes of delays in interfacility transfers by air ambulance.

METHODS: Quantile regression was used to identify patient, paramedic and institutional risk factors for delay. Manual chart review to identify specific causes of delay during interfacility transport.

RESULTS: Characteristics associated with shorter time intervals included nursing station as sending facility, rotor-wing aircraft and critical care paramedic crew. Patients requiring mechanical ventilation or transported from academic centres were all associated with prolonged times. A total of 458 causes of delay were identified. The most frequent delays included refuelling, waiting for land EMS escort, documentation, and delays for intubation, chest tube insertion and diagnostic imaging.

CONCLUSION: Ventilator dependence, paramedic level of care, classification of sending facility and helipad availability are associated with delays to interfacility transport of injured patients.

iii

TABLE OF CONTENTS

1.0 Background ……………………………………………………………………. 1 1.1 Study Objectives ……………………………………………………………. 1 1.2 Trauma Epidemiology ……………………………………………………. 1 1.3 Development of Trauma Systems ……………………………………. 2 1.4 Trauma Centres ……………………………………………………………. 3 1.5 Trauma Prehospital Care ……………………………………………………. 4 Figure 1. Field Trauma Triage Guidelines (Ontario) ……………………………. 5 1.6 Use of Air Ambulance for Transporting Injured Patients ……………. 6 1.7 Overview of Ontario Trauma System ……………..…………………….. 7 Figure 2. Ontario Adult Trauma Centres and Referral Boundaries ……………. 8 1.8 Ornge Air Ambulance …………………………………………………... 9 Figure 3. Base locations of Ornge fixed-wing, rotor-wing and land resources …. 9 1.9 Prehospital Trauma Triage ………………………………………… .. 11 1.10 Delays During Interfacility Transfer ………………………………….. 13 1.11 Limitations of Previous Work ………………………………………….. 15 1.12 Rationale ...………………………………………………………………… 16

2.0 Identifying Patient, Paramedic and Institutional Risk Factors for Delay ….. 18 2.1 Methods ……….…………………………………………………………. 18 2.1.1 Primary Aim …………………………………………………... 18 2.1.2 Study Design …………………………………………………... 18 2.1.3 Data Sources …………………………………………………... 18 Figure 4. Measurement of time intervals during interfacility transport ... 19 2.1.4 Study Population …………………………………………... 19 2.1.5 Exposure Variable …………………………………………... 19 2.1.6 Outcomes …………………………………………...... 20 2.1.7 Data Analysis …………………………………………...... 21 2.2 Results …………………………………………………………………... 22 2.2.1 Patient and Injury Characteristics ……….……………………….. 22 Figure 5. Study flow diagram ……….………………………….. 23

Table 1. Patient characteristics …………………………………... 24 Table 2. Institutional characteristics …………………………………... 25 Table 3. Paramedic characteristics …………………………………... 25 2.2.2 Variability of time intervals …………………………………... 25 Figure 6. Variability of time intervals of interest …………………... 26 Table 4. Duration of time intervals across quantiles of interest ……….. 27 2.2.3 Quantile regression models …………………………………... 27 Table 5. Results of quantile regression model for Interval 1 (Time from call accepted to wheels up time of aircraft) …………………………... 28 Table 6. Results of quantile regression model for Interval 2 (Time from aircraft arriving at sending facility landing site to paramedic arrival at patient bedside) …………………………………………………... 29 Table 7. Results of quantile regression model for Interval 3 (In-hospital time) …………………………………………………………………... 30 Table 8. Results of quantile regression model for Interval 4 (Time from departing patient bedside to arrival back at aircraft) …………………... 31 Table 9. Results of quantile regression model for Interval 5 (Time from aircraft arrival at receiving centre to paramedic handover to trauma team) …………………………………………………………………... 32 2.3 Discussion …………………………………………………………………... 33

3.0 Identified Causes of Delay During Interfacility Transport ……………...... 38 3.1 Methods …………………………………………………………………... 38 3.1.1 Primary Aim …………………………………………...... 38 3.1.2 Secondary Aim …………………………………………...... 38 3.1.3 Study Design …………………………………………...... 38 3.1.4 Data Sources …………………………………………...... 38 Figure 7. Measurement of time intervals and grouping of delays during interfacility transport …………………………………………...... 39 3.1.5 Study Population …………………………………………... 39 3.1.6 Identification and Classification of Delays ……………………..... 40

3.1.7 Attributable Delay and Length of Delay Analysis …………... 42 Figure 8. Measurement of attributable time of delay …………………... 42 3.2 Results …………………………………………………………………... 43 3.2.1 Baseline Characteristics ………………………………………….. 43 3.2.2 Frequency and Total Attributable Time of Delays ………………. 43 Figure 9. Frequency of identified causes of delay …………………... 44 Figure 10. Pareto charts of total attributable time (in min) and cumulative percent for each cause of delay …………………………………... 45 3.2.3 Average Length of Delay…………………………………………. 45 Table 10. Mean length of delay in minutes for each identified cause of delay …………………………………………………………………... 46 3.3 Discussion …………………………………………………………………... 47

4.0 Conclusion …………………………………………………………………... 52 4.1 Directions for Future Research …………………………………………... 52 4.2 Conclusion of Thesis …………………………………………………... 53

5.0 Acknowledgements ...... 54

6.0 References ...... 55

LIST OF FIGURES

Figure 1. Field Trauma Triage Guidelines (Ontario)

Figure 2. Ontario Adult Trauma Centres and Referral Boundaries

Figure 3. Base locations of Ornge fixed-wing, rotor-wing and land resources

Figure 4. Measurement of time intervals during interfacility transport

Figure 5. Study flow diagram

Figure 6. Variability of time intervals of interest

Figure 7. Measurement of time intervals and grouping of delays during interfacility transport

Figure 8. Measurement of attributable time of delay

Figure 9. Frequency of identified causes of delay

Figure 10. Pareto charts of total attributable time (in min) and cumulative percent for each cause of delay

vii

LIST OF TABLES

Table 1. Patient characteristics

Table 2. Institutional characteristics

Table 3. Paramedic characteristics

Table 4. Duration of time intervals across quantiles of interest

Table 5. Results of quantile regression model for Interval 1 (Time from call accepted to wheels up time of aircraft)

Table 6. Results of quantile regression model for Interval 2 (Time from aircraft arriving at sending facility landing site to paramedic arrival at patient bedside)

Table 7. Results of quantile regression model for Interval 3 (In-hospital time)

Table 8. Results of quantile regression model for Interval 4 (Time from departing patient bedside to arrival back at aircraft)

Table 9. Results of quantile regression model for Interval 5 (Time from aircraft arrival at receiving centre to paramedic handover to trauma team)

Table 10. Mean length of delay in minutes for each identified cause of delay

viii

1. BACKGROUND

1.1 Study Objectives

Delays to a trauma centre for definitive care and management of severe injuries have been associated with increased morbidity and mortality.1-3 While interfacility transfers are a known cause of delays to definitive care4 neither the nature of these delays, nor their specific impact, are well understood. The primary objective of this thesis was to examine patient, paramedic, and institutional risk factors for delay during interfacility transfer of injured patients by air ambulance in Ontario. Specifically, I examined the impact of these variables on various time intervals from time of request to transfer a patient through to the time of handover to the trauma team at a trauma centre. To do this I used quantile regression models to estimate how specified quantiles of the distribution of these time variables varied with patient, paramedic and institutional characteristics. The secondary objective of this thesis was to identify specific causes of delay to interfacility transfer of injured patients transported by air ambulance and estimate the attributable time associated with each delay.

1.2 Trauma Epidemiology

Traumatic injuries affect Canadians of all ages, races and socioeconomic backgrounds.

Unintentional injuries are the leading cause of death for Canadians between the ages of 1 and 24 and the second leading cause of death for those aged 24 to 44.5 An estimated 4.27 million Canadians aged 12 or older suffered an injury severe enough to limit their usual activities every year.6 On a daily basis, more than 10,000 Canadians are injured seriously

enough to require medical attention. Of these, approximately 9,567 (93%) are seen in emergency rooms, 43 (0.4%) die, 634 (6%) are hospitalized, and 165 (1.6%) are left partially or totally disabled.7 Nationwide, motor vehicle collisions, falls and suicide are the top three causes of death due to injuries.7 Annually in Ontario, injuries result in he death of nearly 6000 people, over 75,000 hospitalizations and almost 6 billion dollars in direct health care costs.7

1.3 Development of Modern Trauma Systems

Identification of the unique needs and resources required to optimally care for injured patients led to the development of modern trauma systems. The American College of

Surgeons (ACS) first addressed trauma care in 1922 by forming a Committee on Trauma

– initially named the Committee on Treatment of Fractures.8 However, besides some initial military initiatives there was little further interest in civilian injuries until the 1950s and 1960s.8 In 1964, Waller et al. were the first to demonstrate that patients injured in a rural setting were more likely to die despite having less severe injuries.9 This study, along with many subsequent studies10-13, underscored the need for timely medical intervention and prehospital care, which is arguably the guiding principle of modern trauma system. There was significant uptake and development of regionalized trauma systems in the 1990s, with work aimed at developing specialized trauma centres, establishing prehospital trauma care guidelines, and identifying the need for rehabilitation after injury.14 The ACS Committee on Trauma suggests that a comprehensive trauma system should consist of injury prevention, prehospital care, specialized trauma centre care, and post-acute care.14

1.4 Trauma Centres

The first document to recommend categorizing hospitals as trauma centres was published by the ACS Committee on Trauma in 1976.15 Trauma centres are accredited based on clinical and non-clinical criteria by the ACS in the United States or the Trauma

Association of Canada in Canada.14 To pass the clinical accreditation process in Canada, a trauma centre must have a dedicated trauma team composed of: a trauma team leader, general surgery, orthopedic surgery, and anesthesia team members.14 There must also be immediate access to surgical subspecialties such as neurosurgery, cardiothoracic and vascular surgery. An accredited trauma centre also requires a 24-hour emergency department with comprehensive medical imaging facilities, a 24-hour operating room and an appropriately staffed intensive care unit. The non-clinical criteria that must be met include active research, education and quality improvement activities.14

Multiple studies have demonstrated the benefit of regionalized trauma centres.2 The

National Study on the Costs and Outcomes of Trauma showed a 25% reduction in mortality for severely injured patients who received care at a trauma centre compared to patients treated at a non–trauma centre.3 Additionally, a meta-analysis of 14 studies demonstrated an overall 15% decline in mortality caused by the establishment of regionalized trauma care at specialized trauma centres.16

1.5. Prehospital Trauma Care

The role of prehospital care in a trauma system is to transport injured patients to the closest appropriate facility in a timely manner.17 There is the concept of the “golden hour of trauma” which has been engrained in trauma systems and prehospital trauma care.17,18

The golden hour refers to the first 60 minutes after an injury is a critical period to transport patients to a trauma centre to address life-threatening injuries.17,18 Although there is little evidence to support a direct time cut-off, many studies have shown that subgroups of injured patients, mostly those that require emergent surgical intervention, have improved outcomes with short out-of-hospital times.11,17-19

One of the challenges in the prehospital care of injured patients is identifying patients who would benefit from being brought directly to a trauma centre without overburdening these specialized centres with minimally injured patients. The concept of appropriate triage of injured patients has been the focus of a significant amount of work attempting to balance over-triage and under-triage. Most efforts have been aimed at reducing undertriage (the transport of severely injured patients to non trauma centres), which may result in preventable morbidity and mortality owing to a delay in definitive care.20,21 Over- triage (the transport of minimally injured patients to a trauma centre) does not have any deleterious effect to the patient, however can contribute to unnecessary resource utilization and overcrowding.22-25 In 2006 the Center for Disease Control and Prevention worked with the ACS Committee on Trauma to establish the Field Trauma Triage

Guidelines (FTTG).26 The FTTG use physiologic, anatomic, mechanism of injury and special considerations to identify the most severely injured patients that would benefit

from direct transport to a trauma centre; potentially bypassing a closer non-trauma hospital.

Figure 1: Field Trauma Triage Guidelines (Ontario)

Emergency Health Services Branch Paramedic Prompt Card for Field Trauma Triage Standard

This prompt card provides a quick reference of the Field Trauma Triage Standard contained in the Basic Life Support Patient Care Standards (BLS PCS). Please refer to the BLS PCS for the full standard.

1.6 Evidence for Air Ambulance Utilization

Identifying how patients should be transported to a trauma centre is a key aspect of prehospital trauma care. The utilization of air ambulance to expedite transport to trauma centres has become an engrained component of modern trauma systems. Intuitively the use of air ambulance shortens the time to arrival to definitive care, which as mentioned above has been associated with reduced mortality. The evidence supporting the benefit of air ambulance use on outcomes for injured patients however is mixed.

Multiple retrospective studies using a national US database of injured patients (the

National Trauma Data Bank) demonstrated a mortality benefit from the use of helicopter emergency medical services (HEMS) when compared to a cohort of injured patients transported by land EMS.27-29 Likewise, an early literature review of the impact of HEMS demonstrated increased survival in all identified studies, with 2.7 lives saved for every

100 HEMS deployments.30 This evidence was countered by studies questioning the benefit of use of air medical transport, as many patients transported by air ambulance in the United States were minimally injured and therefore the benefit of quick access would be negligible.31-33 A meta-analysis of 22 studies of injured patients transported to trauma centre by air ambulance showed almost 70% of all patients transported were minimally injured and over 25% were discharged home within 24 hours of admission.34 Looking at local evidence, a recent study comparing patients brought to a single trauma centre in

Ontario by either air ambulance or ground EMS showed that patients transported by air ambulance had lower than predicted mortality, whereas patients transported by ground

EMS had higher than predicted mortality.12

Air medical transport carries both significant resource and financial costs to a healthcare system. Air transport is extremely costly; with an estimated a cost of a single patient transport by air ambulance to be $6,500 USD.35 In a time of scrutiny over health care spending and trying to find efficiencies, the reduction in unnecessary use of air transport for minimally injured patients might result in significant cost savings for a healthcare system. Air transport is also associated with significant risks. Fatal accidents in the US air medical transport system are increasing.36 In 2008 alone, air ambulance crashes were responsible for 29 deaths in the US.35

It should be noted that there are significant differences between the American and

Canadian air ambulance systems. In Canada, the air medical transport system is provincially funded and each transport is covered by publically funded provincial healthcare, whereas in the United States, air medical transport is privately run and for- profit.

1.7 The Ontario Trauma System

Ontario has more than 70 different land EMS agencies that are coordinated by upper-tier municipalities (ie. counties, regional municipalities or districts).37 Although there are slight differences in the medical directives between each of these EMS services, there are provincial guidelines to standardize field trauma triage such that if these criteria are met and patients are within 30-45 minutes drive of a trauma centre they should be transported directly and bypass any non-trauma centre.38 Ontario has 9 adult trauma centres and approximately 150 other acute care hospitals that are non trauma centres (Figure 2)7. In

Ontario, 40% of the population lives more than a 60 minutes drive to a trauma centre and

15% are more than a 60 minutes transport by air ambulance to a trauma centre.37 Fewer than half of all severely injured patients are transported directly from the scene to a trauma centre.39 The Ontario air ambulance system provides an essential service to improve trauma care access to patients across the province.

Figure 2: Ontario Adult Trauma Centres and Referral Boundaries7

1.8 Scope of Ornge Air Ambulance

In Ontario, Ornge serves as the sole provider of both air medical transport and critical care transport capability (land and air) for interfacility transfers of severely injured patients. Ornge operates the largest air ambulance fleet in Canada, serving over 13 million people over one million square kilometers of land. They have 9 bases that operate rotor or fixed-wing aircraft; these include Thunder Bay, Timmins, Kenora, Sioux

Lookout, Moosonee, Sudbury, Ottawa, Toronto and London. There is a fleet of twelve

Leonardo AW-139 helicopters and eight Pilatus Next Generation PC-12 airplanes with eight helicopters and four airplanes operational on any given day.

Figure 3: Base locations of Ornge fixed-wing, rotor-wing and land resources

Ornge is staffed by primary, advanced and critical care paramedics. Primary care paramedics have the most restricted scope of practice and can provide oxygen along with limited medications such as toradol, ventolin, nitroglycerin and naloxone. Advanced care paramedics can provide sedation and analgesia with fentanyl, ketamine and midazolam and can administer tranexamic acid. Critical care paramedics have the largest scope of practice with additional medication capabilities such as propofol, esmolol, and vasopressin. Advanced and critical care paramedics are trained in a number of advanced procedures including facilitated intubation and airway management, rapid sequence intubation, needle thoracostomy and cricothyrotomy. Ornge advanced care and critical care paramedics are the only paramedics in the province trained to transfuse blood products, and to run ventilators and infusion pumps. A transport medicine physician provides online medical oversight.

Patients are transported by Ornge to a trauma centre by one of three pathways: a scene call, modified scene call, or interfacility transfer. A scene call is when a patient is transported directly from the scene of injury to a trauma center. In these cases they will bypass the closest hospital to expedite transport to a trauma center. A scene call can be activated based on initial 9-1-1 information obtained by the central ambulance communications center (CACC) or requested by the treating land EMS crew as per

FTTG. A modified scene call occurs when a local ground EMS crew meets the Ornge crew at a site other than the initial place of injury and then transports the patient to a trauma center. A modified scene call occurs when a scene call is activated, but local ground EMS are ready to move the patient prior to the arrival of the air ambulance.

During a modified scene call, the land EMS crew arranges a rendezvous with the air ambulance. This is often done at the local community hospital, but sometimes another location such as a local airport is used. If a patient is brought to a non-trauma center as part of a modified scene response, the physician at this hospital may help stabilize the patient. This may include intubating, placing chest tubes, plain film x-rays, bedside ultrasound, and any other procedures they deem necessary. The intent of a modified scene response is always to expedite transfer to a trauma centre and patients are only brought into a local hospital to rendezvous with the flight paramedics if there is a delay in arrival of the aircraft. An interfacility transfer is a transport where a patient is initially brought to a non-trauma centre where they are assessed, evaluated, stabilized, and then later deemed to require transfer to a trauma centre. Interfacility transfers are not activated through the CACC or land EMS crews but rather by the treating physician who makes a conscious decision to transfer the patient to a trauma centre and requests an interfacility transfer.

1.9 Prehospital Trauma Triage

There are many challenges associated with triaging injured patients appropriately to a trauma centre. To begin with the patient must be identified by EMS that they meet FTTG and then either transport them directly to a trauma centre or request an air medical scene response if their transport time is too long. If the patient is not deemed to meet FTTG they are taken to a non-trauma centre where through further work-up it may be discovered they have injuries requiring transport to a trauma centre and an interfacility transfer is requested.

There are some reasons why a severely injured patient may initially be brought to a non- trauma centre by EMS. Many studies have shown that the timely and proper identification all serious injuries in trauma patients to be challenging.40,41 The decision of whether to transport a patient to a trauma centre or not must be made quickly and without complete information. The FTTG were developed to help aid in this decision making for

EMS providers, but, while they are the best available tool for EMS, they lack both sensitivity and specificity.42,43 Previous studies have identified several factors that are associated with an increased risk of undertriage, including: elderly patients, decreased level of consciousness, presence of intoxication, female sex and falls.44,45

Additionally some severely injured patients are intentionally brought to a non-trauma centre as there is no available trauma centre within an acceptable safe distance to transport. The remoteness of parts of Ontario make this especially challenging as 40% of patients injured in Ontario are further than 60 minutes drive to a trauma centre and 15% are not within a 60 minutes transport by air ambulance.37 These numbers were not taking into account inclement weather or traffic conditions, which could limit timely access even more.

Patients who are later transferred to a trauma centre after initial triage to a non-trauma centre were associated with at least a 30% increase in mortality in the first 48 hours after injury.46 A small study looking at air medical scene calls transferred to two Ontario trauma centres showed that 35% of all air scene calls were cancelled, yet 25% of those patients who were cancelled were still later transferred to a trauma centre.47 These

patients experienced a mean delay of over 2 hours before arriving at a trauma centre compared to patients brought directly from the scene.47 Although it is unclear why these scene calls were initially cancelled, it likely speaks to the difficulty in identifying severely injured patients.47 Poor prehospital identification of severely injured patients results in undertriage; with patients being initially brought to a non-trauma centre and then requiring eventual transport to a trauma centre.47,48 This is one driver of delays that occur during the interfacility transfer process.

1.10 Current Understanding of Delays During Interfacility Transfer

Delays in interfacility transfer are due to failure to immediately recognize the need for transfer, prolonged evaluation or unnecessary interventions, and waiting for transportation.4,49,50

The failure to immediately recognize the need for transfer is multifactorial. As mentioned above, there are some well-known risks for under-recognition of injured patients. These include elderly patients, decreased level of consciousness, presence of intoxication, female sex and falls.44,45 Furthermore, the ACS Committee on Trauma has suggested that patients whom meet any of the following being transferred to a dedicated trauma centre14:

1. Confirmed blood pressure less than 90 mm Hg at any time in adults.

2. Gunshot wounds to the neck, chest, abdomen, or extremities proximal to the

elbow/knee.

3. Glasgow Coma Scale (GCS) score less than 9 with mechanism attributed to

trauma.

4. Transfer patients from other hospitals receiving blood to maintain vital signs.

5. Intubated patients transferred from scene or patients who have respiratory

compromise or are in need of an emergency airway.

One study found that compliance to these recommendations was only 51%-82%.49 The study authors hypothesized that this may be due to lack of identification of patients meeting a criterion for transfer or believing a patient can be adequately cared for at their current institution.49 Failure to apply these criteria results in a delay in the decision to transfer the patient and ultimately the interfacility transfer process.

Another significant cause of delay to interfacility transfer is prolonged evaluation.

Patients transferred from non-trauma centres that have surgical specialties available and access to computed tomography (CT) scans have prolonged in-hospital times compared to hospitals lacking these resources.50 It’s likely the availability of these resources may lead to unnecessary interventions or work-up and increase the time before the patient is transferred to a trauma centre. One study in Ontario found that common causes of in- hospital delays included the sending physician performed a procedure and delays for diagnostic imaging.4

Interfacility transport refers to the time from dispatch of the transporting medical team to arrival of the patient at the receiving facility.4 There have been few studies looking at causes of delay in interfacility transport process within the Ontario trauma system. One

study of 911 patients transported to two Ontario trauma centres by air ambulance showed a median time to complete interfacility transport of 145 minutes (interquartile range of

116-175 minutes) with 5% of patients having a time of over 250 minutes.4 Modifiable causes of delay to arriving at the sending facility included refuelling the aircraft, delays related to crew changes and being cancelled off from transporting a patient and then later called back. Lastly, delays to arriving at the trauma centre included waiting for land EMS escort, the trauma team not being assembled and lack of clarity of who was to receive the patient.4

1.11 Limitations of Prior Work

As mentioned above, there are few studies exploring causes of delay to interfacility transfer. The few studies that have been done have significant limitations when attempting to extrapolate to our provincial trauma system.

Most studies have assessed ways to optimize interfacility transfer within Ontario and aimed to identify ways to improve access of rural patients to trauma care37,51. These studies however do not focus on identifying specific causes of delay. Similarly, many studies have been limited to patients cared for at individual trauma centres. These studies have relied on trauma registry data or individual data-sharing agreements between prehospital and hospital databases but have been limited to one or two trauma centres.4,12,47 Lastly, most of these studies have been limited in their small sample size.

The largest study looking at modifiable causes of delay to air ambulance transport in

Ontario had only 150 delays identified and was a secondary outcome in the study.4

1.12 Rationale

Timely access to definitive care is a critical component of modern trauma systems and has been shown to improve patient outcomes after injury.1-3 Access to a trauma centre is not consistent throughout the world and patients without immediate access have had worse outcomes.2 In Canada almost 66% of severely injured patients are initially brought to a non-trauma centre for initial assessment and stabilization.46 Many of these patients in Ontario are later transferred to a trauma center by our provincial air ambulance, Ornge.

Air ambulance is a costly and limited resource, although has been associated with decreased mortality in the Ontario trauma system.12

Delays to a trauma centre for definitive care and management of severe injuries has been associated with increased morbidity and mortality.46 While interfacility transfers have an inherent delay to definitive care, neither the nature of these delays, nor their specific impact, are well understood. The purpose of this study is to examine patient, paramedic, and institutional-related risk factors for delay and identify specific causes of delays in interfacility transfers by air ambulance.

A detailed analysis of the types and impact of delays to interfacility transport of severely injured patients at a provincial level is essential to evaluate our current trauma system.

This study identified specific causes of modifiable delays and estimated the attributable time associated with each of these delays. The information gained from this study will provide a basis for future quality improvement endeavours and education of frontline

providers (at the physician and paramedic level as well as hospital and air ambulance service level) and ensure our trauma system in Ontario is safe, efficient and timely

2. IDENTIFYING PATIENT, PARAMEDIC AND INSTITUTIONAL RISK

FACTORS FOR DELAY

2.1 Methods

2.1.1 Primary Aim

The primary objective of this study was to examine patient, paramedic, and institutional risk factors for delay during interfacility transport of injured patients by air ambulance in

Ontario.

2.1.2 Study Design

This study was a retrospective cohort study of injured patients undergoing interfacility transport to an Ontario trauma centre who were transported by air ambulance. Ethics approval for this study was obtained from the research ethics board at the University of

Toronto.

2.1.3 Data sources

Data were derived from a database of electronic patient care records (ePCR) at Ornge which includes all patients transported by Ornge paramedics. The ePCR includes data pertaining to patient demographics, reason for transfer, vital signs, mechanism of injury, ventilator settings and parameters, medications administered, interventions performed and information on the level of paramedic care and type of aircraft being used.

There are various times associated with each transport that were entered by paramedics and collected in the ePCR (Figure 1). These include the time that the call was accepted, the time the crew leaves their base, the time they arrive at sending facility landing site, the times they arrive and depart from patient bedside, the time they depart from sending facility landing site, the time they arrive at the receiving trauma centre landing site and the time they handover to the trauma team.

Figure 4: Measurement of time intervals during interfacility transport

Interval 1 Flight Time Interval 2 Interval 3 Interval 4 Flight Time Interval 5

Arrive at Depart Arrive at Arrive at Crew leave sending Depart patient sending receiving Handover to Call accepted patient base hospital/ bedside hospital/ hospital/ trauma team bedside landing site landing site landing site

2.1.4 Study population

We included all emergent interfacility transports for injured patients aged 16 years or greater transported to a trauma centre between January 1, 2013 and December 31, 2017.

Patients who were classified as being an urgent or non-urgent priority transfer, who were transported to a non-trauma centre or who were transported by a land ambulance were excluded from the study. Patients with missing or implausible times were removed from the study.

2.1.5 Exposure variables: Patient, paramedic and institutional characteristics

The primary objective of our study was to assess the impact of patient, paramedic and institutional characteristics on key time intervals during the interfacility transport process.

Patient and injury characteristics that were recorded included: age, sex, mechanism of injury, ventilator dependence at time of request to transfer, the first vitals signs obtained by transporting paramedics (heart rate, respiratory rate, oxygen saturation, systolic blood pressure and Glasgow coma scale [GCS]), time of day and season of transport. Paramedic and transport attributes explored included paramedic level of care (primary, advanced or critical care) and type of aircraft (rotor or fixed-wing). We also assessed institutional characteristics of the sending facility (academic, community with greater than 100 beds, community with less than 100 beds or nursing station), aircraft landing site characteristics of sending and receiving facilities (landing pad at hospital requiring to land ambulance transfer, landing pad remote from hospital requiring land ambulance transport, and no landing pad at hospital requiring landing at local airport with land ambulance transport)) and volume of receiving trauma centre (categorized by tertile of patients arriving by air ambulance).

2.1.6 Outcomes

The primary outcome of interest was the modifiable time to complete the interfacility transport process. Using the times captured in ePCR, we created five unique time intervals for each patient that occurred during their interfacility transport (Figure 1).

Flight times for both the flight to sending facility and flight to receiving centre were not assessed in this study, as they are non-modifiable. In total the five intervals of interest were defined as follows: Interval 1 (time from call accepted to wheels up time of aircraft); Interval 2 (time from aircraft arriving at sending facility landing site to paramedic arrival at patient bedside); Interval 3 (in-hospital time), Interval 4 (time from

departing patient bedside to arrival back at aircraft), Interval 5 (time from aircraft arrival at receiving centre to paramedic handover to trauma team). Time intervals were measured in minutes. The primary outcome of interest – modifiable time to complete interfacility transport – was defined as the sum of Intervals 1-5.

2.1.7 Data analysis

Descriptive statistics were used to assess the distribution for all variables of interest in each group. Continuous variables were assessed for normality by evaluating kurtosis and skewness and were summarized as means and standard deviations or medians and interquartile range for normal and non-normally distributed data, respectively.

Categorical variables were displayed as counts and percentages. P-values less than 0.05 were considered statistically significant for all analyses.

Multivariable analyses of the association between patient, paramedic and institutional characteristics and interfacility transport intervals were conducted by quantile regression.

Commonly used regression models for determining the association of variables with an outcome, such as ordinary least squares (OLS) or linear regression, assess how the mean of a conditional distribution varies with changes in system or patient characteristics.52

However, the mean of a distribution may be a poor indicator of central tendency, and conveys limited information about how the system performs for the majority of patients, which requires an analysis of the tail of a distribution.53 Due to our interest in delays during the interfacility transport process (i.e. the skewed tail of the distribution of interval times), we elected to use quantile regression modelling in our analysis.

We created five quantile regression models, one for each of the time intervals measured.

All exposure variables were considered for inclusion of each model. Variable selection for each model was determined using stepwise selection with significance levels to enter and exit the model set at 0.1. Patients with any missing data for one or more of the variables of interest were excluded from the final model. Missing data resulted in exclusion of less than 1% of observations.

Quantile regression models at the 10th, 30th, 50th, 70th, and 90th percentiles were used to determine the effect of patient, paramedic and institutional characteristics on time intervals to interfacility transport. All variables were assessed for multicollinearity using a variation inflation factor (VIF) of 4 as the cut-off for exclusion.

All statistical analyses were conducted using SAS Studio version 3.71 (SAS Institute,

North Carolina, USA).

2.2 Results

2.2.1 Patient and injury characteristics

There were a total of 24,608 adult emergent interfacility transfers transported by Ornge between January 1, 2013 and December 31, 2017, of which 2,884 met our inclusion criteria (Figure 4). After excluding patients that were transported by critical care land

EMS or had missing/implausible time data, our final study population was 2,178 patients.

The study patient, paramedic and institutional characteristics are summarized in Table 1.

The median patient age was 46 years (interquartile range [IQR] 28-62) and 73.7% were male. The most frequent mechanisms of injury were motor vehicle collisions (36.4%) and falls (23.0%), while penetrating injuries accounted for 4.0% of injuries.

Figure 5: Study flow diagram

Table 1: Patient characteristics

Patient Characteristics (n = 2,178) Age, median (IQR) 46 (28,62) Age in years grouped, n (%) Less than 35 787 (36.1) 35-44 284 (13.0) 45-54 352 (16.2) 55-64 317 (14.6) 65-74 238 (10.9) Greater than 75 200 (9.2) Sex, n (%) Male 1607 (73.7) Female 571 (26.3) Mechanism of injury, n (%) Motor Vehicle 793 (36.4) Fall 500 (23.0) Penetrating 86 (4.0) Other 799 (36.6) Heart rate (beats/min), n (%) Greater than 100 618 (28.4) 50-100 1426 (65.4) Less than 50 134 (6.2) Respiratory rate (breaths/min), n (%) Greater than 29 65 (3.0) 10-29 1931 (88.6) Less than 10 182 (8.4) Systolic blood pressure, n (%) Greater than 180 55 (2.5) 90-180 1963 (90.1) Less than 90 160 (7.4) Oxygen saturation less than 90%, n (%) 238 (10.9) Glasgow coma scale, n (%) 13-15 1467 (67.4) 9-12 42 (1.9) <8 669 (30.7) Mechanically ventilated, n (%) 565 (25.9) Time of transport, n (%) 08:00-16:59 795 (36.5) 17:00-23:59 853 (39.2) 00:00-07:59 530 (24.3) Season of transport, n (%) Winter 346 (15.9) Spring 548 (25.2) Summer 803 (36.8) Fall 481 (22.1)

Table 2: Institutional characteristics

Class of hospital, n (%) Academic 80 (3.7) Community >100 beds 782 (35.9) Community <100 beds 1204 (55.3) Nursing station 112 (5.1) Landing site of sending facility, n (%) Local airport 655 (30.1) At hospital with short drive by land EMS 152 (7.0) At hospital, no land EMS component 1371 (62.9) Landing site of receiving trauma centre, n (%) Local airport 122 (5.6) At hospital with short drive by land EMS 858 (39.4) At hospital, no land EMS component 1198 (55.0) Volume of trauma centre, n (%) High volume (>350 transports) 1229 (56.4) Mid volume (150-300 transports) 704 (32.3) Low volume (<150 transports) 245 (11.3)

Table 3: Paramedic characteristics

Paramedic level of care, n (%) Primary care 28 (1.3) Advanced care 581 (26.7) Critical care 1569 (72.0) Type of aircraft, n (%) Rotor-wing 1551 (71.2) Fixed-wing 627 (28.8)

2.2.2 Variability of time intervals

The variability of time intervals is summarized in Figure 5. The duration of each time interval across the 10th, 30th, 50th, 70th and 90th percentiles are displayed in Table 2.

Interval 3, the in-hospital time, was the longest with a median time of 29 minutes (IQR

17-45 minutes, 90th percentile 71 minutes).

Figure 6: Variability of time intervals of interest

0-15 16-30 31-45 46-60 61-75 76-90 91-105 >105

Time (min)

trauma team

Interval 5: Time from aircraft arrival at receiving centre to paramedic handover to

bedside to arrival back at aircraft Interval 4: Time from departing patient Time Interval Interval 3: In-hospital time arrival at patient bedside Interval 2: Time from aircraft arriving at sending facility landing site to paramedic wheels up time of aircraft Interval 1: Time from call accepted to

80 70 60 50 40 30 20 10 Percent

Table 4: Duration of time intervals across quantiles of interest (in minutes)

Quantile Variable 10th 30th 50th 70th 90th Interval 1 3 8 10 15 26 Interval 2 3 7 10 15 30 Interval 3 9 19 29 41 71 Interval 4 4 9 12 17 31 Interval 5 7 12 17 25 47

2.2.3 Quantile regression models

The results of each quantile regression model are summarized in Tables 3-7.

The characteristics identified through quantile regression that were significantly associated with a shorter time interval at the 90th percentile were nursing station as sending facility and rotor-wing aircraft. By contrast, an academic centre as the sending facility or the need for a land EMS escort were both associated with prolonged times.

The magnitude of effect of these characteristics on time was largest at higher quantiles.

Furthermore, patients that were mechanically ventilated were associated with longer in- hospital times across all quintiles and patients transported with a critical care paramedic crew had shorter in-hospital times compared to advanced care paramedic crews.

Table 5: Results of quantile regression model for Interval 1 (Time from call accepted to wheels up time of aircraft) Quantile Variable 10th 30th 50th 70th 90th Heart rate >100 0.0 -0.3 -0.3 0.0 -2.2 50-100 Ref Ref Ref Ref Ref <50 -2.0 -1.3 -2.0* -2.0 -4.3 Respiratory rate >30 -3.0* -0.5 -0.7 -2.0 -7.8 10-30 Ref Ref Ref Ref Ref <10 3.0* 1.8* 1.7* 3.0* 4.8 GCS 13-15 Ref Ref Ref Ref Ref 9-12 1.0 1.0 0.8 2.0 7.6 <8 -1.0 -0.3 0.3 0.0 1.0 Time of day 08:00-16:59 Ref Ref Ref Ref Ref 17:00-23:59 2.0* 0.8* 0.7* 1.0 2.2 00:00—7:59 2.0* 1.3* 0.7* 1.0 3.3 Class of Hospital Academic centre 0.0 0.2 2.1* 4.0* 18.5* Community >100 beds -1.0 -0.8* -0.7* -2.0* -3.0 Community <100 beds Ref Ref Ref Ref Ref Nursing station 0.0 1.8* 3.7* 5.0* -1.2 Landing site at trauma centre At hospital, no land escort Ref Ref Ref Ref Ref At hospital, land escort -1.0 -0.3 -0.3 0.0 0.9 Local airport -3.0 -3.5* -3.3* -4.0* -5.8* Trauma centre volume Highest tertile Ref Ref Ref Ref Ref Middle tertile -1.0 0.3 0.0 1.0 0.9 Lowest tertile 2.0 2.0* 3.0* 4.0* 5.0 Level of care Primary care Ref Ref Ref Ref Ref Advanced care -1.0 -1.5 -3.7* -4.0 -24.1* Critical care 0.0 -2.5 -5.7* -7.0* -28.2* Type of aircraft Rotor-wing 1.0 -1.5* -2.3* -3.0* -7.2* Fixed-wing Ref Ref Ref Ref Ref Coefficient estimates are reported as change to time interval in minutes GCS = Glasgow coma scale, Ref=reference, *p-value<0.05

Table 6: Results of quantile regression model for Interval 2 (Time from aircraft arriving at sending facility landing site to paramedic arrival at patient bedside) Quantile Variable 10th 30th 50th 70th 90th Heart rate >100 1.0 0.3 0.7 0.8 0.0 50-100 Ref Ref Ref Ref Ref <50 1.0 0.3 0.7 0.6 0.0 Mechanically ventilated 1.0* 1.3* 1.7* 2.4* 3.0* Class of Hospital Academic centre -1.0 1.7* 5.0* 9.0* 26.0* Community >100 beds 1.0* 1.3* 1.7* 1.6* 0.0 Community <100 beds Ref Ref Ref Ref Ref Nursing station -3.0* -8.0* -13.0* -19.3* -22.0* Landing site at sending facility At hospital, no land escort Ref Ref Ref Ref Ref At hospital, land escort 1.0* 3.0* 2.7* 2.5* 6.0* Local airport 2.0* 4.3* 5.0* 7.3* 11.0* Trauma centre volume Highest tertile Ref Ref Ref Ref Ref Middle tertile 0.0 0.3 0.3 0.3 3.0* Lowest tertile 0.0 0.0 1.3* 1.1 3.0 Season Summer 0.0 0.0 0.3 -0.1 0.0 Fall 0.0 0.7* 1.3* 1.0 3.0* Winter 0.0 0.3 1.7* 1.1* 1.0 Spring Ref Ref Ref Ref Ref Type of aircraft Rotor-wing 1.0* -4.3* -11.7* -18.1* -27.0* Fixed-wing Ref Ref Ref Ref Ref Coefficient estimates are reported as change to time interval in minutes Ref=reference, *p-value<0.05

Table 7: Results of quantile regression model for Interval 3 (In-hospital time) Quantile Variable 10th 30th 50th 70th 90th Age Less than 35 Ref Ref Ref Ref Ref 35-44 1.0 1.4 1.4 0.4 2.0 45-54 0.0 2.6* 0.7 1.5 7.6* 55-64 2.5* 3.0* 2.0 3.3 4.1 65-74 0.5 2.6 3.1* 6.8* 6.8 Greater than 7 0.5 0.9 0.2 0.3 6.4 Heart rate >100 1.0 1.5 3.1* 2.6* 6.4* 50-100 Ref Ref Ref Ref Ref <50 -2.0 -0.7 1.2 -3.8 -9.4 Oxygen saturation <90% 1.5 2.1 1.8 8.6* 10.2* Mechanically ventilated 12.5* 18.1* 22.9* 29.5* 37.1* Mechanism of injury Motor vehicle 1.5 2.3* 3.3* 3.0* 3.4 Fall 0.5 -1.3 -1.6 -2.4 -4.3 Penetrating -1.0 -3.2 -3.9 -6.1* -11.5* Other Ref Ref Ref Ref Ref Class of Hospital Academic centre 3.5* 14.1* 22.2* 21.8* 32.6* Community >100 beds 0.5 0.5 0.3 -1.0 -1.6 Community <100 beds Ref Ref Ref Ref Ref Nursing station 2.0 -2.3* -2.3 -5.5 -13.5* Landing site at sending facility At hospital, no land escort Ref Ref Ref Ref Ref At hospital, land escort -2.0 -3.7* -2.9 -4.1 -3.4 Local airport -4.5* -4.3* -5.4* -7.3* -5.8* Landing site at trauma centre At hospital, no land escort Ref Ref Ref Ref Ref At hospital, land escort 1.5* 2.0* 2.2* 3.4* 1.4 Local airport 3.0 6.7* 6.1* 9.8* 16.0* Trauma centre volume Highest tertile Ref Ref Ref Ref Ref Middle tertile 2.0* 1.8 3.9* 5.0* 5.6* Lowest tertile -1.0 0.0 -0.8 -1.4 -3.4 Level of care Primary care Ref Ref Ref Ref Ref Advanced care 1.5 6.6 15.8* 18.2* 22.1* Critical care 0.5 4.4 10.3* 11.9* 14.8* Type of aircraft Rotor-wing 5.0* 2.3* -5.0* -16.6* -25.8* Fixed-wing Ref Ref Ref Ref Ref

Coefficient estimates are reported as change to time interval in minutes Ref=reference, *p-value<0.05

Table 8: Results of quantile regression model for Interval 4 (Time from departing patient bedside to arrival back at aircraft) Quantile Variable 10th 30th 50th 70th 90th Mechanically ventilated 3.0* 3.5* 4.0* 3.0* 7.0* Class of Hospital Academic centre 0.5 3.1* 5.1* 7.5* 11.6* Community >100 beds 1.5* 2.3* 2.0* 2.0* 0.0 Community <100 beds Ref Ref Ref Ref Ref Nursing station 0.5 -4.0* -10.0* -13.0* -16.0* Landing site at sending facility At hospital, no land escort Ref Ref Ref Ref Ref At hospital, land escort 0.5 0.3 1.0 1.0 4.0* Local airport 0.5 2.3* 4.0* 5.0* 7.0* Landing site at trauma centre At hospital, no land escort Ref Ref Ref Ref Ref At hospital, land escort 0.0 0.3 1.0* 1.0* 1.0 Local airport 1.5 2.7* 4.0* 3.0* -1.0 Trauma centre volume Highest tertile Ref Ref Ref Ref Ref Middle tertile -0.5 0.8 1.0* 2.0* 4.0* Lowest tertile 0.5 0.3 0.0 0.0 3.0 Level of care Primary care Ref Ref Ref Ref Ref Advanced care 2.0 4.3* 9.0* 9.0* 10.0* Critical care 2.5* 4.5* 9.0* 9.0* 9.0* Type of aircraft Rotor-wing 3.0* -2.5* -13.0* -18.0* -22.0* Fixed-wing Ref Ref Ref Ref Ref Coefficient estimates are reported as change to time interval in minutes Ref=reference, *p-value<0.05

Table 9: Results of quantile regression model for Interval 5 (Time from aircraft arrival at receiving centre to paramedic handover to trauma team) Quantile Variable 10th 30th 50th 70th 90th Mechanically ventilated 1.0 2.0* 2.3* 2.0* 2.3 Mechanism of injury Motor vehicle -1.0 0.0 -0.8 -1.0 -3.3* Fall -1.0 0.0 0.0 0.0 -1.3 Penetrating 0.0 -2.0* -2.5* -2.0 -5.7* Other Ref Ref Ref Ref Ref Time of day 08:00-16:59 Ref Ref Ref Ref Ref 17:00-23:59 0.0 0.0 -1.5* -2.0* -3.7* 00:00—7:59 -1.0 -1.0* -2.0* -2.0* -3.3* Class of Hospital Academic centre 4.3* 7.0* 9.9* 9.0* 5.6 Community >100 beds 0.0 0.0 0.3 1.0 2.7* Community <100 beds Ref Ref Ref Ref Ref Nursing station 4.0* -1.0 -1.8 -3.0 -1.7 Landing site at trauma centre At hospital, no land escort Ref Ref Ref Ref Ref At hospital, land escort 3.0* 4.0* 4.5* 5.0* 5.3* Local airport 14.0* 20.0* 24.5* 31.0* 54.0* Trauma centre volume Highest tertile Ref Ref Ref Ref Ref Middle tertile 0.0 1.0* 2.5* 4.0* 4.7* Lowest tertile 0.0 -2.0* 0.5 0.0 0.3 Level of care Primary care Ref Ref Ref Ref Ref Advanced care 7.0* 18.0* 11.5* 8.0* 4.7 Critical care 8.0* 19.0* 13.0* 10.0* 6.3 Type of aircraft Rotor-wing -5.0* -18.0* -22.3* -27.0* -35.3* Fixed-wing Ref Ref Ref Ref Ref Coefficient estimates are reported as change to time interval in minutes Ref=reference, *p-value<0.05

2.3 Discussion

This objective evaluated risk factors for delays during interfacility air transport. There were three key findings identified. First, the use of rotor-wing aircraft and hospital-based helipads was associated with substantially lower transport times. Second, transports from academic centres were associated with longer transport times compared to those that originated at community hospitals or nursing stations. Third, interfacility transport times are heavily skewed and delays disproportionately affect longer patient transports.

Our study demonstrates the large variability of transport times in our air ambulance system. There were heavily skewed distributions across all transport time intervals with the in-hospital time interval being the longest. The presence of heavily skewed distributions suggests the potential for improvement to reduce the variation across interfacility transports. This is in keeping with a previous study from Ontario that found wide variability of time to complete interfacility transports.4

We used quantile regression modeling to explore the skewed tail end of transport times to better assess risk factors for delay. Previous studies in a prehospital setting have demonstrated the benefits of using quantile regression modeling over OLS or linear regression. 53,54 The flexibility of quantile regression makes it well suited for the non- uniformity, skewed or asymmetrical distribution of data that would violate the assumptions of OLS regression techniques.52 Our results demonstrate that the association of delays due to patient, paramedic and institutional factors are not uniform, but worse at

the tail end of transport intervals. Put another way, delays during interfacility transport disproportionally affect patients who already have longer transport times.

Our findings expand on the limited understanding of interfacility transport delays by identifying patient, paramedic and institutional risk factors associated with delays. We found that being transported by a critical care or advanced care paramedic was associated with shorter times from the call being accepted to wheels up time of the aircraft. At the

90th percentile, the time benefit of an advanced or critical care medic in reference to a primary care crew was 24.1 and 28.2 minutes respectively. A small study examining delays to interfacility transport to two Canadian trauma centres by air ambulance identified refuelling, mechanical and weather issues as being frequent causes of delays to launching an aircraft.4 Our study suggests that we may be able to expedite launching of an aircraft by ensuring we have advanced and critical care crews transporting our severely injured patients. Furthermore, up-training paramedics to a critical care level across the organization could also reduce the time to launch an aircraft as critical care paramedics had shorter transport times compared to advanced care crews.

Our study identified the type of aircraft landing site being associated with interfacility transport delays. Many sending facilities in our trauma system do not have a helipad on site or require the aircraft to land at a local airport away from the hospital and then have a local land EMS crew pick them up from the airport and bring them to the sending facility.

This effort requires coordination from our air ambulance service as well as local EMS to ensure an ambulance is available when the aircraft lands. Furthermore, the use of rotor-

wing aircraft, even when controlling for landing sites was consistently faster than use of a fixed-wing aircraft for every transport time interval measured in our study. Advocating for hospital-based helipads and optimizing coordination between our air ambulance provider and local land EMS when a land escort is required may help reduce interfacility transport times.

Another finding of our study was that patients being transferred from a nursing station had shorter in-hospital times compared to patients being transported from an academic centre. This relationship may be partially explained by the higher level resources available at academic centres. Gomez et al. demonstrated that patients being transported from “resource rich” centres, defined by the presence of surgical specialties, CT scanners and intensive care capabilities, are associated with longer emergency department length of stays compared to centres lacking these resources.50 Furthermore, previous work has identified patients undergoing diagnostic imaging at the sending facility an important cause of in-hospital and patient contact delays.4 In this study, over 14% of in-hospital delays were a result of patients undergoing further diagnostic imaging after the arrival of transporting paramedics.4 Nursing stations by comparison have very limited resources, often with no access to blood products, x-rays or CT scans.55 Since there is less than can be done for patients at nursing stations, it may propagate a mentality of “load and go” as little can be done to stabilize patients in these settings. Improved communication between the sending and receiving physicians to ensure only essential diagnostic imaging and medically necessary procedures are completed prior to transport may help reduce interfacility transport times.

Patient characteristics mostly influenced the in-hospital time interval in our study. The need for mechanical ventilation prolonged in-hospital times by over 37 minutes at the

90th percentile. Intubated patient require infusions for sedation and ventilator manipulation, which will naturally take time to accomplish. Furthermore, hypoxic patients may have required interventions such as intubation, placement of airways or insertion of chest tubes to optimize their oxygenation prior to air transport, resulting in longer in-hospital times. The hypobaric effects of air transport often make it necessary to place chest tubes before insertion and improve oxygenation prior to transport as it is difficult to address hypoxia at altitude.56,57

From a patient perspective, a clinically meaningful delay that can result in increased mortality may be as short as 15-30 minutes.58,59 Therefore, based on our results, the decision to send an advanced care or critical care paramedic crew, or the availability of rotor-wing aircraft or the placement of a hospital helipad could be a matter of life or death.

Our study has several potential limitations. Paramedics entered the times used to calculate the transport intervals manually. These transport times were recorded either in real time during the patient transport or retrospectively after the transport was completed, leaving the potential for measurement error. Additionally, our study was unable to measure the time a patient spent at a sending facility prior to the request to transfer the patient. The time to make the decision to transport a critically injured patient plays an important role in overall delays to interfacility transfer and warrants future study.

Furthermore, as our study was limited to the air ambulance database, we were unable to directly assess the impact of longer transport times on patient outcomes. We can, however, infer from other studies that these delays experienced by our patients would be associated with worse outcomes.58,59 Finally, the use of an air ambulance database precluded our ability to include variables such as injury severity scores and comorbidity indices commonly presented in trauma literature. We hope that future relationships between our air ambulance service and provincial trauma registries may allow for this sharing of data to enrich this understanding.

In summary, we have demonstrated that ventilator dependence, paramedic level of care, classification of sending facility and helipad availability are associated with delays to interfacility transport of injured patients. Efforts can be made at both the air ambulance and institutional levels to ensure timely and efficient transports.

3. IDENTIFIED CAUSES OF DELAY DURING INTERFACILITY TRANSPORT

3.1 Methods

3.1.1 Primary Aim

The primary aim of this objective was to identify specific causes of modifiable delays during the interfacility transport process for injured patients transported to a trauma centre by air ambulance in Ontario.

3.1.2 Secondary Aim

The secondary aim of this objective was to estimate the attributable time associated with delays identified during interfacility transport of injured patients transported to a trauma centre by air ambulance in Ontario.

3.1.3 Study Design

This study was a retrospective cohort study of injured patients undergoing interfacility transfer to a trauma centre who were transported by air ambulance in Ontario. Ethics approval for this study was obtained from the research ethics board at the University of

Toronto.

3.1.4 Data Sources

Data were derived from a database of ePCRs at Ornge which includes all patients transported by Ornge paramedics. The ePCR includes data pertaining to patient demographics, reason for transfer, vital signs, medications given and interventions

performed. Paramedics also complete a narrative text of the transport and could assign standardized delay codes to the call. In addition, there are various times associated with each transfer that were entered by paramedics and collected in the ePCR (Figure 2).

These include the time of dispatch, the time the crew leaves their base, the time they arrive at sending facility landing site, the times they arrive and depart from patient bedside, the time they depart from sending facility landing site, the time they arrive at the receiving trauma centre landing site and the time they handover to the trauma team.

Figure 7: Measurement of time intervals and grouping of delays during interfacility transport

Overall Time

Time-to-sending delays In-hospital delays Time-to-receiving delays

Arrive at Depart Arrive at Depart Arrive at Handover Call Crew leaves landing landing patient patient trauma to trauma accepted base site/ site/ bedside bedside centre team hospital hospital

Flight time Flight time

3.1.5 Study Population

The study population included all emergent interfacility transfers for injured patients transported to a trauma centre by either fixed or rotor-wing resources between January 1,

2014 and December 31, 2016. Patients with a primary medical reason for transfer, those who were transported to a non-trauma centre or were transported by a land ambulance were excluded from the study.

3.1.6 Identification and classification of delays

The secondary objective of this study was to identify the frequency and causes of delays during interfacility transport. In addition, we evaluated the total attributable time for each delay.

Using the times captured in ePCR, we created three time intervals for each patient transport. These times included: i) the time-to-sending interval, which was measured from the time of dispatch to arrival to patient bedside; ii) the in-hospital time interval, defined by the time from paramedic arrival to patient bedside to departure with patient; and iii) the time-to-receiving/handover interval, which was measured from the time of departure with patient to handover to the trauma team (Figure 1). Since we were interested in the modifiable aspect of interfacility transport, the flight times for both the time-to-sending and time-to-receiving/handover intervals were not included in the calculation of these times.

Given the large number of records and the need for manual review of the ePCR, we used a screening process to identify patients that were likely to have experienced a delay during their interfacility transport. The screening process involved using three approaches. First, we identified charts for review if there was a standardized delay code entered by paramedics. These delay codes are pre-determined and can be added by paramedics to the patient care record at any point if they deem appropriate. Second, the free-text narrative field of each patient record was searched for the terms “delay”

“prolong” “wait” or “duty out”, including common misspellings of these words. Any

patient record containing these terms was then flagged for review. Lastly, all patient records that had transport times exclusive of flight times above the 75th percentile for overall time to complete interfacility transfer, time-to-sending-hospital, in-hospital or time-to-receiving/handover (excluding flight times) were also manually reviewed.

Any patient identified through any one of these screening methods had their entire Ornge electronic patient care record manually reviewed to search for causes of the delay. In the case that a patient was positively screened but no reason for delay was identified, no delay reason was recorded for that patient. Likewise, if a patient had a delay code entered by the paramedics but there was nothing to substantiate the reason for delay, no delay reason was recorded. A delay was defined as anything the paramedics identified in their charting that hindered or postponed transport. Identified causes of delay were then coded and categorized into time-to-sending, in-hospital and time-to-receiving/handover delays. The frequency of each type of delay was recorded. A 10% random sample of patient records that were not identified through our search strategy were also manually reviewed to validate our screening methods and to inform if these search parameters should be extended. Our screening approached proved to be effective, with no additional incidents or causes of delay identified in the sample.

3.1.7 Attributable delay and length of delay analysis

Having categorized causes of delay, we then sought to evaluate the mean time attributable to each cause of delay. Mean times for time-to-sending, in-hospital and time- to-receiving/handover were calculated for each sending facility using records where no delay had been identified. Similarly, times for the interval where a delay was identified were determined. The difference between the two was the “attributable time of delay” for that type of delay (Figure 3). We then calculated the “total attributable time” for each delay type as the product of its duration and its frequency such that it represents the cumulative time (in minutes) that a delay was responsible for. This was done for each sending facility, and then summed across all facilities. Ultimately, the average length of delay was calculated by dividing the total attributable time by the frequency of delay type.

Figure 8: Measurement of attributable time of delay

Mean in-hospital time for Hospital A 30 minutes

In-hospital time for Delay Z that occurred in Hospital A 40 minutes

Attributable time of delay 10 minutes

3.2 Results

3.2.1 Baseline characteristics

There were 932 injured patients emergently transported by air ambulance from a community hospital to a trauma centre over the 3-year study period. Our screening method identified 552 (59%) patients whom required manual review of their electronic patient records and from which 329 (35%) patients were identified as having at least one delay during their transport. There were a total of 458 unique causes of delay that were identified. Of the 329 patients who experienced a delay during interfacility transport, there were 234 (71%) patients with a single delay during their transport, 67 (20%) patients with two delays, 24 (7%) patients with three delays and 2 (1%) each with four and five delays, respectively.

3.2.2 Frequency and total attributable time of delays

The most frequent cause of delays to sending facility were refuelling (38%), waiting for land EMS escort (25%) and weather (12%) (Figure 6). The most common in-hospital delays included waiting for documentation (32%), delay to intubate (15%), medically unstable patient (13%) and waiting for diagnostic imaging (DI) (12%). The most frequent delays to receiving/handover included waiting for land EMS escort (31%), trauma team not assembled (24%) and weather (17%).

Figure 9: Frequency of identified causes of delay

120

100

60 (Count) 40 Delays to sending 20

0 Refuel Waiting for land Weather Mechanical Crew change Triage Cancelled and Restocking Dispatch issues Other EMS escort called back aircraft 60

30 (Count) 20 In hospital delays 10

10 8 6 4 2

handover (Count) 0

Delays to receiving/ Waiting for land EMS Trauma team not Weather Equipment issues Refuel Mechanical Other escort assembled

The delays to sending facility with the highest total attributable time were refuelling

(1249 minutes), waiting for land EMS (898 minutes) and weather (478 minutes) (Figure

7). The in-hospital delays with the highest total attributable time included delay to intubate (1226 minutes), delays for diagnostic imaging (911 minutes), delays waiting for documentation (801 minutes) and medically unstable (693 minutes). The delays to receiving/handover with the highest attributable time were trauma team not assembled

(153 minutes), waiting for land EMS escort (115 minutes) and weather (113 minutes).

We examined the individual cases involved in delays due to the trauma team not being assembled and found that the mean delay is significantly skewed by two patients. These

two patients both sustained isolated head injuries and had a 100-minute and 40-minute delay due to handing over to the neurosurgical team at the receiving trauma centre. All other delays waiting for the trauma team to assemble were less than 10 minutes.

Figure 10: Pareto charts of total attributable time (in min) and cumulative percent for each cause of delay

1400 100% 1200 80% 1000 800 60%

600 40% Time (min) 400 Delays to sending 20%

200 Cumulative percent 0 0% Refuel Waiting for land Weather Crew change Mechanical Restocking Cancelled and Dispatch issues Triage Other EMS escort aircraft called back

1400 100% 1200 80% 1000 800 60%

600 40% Time (min) 400 In hospital delays

20% Cumulative percent 200 0 0%

200 100%

150 80% 60% 100

Time (min) 40% 50 20%

0 0% Cumulative percent

Delays to receiving/handover Trauma team not Waiting for land EMS Weather Equipment issues Refuel Mechanical Other assembled escort

3.2.3 Average length of delay

Delays to sending facility with the highest average length of delay were dispatch issues

(23 minutes), restocking aircraft (21 minutes) and crew change (20 minutes) (Table 8).

In-hospital delays with the longest average length of delay included stabilization of patient in the operating room (77 minutes), chest tube insertion (53 minutes), multi- casualty incident (50 minutes), delay to intubate (49 minutes) and delays for diagnostic

imaging (46 minutes). Delays to receiving/handover with the highest average length of delay were weather (23 minutes), trauma team not assembled (22 minutes) and equipment issues (15 minutes).

Table 10: Mean length of delay in minutes for each identified cause of delay

Delay Mean delay in min (SD) Delay to sending facility Dispatch issues 23.5 (42.7) Restocking aircraft 21.2 (18.5) Crew change 20.6 (19.4) Cancelled and called back 16.2 (12.8) Weather 15.4 (37.3) Mechanical 14.3 (29.1) Waiting for land EMS escort 13.6 (23.0) Refuel 12.4 (22.8) Triage 7.7 (13.7) Other 4.8 (13.8) In-hospital delays Stabilization in operating room 76.7 (69.6) Other 54.0 (55.2) Delay for chest tube 53.4 (52.8) Multi casualty incident 50.0 (46.7) Delay to intubate 49.0 (37.6) Delay for diagnostic imaging 45.6 (41.3) Equipment issues 43.4 (38.9) Medically unstable 31.5 (36.5) Confirming disposition/receiving 28.8 (20.4) Delay for cast/splint 23.7 (50.2) Waiting for blood products 17.2 (19.9) Waiting for documentation 15.1 (29.0) Delay to sending facility Weather 22.6 (17.3) Trauma team not assembled 21.9 (43.3) Equipment issues 15.5 (30.7) Waiting for land EMS escort 12.8 (13.9) Refuel 5.5 (2.1) Mechanical 0.5 (0.5) Other 0.5 (0.5) SD = standard deviation

3.3 Discussion

In this study, we identified multiple modifiable causes of delay during the process of interfacility transport of injured patients transported by air ambulance. There are three key findings in our study. First, it is important to assess both the frequency and duration of delay, as many high frequency delays were short in duration. Second, patients who had invasive procedures (ie. intubation, chest tube insertion) and advanced DI at the sending facility experienced the longest delays. Third, improving communication between local EMS and air ambulance can reduce delays incurred by waiting for land

EMS escorts.

Our findings on in-hospital delays highlight the importance of understanding both the frequency and duration of delays. For example, the most common delay experienced in- hospital was waiting for documentation and although was responsible for 32% of all in- hospital delays it had the lowest impact on length of delay; resulting in an average delay of 15 minutes. Likewise, both the stabilization of a patient in the operating room and mass-casualty incidents were some of the least frequent delays encountered in hospital however had significant impacts on time resulting in, respectively, an average delay of 77 minutes and 50 minutes. Pareto charts provide a helpful visual analysis of this relationship between the sum total and cumulative impact of delays (Figure 7). This approach can be useful to understand where to put efforts into improving the trauma transport system. For example, efforts to reduce frequent yet smaller delays such as waiting for documentation could help our overall trauma system efficiency. On the other

hand, rare delays such as mass casualty incidents, while significant on a patient level are a poor focus for systemic improvements.

Another finding in our study was that invasive procedures done at a non-trauma centre result in some of the longest delays to interfacility transfer. If a patient needed to be intubated once the flight paramedics arrived (15% of all in-hospital delays), it increased the in-hospital time by 49 minutes. Likewise chest tube insertion resulted in an average delay of 53 minutes. There are many sending facilities in the trauma system that have a low volume of acutely injured patients which may be a contributing factor to the resultant delay these procedures cause as physicians who are unfamiliar with technique or equipment available in these high-risk situations may be uncomfortable proceeding without the backup of another physician or the flight paramedics. Furthermore, it is possible that patients may continue to deteriorate or previously unidentified injuries are recognized, such as worsening pneumothorax or hemothorax resulting in a delay to initiative these procedures. Another cause related to delays from procedures may be from a lack of familiarity with the physiologic changes and hypobaric environments associated with air transport.56 Sending physicians may be unfamiliar with the need to place chest tubes for minimal pneumothoraces or the challenges associated with intubation in an aircraft, which could result in a delay to initiate these procedures until the paramedic crew arrives. Communication between the sending physician and receiving trauma team leader or transport medicine physician could help optimize patients for transport prior to arrival of the transporting paramedics and reduce these in-hospital delays. The use of a checklist to optimize patients prior to air transport has previously been suggested.60

Additionally, airway management in injured patients is inherently challenging and may also precipitate a delay for appropriate preparation and execution.10 Another significant cause of in-hospital delay is waiting for DI. Delays due to DI accounted for 12% of all in- hospital delays and resulted in an average delay of 46 minutes. One study found that

60% of all interfacility transfers that have CT scans imaging done at the sending facility have at least one CT scan repeated at the trauma centre.11 Efforts to reduce delays caused by diagnostic imaging may include a discussion between the sending and trauma physicians to clarify the necessity of advanced DI prior to transport.

Our findings expand on the limited understanding of interfacility delays and serves to better characterize modifiable delays at a systemic level. A small study examining delays to interfacility transport to two Canadian trauma centres by air ambulance identified refuelling, mechanical and weather issues as being frequent causes of delay to arriving at sending facility.7 Our findings were consistent, yet we also identified a significant number of transports that were delayed as a result of waiting for a land EMS escort.

Many sending facilities in our trauma system do not have a helipad on site or require the aircraft to land at a local airport away from the hospital and then have a local land EMS crew pick them up from the airport and bring them to the sending facility. This effort requires coordination from our air ambulance services as well as local EMS systems to ensure an ambulance is available when the aircraft lands. We found there is often a breakdown of this coordination resulting in the flight paramedics waiting an average of

14 minutes for a land EMS escort to arrive. Furthermore, we found that dispatching

issues, having to restock aircraft and delays surrounding crew changes although occurring less frequently, had the greatest impact as measured by minutes per delay.

Overall, delays to receiving trauma centre and handover were relatively uncommon. All causes of delay to receiving accounted for only 29 of all 458 delays identified the study.

As discussed above, like many of our sending facilities, some of our receiving trauma centres do not have a rooftop helipad and require a land EMS escort from the landing site to the trauma bay. Waiting for a land EMS escort was the most common cause of delay to receiving/handover, resulting in 31% of all handover delays and had an average delay of 13 minutes. Once again, improved communication between air ambulance and land

EMS services may improve coordination and lessen the impact of this delay.

It should be noted that almost 30% of patients identified as having a delay during interfacility transport experienced more than one delay. This is significant because having even two or three shorter delays will lead to clinically significant total delay in transfer. For example a patient who three of the most common but shortest delays; such as refuelling, waiting for land EMS escort and delay to receiving documentation would incur around 45 minutes of total delay time during their transport. That may be long enough to cause patient harm due to delay to definitive care at a trauma centre.

This study is the largest of it’s kind to examine causes of delay to interfacility transport of injured patients within our trauma system. It provides useful information for targeting interventions that can reduce the frequency or impact of these delays.

There are several limitations to this objective that warrant discussion. This study relied on delays that were identified by paramedics by either delay codes or written text describing the delay that occurred. As such there are likely cases where a delay did occur but no documentation was done and thus we would not have captured those delays in this study. Additionally, paramedics may have been less likely to report causes of delay that resulted from their actions. It was not feasible to obtain individual medical records from each sending facility to assess the physician or nursing notes to see if they documented any delays incurred on the paramedic side that we did not capture. However, our study does hold face validity with previous work identifying causes of delay to interfacility transfer.7 Another limitation to this study is the potential for measurement error in calculating the attributable delay time and average time per delay. Delay times were estimated using time stamps of a patient transport entered manually by paramedics, something that may be done in real time or retrospectively after the patient is transported.

This approach could lead to either an overestimate or underestimate of time of delay, however is unlikely to result in a significant bias in our results.

4. CONCLUSION

4.1 Directions for Future Research

The results of this study suggest several directions for future research. As noted above, this study was limited to data available in an air ambulance database and thus in-hospital outcomes, such as mortality could not be assessed. Data sharing agreements between

Ornge and a provincial health administrative database at the Institute for Clinical

Evaluative Sciences are underway. This will allow future studies to assess the impact of interfacility transport delays on mortality, hospital length of stay, blood product usage and other patient-centric outcomes.

Also noted in the previous section is the potential inability of our study to have identified causes of delay from the perspective of the sending or receiving centres. By limiting our methods to the paramedic patient record, we may have been unable to identify causes of delay identified by the sending or receiving centres. A future study could reach out to these centres for feedback or specifically collect causes of delay perceived by these hospitals.

Futhermore, armed with the knowledge gained from this study in identifying risk factors and causes of delay during interfacility transport, future endeavours to reduce these delays can be considered. Another study could assess the effectiveness of these strategies on reducing the frequency or duration of identified delays.

4.2 Conclusion of thesis

This thesis of injured patients transported by air ambulance to a trauma centre was able to identify both risk factors for and specific modifiable causes of delay that occur during the interfacility transport process. This thesis demonstrated that ventilator dependence, paramedic level of care, classification of sending facility and helipad availability are associated with delays to interfacility transport of injured patients. Furthermore, efforts to improve communication between air ambulance service and local land EMS services should be made in an effort to reduce the impact of delays to both sending a receiving hospitals caused by a lack of land EMS escort. Patients requiring intubation or chest tubes experience delays of more then 50 minutes. Ensuring physicians are comfortable with and equipment is readily available for these life saving interventions may help expedite transport. Patients undergoing advanced diagnostic imaging after the decision to transfer had been made should ensure the timing does not affect the patient’s transport and deferral of further DI until arrival at the trauma centre should be considered. Future efforts can be made at both the air ambulance and institutional levels to ensure timely and efficient transports.

5. ACKNOWLEDGEMENTS

I would like to acknowledge the following people and organizations for their support in the development of this thesis.

• Thesis committee: Dr. Avery Nathens (thesis supervisor), Dr. Barbara Haas, Dr.

Homer Tien, Refik Saskin

• Queen Elizabeth II/Sunnybrook Prehospital Care Program Graduate Scholarships

in Science and Technology at the University of Toronto

• Canadian Association of Emergency Physicians

• Institute for Health Policy, Management & Evaluation (IHPME)

• Clinical Epidemiology and Health Care Research program at IHPME

• My wife (Julia) and dog (Oliver)

6. REFERENCES

1. Piontek FA, Coscia R, Marselle CS, Korn RL, Zarling EJ, American College of Surgeons. Impact of American College of Surgeons verification on trauma outcomes. J Trauma 2003;54(6):1041–6–discussion1046–7.

2. Demetriades D, Martin M, Salim A, et al. Relationship between American College of Surgeons trauma center designation and mortality in patients with severe trauma (injury severity score > 15). J Am Coll Surg 2006;202(2):212–5–quizA45.

3. MacKenzie EJ, Rivara FP, Jurkovich GJ, et al. A national evaluation of the effect of trauma-center care on mortality. N Engl J Med 2006;354(4):366–78.

4. Nolan B, Tien H, Sawadsky B, et al. Comparison of Helicopter Emergency Medical Services Transport Types and Delays on Patient Outcomes at Two Level I Trauma Centers. Prehosp Emerg Care 2017;21(3):327–33.

5. Canada S. Leading causes of death [Internet]. 2011 [cited 2018 Dec 20];Available from: https://www.statcan.gc.ca/pub/82-625-x/2014001/article/11896/c-g/c-g01- eng.htm

6. Canada S. Injuries in Canada: Insights from the Canadian community health survey [Internet]. httpswww.statcan.gc.capub--xarticle-eng.htm. 2011 [cited 2018 Dec 20];Available from: https://www.statcan.gc.ca/pub/82-624- x/2011001/article/11506-eng.htm

7. Parachute. The cost of injury in Canada [Internet]. httpwww.parachutecanada.orgdownloadsresearchCostofInjury-.pdf. 2015 [cited 2018 Dec 20];Available from: http://www.parachutecanada.org/downloads/research/Cost_of_Injury-2015.pdf

8. Cales RH, Trunkey DD. Preventable trauma deaths. A review of trauma care systems development. JAMA 1985;254(8):1059–63.

9. Waller JA, Curran R, Noyes F. Traffic Deaths. A preliminary study of urban and rural fatalities in Californi. Calif Med 1964;101(4):272–6.

10. Sampalis JS, Denis R, Lavoie A, et al. Trauma care regionalization: a process- outcome evaluation. J Trauma 1999;46(4):565–79–discussion579–81.

11. Newgard CD, Schmicker RH, Hedges JR, et al. Emergency medical services intervals and survival in trauma: assessment of the “golden hour” in a North American prospective cohort. Ann Emerg Med 2010;55(3):235–246.e4.

12. Buchanan IM, Coates A, Sne N. Does Mode of Transport Confer a Mortality Benefit in Trauma Patients? Characteristics and Outcomes at an Ontario Lead Trauma Hospital. CJEM 2016;:1–7.

13. McCoy CE, Menchine M, Sampson S, Anderson C, Kahn C. Emergency medical services out-of-hospital scene and transport times and their association with mortality in trauma patients presenting to an urban Level I trauma center. Ann Emerg Med 2013;61(2):167–74.

14. Committee on Trauma ACOS. Resources for optimal care of the injured patient [Internet]. httpswww.facs.orgmediafilesqualityprogramstraumavrcresourcesresourcesforopti malcare.ashx. 2014 [cited 2019 Feb 2];Available from: https://www.facs.org/~/media/files/quality%20programs/trauma/vrc%20resources/ resources%20for%20optimal%20care.ashx

15. Committee on Trauma ACOS. Optimal hospital resources for care of the seriously injured. Bulletin of American College of Surgeons 1976;(61):15–22.

16. Celso B, Tepas J, Langland-Orban B, et al. A systematic review and meta-analysis comparing outcome of severely injured patients treated in trauma centers following the establishment of trauma systems. J Trauma 2006;60(2):371–8–discussion378.

17. Lerner EB, Moscati RM. The golden hour: scientific fact or medical "urban legend"? Acad Emerg Med 2001;8(7):758–60.

18. Newgard CD, Meier EN, Bulger EM, et al. Revisiting the “Golden Hour”: An Evaluation of Out-of-Hospital Time in Shock and Traumatic Brain Injury. Ann Emerg Med 2015;66(1):30–3.

19. Harmsen AMK, Giannakopoulos GF, Moerbeek PR, Jansma EP, Bonjer HJ, Bloemers FW. The influence of prehospital time on trauma patients outcome: a systematic review. Injury 2015;46(4):602–9.

20. Wigman LD, van Lieshout EMM, de Ronde G, Patka P, Schipper IB. Trauma- related dispatch criteria for Helicopter Emergency Medical Services in Europe. Injury 2011;42(5):525–33.

21. Thomas SH, Brown KM, Oliver ZJ, et al. An Evidence-based Guideline for the air medical transportation of prehospital trauma patients. Prehosp Emerg Care 2014;18 Suppl 1:35–44.

22. Vercruysse GA, Friese RS, Khalil M, et al. Overuse of helicopter transport in the minimally injured: A health care system problem that should be corrected. J Trauma Acute Care Surg 2015;78(3):510–5.

23. Cheung BH, Delgado MK, Staudenmayer KL. Patient and trauma center characteristics associated with helicopter emergency medical services transport for patients with minor injuries in the United States. Acad Emerg Med 2014;21(11):1232–9.

24. Andruszkow H, Hildebrand F, Lefering R, Pape H-C, Hoffmann R, Schweigkofler

U. Ten years of helicopter emergency medical services in Germany: do we still need the helicopter rescue in multiple traumatised patients? Injury 2014;45 Suppl 3:S53–8.

25. Doucet J, Bulger E, Sanddal N, Fallat M, Bromberg W, Gestring M. Appropriate use of Helicopter Emergency Medical Services for transport of trauma patients. Journal of Trauma and Acute Care Surgery 2013;75(4):734–41.

26. Center for Disease Control, Prevention. Guidelines for field triage of injured patients. MMWR Morb Mortal Wkly Rep 2009;58:2–12.

27. Brown JB, Stassen NA, Bankey PE, Sangosanya AT, Cheng JD, Gestring ML. Helicopters and the civilian trauma system: national utilization patterns demonstrate improved outcomes after traumatic injury. J Trauma 2010;69(5):1030–4–discussion1034–6.

28. Sullivent EE, Faul M, Wald MM. Reduced mortality in injured adults transported by helicopter emergency medical services. Prehosp Emerg Care 2011;15(3):295– 302.

29. Stewart KE, Cowan LD, Thompson DM, Sacra JC, Albrecht R. Association of direct helicopter versus ground transport and in-hospital mortality in trauma patients: a propensity score analysis. Acad Emerg Med 2011;18(11):1208–16.

30. Ringburg AN, Thomas SH, Steyerberg EW, van Lieshout EMM, Patka P, Schipper IB. Lives saved by helicopter emergency medical services: an overview of literature. Air Med J 2009;28(6):298–302.

31. Cunningham P, Rutledge R, Baker CC, Clancy TV. A comparison of the association of helicopter and ground ambulance transport with the outcome of injury in trauma patients transported from the scene. J Trauma 1997;43(6):940–6.

32. Shatney CH, Homan SJ, Sherck JP, Ho C-C. The utility of helicopter transport of trauma patients from the injury scene in an urban trauma system. J Trauma 2002;53(5):817–22.

33. Eckstein M, Jantos T, Kelly N, Cardillo A. Helicopter transport of pediatric trauma patients in an urban emergency medical services system: a critical analysis. J Trauma 2002;53(2):340–4.

34. Bledsoe BE, Wesley AK, Eckstein M, Dunn TM, O'Keefe MF. Helicopter scene transport of trauma patients with nonlife-threatening injuries: a meta-analysis. J Trauma 2006;60(6):1257–65–discussion1265–6.

35. Delgado MK, Staudenmayer KL, Wang NE, et al. Cost-effectiveness of helicopter versus ground emergency medical services for trauma scene transport in the United States. Ann Emerg Med 2013;62(4):351–364.e19.

36. Bledsoe BE, Smith MG. Medical helicopter accidents in the United States: a 10- year review. J Trauma 2004;56(6):1325–8–discussion1328–9.

37. Gomez D, Haas B, Doumouras AG, et al. A population-based analysis of the discrepancy between potential and realized access to trauma center care. Ann Surg 2013;257(1):160–5.

38. Health OMO, Care LT. Field trauma triage and air ambulance utilization [Internet]. httpwww.health.gov.on.caenproprogramsehsdocsehstrainingblltnen.pdf. 2014 [cited 2018 Dec 20];Available from: http://www.health.gov.on.ca/en/pro/programs/ehs/docs/ehs_training_blltn113_en.p df

39. Gomez D, Berube M, Xiong W, et al. Identifying targets for potential interventions to reduce rural trauma deaths: a population-based analysis. J Trauma 2010;69(3):633–9.

40. Newgard CD, Rudser K, Hedges JR, et al. A critical assessment of the out-of- hospital trauma triage guidelines for physiologic abnormality. J Trauma 2010;68(2):452–62.

41. Brown JB, Forsythe RM, Stassen NA, Gestring ML. The National Trauma Triage Protocol: can this tool predict which patients with trauma will benefit from helicopter transport? J Trauma Acute Care Surg 2012;73(2):319–25.

42. Lehmann RK, Arthurs ZM, Cuadrado DG, Casey LE, Beekley AC, Martin MJ. Trauma team activation: simplified criteria safely reduces overtriage. Am J Surg 2007;193(5):630–4–discussion634–5.

43. van Rein EAJ, Houwert RM, Gunning AC, Lichtveld RA, Leenen LPH, van Heijl M. Accuracy of prehospital triage protocols in selecting severely injured patients: A systematic review. J Trauma Acute Care Surg 2017;83(2):328–39.

44. Ciesla DJ, Pracht EE, Tepas JJ, et al. Measuring trauma system performance: Right patient, right place-Mission accomplished? J Trauma Acute Care Surg 2015;79(2):263–8.

45. Staudenmayer KL, Hsia RY, Mann NC, Spain DA, Newgard CD. Triage of elderly trauma patients: a population-based perspective. J Am Coll Surg 2013;217(4):569– 76.

46. Haas B, Stukel TA, Gomez D, et al. The mortality benefit of direct trauma center transport in a regional trauma system: a population-based analysis. J Trauma Acute Care Surg 2012;72(6):1510–5–discussion1515–7.

47. Nolan B, Ackery A, Nathens A, Sawadsky B, Tien H. Canceled to Be Called Back: A Retrospective Cohort Study of Canceled Helicopter Emergency Medical Service Scene Calls That Are Later Transferred to a Trauma Center. Air Med J

2018;37(2):108–14.

48. Giannakopoulos GF, Bloemers FW, Lubbers WD, et al. Criteria for cancelling helicopter emergency medical services (HEMS) dispatches. Emerg Med J 2012;29(7):582–6.

49. Tignanelli CJ, Vander Kolk WE, Mikhail JN, Delano MJ, Hemmila MR. Non- compliance with ACS-COT recommended criteria for full trauma team activation is associated with undertriage deaths. J Trauma Acute Care Surg 2017;84(2):287– 94.

50. Gomez D, Haas B, De Mestral C, et al. Institutional and provider factors impeding access to trauma center care: an analysis of transfer practices in a regional trauma system. J Trauma Acute Care Surg 2012;73(5):1288–93.

51. Gomez D, Haas B, Larsen K, et al. A novel methodology to characterize interfacility transfer strategies in a trauma transfer network. J Trauma Acute Care Surg 2016;81(4):658–65.

52. Austin PC, Schull MJ. Quantile regression: a statistical tool for out-of-hospital research. Acad Emerg Med 2003;10(7):789–97.

53. Schull MJ, Morrison LJ, Vermeulen M, Redelmeier DA. Emergency department overcrowding and ambulance transport delays for patients with chest pain. CMAJ 2003;168(3):277–83.

54. Do YK, Foo K, Ng YY, Ong MEH. A quantile regression analysis of ambulance response time. Prehosp Emerg Care 2013;17(2):170–6.

55. Nolan B, Ackery A, Mamakwa S, et al. Care of the Injured Patients at Nursing Stations and during Air Medical Transport. Air Med J 2018;37(3):161–4.

56. Grocott M, Montgomery H, Vercueil A. High-altitude physiology and pathophysiology: implications and relevance for intensive care medicine. Crit Care 2007;11(1):203.

57. Braude D, Tutera D, Tawil I, Pirkl G. Air transport of patients with pneumothorax: is tube thoracostomy required before flight? Air Med J 2014;33(4):152–6.

58. Harrington DT, Connolly M, Biffl WL, Majercik SD, Cioffi WG. Transfer times to definitive care facilities are too long: a consequence of an immature trauma system. Ann Surg 2005;241(6):961–6–discussion966–8.

59. Pham H, Puckett Y, Dissanaike S. Faster on-scene times associated with decreased mortality in Helicopter Emergency Medical Services (HEMS) transported trauma patients. Trauma Surg Acute Care Open 2017;2(1):e000122.

60. Lockwood J, Ackery A. Improving air medical transport of the trauma patient from

the ground. CJEM 2014;16(3):243–6.