Prehospital Emergency Care

ISSN: 1090-3127 (Print) 1545-0066 (Online) Journal homepage: https://www.tandfonline.com/loi/ipec20

A Statewide Analysis of EMS’ Pediatric Transport

Destination Decisions

Kayla McManus, Erik Finlay, Sam Palmer, Jennifer F Anders, Phyllis Hendry &
Jennifer N Fishe

To cite this article: Kayla McManus, Erik Finlay, Sam Palmer, Jennifer F Anders, Phyllis Hendry &
Jennifer N Fishe (2019): A Statewide Analysis of EMS’ Pediatric Transport Destination Decisions,
Prehospital Emergency Care, DOI: 10.1080/10903127.2019.1699211

To link to this article: https://doi.org/10.1080/10903127.2019.1699211

Accepted author version posted online: 09

Dec 2019.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=ipec20
 A Statewide Analysis of EMS’ Pediatric Transport Destination Decisions

Kayla McManus, DO,1 Erik Finlay, MPH,2 Sam Palmer, MAURP,2 Jennifer F Anders, MD,3
Phyllis Hendry, MD,4 Jennifer N Fishe, MD4*
1
Department of Pediatrics, University of Florida College of Medicine, Jacksonville
2
GeoPlan Center, University of Florida College of Design, Construction, and Planning
3
Department of Pediatrics, Division of Pediatric Emergency Medicine, Johns Hopkins University
4
Department of Emergency Medicine, University of Florida College of Medicine, Jacksonville

*Address correspondence and reprint requests to: Dr. Jennifer N. Fishe, MD, Department of

t
ip
Emergency Medicine, University of Florida - Jacksonville, 655 West 8th Street, Jacksonville, FL
32209. Tel: (904) 244-4046, Email: jennifer.fishe@jax.ufl.edu

cr
Presentation: This work was accepted as a poster presentation to the National Association of
EMS Physicians Annual Meeting 2020, to be presented during the meeting in San Diego

us
between January 6-11, 2020.

an
Keywords: Transport Destination, Pediatrics, Emergency Medical Services, Rural
M
ABSTRACT
d

Introduction: Deciding where to transport a patient is a key decision made by emergency

e

medical services (EMS), particularly for children because pediatric hospital resources are
pt

regionalized. Since evidence-based guidelines for pediatric transport destinations are being
ce

developed, the purpose of this study was to use a large statewide EMS database to describe
Ac

current patterns of EMS providers’ transport destination decisions for pediatric patients.

Methods: This is a retrospective study of pediatric transports from 2011-2016 in EMS Tracking

and Reporting System (EMSTARS), Florida’s statewide EMS database. We included patients

greater than 1 day and less than or equal to 18 years who were primary EMS scene transports.

1
Our primary outcome variable was ‘reason for choosing destination.’ We performed descriptive

and comparative analysis between closest facility and all other ‘reason for choosing destination’

choices. We used geospatial analysis to examine destination choice in urban and rural counties.

Results: Our final study sample was 446,274, and 48.2% of patients had closest facility as their

‘reason for choosing destination.’ The next largest category was patient / family choice (154,035

patients, 35.7%). Closest facility patients were older (median age 12 versus 10 years, p<0.0001)

t
ip
and had shorter median EMS transport times (11.3 versus 15 minutes, p<0.0001) compared to all

cr
other destination decisions. Notably, 60% of respiratory distress patients’ and 44% of seizure

patients’ reason for choosing destination was something other than closest facility. Geospatial

us
analysis revealed that fewer rural patients were documented as closest facility compared to urban
an
(43.9% versus 47%, p< 0.0001). Correspondingly, more rural patients’ destination decision was

patient / family choice than urban patients (36.3% versus 34.3%, p<0.0001).
M
Conclusions: This large, statewide study describes EMS’ reason for choosing destination for
d

pediatric patients. We found that just under half of patients were documented as closest facility,
e

and over one-third as patient / family choice. Significant differences in destination reasons were
pt

noted for rural versus urban counties. This study can help those currently developing pediatric
ce

EMS destination guidelines by revealing a high proportion of patient / family choice and
Ac

identifying conditions with high proportions of destination reasons other than closest facility.

2
Introduction

Deciding where to transport a patient is a key decision made by emergency medical services

(EMS). That decision is particularly important for children because regionalization of care has

concentrated pediatric resources in select hospitals.1 Despite regionalization, there are no widely-

accepted evidence-based guidelines for EMS’ transport destination for pediatric patients, save for

those with trauma.2

t
ip
cr
In order to safely transport children to an optimal destination without overburdening EMS

us
resources, work has begun to develop transport guidelines for non-trauma pediatric patients. A
an
multidisciplinary group of EMS and pediatric stakeholders constructed a 13-item consensus-

based criterion for EMS to identify children in need of higher-level pediatric services such as
M
specialty and/or critical care.3 Retrospective studies have identified specific non-trauma pediatric

conditions at risk for requiring interfacility transport to definitive care after initial EMS transport
d

to a local hospital.4 Since prehospital identification of specific medical conditions is not always
e
pt

feasible, an abnormal Pediatric Assessment Triangle (PAT) has shown promise in identifying

children in need of specialty care.5

ce
Ac

Meanwhile, while those studies help formulate future guidelines for pediatric medical patients,

the current patterns of EMS providers’ transport destination decision-making for those patients

are not yet described. For adult trauma patients, studies indicate that EMS providers use “gut

instinct” or patient/family input to guide transport decisions despite evidence-based trauma

3
transport guidelines.6-8 Understanding current EMS pediatric destination decision-making is

important for the development, dissemination, and implementation of the guidelines currently in

development. Therefore, our objective was to describe EMS providers’ current transport

destination decisions for pediatric patients using a large statewide EMS database.

t
Methods

ip
Study Setting and Data Source

cr
This is a retrospective study of pediatric transports from EMS Tracking and Reporting System

us
(EMSTARS), Florida’s statewide EMS database. Florida is the state with the fourth-largest
an
pediatric population.9 During the study period, EMSTARS captured approximately 74% of 9-1-1

calls in Florida.10 EMSTARS’ characteristics and data elements have been described in-depth in
M
a prior publication.11 EMSTARS uses National EMS Information Systems-National Highway

Traffic Safety Administration (NEMSIS) Version 2.2.1 Data Dictionary elements,12 as well as
e d

data elements custom to Florida.

pt
ce

Inclusion and Exclusion Criteria

Ac

Our original dataset was EMS encounters in EMSTARS from 2011-2016. Included encounters

were primary EMS scene transports of patients ages greater than 1 day and less than or equal to

18 years, excluding those whose age was unknown. We excluded interfacility transports. Our

reason for excluding patients aged 1 day or less was the assumption that those patients would

4
likely be out-of-hospital births. Since out-of-hospital births would likely be transported with a

second patient (the mother), we reasoned that EMS’ transport destination decision would also be

influenced by the medical needs of that second patient.

Data Variables

t
Our primary outcome variable was ‘reason for choosing destination’ (a NEMSIS 2.2.1 compliant

ip
variable).12 EMSTARS only provides for one selection for that variable, meaning that the options

cr
(closest facility, patient choice, family choice, protocol, specialty resource center, patients’

us
physicians’ choice, other, law enforcement choice, diversion, online medical direction, insurance

status, not known, not recorded, and not applicable) are mutually exclusive.10,12 We merged
an
patient choice and family choice into one variable - ‘patient / family choice’ for simplicity since
M
we reasoned that patient choice and family choice were interchangeable for pediatric patients.
e d

Other data variables included patient demographics, clinical characteristics, and EMS variables.
pt

We categorized the EMS agency’s home county as urban or rural based on the Florida
ce

Department of Health’s classification.13 We used the first recorded vital sign value by EMS since
Ac

evidence suggests many EMS decisions are made based on the initial assessment, including vital

sign measurements.14 We created categories of abnormal or normal vital sign values based on

patient age using Pediatric Advanced Life Support ‘General Vital Signs and Guidelines’ table. 15

We grouped EMS-administered medications and procedures into a priori-determined clinical

categories, detailed in Online Supplement e1. We also included EMSTARS’ ‘possible injury’
5
variable, which is defined as “…the reason for the EMS encounter was related to an injury or

traumatic event…based on mechanism and not on actual injury.”10

Statistical Analysis

The primary outcome and other variables were analyzed with descriptive statistics (frequencies,

t
percentages, mean and standard deviation or median and interquartile ranges, as appropriate). A

ip
priori based on our experience and studies finding that closest facility is a common choice,7,8 we

cr
decided to compare patients whose ‘reason for choosing destination’ was closest facility versus

us
those with any other ‘reason for choosing destination’ (‘Other’). A priori, we chose to include
an
patients whose reason for choosing destination was marked as not known, not recorded, or not

applicable in the ‘Other’ category. We performed univariate analysis comparing characteristics

M
of closest facility versus other patients using unpaired t test or Wilcoxon Rank Sum tests (as

appropriate for continuous variables), and Chi-square test for categorical variables. We used the
d

Kolmogorov-Smirnov test for normality given that our sample size was much greater than
e
pt

2,000.16 Missing data are noted in the results and were excluded from individual variable

univariate analysis (e.g., patients with missing data for ethnicity were excluded from the chi-
ce

square analysis of ethnicity between closest facility and other patients). Statistical analysis was
Ac

performed using SAS® version 9.4 (Cary, NC).

6
Geospatial Analysis

Using ArcGIS Desktop 10.4.1 (Redlands, CA), we mapped each patient to the home county of

the treating EMS agency. We created shaded thematic (“choropleth”) maps showing descriptive

statistics of patient demographic and EMS clinical variables aggregated to county-level units, as

well as each county’s rural/urban classification.13 We mapped the spatial patterns of pediatric

admitting hospitals in relation to the study populations’ county-level statistics. We identified

t
ip
pediatric admitting hospital facilities from a previous study of asthma exacerbations.17 Since

cr
asthma is a common pediatric condition requiring inpatient admission, we assumed the pattern of

admitting children for asthma should mimic the pattern for pediatric inpatient admissions in

us
general.
an
M
Results
d

From 2011-2016 there were 3,491,241 unique patient encounters of all ages in EMSTARS. After
e

application of inclusion and exclusion criteria, our final study sample was 446,274 patient
pt

encounters (Figure 1). Nearly half of patients (48.2%, N=208,323) had closest facility as their
ce

‘reason for choosing destination.’ The next largest category was patient / family choice (35.7%,

N=154,035). Table 1 details all ‘reason for choosing destination’ categories for the study sample
Ac

with select demographics and EMS transport times. Of note, patient / family choice, patients’

physicians’ choice, specialty resource center, and online medical direction had younger median

ages. The longest median transport times were seen with specialty resource center, patients’

7
physicians’ choice, and diversion. Nearly half of ‘possible injury’ patients were marked as

closest facility.

Demographic, clinical, and EMS encounter characteristics are detailed in Table 2. About half of

patients were male (48.8%), and the median age was 11 years. Age was bimodally distributed

t
(data not shown). A plurality of incidents occurred at a home or residence (45.4%), and 29.2%

ip
overall were categorized as a possible injury. A minority of patients had abnormal vital signs. 15

cr
Table 2 also displays differences between the closest facility and other destination choice

us
patients. Notably, closest facility patients were significantly older and had significantly shorter

EMS transport times. Most vital signs were statistically, but not clinically, different between the
an
two groups. Overall most patients did not receive any medications or procedures from EMS, save
M
for oxygen (15.3% overall; 16.4% other vs 14% closest facility, p<0.0001).
e d

Table 3 displays ‘reason for choosing destination’ by EMS provider’s primary impression. Most
pt

cardiac arrest patients’ (73.5%) were marked as closest facility. However, most patients with
ce

other critically ill provider primary impressions (e.g., sepsis, airway obstruction,

hypovolemia/shock) had other than closest facility as their reason for choosing destination.
Ac

Notably, 60% of respiratory distress patients and 44% of seizure patients’ reason for choosing

destination was other, rather than closest facility.

8
We were able to map patients to 65 of Florida’s 67 counties based on EMS agency home county

data. The two counties without data were Hardee and Columbia, both rural. Most (93.2%,

N=415,956) patient encounters were in urban counties and 6.2% (N=27,490) in rural counties

(N=2,828, <1%, missing EMS agency county information). The maps in Figure 2 highlight rural

counties, displaying locations of pediatric admitting hospitals and county-level percentages for

the closest facility (Figure 2a), and patient / family choice ‘reason for choosing destinations’

t
(Figure 2b). Overall, fewer rural patients’ transport destination decision was closest facility

ip
compared to urban (43.9% versus 47%, p<0.0001). Correspondingly, more rural patients’

cr
destination decision was patient / family choice than urban patients (36.3% versus 34.3%,

us
p<0.0001). EMS median transport times were materially different between urban and rural
an
locations (13 minutes urban versus 18 minutes rural, p< 0.0001, Online Supplement e2).
M
Discussion
d

This large, statewide study describes EMS’ reason for choosing destination facilities for pediatric
e
pt

patients. We found that just under half of patients were documented as closest facility, and a

large percentage (35.7%) were designated as patient / family choice. Our results include both
ce

trauma and non-trauma patients, which may explain differences compared to a previous study of
Ac

adult and pediatric trauma patients’ destinations.8 That study found that 50.6% of patients’

destinations were patient or family choice, with 20.7% being closest facility.8 However, there is

geographic variation in those results,8 which we also found with more rural patients having

destination decisions other than closest facility compared to urban, including patient / family

9
choice. It is notable that our and others results from the United States contrast greatly with a

study from Australia which found that 82% of rural pediatric prehospital patients were taken to

the closest facility.18

Given widely operative pediatric trauma protocols, possible injury may also influence EMS’

t
destination choice. A previous analysis from three EMS agencies whose catchment areas

ip
included large children’s hospitals found that patients with injuries were more likely to be

cr
transported to higher-level pediatric resources than non-injury patients.19 We cannot compare our

us
study directly to those results as our data does not include whether closest facility or patient /

family choice led to transport to a higher-level pediatric resource or trauma center. However,
an
compared to that study of three EMS agencies,19 this study has a much larger sample size from a
M
state where 30 counties are rural and 37 are urban.13 Perhaps reflecting that more heterogeneous

sample, we found most possible injury patients’ reasons for choosing destination were closest
d

facility or patient / family choice (80%), and not protocol (e.g., trauma protocol) or specialty
e

resource center. This may reflect areas where there is no tertiary care children’s hospital within a
pt

feasible EMS transport time. Such areas may be rural, suburban, or even urban, as not all cities
ce

have a tertiary care children’s hospital.

Ac

It appears that age may also influence destination decisions, with more marked effects for

younger children. We found younger median ages in the other group compared to closest facility.

Similarly, the study by Lerner et al found that patients above 15 years of age were more likely to

10
go to general EDs than pediatric EDs.19 Further to this, as age decreased, increasing proportions

of patients had destination decisions other than closest facility. At a certain teenage age range,

EMS likely treats pediatric patients as they would adults and transports to the closest facility.

One of our most notable findings was that 60% of respiratory distress and 44% of seizure

patients’ destination decisions were other than closest facility. Although we were unable to

t
ip
examine the free text narratives in the EMS records to see if respiratory distress patients were

cr
hemodynamically unstable or if seizure patients were actively seizing (both of which would

us
prompt concern for needing to transport to the closest facility), this finding is notable for several

reasons. First, both of those conditions / provider impressions have been previously identified as
an
high-risk for secondary transport and interfacility transport.4,20 This study provides additional
M
evidence that a significant proportion of pediatric patients with respiratory distress and seizures

are thought by EMS to need transport to specialized care. EMS’ “gut instinct” or prior
d

experience may be that certain high-risk respiratory distress patients will require intensive care.
e

Seizure patients’ families may express to EMS that they wish to be cared for in a center which
pt

knows their child’s epilepsy history, prompting transport to a specialty resource center (i.e.,
ce

children’s hospital). What is not revealed by this study is EMS’ comfort and proficiency in
Ac

caring for children in respiratory distress and with seizures over longer transport distances to get

to those specialty centers. Additionally, only 3.5% of destination decisions overall were marked

as protocol, which may reflect the need for more protocolized destination guidance for pediatric

conditions other than trauma. Conversely, this may also reflect that more common destination

choices selected by EMS providers (e.g., closest facility), are also specified by protocols.
11
Whether a pediatric patient goes to the closest facility or not affects EMS operations, particularly

if the non-closest facility choice is much further away. Indeed, we found significantly longer

transport times for other versus closest facility patients. Those longer transport times could stem

from bypassing the closest facility to an alternative destination such as a pediatric specialty

hospital. A study of non-trauma pediatric EMS bypass of closest facilities in Maryland also

t
ip
discovered a high (43%) rate of bypass of the closest facility and corresponding longer transport

cr
times and distances for bypass patients.21 We also found longer emergency department (ED)

turnaround times for other versus closest facility patients, which may reflect patient complexity

us
or acuity resulting in longer EMS-to-ED handoffs. Similarly, a study from a large EMS system
an
in North Carolina found longer turnaround times for those transported to academic or tertiary

care facilities versus community hospitals.22

M
d

A combination of individual patient, caregiver, EMS, and geographic factors likely coalesce to
e
pt

inform the destination decisions for pediatric patients. Regardless, this study reveals a large

proportion of other than closest facility destination decisions and significant differences between
ce

urban and rural counties. Therefore, it seems appropriate for future work to combine clinical and
Ac

geospatial analysis from multiple studies and locations to produce evidence-based non-trauma

pediatric transport guidelines.

12
Limitations

This study has limitations that merit consideration. Overall, interpretation of closest facility

versus other should be tempered by our lack of precise coordinates for EMS scene and

destination locations. Therefore, we are unable to say if closest facility patients went to the actual

closest facility, and if other patients bypassed a closer facility in favor of an alternative

destination. Patient / family choice may also be the geographically closest facility, specialty

t
ip
resource center may not be a children’s hospital, etc. Therefore, the reader should interpret the

cr
variable as the reason for choosing destination, and not the actual destination itself. Additionally,

since ‘reason for choosing destination’ choices in EMSTARS are mutually exclusive, we are

us
unable to discern if EMS providers’ choice was multifactorial. Since we did not have EMS scene
an
and hospital locations, we were unable to determine if any transports were bypasses of the

closest facility. Online medical direction is a value for 'reason for choosing destination',
M
however, the EMSTARS data dictionary does not specify if online medical direction was
d

attempted, provided, and/or whether the EMS providers transported to the advised destination.10
e
pt

Regarding the 29% of patients noted by EMS to have a possible injury, there are widely
ce

operative guidelines for trauma patient destination nationwide and in Florida. However, only
Ac

3.2% of patients’ reason for choosing destination in this study was protocol, and only 5.4% was

specialty resource center. A limitation of this study is that the ‘possible injury’ variable is a poor

surrogate for a trauma activation patient, and our IRB-approved study design did not include a

13
specific trauma patient analysis, therefore we were precluded from a pediatric trauma sub-group

analysis.

EMSTARS, while containing a wealth of data, is a database including only Florida; as such, our

results may not be generalizable to all areas. However, Florida is a very populous state with a

t
mix of urban and rural counties, therefore our results (including geospatial analysis) and large

ip
sample size can provide more generalizable information than single-agency studies from urban

cr
locations. However, Florida DOH’s method of classifying urban and rural is at the unit of a

us
county, which limits identification of areas in urban counties that may actually be rural or vice

versa. Since our dataset was limited to county-level geographic variables, we were unable to use
an
more spatially resolute descriptions of urban and rural areas. EMSTARS does not contain
M
information from all EMS agencies in Florida, which is reflected in our geospatial analyses, and

may impact how certain county-level variables appear on the maps presented. As with other
d

large EMS databases, there are variables which contain large amounts of missing data. For our
e

univariate comparisons between closest facility and other patients, we did not include patients
pt

with missing values for those data (e.g., for comparing transport time, those whose transport
ce

times were not recorded were excluded for only the transport time comparison). Some variables
Ac

have higher degrees of missing data (e.g., ethnicity), which may bias the results of univariate

comparisons for those variables. Additionally, for some counties, there was an unusual

breakdown of gender amongst transported patients. Specifically, an overwhelming majority of

transported patients were female (Collier 92.8%, Lee 78.3%, Hernando 78%, and Seminole

14
75%). We did not have access to the free text narrative in the EMS patient care reports, which

would provide additional information on provider’s reason for destination choice.

Conclusions

In this large statewide study of EMS’ reason for choosing destination for pediatric patients,

t
nearly one-half of patient encounters were documented as closest facility, and over one-third as

ip
patient / family choice. Significant differences in destination decision-making were noted for

cr
rural versus urban counties. Additionally, in this study of nearly a half-million pediatric EMS

us
transports, there was a high proportion of patient / family choice, and specific conditions with
an
high proportions of reason for choosing destination other than closest facility. Those results

should be noted within the context of regionalization of pediatric care into specialized urban
M
centers. These results can therefore aid EMS agencies in operations planning and those currently

developing pediatric EMS destination guidelines.

e d
pt

Funding Source
ce

Research reported in this publication was supported by the National Center for Advancing
Ac

Translational Sciences of the National Institutes of Health under University of Florida Clinical

and Translational Science Awards KL2TR001429 and UL1TR001427. The information, content,

and conclusions are those of the authors and should not be construed as the official position,

policy, or endorsement by the National Institutes of Health.

15
Author Contributions

K.M. contributed to the study concept and design, interpretation of the data, drafting of the

manuscript, and critical revision of the manuscript for important intellectual content. E.F. and

S.P. contributed to the statistical analysis, interpretation of the data, and critical revision of the

manuscript for important intellectual content. J.F.A. and P.H. contributed to the study concept

and design, interpretation of the data, and critical revision of the manuscript for important

t
ip
intellectual content. J.N.F. contributed to the study concept and design, acquisition of funding

and the data, statistical analysis, interpretation of the data, drafting of the manuscript, and critical

cr
revision of the manuscript for important intellectual content.

us
Conflict of Interest an
No authors have conflicts of interest or financial disclosures.
M
Acknowledgements

The study investigators acknowledge Steve McCoy, Brenda Clotfelter, Karen Card, DrPH and
e d

Joshua Sturms from the Florida Department of Health’s Bureau of Emergency Medical
pt

Oversight for their assistance and data management.

ce
Ac

16
References

1. França UL, McManus ML. Availability of definitive hospital care for children. JAMA Pediatr.

2017;171(9):e171096. doi: 10.1001/jamapediatrics.2017.1096.

2. McCoy CE, Chakravarthy B, Lotfipour S. Guidelines for the Field Triage of Injured Patients:

In conjunction with the Morbidity and Mortality Weekly Report published by the Center for

t
Disease Control and Prevention. West J Emerg Med. 2013;14(1):69-76.

ip
3. Studnek JR, Lerner EB, Shah MI, Browne LR, Brousseau DC, Cushman JT, Davan PS,

cr
Drayna PC, Drendel AL, Gray MP, Kahn CA, Meyer MT, Shah MN, Stanley RM. Consensus‐

us
based Criterion Standard for the Identification of Pediatric Patients Who Need Emergency

Medical Services Transport to a Hospital with Higher‐level Pediatric Resources. Acad Emerg
an
Med. 2018;25(12):1409-1414.
M
4. Fishe JN, Psoter KJ, Klein BL, Anders JF. Retrospective Evaluation of Risk Factors for

Pediatric Secondary Transport. Prehosp Emerg Care. 2017;22(1):41-49.

d

5. Fratta KA, Fishe JN. EMS, Pediatric Transport Safety and Secondary Transport. StatPearls
e

[Internet]. Treasure Island (FL): StatPearls Publishing; 2019 Jan-2019 Feb 25.
pt
ce

6. Jones, CM, Cushman, JT, Lerner, EB, Fisher SG, Seplaki CL, Veazie PJ, Wasserman EB,

Dozier A, Shah MN. Prehospital Trauma Triage Decision‐making: A Model of What Happens
Ac

between the 9‐1‐1 Call and the Hospital. Prehosp Emerg Care 2016;20(1):6–14.

7. Newgard CD, Nelson MJ, Kampp M, Saha S, Zive D, Schmidt T, Daya M, Jui J, Wittwer L,

Warden C, Sahni R, Stevens M, Gorman K, Koenig K, Gubler D, Rosteck P, Lee J, Hedges JR.

17
Out-of-Hospital Decision Making and Factors Influencing the Regional Distribution of Injured

Patients in a Trauma System. J Trauma. 2011;70(6):1345-53.

8. Newgard, CD, Mann NC, Hsia RY, Bulger EM, Ma OJ, Staudenmayer K, Haukoos JS, Sahni

R, Kuppermann N, Western Emergency Services Translational Research Network (WESTRN)

Investigators. Patient choice in the selection of hospitals by 9-1-1 emergency medical services

providers in trauma systems. Acad Emerg Med. 2013;20(9):911–9.

t
ip
9. United States Census Bureau. Quick Facts: Florida. 2018 [Accessed 2019 November 2].

Available from: https://www.census.gov/quickfacts/fact/table/fl/PST045217.

cr
10. Florida’s Prehospital EMS Tracking and Reporting System (EMSTARS) Data Dictionary

us
Version 1.4.2018. 2018. [Accessed 2019 November 2]. Available from:

http://www.floridaemstars.com.
an
11. Fishe JN, Lynch S. Pediatric Behavioral Health-Related EMS Encounters: A Statewide
M
Analysis. Prehosp Emerg Care. 2019;23(5):654-62.

12. NEMSIS NHTSA Version 2.2.1 Data Dictionary. 2018. [Accessed 2019 November 2].
e d

Available from https://nemsis.org/technical-resources/version-2/version-2-dataset-dictionaries/.

pt

13. Florida’s Rural Counties. 2018. [Accessed 2019 November 2]. Available from:
ce

http://www.floridahealth.gov/programs-and-services/community-health/rural-

health/_documents/ruralcountiespdf.12.pdf.
Ac

14. Gausche-Hill M, Eckstein M, Horeczko T, McGrath N, Kurobe A, Ullum L, Kaji AH, Lewis

RJ. Paramedics accurately apply the pediatric assessment triangle to drive management. Prehosp

Emerg Care. 2014;18(4):520-30.

18
15. General Vital Signs and Guidelines, Pediatric Advanced Life Support. [Accessed 2019

November 2]. Available from: https://www.acls-pals-bls.com/algorithsm/pals.

16. Razali NM. Power comparisons of Shapiro-Wilk, Kolmogorov-Smirnov, Lilliefors and

Anderson-Darling tests. Journal of Statistical Modeling and Analytics. 2011;2(1):21-33.

17. Fishe J, Finlay E, Palmer S, Hendry P. A Geospatial Analysis of Distances to Hospitals that

t
Admit Pediatric Patients. Prehosp Emerg Care. 2019;23(6):882-86.

ip
18. Kloot K, Salzman S, Kilpatrick S, Baker T, Brumby SA. Initial destination hospital of

cr
paediatric prehospital patients in rural Victoria. Emerg Med Australas. 2016;28(2):205-10.

us
19. Lerner EB, Studnek JR, Fumo N, Banerjee A, Arapi I, Browne LR, Ostermayer DG,
an
Reynolds S, Shah MI. Multicenter Analysis of Transport Destinations for Pediatric Prehospital

Patients. Acad Emerg Med. 2019;26(5):510-16.

M
20. Aledhaim A, Fishe JN, Hirshon JM, Anders JF. Pediatric Conditions Requiring Interfacility
d

Transport From Emergency Departments: A Statewide Study of Regionalization. Pediatr Emerg

e

Care. 2018. DOI: 10.1097/PEC.0000000000001578. [Epub ahead of print].

pt

21. Fishe JN, Psoter KJ, Anders JF. Emergency Medical Services Bypass of the Closest Facility
ce

for Pediatric Patients. Prehosp Emerg Care. 2019;23(4):485-90.

Ac

22. Vandeventer S, Studnek JR, Garrett JS, Ward SR, Staley K, Blackwell T. The association

between ambulance hospital turnaround times and patient acuity, destination hospital, and time

of day. Prehosp Emerg Care. 2011;15(3):366-70.

19
Table 1: Reason for Choosing Destination for Ages > 1 day and ≤ 18 years, Florida EMSTARS
2011-2016

Reason for Choosing Overall Age in Years EMS Transport Possible Injury
Destination Time in minutes
N=446,274 Median (IQR) Median (IQR) N=130,369
Closest Facility 208,323 (48.2%) 12 (4-16) 11.3 (7.8-17) 62,042 (47.6%)
Patient / Family Choice 154,035 (35.7%) 10 (3-16) 14 (9.2-22) 41,788 (32.1%)

t
Protocol 35,567 (3.2%) 11 (3-16) 14.7 (9-22) 11,793 (9.0%)

ip
Specialty Resource Center 23,119 (5.4%) 9 (2-15) 20.5 (13-31) 9,158 (7.0%)
Patients Physicians Choice 4,381 (1%) 6 (1-14) 23.9 (14-37.4) 432 (<1%)

cr
Other 3,322 (<1%) 12 (4-16) 17.3 (10.7-26.3) 1,359 (1.0%)
Law Enforcement Choice 2,212 (<1%) 16 (15-17) 14 (9-20) 670 (<1%)
Diversion 671 (<1%) 13 (5-17) 21 (14-30) 231 (<1%)

us
Online Medical Direction 410 (<1%) 10 (3-15) 17.3 (9-33) 181 (<1%)
Insurance Status 63 (<1%) 14 (4-17) 16.7 (9.4-26) 11 (<1%)
Reason for Choosing Destination = Not Known, Not Recorded, or Not Applicable = 14,171 (3.2% of
446,274)
an
M
e d
pt
ce
Ac

20
Table 2: Demographics, EMS Encounter, Vital Signs, and EMS-Administered Medications and
Procedures for EMSTARS patients < 1 day and ≤ 18 years, 2011-2016
Variable Overall Missing Closest Facility Other^ p value
N=446,274 N=208,323 N=237,951
Patient Demographics
Gender – Male 217,545 259 102,546 114,999 (52.9%) < 0.0001
(48.8%) (<1%) (47.1%)
Race 36,847
White 194,674 (8.3%) 92,404 (44.3%) 102,270 (43.0%) < 0.00011
Black or African (43.6%) 63,666 (30.6%) 79,424 (33.4%)
American 143,090 (32%) 34,698 (16.7%) 36,965 (15.5%)
Other2 71,663
(16.1%)

t
Ethnicity – Not Hispanic 291,462 89,567 135,699 155,763 (65.5%) 0.024

ip
or Latino3 (65.3%) (20.1%) (65.1%)
Age in years, Median 11 (3-16) 0 12 (4-16) 10 (3-16) < 0.0001

cr
(IQR)
Age < 5 years 133,966 0 57,374 (27.5%) 76,592 (32.2%) < 0.0001
(30.0%)

us
Age < 10 years 199,299 0 86,972 (41.7%) 112,327 (47.2%) < 0.0001
(44.7%)
Age < 15 years 260,921 0 115,513 145,408 (61.1%) < 0.0001
(58.5%) an (55.4%)
EMS Encounter Characteristics
Incident at Home / 202,768 21,712 87,630 (42.1%) 115,138 (48.4%) < 0.0001
Residence4 (45.4%) (4.9%)
M
EMS Response Mode to 360,274 0 165,119 195,155 (82.0%) < 0.0001
Scene: Lights and Sirens5 (79.0%) (79.3%)
EMS Transport Mode 177,762 0 82,492 (39.6%) 95,270 (40.0%) 0.0028
from Scene: Lights and (39.8%)
d

Sirens5
Time Intervals in
e

Minutes6
Median (IQR) 5.8 (4 – 8.1) 803 5 (3.8 – 8) 6 (4 – 9) < 0.0001
pt

EMS Arrival 12 (8 – 16) (<1%) 12 (8 – 16) 12 (8 – 16) 0.27

Scene 13 (8.3 – 20) 911 11.3 (7.8 – 17) 15 (9.5 – 23) < 0.0001
Transport 19 (13 – 27) (<1%) 19 (12.7 – 26) 20 (13.7 – 28) < 0.0001
ce

ED Turnaround 428
(<1%)
1023
(<1%)
Ac

Possible Injury – Yes 130,369 259 62,042 (29.8%) 68,327 (28.7%) < 0.0001
(29.2%) (<1%)
Vital Sign7
Heart Rate (Mean (SD)) 109.5 (28.9) 4,353 107.8 (28.5) 110.9 (29.2) < 0.0001
Tachycardia 138,600 (<1%) 63,118 (30.3%) 75,482 (31.7%) < 0.0001
(31.1%)
Respiratory Rate (Mean 22.2 (9.9) 11,226 21.9 (10.1) 22.5 (9.7) < 0.0001
(SD)) 75,415 (2.5%) 34,209 (16.4%) 41,206 (17.3%) < 0.0001
Tachypnea (16.9%)
Pulse Oximetry (Median 99 (98-100) 41,553 99 (98-100) 99 (98-100) < 0.0001
21
(IQR)) (9.3%)
Glasgow Coma Score 15 (15-15) 31,206 15 (15-15) 15 (15-15) < 0.0001
(Median (IQR)) (7%)
<15 59,712 26,487 (12.7%) 33,225 (14.0%) < 0.0001
<14 (13.4%) 20,133 (9.7%) 26,295 (11.1%) < 0.0001
46,428
(10.4%)
Systolic Blood Pressure 119.2 (21.8) 65,650 119.8 (21.1) 118.7 (22.4) < 0.0001
(Mean (SD)) (14.7%)
<90 16,086 (3.6%) 6,765 (3.2%) 9,321 (3.9%) < 0.0001
<80 7,634 (1.7%) 3,098 (1.5%) 4,536 (1.9%) < 0.0001
Diastolic Blood Pressure 74.5 (16) 77,543 74.9 (15.4) 74.1 (16.4) < 0.0001
(Mean (SD)) (17.4%)
<40 5,518 (1.2%) 2,142 (1.0%) 3,376 (1.4%) < 0.0001

t
<30 4,042 (1.0%) 1,495 (<1%) 2,547 (1.1%) < 0.0001

ip
8
EMS-Administered Medications & Procedures
Airway Procedure 4,945 (1.1%) n/a 2,454 (1.2%) 2,491 (1.0%) < 0.0001

cr
Other Critical Procedure 5207 (1.2%) n/a 2,760 (1.3%) 2,447 (1.0%) < 0.0001
Contact Medical Control 831 (<1%) n/a 58 (<1%) 773 (<1%) < 0.0001
Resuscitation Medication 3846 (<1%) n/a 2,242 (1.1%) 1,604 (<1%) < 0.0001

us
Nebulized Medication 12,016 (2.7%) n/a 5,353 (2.6%) 6,663 (2.8%) < 0.0001
Benzodiazepine 2,445 (<1%) n/a 1,208 (<1%) 1,237 (<1%) 0.0067
Oxygen 68,105 n/a 29,175 (14.0%) 38,930 (16.4%) < 0.0001
(15.3%) an
Fluids 15,605 (3.5%) n/a 6,126 (2.9%) 9,479 (4.0%) < 0.0001
Dextrose 636 (<1%) n/a 297 (<1%) 339 (<1%) 0.99
Pain Medication 6,240 (1.4%) n/a 2,582 (1.2%) 3,658 (1.5%) < 0.0001
^
Denotes all other destination choices besides closest facility as per Table 1, includes ‘Not Known, Not Applicable,
M
and Not Recorded’
1
Chi Square test performed as White vs all others
d

2
Other includes American Indian or Alaska Native, Asian, Native Hawaiian or Other Pacific Islander, and Other
e

Race
3
Chi Square test performed as Not Hispanic or Latino vs Hispanic or Latino
pt

4
Other includes: Bathtub, Bucket, Farm, Forest/Trail/Wilderness, Health Care Facility, Industrial Place and
Premises, Lake/River/Ocean, Mine or Quarry, Other, Place of Recreation or Sport, Pool, Public Building, Railroad,
ce

Residential Institution, Spa, Street or Highway, Trade or service

5
Chi Square test done as lights and sirens versus all other modes of response/transport
Ac

6
EMS Arrival = Time from EMS unit en route to scene arrival, Scene = Time from EMS scene arrival to EMS scene
departure, Transport = Time from EMS scene departure to EMS emergency department (ED) arrival, Turnaround =
Time from EMS ED arrival to EMS unit back in service, Total = Time from EMS notification by dispatch to EMS
unit back in service

Negative time intervals were treated as errors and classified as missing

7
Created specific for study using PALS 2019 as a reference
8
See online Supplement e1 for list of included medications and procedures

22
Table 3: Reason for Choosing Destination by Provider Primary Impression
Impression Overall Closest Other
N=446,274 N=208,323 N=237,951
Trauma / Injury
Traumatic Injury 122,702 (27.5%) 59,523 (28.6%) 63,179 (26.6%)
Stings / venomous bites 840 (<1%) 401 (<1%) 439 (<1%)
Electrocution 64 (<1%) 25 (<1%) 39 (<1%)
Respiratory

t
Respiratory Distress 34,715 (7.8%) 13,896 (6.7%) 20,819 (8.7%)

ip
Neurologic
Seizure 42,785 (9.6%) 23,969 (11.5%) 18,816 (7.9%)
Behavioral/psychiatric disorder 14,269 (3.2%) 6,730 (3.2%) 7,539 (3.2%)

cr
Stroke/Cerebrovascular accident 501 (<1%) 278 (<1%) 223 (<1%)
Altered level of consciousness 13,188 (3.0%) 6,373 (3.1%) 6,815 (2.9%)

us
Alcohol Related Problem / Delirium tremens 1,406 (<1%) 844 (<1%) 562 (<1%)
Intentional Drug Use / Related Problems 2,010 (<1%) 1,213 (<1%) 797 (<1%)
Endocrine
Diabetic Symptoms 2,186 (<1%)
an 969 (<1%) 1,217 (<1%)
Critically Ill
Cardiac Arrest 1,835 (<1%) 1,349 (<1%) 486 (<1%)
Cardiac Rhythm Disturbance 1,792 (<1%) 753 (<1%) 1,039 (<1%)
Airway Obstruction 2,478 (<1%) 1,080 (<1%) 1,398 (<1%)
M
Hypovolemia / shock 384 (<1%) 149 (<1%) 235 (<1%)
Obvious 122 (<1%) 92 (<1%) 30 (<1%)
Respiratory arrest 550 (<1%) 281 (<1%) 269 (<1%)
Sepsis 403 (<1%) 61 (<1%) 342 (<1%)
d

Other
Syncope / fainting 11,171 (2.5%) 5,335 (47.8%) 5,836 (52.2%)
e

Vaginal Hemorrhage 706 (<1%) 300 (42.5%) 406 (57.5%)

Pregnancy / Obstetric delivery 4,211 (1.0%) 1,342 (31.9%) 2,869 (68.1%)
pt

Hyperthermia 4,071 (<1%) 1,968 (48.4%) 2,103 (51.6%)

Hypothermia 93 (<1%) 48 (51.6%) 45 (48.4%)
ce

Chest pain 7,739 (1.7%) 3,334 (43.1%) 4,405 (56.9%)

Abdominal pain / problems 28,616 (6.4%) 14,607 (51.0%) 14,009 (49.0%)
Sexual assault / rape 351 (<1%) 171 (48.7%) 180 (51.3%)
Poisoning / drug ingestion 10,287 (2.3%) 5,007 (48.7%) 5,280 (51.3%)
Ac

Inhalation injury (toxic gas) 159 (<1%) 77 (48.4%) 82 (51.6%)

Smoke inhalation 227 (<1%) 55 (24.2%) 172 (75.8%)
Allergic reaction 10,075 (2.3%) 5,039 (50.0%) 5,036 (50.0%)
Blood pressure related problems 254 (<1%) 101 (39.8%) 153 (60.2%)
Fever Related Symptoms / Problems 10,798 (2.4%) 3,516 (32.6%) 7,282 (67.4%)
General Malaise 16,787 (3.7%) 7,511 (44.7%) 9,276 (55.3%)
Heart Related Illness 702 (<1%) 241 (34.3%) 461 (65.7%)
Other, Non Traumatic Pain 24,988 (5.6%) 16,423 (65.7%) 8,565 (34.3%)
Sickle Cell Crisis 148 (<1%) 49 (33.1%) 99 (66.9%)
Missing Provider Primary Impression = 72,661 (16.3%)

23
Figure 1: Inclusion and Exclusion Criteria Producing Final Study Sample

• EMSTARS 2011 - 2016 patients of all ages

3,491,241

• Age 18 years or less, 911 call, transported

466,075 to ED

t
ip
• Removed Age < 1 day (N=1,527)

cr
466,075

us
• Removed Age missing (N=4,298)
464,548
an
• Removed interfacility transports
M
460,250 (N=13,976)
d

• Final study sample

446,274
e
pt
ce
Ac

24