Canadian Journal of Rural Medicine

ORIGINAL ARTICLE
Year
: 2019  |  Volume : 24  |  Issue : 3  |  Page : 83--91

Standardised early warning scores in rural interfacility transfers: A pilot study into their potential as a decision-making aid


Andrew Stanley1, Holly Buhler2, Brent Hobbs3, Jude Kornelsen4, Scott Lamont3, Rebecca Kaus3, Kari Grant3, Stefan Grzybowski4,  
1 Faculty of Medicine, Southern Medical Program, University of British Columbia, Okanagan campus, Kelowna, BC, Canada
2 Research Department, Interior Health Authority, Kelowna, BC, Canada
3 Patient Transport Services, Interior Health Authority, Kelowna, BC, Canada
4 The Centre for Rural Health Research, Department of Family Practice, University of British Columbia Vancouver, BC, Canada

Correspondence Address:
MMedSc, MGIS, CE Holly Buhler
Research Department, Interior Health Authority, Kelowna, BC
Canada

Abstract

Introduction: While 12.4% of British Columbians live rurally, only 2.0% of specialists practise rurally, making interfacility transport of high-acuity patients vital. Decision-making aids have been identified as a way to improve the interfacility transfer process. We conducted a pilot study to explore the potential of the Standardised Early Warning Score (SEWS) as a decision-making aid for staff at sending facilities. Methods: SEWSs were calculated from a database of 418 transfers from sending facilities in rural, small and medium population centres to larger receiving facilities. The SEWSs were compared against one another over time using McNemar's and the Wilcoxon signed-ranks tests. The SEWSs were then tested for their association with six outcomes using Pearson's or Fisher's Chi-squared test and the Mann–Whitney U-test. Results: While at the sending facility, both the number of SEWSs that was four or greater and the average SEWS decreased over time (P < 0.001 for both). A first SEWS of four or greater was predictive of more intervention categories during transport (P = 0.047), an adverse event during transport (P = 0.004), an adverse event within 30 min of arrival at the receiving facility (P = 0.004) and death before discharge from the receiving facility (P = 0.043) but not deterioration during transport, or the length of stay at the receiving facility. Conclusion: Overall, the performance of the SEWS in the context of rural interfacility transport suggests that the tool will have utility in supporting decision-making.



How to cite this article:
Stanley A, Buhler H, Hobbs B, Kornelsen J, Lamont S, Kaus R, Grant K, Grzybowski S. Standardised early warning scores in rural interfacility transfers: A pilot study into their potential as a decision-making aid.Can J Rural Med 2019;24:83-91


How to cite this URL:
Stanley A, Buhler H, Hobbs B, Kornelsen J, Lamont S, Kaus R, Grant K, Grzybowski S. Standardised early warning scores in rural interfacility transfers: A pilot study into their potential as a decision-making aid. Can J Rural Med [serial online] 2019 [cited 2021 Feb 26 ];24:83-91
Available from: https://www.cjrm.ca/text.asp?2019/24/3/83/261315


Full Text

 Introduction



Rural British Columbians, defined as those living outside census metropolitan areas or census agglomerations, make up 12.4% of the provincial population,[1] and while 11.6% of family physicians practise rurally, only 2% of specialists in the province do.[2] Only 5% of rural BC hospitals have an intensive care unit (ICU) and just 3% have 24/7 access to a computed tomography (CT) scanner, making interfacility transport of paramount importance in this province.[3] One study found that 31% of trauma patients presenting to rural facilities in Northwestern BC required secondary transfer for higher level care,[4] a pattern repeated in rural settings across the country.[5],[6]

A qualitative analysis of the interfacility transfer process from three hospitals in the United States, ranging from rural to urban, found that the availability of a decision-making aid improved identification of patients requiring transfer and streamlined the process of arranging the transfer. The authors of that study suggest that increased use of decision-making aids could reduce the overall time required for the transfer process,[7] an idea which has also been identified in a recent review of rural transport structures in BC.[8]

An example of a decision-making aid already implemented in other settings is the Early Warning Score (EWS). EWSs are tools allowing a set of vital signs to be condensed into a single numerical value, simplifying determination of a patient's severity and risk of subsequent adverse events. Vital sign components within an acceptable range are assigned the value zero, with increasingly large values assigned as the components become increasingly deranged. The component scores are then summed, such that an EWS higher than zero is associated with an increased risk of adverse events.[9] As the EWS rises, so does the risk, and a cut-off score is used to determine when the patient's risk has reached a level which requires intervention.

EWSs are most commonly used in track and trigger systems to recognise deteriorating patients on medical wards,[10] and their ability to predict outcomes including cardiac arrest, death, increased length of stay in hospital and upgrade to intensive care has been well validated in this setting.[11],[12],[13] In particular, a prospective study of 848 patients admitted to a high-acuity ward in the United Kingdom found that EWSs of four or greater were predictive of increased mortality and length of stay.[11]

The utility of EWSs in other settings has also been examined. A 2015 European study of 300 emergency room patients[14] found that an EWS was able to predict risk of admission, length of stay and death with greater accuracy than traditional triage methods. A 2012 retrospective chart review from the UK examined the ability of EWSs combined with clinical judgment in the prehospital setting to predict adverse patient outcomes.[15] The combination was found to have a sensitivity of 72.4% and specificity of 84.8% using a cut-off of four. Another retrospective chart review from the UK found that EWSs of seven or greater in the pre-hospital environment were associated with an increased risk of 48-h mortality, 30-day mortality and ICU admission.[16] Finally, two studies of transfers escorted by either nurses or physicians between large hospitals in Hong Kong found that higher EWSs are associated with physiological deterioration during transport.[17],[18]

Following from previous, unpublished work by some of the present authors,[19],[20] here we have conducted a pilot study examining the ability of an EWS to predict a selection of adverse outcomes in patients who presented to sending facilities in rural areas, as well as small and medium population centres, throughout Southeastern BC and were eventually transferred to larger receiving facilities for higher level care. Establishing the validity of EWSs in this setting is the first step in creating an EWS-based decision-making tool which can be used by both sending facilities and transport personnel.

 Methods



We conducted a secondary analysis using a dataset created via retrospective chart review. The dataset contains a broad range of information from acutely ill patients transferred between facilities spread across 215,000 square kilometres of the southern BC interior (population approximately 743,000).[21] All patients were transferred to one of the five receiving facilities for higher level care than was available locally and were transported by one of the four methods: ground basic life support (BLS) ambulance crew, air and ground critical care paramedic (CCP) ambulance crew, ground BLS ambulance crew with a doctor and nurse escort from the sending facility or a dedicated transfer team made up of a Registered Nurse (RN) and Registered Respiratory Therapist working with a ground BLS ambulance crew, known as a High Acuity Response Team (HART).[22],[23]

Criteria for inclusion were being 16 years or older on admission, direct admission to the sending site emergency department, a Canadian Triage and Acuity Scale (CTAS) score of one or two at the receiving facility, transport between 20th November 2010 and 6th October 2014 and categorisation as neurological, cardiac, respiratory, sepsis or trauma. Eligible records were identified through an electronic patient database and selected randomly for data extraction, which was carried out at the five receiving facilities by a researcher and a clinician. When information was unavailable in the original patient care record, a RN and a paramedic, both with experience in interfacility transfers, attempted to infer it. If inference was not possible, the entry was left blank with no further data imputation. Validation was carried out on 20% of the data collected from one of the receiving facilities. The charts for validation were selected at random, and two research team members then compared them against the information in the database to confirm accuracy. Data collection and cleaning was completed using Microsoft Excel.

Standardised EWSs (SEWSs) were calculated from the patients' vital signs as per the method described by Paterson et al.[11] In brief, numerical scores were assigned to the patients' respiratory rate, oxygen saturation, temperature, blood pressure, heart rate and level of consciousness, with normal values scored zero and abnormal values scored from one to three, depending on the degree of abnormality. The sum of these six individual scores produced the SEWS. Patients were then grouped according to whether their SEWSs were <4 or were four or greater. Within-patient comparisons were made between the SEWS calculated from the first and last vital signs recorded by the sending facility, as well as the SEWSs calculated from the last vital signs recorded by the sending facility and the first recorded by the receiving facility. McNemar's test was used to determine if the number of patients with a SEWS of four or greater changed between these times, and Wilcoxon signed-ranks tests were used to determine if the average SEWS changed between these times.

The SEWS study groups (i.e., those with scores <4 and those with scores four or greater) were then compared against the following outcomes: the number of intervention categories during transport, an adverse event during transport, deterioration during transport, an adverse event within 30 min of arrival at the receiving facility, the length of stay at the receiving facility and death before discharge from the receiving facility [[Box 1] for detailed definitions].[INLINE:1]

Chi-square tests were used to investigate the relationship between the SEWS and the following outcomes: an adverse event during transport, deterioration during transport, an adverse event within 30 minutes of arrival at the receiving facility, or death before discharge from the receiving facility. Pearson's chi-square test was used when the expected incidence in the sample was five or greater and two-tailed Fisher's exact test when the expected incidence was below five. The relationship between SEWS and outcome was described using relative risk. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated for statistically significant results. The number of intervention categories required during transport and the length of stay at the receiving facility were both investigated using Mann–Whitney U-tests. The relationship between SEWS and outcome was described by comparing means.

All statistical analyses were carried out in IBM SPSS Statistics for Windows, version 23.0 (IBM Corp., Armonk, NY, USA), and P < 0.05 was considered as statistically significant.

The original data collection was carried out under an ethics application approved by the University of British Columbia in harmonisation with Interior Health.

 Results



The characteristics of the transfers contained within the dataset are described in [Table 1][24],[25],[26],[27],[28],[29],[30],[31],[32] [[Box 2] for an explanation of the 2011 Canadian census definitions].[33],[34] Of the 418 patient entries in the dataset, 98 (23% of total) were missing one or more data points required for statistical analysis and were subsequently excluded on a test-by-test basis.{Table 1}[INLINE:2]

During the time between the first and final SEWS calculated at the sending site, the proportion of scores which were four or greater decreased from 24.8% to 14.8% (P < 0.001). Between the second and third SEWS, there was no significant change (14.8% and 17.0%, respectively, P = 0.358) [Figure 1].{Figure 1}

A first SEWS of four or greater, calculated from the first vital signs recorded at the sending site, was associated with more intervention categories during transport (1 vs. 0, P = 0.047), an adverse event during transport (P = 0.004, risk ratio [RR] = 2.41, 95% confidence interval [CI] 1.32–4.39, sensitivity = 42%, specificity = 79%, PPV = 17%, NPV = 93%), an adverse event within 30 min of arrival at the receiving facility (P = 0.004, RR = 5.85, 95% CI 1.75–19.61, sensitivity = 64%, specificity = 78%, PPV = 7%, NPV = 99%) and death before discharge from the receiving facility (P = 0.043, RR = 2.29, 95% CI 1.01–5.18, sensitivity = 41%, specificity = 78%, PPV = 10%, NPV = 96%) as compared to a SEWS <4. There was no difference in deterioration during transport (P = 0.105, RR = 2.29, 95% CI 0.94–2.06) or the length of stay at the receiving facility (4.5 vs. 4, P = 0.236) between a SEWSs of four or greater or <4 [Table 2] and [Table 3].{Table 2}{Table 3}

 Discussion



The vast size of BC combined with ongoing regionalisation of health care has meant that the services required by rural patients with complex health problems are often not available at the closest hospital.[8],[35] Accordingly, rural health care in BC is heavily dependent on interfacility transfers provided by ground or air ambulances.[3],[36] This system is well used, with one study showing that 31% of trauma patients in Northwestern BC required interfacility transport to receive higher care.[4] The need for improvement in this system, however, has been identified by two recent works[8],[23] and is signalled by the fact that rural British Columbians are still more likely to die after a traumatic injury than their non-rural counterparts.[4],[37]

One possible approach to improving the outcomes of patients who require interfacility transport to receive higher level care is the implementation of decision-making aids for various steps in the transfer process.[7] While there are a variety of predictive tools to choose from, some, such as the Acute Physiology And Chronic Health Evaluation II score,[38] require information from laboratory investigations that may not be readily available in rural BC hospitals or during transport.[3] EWSs, on the other hand, require only the patient's vital signs, which are easily obtained at any health facility, or by any transport crew. Here, we have evaluated the ability of the SEWS[11] to predict a selection of outcome measures among patients who underwent interfacility transfers from sending facilities throughout the southern BC interior.

While numerous studies have used the area under the receiver operating characteristic curve to identify the optimal cut-off score to use with a particular EWS,[15],[16],[17],[39],[40],[41],[42] the objective of this pilot study was to evaluate the usefulness of a previously validated EWS and cut-off value. The SEWS was chosen because it has performed well when compared against other EWSs[41] and could be calculated from the available data.

We began by examining the proportion of patients who had a SEWS of four or greater over three time periods: two at the sending site and one at the receiving facility. The decrease in proportion between the first and second time periods from 24.8% to 14.8% represents a drop of 40.3%. We view this as an indication that sending facilities in rural areas, as well as small and medium population centres, have an important ability to improve a patient's condition before transport despite not being able to provide the full level of care required. This interpretation is consistent with the findings of Belway et al.[43] who examined the outcomes of patients transferred from smaller hospitals to the ICU or critical care unit (CCU) of larger hospitals in BC. They found increased time at the sending facility was associated with a decreased length of stay in the ICU or CCU, an outcome they attributed to improved pre-transport stabilisation. The potential for the sending site to improve the patient's condition before transport also highlights the importance of education and support for rural staff, previously identified by others.[4],[8] Our failure to detect a difference between the final SEWS from the sending site and the first SEWS from the receiving site suggests that crews are maintaining the improvements achieved at the sending facility and adequately managing new problems as they arise.

A SEWS of four or greater was associated with all outcomes except deterioration during transport and length of stay at the receiving facility. The definition of deterioration during transport used here [Box 1] was a novel one we have not seen elsewhere in the literature. Because the SEWSs used to calculate deterioration were temporally proximate to the final SEWS from the sending site and the first SEWS from the receiving site, it is not surprising that this measure also failed to show a change over time. It seems likely that the transport crews are successfully correcting any deterioration of the patient's condition as it happens, and as a result, the final transport SEWS ends up very similar to the initial transport SEWS, regardless of brief variations in between. The inability of the SEWS to predict length of stay at the receiving facility was unexpected, as it has previously been shown to perform well on this measure in a selection of 848 patients admitted to a medical/surgical combined assessment area.[11] It is possible that it was unable to do so here because of complicating factors unique to rural interfacility transfers. These may include the lack of clinical resources required to support long-distance repatriations to the communities represented in this dataset, or the geography over which the transfers must take place.[8]

The association between a SEWS of four or greater and both adverse events during transport and the number of interventions required is consistent with the findings of Wong et al. in an urban setting.[17] They examined the ability of the Modified EWS (MEWS) to predict complications during interfacility transfers originating in the accident and emergency department of Pok Oi Hospital in Hong Kong, and showed that the MEWS calculated from vital signs obtained at triage were best able to do this. Here, we have found the relative risk of an adverse event during transport for patients with a SEWS of four or greater to be 2.41, with a sensitivity of 42% and specificity of 79%. These values are similar to those found by Lee et al.[18] using the MEWS with a cut-off score of 3 to predict deterioration during interfacility transfers originating in the emergency room of Alice Ho Miu Ling Nethersole Hospital in Hong Kong (43% and 78%, respectively). With 68% of rural BC hospitals more than 300 km from a level II trauma centre,[3] prior knowledge of a patient's risk for these outcomes could form a valuable part of the transport team's pre-planning, as well as play a role in decisions about what level of transport team to activate.

The SEWS was also able to predict two receiving facility outcomes. An adverse event within 30 min of arrival at the receiving facility was the outcome most strongly associated with a SEWS of four or greater, with a relative risk of 5.85, a sensitivity of 64% and a specificity of 78%. These values are similar to those reported by Fullerton et al. (71% and 76%, respectively)[15] using the MEWS with a cut-off of 3 in the pre-hospital environment to predict an adverse event within 24 h of arrival. Patients with a SEWS of four or greater also had an elevated risk of death before discharge from the receiving facility, with a relative risk of 2.29, sensitivity of 41% and specificity of 78%. While not as impressive as the 8-fold increase in mortality among these patients observed by Paterson et al.,[11] any advanced warning about an outcome of this severity can be helpful in making decisions about patient care.

Overall, the performance of the SEWS here is similar to other EWSs in other patient transport environments. Because this study uses a database of patients who were transferred from one facility to another, clinical judgment had already been applied to recognise the need for higher level care and the risk of deterioration. What we have demonstrated is that, among these pre-selected patients, using the SEWS with a cut-off of four offers a reasonably specific means of identifying those with the greatest and most immediate need for close monitoring and care. This points towards several potential uses for the SEWS in the rural setting, including confirming clinical gestalt about the urgency of a transfer, aiding in decisions about which type of transport crew to dispatch and alerting transport crews and receiving facilities to those patients likely to require the most care.

Limitations

As a pilot study, the sample size here (n = 418) is relatively small, limiting its power, particularly when examining rare outcomes such as death before discharge.

While the modes of transport included in this dataset represent the full spectrum of acute interfacility transfers within the study area, the frequency of each within the study area was not determined, so the proportions in this sample are likely not representative. This has the most bearing on the number of intervention categories required during transport because these were generally above the scope of practice of ground BLS crews, introducing the possibility of values >0 being inappropriately rare or common.

The data used in this study were extracted from clinical records rather than documents prepared specifically for research. Consequently, there were variations in both the recording style and completeness of the information. While experienced clinicians attempted to infer unclear or missing values where reasonable, this was often not possible, requiring cases to be excluded for particular tests, and it is unknown what effect this has had on the results.

Moving forward, it is important to validate these findings with a larger, prospective study conducted in the same environment and utilising a random sampling of all comers to sending facilities in rural areas, as well as small and medium population centres. This will allow evaluation of the SEWS's ability to inform the initial decision to transport patients for a higher level of care in addition to the subsequent decisions considered here.

 Conclusion



Here, we have examined the potential of the SEWS to inform decisions about patient care and disposition when transferring patients from sending facilities in rural areas, as well as small and medium population centres, to larger receiving facilities, to receive a higher level of care. Within the limitations of this pilot study, we have found its performance to be similar to that of other EWSs in urban interfacility or pre-hospital environments. This opens the door to using the SEWS as part of the planning process and risk stratification for transfers from these sending facilities. In addition, we have found that the average SEWS generally decreases over the time patients spend at the sending facility, highlighting the important role sending facility staff play in patient stabilisation.

Acknowledgement: We would like to thank Dr. Lawrence Takeuchki M. D., FRCPC, who originally proposed the potential utility of an EWS in making interfacility transport decisions to this team and Jodi Siever MSc, who provided invaluable statistical consultation for this project.

Financial support and sponsorship: Funding for this project was provided by the B.C.'s Interior Health Authority.

Conflicts of interest: There are no conflicts of interest.

References

1Statistics Canada. Focus on Geography Series, 2011 Census – Province of British Columbia. Statistics Canada Catalogue No. 98-310-XWE2011004. Statistics Canada; 2012. Available from: http://www12.statcan.gc.ca/census-recensement/2011/as-sa/fogs-spg/Facts-pr-eng.cfm?Lang=Eng&GK=PR&GC=59. [Last accessed on 2017 May 02].
2Canadian Medical Association. Physicians Within and Outside Census Metropolitan Areas (CMA) and Census Agglomerations (CA) – 2015. Ottawa, ON: Canadian Medical Association; 2015. Available from: https://www.cma.ca/Assets/assets-library/document/en/advocacy/13cma_ca_outside.pdf. [Last accessed on 2017 May 02].
3Fleet R, Audette LD, Marcoux J, Villa J, Archambault P, Poitras J. Comparison of access to services in rural emergency departments in Quebec and British Columbia. CJEM 2014;16:437-48.
4Simons R, Brasher P, Taulu T, Lakha N, Molnar N, Caron N, et al. A population-based analysis of injury-related deaths and access to trauma care in rural-remote Northwest British Columbia. J Trauma 2010;69:11-9.
5Rourke JT, Kennard MA. Emergency patient transfers from rural hospitals: A regional study. Can J Emerg Med 2001;3:296-301.
6De Freitas TL, Spooner GR, Szafran O. Admissions and transfers from a rural emergency department. Can Fam Physician 1998;44:789-95.
7Bosk EA, Veinot T, Iwashyna TJ. Which patients and where: A qualitative study of patient transfers from community hospitals. Med Care 2011;49:592-8.
8Kornelsen J, McCartney K, Newton L, Butt E, Sax M. Rural Patient Transport and Transfer : Findings from a Realist Review. Vancouver, BC; 2016. Available from: https://www.divisionsbc.ca/CMSMedia/WebPageRevisions/PageRev-12516/Rural Patient Transport Report-FINAL.pdf. [Last accessed on 2017 Dec 28].
9Royal College of Physicians. National Early Warning Score – Standardising the Assessment of Acute-Illness Severity in the NHS. Report of a Working Party. London: Royal College of Physicians; 2012. p. 47.
10McNeill G, Bryden D. Do either early warning systems or emergency response teams improve hospital patient survival? A systematic review. Resuscitation 2013;84:1652-67.
11Paterson R, MacLeod DC, Thetford D, Beattie A, Graham C, Lam S, et al. Prediction of in-hospital mortality and length of stay using an early warning scoring system: Clinical audit. Clin Med (Lond) 2006;6:281-4.
12Kellett J, Kim A. Validation of an abbreviated vitalpac™ early warning score (ViEWS) in 75,419 consecutive admissions to a canadian regional hospital. Resuscitation 2012;83:297-302.
13Smith GB, Prytherch DR, Meredith P, Schmidt PE, Featherstone PI. The ability of the national early warning score (NEWS) to discriminate patients at risk of early cardiac arrest, unanticipated intensive care unit admission, and death. Resuscitation 2013;84:465-70.
14Alam N, Vegting IL, Houben E, van Berkel B, Vaughan L, Kramer MH, et al. Exploring the performance of the national early warning score (NEWS) in a European emergency department. Resuscitation 2015;90:111-5.
15Fullerton JN, Price CL, Silvey NE, Brace SJ, Perkins GD. Is the modified early warning score (MEWS) superior to clinician judgement in detecting critical illness in the pre-hospital environment? Resuscitation 2012;83:557-62.
16Silcock DJ, Corfield AR, Gowens PA, Rooney KD. Validation of the national early warning score in the prehospital setting. Resuscitation 2015;89:31-5.
17Wong YK, Lui CT, Li KK, Wong CY, Lee MM, Tong WL, et al. Prediction of en-route complications during interfacility transport by outcome predictive scores in ED. Am J Emerg Med 2016;34:877-82.
18Lee LL, Yeung KL, Lo WY, Lau YS, Tang SY, Chan JT. Evaluation of a simplified therapeutic intervention scoring system (TISS-28) and the modified early warning score (MEWS) in predicting physiological deterioration during inter-facility transport. Resuscitation 2008;76:47-51.
19Takeuchi L, Watson D, Kaus R, Brayman C, Hobbs B, Hill W, et al. Standardized Early Warning Scores in Trauma. Whistler, BC: Trauma Association of Canada Scientific Conference; 2013.
20Hobbs B, Kornelsen J, Grzybowski S, Kaus R, Grant K, Lamont S, et al. Mixed-Methods Evaluation of the Interior Health Authority “High Acuity Response Team.” Winnipeg, MB: Dynamics of Critical Care Conference, Canadian Association of Critical Care Nurses; 2015.
21Interior Health Authority. Quick Facts – About Us – Interior Health. Interior Health Authority; 2017. Available from: https://www.interiorhealth.ca/AboutUs/QuickFacts/Pages/default.aspx. [Last accessed on 2017 May 08].
22Brayman C, Hobbs B, Hill W, Watson D., Kaus R, Lamont S, et al. ICU Without walls-interprofessional high acuity response teams (HARTs) improve access to higher level of care in rural and remote communities. Can J Respir Ther 2012;48:14-9.
23Kornelsen J, Hobbs B, Buhler H, Kaus R, Grant K, Lamont S. High acuity rural transport: findings from a qualitative investigation. Rural Remote Health 2018;18:4316.
24Statistics Canada. Population Counts, for Census Metropolitan Areas, Census Agglomerations, Population Centres and Rural Areas, 2011 Census. Salmon Arm: Statistics Canada; 2016. Available from: https://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=209&CMA=920&S=51&O=A& RPP=25. [Last accessed on 2018 Oct 19].
25Statistics Canada. Population Counts, for Census Metropolitan Areas, Census Agglomerations, Population Centres and Rural Areas, 2011 Census. Penticton: Statistics Canada; 2016. Available from: https://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=209&CMA=913&S=51&O=A& RPP=25. [Last accessed on 2018 Oct 19].
26Statistics Canada. Population Counts, for Census Metropolitan Areas, Census Agglomerations, Population Centres and Rural Areas, 2011 Census. Vernon: Statistics Canada; 2016. Available from: https://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=209&CMA=918&S=51&O=A& RPP=25. [Last accessed on 2018 Oct 19].
27Statistics Canada. Population Counts, for Census Metropolitan Areas, Census Agglomerations, Population Centres and Rural Areas, 2011 Census. Cranbrook: Statistics Canada; 2016. Available from: https://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=209&CMA=905&&S=51&O=A& RPP=25. [Last accessed on 2018 Oct 19].
28Statistics Canada. Population Counts, for Census Metropolitan Areas, Census Agglomerations, Population Centres and Rural Areas, 2011 Census. Williams Lakel: Statistics Canada; 2016. Available from: https://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=209&CMA=950&S=51&O=A& RPP=25. [Last accessed On 2018 Oct 19].
29Statistics Canada. Population Counts, for Census Metropolitan Areas, Census Agglomerations, Population Centres and Rural Areas, 2011 Census. Kamloops: Statistics Canada; 2016. Available from: https://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=209&CMA=925&S=51&O=A& RPP=25. [Last accessed on 2018 Oct 19].
30Statistics Canada. Population and Dwelling Counts, for Canada, Provinces and Territories, Census Metropolitan Areas and Census Agglomerations, 2011 and 2006 Censuses. Statistics Canada; 2016. Available from: https://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=202&SR=1&S=3&O=D&RPP=9999& CMA=0&PR=59. [Last accessed on 2018 Oct 19].
31Statistics Canada. Population and Dwelling Counts, for Canada, Provinces and Territories, and Population Centres, 2011 and 2006 Censuses. Statistics Canada; 2016. Available from: http://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=802&SR=1&S=3& O=D&RPP=9999&CMA=0&PR=59. [Last accessed on 2017 May 08].
32Statistics Canada. Population and Dwelling Counts, for Canada, Provinces and Territories, and Census Subdivisions (Municipalities), 2011 and 2006 Censuses. Statistics Canada; 2016. Available from: http://www12.statcan.gc.ca/census-recensement/2011/dp-pd/hlt-fst/pd-pl/Table-Tableau.cfm?LANG=Eng&T=302&SR=1&S=51&O=A&RPP=9999& PR=59. [Last accessed on 2017 May 08].
33Statistics Canada. Archived – From Urban Areas to Population Centres. Statistics Canada; 2011. Available from: http://www.statcan.gc.ca/eng/subjects/standard/sgc/notice/sgc-06./. [Last accessed on 2017 May 08].
34Statistics Canada. Census Metropolitan Area (CMA) and Census Agglomeration (CA). Statistics Canada; 2015. Available from: https://www12.statcan.gc.ca/census-recensement/2011/ref/dict/geo009-eng.cfm. [Last accessed on 2018 Oct 19].
35Schuurman N, Fiedler RS, Grzybowski SC, Grund D. Defining rational hospital catchments for non-urban areas based on travel-time. Int J Health Geogr 2006;5:43.
36Hameed SM, Schuurman N, Razek T, Boone D, Van Heest R, Taulu T, et al. Access to trauma systems in Canada. J Trauma 2010;69:1350-61.
37Bell N, Simons RK, Lakha N, Hameed SM. Are we failing our rural communities? Motor vehicle injury in British Columbia, Canada, 2001-2007. Injury 2012;43:1888-91.
38Knaus WA, Draper EA, Wagner DP, Zimmerman JE. APACHE II: A severity of disease classification system. Crit Care Med 1985;13:818-29.
39Subbe CP, Kruger M, Rutherford P, Gemmel L. Validation of a modified early warning score in medical admissions. QJM 2001;94:521-6.
40Prytherch DR, Smith GB, Schmidt PE, Featherstone PI. ViEWS – Towards a national early warning score for detecting adult inpatient deterioration. Resuscitation 2010;81:932-7.
41Jarvis SW, Kovacs C, Briggs J, Meredith P, Schmidt PE, Featherstone PI, et al. Are observation selection methods important when comparing early warning score performance? Resuscitation 2015;90:1-6.
42Singh JM, Gunz AC, Dhanani S, Aghari M, MacDonald RD. Frequency, composition, and predictors of in-transit critical events during pediatric critical care transport. Pediatr Crit Care Med 2016;17:984-91.
43Belway D, Dodek PM, Keenan SP, Norena M, Wong H. The role of transport intervals in outcomes for critically ill patients who are transferred to referral centers. J Crit Care 2008;23:287-94.