Abstract
Objective
To describe remote triage of ‘potentially’ critically ill or injured children in a western Canadian province and to examine the associated factors with ‘missings’ in vital sign items recorded in centralized telephone triage consultations.
Methods
This is a provincial-wide prospective cohort study. We included all children under 17 years of age consulted through the central transport coordination centres in Alberta from June 2016 to July 2017. We labeled a value as ‘missing’ when the actual value was not identified in the audio records.
Results
In total, 429 cases were included in this study. The median duration of triage calls was 6.8 minutes. Although the patients’ demographics and primary diagnoses were similar, backgrounds of the referring physicians and hospitals were significantly different between the two cohorts (i.e., patients referred to Calgary versus Edmonton). The proportion of ‘missings’ among the vital sign items varied significantly, in which capillary refilling time (60%), pupils (86%), Glasgow Coma Scale (GCS) (79%), and level of respiratory effort (50%) were not well recorded, whereas heart rate (proportion of ‘missings’: 12%), SpO2 (20%), and respiratory rate (26%) were recorded reasonably well. The lower proportion of ‘missings’ was observed in older aged patients for several vital sign items including systolic blood pressure and GCS.
Conclusions
The proportion of missing vital signs recorded varied significantly. The ‘missings’ could be associated with referring physician’s background and patients’ demographics such as ‘age’ that should be considered for the improvement of triage quality in the future.
Keywords: Paediatric critical care, Quality, Triage, Vital signs
Alberta has a population of about 4.3 million people in 2017, including approximately 940,000 children under 17 years of age. Paediatric intensive care units (PICU) in Edmonton and Calgary provide critical care support for the province of Alberta as well as for the Northwest Territories, eastern Yukon, and western Nunavut. This large geographic area creates a challenge for our paediatric retrieval system. Regional patient transport in Alberta is coordinated by RAAPID (Referral, Access, Advice, Placement, Information and Destination). RAAPID connects health care practitioners from remote centres (i.e., clinics, regional hospitals), with specialists in tertiary centres in Edmonton (Stollery Children’s Hospital [STOL]) and Calgary (Alberta Children’s Hospital [ACH]) by tele-consult. Both PICUs operate dedicated PICU transport teams, which are staffed mainly by respiratory therapists and nurses, all with extensive experience and training in the assessment and management of critically ill or injured children (1). Both teams usually send physician-less PICU transport teams based upon discussion during the triage calls; however, physicians trained in either paediatric critical care, anaesthesiology, or emergency medicine occasionally accompany the team (2).
Remote patient triage in Alberta currently relies on individual physician practice often in the absence of a systematic approach to patient assessment to support triage decision-making. The proper triage of critically ill or injured children is contingent upon appropriate categorization and referral. Several paediatric triage scoring systems have been evaluated for their effectiveness in paediatric transport triage (3–5). Objective patient triage tools might help in making consistent and efficient transport triage decisions. These tools may help us not only with decisions regarding transport team composition, but also with determining the disposition of patients (2).
To achieve consistent and efficient decision-making, either with or without applying a triage scoring system, it is necessary to understand the challenges in our current remote triage practice. In this study, we described the remote triage of potentially critically ill or injured children in a western Canadian province; specifically, we examined the associated factors with vital signs that were not recorded, that is, ‘missings’.
METHODS
Study population
We included all referral calls coordinated through RAAPID involving children under 17 years of age, between health care professionals in peripheral hospitals, clinics, and tertiary care paediatric intensivists via RAAPID between June 2016 and July 2017.
Data sources
We used RAAPID summary transcripts and audio record files. The RAAPID summary transcripts included patients’ identifiers, summaries of each transport referred through RAAPID including the name of the physician involved, modality of transport, and a brief summary note of triage call/discussion. The audio-record files contained all the conversations among practitioners (i.e., intensivists, PICU trainees, referring physicians, transport nurses). The duration of the triage conversation defined as the time when the paediatric intensivists (i.e., triage physicians) joined the consult conversations to the end of discussion about the patient’s clinical conditions. We excluded the time needed to address transport modality or weather. These data were provided prospectively and monthly by RAAPID over the study period. We also accessed the website of College of Physician and Surgeons of Alberta (CPSA) and collected the publically available data regarding referring physicians’ educational background and practice history (6).
Statistical analysis
We first described the demographics of patients and the triage calls, with comparisons between the two cohorts consulted to ACH and STOL. We then applied a cohort study design to examine the factors associated with the ‘missings’. In this study, we defined and labeled a value as ‘missing’ when the actual value was not identified in the audio records. Each variable’s distribution was described by its median and interquartile range (IQR). Mann–Whitney U test and chi-square test (or Fisher’s exact test if needed) were used to compare the continuous and nominal variables, respectively. To profile the triage physician’s individual effects, a multivariate regression with a random effect model, with triage physician as a random effect, was applied. We have included the following potential confounding variables, all of which were kept in the analyses due to their clinical importance: accepting hospital (STOL or ACH); referring hospital level of care (with adult ICU or not); referring physicians’ background demographics (graduated from a Canadian medical school or not, years since the graduation from a medical school); patient age (<28 days, 1 to 12 months, 1 to 4 years, 4 to 8 years, and >8 years), time of triage (i.e., day: 07:00 to 19:00 or night: 19:00 to 07:00). We have performed the analyses for each vital sign item except heart rate (HR) due to its small rate of missing. We also conducted subgroup analyses for the cases with (1) respiratory distress or respiratory disease(s) and (2) seizures or altered level of continuousness (LOC). For the study purpose, we created a variable ‘total missings’ consisting of values 0 to 10. Points were added depending on whether there was a missing value for each vital sign item (missing=1 and no missing=0) and the total score was calculated. For instance, if a triage call filed six missing items out of the ten, six points were assigned to this case. We performed a multivariate linear regression with random effect model adjusting for the same potential confounding variables noted before. We did not include disease category in the regression analyses, because the number of cases in each disease category was not large enough except the cases with respiratory distress or disease(s) and seizure or altered LOC. No interaction terms were included in the analyses. A two-sided P-value of less than 0.05 was considered statistically significant. All statistical analyses were conducted with Stata version 13 (Stata Corp LP, 2013). This study was approved by the Health Research Ethics Board of University of Alberta, Canada.
RESULTS
Patients and triage demographics
In total, 429 cases (179 cases in ACH and 250 cases in STOL) were included in the analyses of this study. The proportion of excluded cases mainly due to incomplete audio records was 8% from the ACH cohort and 13% from the STOL cohort, respectively (Figure 1). The numbers of triage calls were lowest between 01:00 and 06:00 and increased from the early morning with a peak at midnight. Spikes in call activity were observed at 11:00, 15:00, and 23:00 (Supplementary Figure 1). The median triage duration was 6.8 minutes (IQR: 6 to 9 minutes, range: 1 to 12.8 minutes) (Supplementary Figure 2). Ten different physicians engaged in patient triage at STOL and 13 at ACH during the study period. The number of triage calls per physician during the study period was a median of 19 (IQR: 9 to 24). Physicians in STOL took a significantly higher number of triage calls than the ACH (STOL: 37 triages [IQR: 22 to 40], ACH: 14 triages [IQR: 9 to 20]) (Supplementary Figure 3). The triage calls of seven physicians were merged into ‘other physician’ because they were each involved in less than ten calls during the study period. No significant variability to have ‘missings’ among the 17 triage physicians (i.e., 16 individual triage physicians + one ‘other physician’) was observed in the majority of vital sign items, whereas noticeable variability was seen in capillary refilling time (CRT) (Figure 2).
Figure 1.
Patients flow chart. *Either different transport triage physician/team (e.g., NICU) was called or audio data was missing.
Figure 2.
Trend of having ‘missings’ for individual triage physicians. Y-axis: 17 triage physicians (16 individual triage physicians + 1 ‘other physician’). X-axis: Adjusted log odds ratios to have ‘missings’ in each vital sign variable compared to the sample average. *Odds=1 is the average of the sample. Each bar and marker indicates lower 90% confidence limit, mean, and upper 90% confidence limit, respectively.
The patients’ demographics and the primary diagnoses were similar between the two cohorts; however, ACH received a significantly higher number of referral calls from general paediatricians (i.e., paediatricians accredited by College of Physicians and Surgeons of Alberta), family physicians, and physicians graduated from Canadian universities. ACH also received more calls from tertiary-level health care facilities with ICU (Table 1).
Table 1.
Demographics of patients and triage
| Variables | Entire cohort N=429 | ACH N=179 | STOL N=250 | P-values |
|---|---|---|---|---|
| Gender Male (%) | 248 (58) | 106(59) | 142 (57) | 0.62 |
| Age: year (IQR) | 1.9 (0.7 6.5) | 2.1 (0.6 6.8) | 1.7 (0.7 6.0) | 0.81 |
| Body Weight: kg (IQR) | 11.9 (8.0 20.0) | 12.0 (8.0 20.0) | 11.5 (8.0 22.0) | 0.93 |
| Diagnosis‡ (%) | ||||
| Diabetic Ketoacidosis (%) | 24 (6) | 10 (6) | 14 (6) | 1.00†† |
| Respiratory Distress/Diseases (%) | 214 (50) | 89 (50) | 125 (50) | 1.00†† |
| Seizure or/and Altered LOC (%) | 77 (18) | 31 (17) | 46 (18) | 0.80†† |
| Intoxication (%) | 21 (5) | 10 (6) | 11 (4) | 0.65†† |
| Trauma (%) | 21 (5) | 8 (4) | 13 (5) | 0.82†† |
| Sepsis or/and Shock (%) | 58 (14) | 23 (13) | 35 (14) | 0.78†† |
| Cardiac Disease (%) | 20 (5) | 10 (6) | 10 (4) | 0.49†† |
| Cardiac Arrest (%) | 7 (2) | 2 (1) | 5 (2) | 0.70†† |
| Others (%) | 21 (5) | 8 (4) | 13 (5) | 0.82†† |
| Referring Physicians’ Demographics | ||||
| Referring Physicians Specialty | ||||
| General Pediatrics (%) | 113 (26) | 70 (39) | 43 (17) | <0.001 |
| Family Medicine (%) | 152 (35) | 79 (44) | 73 (29) | |
| General Practice (%) | 138 (32) | 15 (8) | 123 (49) | |
| Emergency Medicine (%) | 19 (4) | 13 (7) | 6 (2) | |
| Others (%) | 7 (2) | 2 (1) | 5 (2) | |
| Experience in Alberta: years (IQR)* | 5 (2 12) | 5 (2 12) | 5 (2 12) | 0.40 |
| Experience of Medicine: years (IQR)† | 15 (9 26) | 12 (6 23) | 17 (10 28) | <0.001 |
| Country of Medical School Graduated# | ||||
| Canada (%) | 187 (44) | 116 (65) | 71 (28) | <0.001** |
| South Africa (%) | 128 (30) | 28 (16) | 100 (40) | |
| Others or Unknown (%)‡‡ | 114 (26) | 35 (19) | 79 (32) | |
| Transport and Hospital Demographics | ||||
| Duration of Triage call: seconds (IQR) | 410 (300 540) | 400 (300 530) | 429 (310 540) | 0.35 |
| Transport Team Used | ||||
| PICU Transport Team (%) | 333 (78) | 154 (86) | 179 (72) | 0.005 |
| Emergency Medical Service (%) | 67 (16) | 17 (10) | 50 (20) | |
| STARS (%) | 6 (1) | 1 (1) | 5 (2) | |
| Not Transported or Unknown (%) | 23 (5) | 7 (4) | 16 (6) | |
| First Triage call: Day 07:00 to 19:00 | 223 (52) | 87 (49) | 136 (54) | 0.24 |
| Transport from hospital with ICU# (%) | 191 (45) | 106 (59) | 85 (34) | <0.001 |
AOH Alberta Children’s Hospital; IQR Interquartile range; LOC Level of continuousness; PICU Paediatric intensive care unit; STOL Stollery Children’s Hospital.
*Years from the starting year of practice in Alberta, †Years from graduation from medical school, ‡Some patients had more than one diagnosis, #ICU: including Level 1-3, **P is also <0.001, when comparing Canadian graduate vs. non-Canadian, ††Fisher’s exact test was applied. ‡‡Physicians graduated from the UK, Nigeria, India, and Iran consisted of 46% (52/114) in the entire cohort, 29% (10/35) in ACH cohort, 53% (42/79) in STOL cohort.
Proportion of ‘missings’
The percentages of ‘missings’ were inconsistent in each vital sign item. CRT, neurological symptoms including pupil size and reactivity and Glasgow Coma Scale (GCS), and level of respiratory effort were not well recorded (‘missings’ in CRT: 60% GCS: 79%; pupil size and reactivity: 86%; and level of respiratory effort: 50%), whereas, HR, oxygen saturation (SpO2), and respiratory rate (RR) were recorded reasonably well (‘missings’ in HR: 12%; SpO2: 20%; RR: 26%; Table 2). In 68 cases (16%), systolic blood pressure (SBP) was not measured by referring physicians despite being asked for by triage physicians.
Table 2.
Proportions of missings in each vital sign item
| Variables | Entire cohort N=429 | ACH N=179 | STOL N=250 | P-values |
|---|---|---|---|---|
| Heart Rate | 52 (12)* | 24 (13) | 28 (11) | 0.55¶ |
| Systolic Blood Pressure | 181 (42) | 86 (48) | 95 (38) | 0.038 |
| Capillary Refilling Time | 258 (60) | 106 (59) | 152 (61) | 0.74 |
| Respiratory Rate | 113 (26) | 57 (32) | 56 (22) | 0.029 |
| Respiratory Effort† | 215 (50) | 93 (52) | 122 (49) | 0.52 |
| SpO2 | 87 (20) | 36 (20) | 51 (20) | 0.94 |
| Amount of Oxygen given‡ | 105 (24) | 49 (27) | 56 (22) | 0.24 |
| Glasgow Coma Scale | 337 (79) | 143 (80) | 194 (78) | 0.57 |
| LOC with AVPU | 171 (40) | 72 (40) | 99 (40) | 0.90 |
| Overall LOC# | 128 (30) | 50 (28) | 78 (31) | 0.47 |
| Pupils | 370 (86) | 160 (89) | 210 (84) | 0.12 |
| Body Temperature** | 127 (30) | 53 (30) | 74 (30) | 0.99 |
AOH Alberta Children’s Hospital; AVPU alert, verbal, pain, and unresponsive; LOC Level of continuousness; STOL Stollery Children’s Hospital.
*The number (%) of vital sign item which was not recorded/measured even when it was asked/discussed during triage calls: HR: 2 (1), SBP: 68 (16), CRT: 3(1), RR: 3 (1), SpO2: 3 (1), Amount of Oxygen: 0 (0), GCS: 1 (1), Pupils: 0 (0), Temperature: 2 (1). †Description of the conditions such as mild or severe increased work of breathing. ‡including room air, #No LOC was recorded including rough description of LOC such as ‘lethargic, crying, sleeping’, **categorized as no missing for the cases described as ‘no fever’ or ‘normal temperature’.
Duration of the triage calls and the associated items
There were independent associations noted between the median duration of triage call and the referring physician’s background demographics such as whether they were an international medical graduate or not (+59 seconds, 95% CI: 22 to 97 seconds, P=0.002) and triage calls between 01:00 and 06:00 (+46 seconds, 95% CI: −0.1 to 99 seconds, P=0.09; Supplementary Table 1).
Factors associated with ‘missings’
There was a lower likelihood of having ‘missings’ observed in older aged patients for several vital sign items including SBP (Ref. neonates, 4 to 8 years: OR=0.14 (95% CI: 0.04 to 0.51), >8 years: OR=0.06, 95% CI: 0.02 to 0.23), GCS (>8 years: OR=0.09, 95% CI: 0.01 to 0.77). We also found that accepting hospitals (ACH or STOL) were independently associated with the likelihood of recording ‘missings’ such as in SBP (OR=0.58, 95% CI: 0.36 to 0.93) and pupil size and reactivity (OR: 0.52, 95% CI: 0.27 to 1.01). There was no major difference in the results when we did the same analyses adjusting for the referring physicians’ specialty as an additional confounding variable (i.e., GP, family physicians, paediatrics; Supplementary Table 2). In the subgroup analyses for cases with respiratory distress or respiratory disease(s), seizures or altered LOC, we observed similar trends as those seen in the whole cohort.
DISCUSSION
This study explores the remote triage of children for potential transfer to a higher-level facility in a Canadian province. Published studies to date have examined the ability of existing paediatric triage scores to be applied in the paediatric critical care transport settings (3,4,7–10). However, no single triage scoring system has been validated and proved to be leading a better patient outcome or more efficient triage than others. It should be essential to collect and integrate all the potentially necessary vital signs and other clinical information in each triage to better capture the patient condition. Also, this triage procedure should be standardized among triage providers to provide most possibly efficient and effective transport outcome in each case.
Although some differences might exist in the triage and transport standard and practice when comparing the two programs (ACH and STOL), many components such as regional air ambulances and the third-party propeller retrieval services (i.e., Shock Trauma Air Rescue Service [STARS]) are part of the same common public provincial health care system (11). As expected, patient demographics and primary diagnoses were similar between the two groups. STOL, however, received more calls from the physicians with longer duration from the graduation and IMGs, general practitioners (i.e., physicians who were not recognized as a specialist of family medicine or emergency medicine), as well as from nontertiary hospitals. It has been well recognized that physician numbers, particularly specialist physicians such as general paediatricians, are condensed in more southern geographical area in Canada close to the border of the USA. Alberta is no exception and the province has had to actively recruit general practitioners to work in rural and northern underserviced areas (12–16). Our data support this and may infer the need for a tailored approach to triage given the potential heterogeneity of referring physician background.
Another interesting finding was the three distinct notable spikes of triage call activity and the timing in a typical day. These spikes might coincide with the timings before lunch, evening handover, and bedtime of referring medical staff. This finding should be considered as part of future staffing plans to insure efficient use of transport and triage resources.
Among vital signs such as CRT, neurological findings such as GCS, and degree of respiratory distress were not well recorded. We assume that this might be due to the difficulty in acquiring the skills necessary to measure the vital signs of critically ill or injured children when compared to adult patients. There did not appear to be significant variability between triage physicians in having ‘missings’ in most vital signs, while there were individual effects (i.e., effects in triage physicians) seen in some items such as CRT. To provide a ‘better’ triage, the consistent collection of necessary information is important. We could achieve this by applying techniques such as a data collection sheet or ‘live’ sharing of electronic medical records of the patients in a standardized triage system.
The age of patients may be associated with ‘missings’. ‘Missings’ were observed less frequently in SBP and LOC during the triages of older children. This may be due to the inherent challenges in measuring clinical signs in smaller children. The devices necessary for vital sign measurement such as small size blood pressure cuffs were not consistently available, or referring physicians were not consistently comfortable with their use. We also found that the ‘missings’ in some vital sign items such as SBP and pupil assessment were significantly different when comparing both cohorts (ACH and STOL). This may indicate a specific (tertiary care) institution’s practice of asking for (or not) specific kinds of patient information when triaging patients, and should be further studied.
Referred calls from Canadian medical graduates tended to be shorter in duration. Duration of calls after midnight was significantly longer than that from 06:00 to 12:00 for all the referred physicians. The difference between the IMG and non-IMG triage practice may stem from a difference in medical education or ‘culture’ of practice (17). Longer triage calls during night shifts could occur as a result of decreased cognitive performance of triage physicians (18,19). Triage during the night was performed by ‘on-call’ physicians both from home or in-hospital with a primary responsibility for PICU coverage. Different results might be found when studying transport programs with dedicated triage physicians (20).
There are limitations in the findings of study. First, although the majority of consult calls during the study period were captured, some of consults came directly to the PICU without going through RAAPID. This would lead to potential information bias. However, the number is unlikely to be significant as this system has been widely accepted by Alberta healthcare providers as the best way to access triage advice in the care of ‘potentially’ critically ill or injured children. Information bias could also have occurred with the data extraction. When measuring the duration of transport calls, the length of nontriage discussion was not recorded as part of the audio record, omitting some of the nonrelevant parts of the discussion. In this study, one investigator (AK) extracted all the data from the audio records so as to minimize bias and to make the procedure as consistent as possible. Unknown differences in patient or transport demographics might exist between the two programs, such as geographical differences that could lead to differences in transport time, impacting over observed associations. Due to the small sample size, we did not examine the interactions between referring and accepting physicians; those interaction effects could have affected the triage quality. For example, if an experienced emergency physician refers a case to a trusting triage physician, the triage might become an abbreviated version based on the trust-relationship between the two physicians, which should not be happening in triage practice.
CONCLUSIONS
To sum up, the proportion of vital sign items discussed and recorded during the triage of referred patients significantly varied. The ‘missings’ could be associated with referring physician background and patient demographics such as age, which should be considered in the improvement of triage quality. Our study could suggest that there is room for improvement in our triage practice. As a next step, we need to explore how ‘missings’ can affect patient outcomes and how to improve triage quality. This might include the implementation of triage guidelines or a novel system such as remote live sharing of patients’ information to discuss from the triage (e.g., video-consultation).
SUPPLEMENTARY DATA
Supplementary data are available at Paediatrics & Child Health Online.
Supplemental Figure 1. Times that event occurred.
Supplemental Figure 2. Distribution of triage call duration.
Supplemental Figure 3. Distribution of the numbers of triage in STOL and ACH.
Supplemental Table 1. Duration of Triage Call and the Associated Items.
Supplemental Table 2. Associations between Vital Sign Items and Patients’ and Transport Relevant Factors. *Every five years, bold: statistically significant. There was no major difference in the results when we did the same analyses adjusting for the referring physician’s specialty as an additional confounding variable (i.e., GP, family physicians, paediatrics).
ACKNOWLEDGEMENTS
The authors thank the following individuals for their contribution to the study: Ms. Carla Easter, Ms. Jessica Ewing, and Mr. John Montpetit director of RAAPID.
Contributors’ Statement: AK conceptualized and designed the study, conducted data collection and cleaned the data, carried out analysis, drafted the initial manuscript, revised the initial manuscript, and approved the final manuscript as submitted. GGG helped to discuss the results, revised the initial manuscript, and approved the final manuscript as submitted. EG revised the initial manuscript, and approved the final manuscript as submitted. PJ helped to coordinate the data management from RAAPID, revised the initial manuscript, and approved the final manuscript as submitted. AD conceptualized and designed the study, revised the initial manuscript, and approved the final manuscript as submitted.
Funding: AK is supported by a postdoctoral research grant from the Fonds de recherché du Québec (FRQS).
Potential Conflicts of Interest: All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.
References
- 1. Kawaguchi A, Nielsen CC, Guerra GG, Saunders LD, Yasui Y, DeCaen A. Epidemiology of pediatric critical care transport in Northern Alberta and the Western Arctic. Pediatr Crit Care Med. 2018; 19(6):e279–e285. [DOI] [PubMed] [Google Scholar]
- 2. Kawaguchi A, Nielsen CC, Saunders LD, Yasui Y, de Caen A. Impact of physician-less pediatric critical care transport: Making a decision on team composition. J Crit Care 2018;45:209–14. [DOI] [PubMed] [Google Scholar]
- 3. Kandil SB, Sanford HA, Northrup V, Bigham MT, Giuliano JS Jr. Transport disposition using the Transport Risk Assessment in Pediatrics (TRAP) score. Prehosp Emerg Care 2012;16(3):366–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Orr RA, Venkataraman ST, Cinoman MI, Hogue BL, Singleton CA, McCloskey KA. Pretransport Pediatric Risk of Mortality (PRISM) score underestimates the requirement for intensive care or major interventions during interhospital transport. Crit Care Med 1994;22(1):101–7. [DOI] [PubMed] [Google Scholar]
- 5. Warren DW, Jarvis A, LeBlanc L, Gravel J, Group CNW, Canadian Association of Emergency Physicians , et al. Revisions to the Canadian triage and acuity scale paediatric guidelines (PaedCTAS). CJEM. 2008;10(3):224–43. [PubMed] [Google Scholar]
- 6. College of Physicians and Surgeons of Alberta. http://www.cpsa.ca (Accessed November 20, 2018).
- 7. Holt T, Prodanuk M, Hansen G. Utilizing pediatric scoring systems to predict disposition during interfacility transport. Prehosp Emerg Care. 2019;23(2):249–53. [DOI] [PubMed] [Google Scholar]
- 8. Kanter RK, Edge WE, Caldwell CR, Nocera MA, Orr RA. Pediatric mortality probability estimated from pre-ICU severity of illness. Pediatrics 1997;99(1):59–63. [DOI] [PubMed] [Google Scholar]
- 9. Orr RA, Venkataraman ST, McCloskey KA, et al. Measurement of pediatric illness severity using simple pretransport variables. Prehosp Emerg Care 2001;5(2):127–33. [DOI] [PubMed] [Google Scholar]
- 10. Petrillo-Albarano T, Stockwell J, Leong T, Hebbar K. The use of a modified pediatric early warning score to assess stability of pediatric patients during transport. Pediatr Emerg Care 2012;28(9):878–82. [DOI] [PubMed] [Google Scholar]
- 11. SHOCK TRAUMA AIR RESCUE SERVICE (STARS). https://www.stars.ca/ab/ (Accessed March 1, 2018).
- 12. Geographic Distribution of Physicians in Canada: Beyond How Many and Where. https://secure.cihi.ca/free_products/Geographic_Distribution_of_Physicians_FINAL_e.pdf. (Accessed November 20, 2018).
- 13. Joudrey R, Robson K. Practising medicine in two countries: South African physicians in Canada. Sociol Health Illn 2010;32(4):528–44. [DOI] [PubMed] [Google Scholar]
- 14. McDonald T, Seidel JE, Patel AB, et al. High-volume general practitioners in Alberta: A descriptive analysis. CMAJ Open 2018;6(3):E254–60. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Supply, Distribution and Migration of Physicians in Canada, 2016. https://www.cihi.ca/sites/default/files/document/smdb_data_release_methodology_notes_phys2016_en.pdf. (Accessed November 20, 2018).
- 16. Grant HM. From the transvaal to the prairies: The migration of South African physicians to Canada. J Ethn Migr Stud 2006;32(4):681–95. [Google Scholar]
- 17. Triscott JA, Szafran O, Waugh EH, Torti JM, Barton M. Cultural transition of international medical graduate residents into family practice in Canada. Int J Med Educ 2016;7:132–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Griffiths JD, McCutcheont C, Silbert BS, Maruff P. A prospective observational study of the effect of night duty on the cognitive function of anaesthetic registrars. Anaesth Intensive Care 2006;34(5):621–8. [DOI] [PubMed] [Google Scholar]
- 19. Kahol K, Leyba MJ, Deka M, et al. Effect of fatigue on psychomotor and cognitive skills. Am J Surg 2008;195(2):195–204. [DOI] [PubMed] [Google Scholar]
- 20. National Health Services, Annual Resports from Chidlren’s Acute Transport System in National health Services. http://site.cats.nhs.uk (Accessed March 1, 2018).
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.