Prevalence of Low Acuity Pediatric Emergency Medical Services (EMS) Transports to a Pediatric Emergency Department (ED) in an Urban Area

Abstract

Objectives

Many patients transported by Emergency Medical Services (EMS) do not have emergent resource needs. Estimates for the proportion of pediatric EMS calls for low acuity complaints, and thus potential candidates for alternative dispositions, vary widely and are often based on physician judgment. A more accurate reference standard should include patient assessments, interventions, and dispositions. The objective of this study was to describe the prevalence and characteristics of low acuity pediatric EMS calls in an urban area.

Methods

This is a prospective observational study of children transported by EMS to a tertiary-care pediatric emergency department. Patient acuity was defined using a novel composite measure that included physiologic assessments, resources used, and disposition. Bivariable and multivariable logistic regression was conducted to assess for factors associated with low acuity status.

Results

996 patients were enrolled, of whom 32.9% (95%CI 30.0,36.0) were low acuity. The majority of the sample was Black, Non-Hispanic with a mean age of 7 years. When compared to adolescents, children <1 year were more likely to be low acuity (aOR 3.1 [1.9,5.1]). Patients in a motor vehicle crash were also more likely to be low acuity (aOR 2.4 [1.2,4.6]). All other variables, including race, insurance status, chief complaint, and dispatch time, were not associated with low acuity status.

Conclusions

One-third of pediatric patients transported to the pediatric ED by EMS in this urban area are for low acuity complaints. Further research is needed to determine low acuity rates in other jurisdictions and whether EMS providers can accurately identify low acuity patients to develop alternative EMS disposition programs for children.

INTRODUCTION

Background and Importance

Emergency Medical Services (EMS), originally designed for the rapid transport of patients with imminently life-threatening illnesses and injuries, now receives more than 28 million requests for 911-scene response each year.¹ Many patients transported by EMS do not need emergent interventions.² This misallocation of EMS resources is unsafe and inefficient.³ There have been calls to action from both the federal government and industry leaders to develop alternative EMS disposition processes.^4,5 These alternatives include leaving patients at the scene, transporting patients to urgent care facilities or clinics, or substituting rideshare services for ambulances. Children, however, have been excluded from most of these alternative disposition options. The reasons for excluding children are multifactorial but include a lack of evidence accurately describing the prevalence and characteristics of low acuity pediatric EMS transports that would be candidates for alternative EMS dispositions.⁶

Estimates for the proportion of pediatric EMS transports that could be considered low acuity vary widely from 16 – 61%.^7–9 The use of EMS for non-emergent complaints is more likely for patients with public insurance,^8–11 non-white race,^8,9,11 young age,¹¹ arrival in the ED overnight,^9,10,12,13 and specific chief complaints (including fever¹⁴ and behavior problems¹⁰). These studies, however, have used problematic approaches to define low acuity EMS transports, including ESI triage,⁹ final ED diagnosis,^8,15 and physician judgment.^7,15 These approaches lack validity in determining the need for emergent ambulance transport to the emergency department (ED). Furthermore, none of these approaches are consistent with the best practice principles established by the Neely Conference: Developing Research Criteria to Define Medical Necessity in EMS.¹⁶ Conference participants agreed that when defining medical necessity in EMS, it is important to include both triage criteria (such as vital signs) and outcome measures (including patient care interventions and disposition).

Study Objectives

An accurate understanding of the prevalence and characteristics of non-emergent pediatric EMS calls is a prerequisite to assessing the accuracy of EMS providers’ determinations of medical necessity in children and ultimately developing alternative disposition triage protocols. Therefore, this study was conducted to determine the prevalence of non-emergent pediatric EMS transports within a single urban region. The secondary objective was to identify patient and encounter characteristics associated with non-emergent EMS transport.

METHODS

Study Design and Setting

This is cross-sectional study of children transported by EMS to an urban academic pediatric ED. This facility is a freestanding children’s hospital with a Level 1 Pediatric Trauma Center. The hospital provides EMS medical control for pediatrics and receives approximately 5,000 pediatric EMS transports per year. This represents almost all pediatric EMS transports from the District of Columbia (DC), and most pediatric EMS transports from two neighboring counties in Maryland. The vast majority of EMS transports to the study site are from four EMS agencies, all of which are two-tiered ALS and BLS systems. Three of these agencies are entirely staffed by career EMS clinicians, and one is comprised of both career and volunteer staff. An a priori decision was made to collect an initial sample of 1000 patients to enable multivariable modeling with 15 predictor variables for the outcome of low acuity status (assuming 15% of enrolled patients were low acuity). The local Institutional Review Board approved this study.

Selection of Participants

The data for this study were collected from the electronic health care records of children transported to the study ED by EMS. Patients were eligible for enrollment if they were less than 18 years old. Interfacility transports were excluded. Patients with an Emergency Severity Index (ESI) score of 1 were also excluded (requiring immediate life-saving interventions),¹⁷ as this study was part of a larger investigation that involved approaching caregivers to complete a survey. Children were eligible for enrollment even if they had previously been transported to the ED by EMS during the study window. Caregivers of eligible children were approached consecutively for consent by research staff during open enrollment periods when research staff were available (Monday – Friday from 8 am – 11 pm, and Saturday – Sunday from 2 pm to 10 pm).

Data Source

Trained research staff reviewed the ED medical record and EMS care report (where available). Data from the EMS care report was retrieved by manual chart review by trained study staff. Data from ED medical record was extracted electronically. Data extracted from the EMS care report included: the chief complaint, all vital signs charted, medications administered, and procedures performed. Data extracted from the ED medical record included: arrival time, patient demographics, zip code of residence, insurance status, all vital signs charted, ESI triage level, medications administered, radiographic procedures performed, procedures completed, disposition and any return visits with five days of the index visit.

Outcome Measures

The primary outcome variable was patient acuity. Enrolled patients were classified as “low acuity” based on a novel composite measure that included physiologic assessments, resources used, and final disposition. For this dichotomous variable, low acuity was defined as any EMS transport where the patient had stable vital signs, did not receive any medications or procedures from EMS, did not receive any radiographic imaging or blood based testing the ED, and were discharged home from the ED (without any return visits within 5 days that led to admission). Encounters were excluded from the low acuity group if they met one or more of the following criteria: unstable vital signs; required any procedures or medications from EMS; required blood-based laboratory tests, radiographic imaging, or procedures in the ED; or had a final disposition of admission to the hospital (either on the date of the initial EMS encounter or as a result of a return ED visit within 5 days) or death. This composite definition, which includes both triage criteria and outcome measures, is consistent with the principles agreed on at the Neely Conference: Developing Research Criteria to Define Medical Necessity in EMS.¹⁶ Unstable vital signs were defined as any oxygen saturation less than 96%, systolic blood pressure less than the 5^th percentile for age,¹⁸ or vital signs that triggered the automated sepsis alert in the ED (details of the sepsis screening tool used as the study institution are provided in Supplemental Table 1).¹⁹ For both the oxygen saturation and systolic blood pressure all recorded vital signs in both the EMS and ED records were reviewed. EMS procedures included any airway intervention, placement of a peripheral intravenous catheter, and application of a splint or cervical collar. EMS medications included all medications administered, regardless of route of administration. Procedures in the ED included laceration repair, fracture reduction, and procedural sedation. Patients were excluded from the low acuity group if they received radiographs or blood based tests regardless of whether these results were abnormal. Similarly, patients who had a cervical spine collar placed by EMS were excluded, regardless of whether the child had an actual injury or even if this decision was correct per local protocol. While some of these patients may ultimately be considered to have a low acuity problem (for example a child with ankle pain and a negative radiograph), the treating clinician provided resources that are not routinely available if a child was transported by private vehicle or seen in a primary care setting. Based on previous studies,^8–15 the following variables were included as potential factors associated with low acuity: age, gender, race/ethnicity, use of an interpreter, insurance status, household zip code income quartile, weekday and hour of arrival, chief complaint organ system, injury status, whether a patient was involved in a motor vehicle crash, and ESI triage level.

Analysis

The prevalence of all eligible transports that were low acuity was calculated and descriptive statistics generated to summarize the study population. The variables in the regression analysis with missing data were patient race (1.4%), residential zip code (8.3%), and insurance status (5.6%). Imputation techniques were not used for these missing data elements. Bivariable logistic regression was used to assess for potential factors associated with low acuity patient status. An a priori decision was made to adjust the final multivariable regression models for age, race/ethnicity, insurance status, and household zip code income quartile. All statistical analyses were conducted using SAS software, Version 9.3 (SAS Institute, Inc, Cary, NC).

RESULTS

Characteristics of Study Subjects

996 subjects were enrolled between August 2020 and September 2021. 98.1% of caregivers that were approached agreed to participate in the study. There was missing EMS documentation for 11.1% of all enrolled patients. The mean age was 7 years (standard deviation of 5.5 years), and 52.1% were male. The most common race/ethnicity responses were non-Hispanic Black (62.6%) and Hispanic (17.5%). Most patients were enrolled in public insurance programs (69.1%). These sample characteristics are similar to the study institution’s overall ED patient demographics. 39.5% of patients arrived in the ED outside of usual primary care office hours. (Table 1)

Table 1:

Characteristics of enrolled children transported to the emergency department (ED) by Emergency Medical Services (EMS) (Total n = 996)

Variable
Age (years)	N	%
< 1	125	12.6
1 – 3	297	29.8
4 – 6	133	13.4
7 – 12	230	23.1
13 – 18	211	21.2
Mean age (std)	7.0 (5.5)
Sex
Male	519	52.1
Female	477	47.9
Race
Black/African-American, NH	623	62.6
Caucasian, NH	111	11.1
Hispanic	174	17.5
Other	74	7.4
Not Documented	14	1.4
Interpreter
Yes	85	8.5
No	911	91.5
Insurance status
Private insurance	195	19.6
Public	688	69.1
No insurance	58	5.8
Unknown	55	5.5
Median household income by zip code
Fourth quartile	237	23.8
Third quartile	224	22.5
Second quartile	219	22.0
First quartile	233	23.4
Missing	83	8.3
Date/time arrival
Office Hours (Mon-Fri, 8 am – 5 pm)	603	60.5
Out of Office Hours	393	39.5
Chief Complaint
Behavioral/psychiatric	46	4.6
CNS/Neurologic	148	14.9
Global/general	171	17.2
Musculoskeletal/skin	370	37.2
Pulmonary	136	13.7
Other	51	5.1
Injury
Yes	386	38.8
No	610	61.2
Motor vehicle crash victim
Yes	67	6.7
No	929	93.3
ESI triage level
2	132	13.2
3	520	52.2
4	319	32.0
5	25	2.5

Main Results

32.9% (95% CI 30.0, 36.0) of enrolled patients were classified as low acuity. The most common reasons patients were excluded from the low acuity group were interventions in the ED, including radiographic testing (35.8%) and blood testing (29.5%). The most common vital sign abnormality that excluded children from the low acuity group was hypoxia (14.8%). For the complete cohort of enrolled children, 23.1% of children were admitted to the hospital (the admission rate for all children in the study ED varies from 12 – 16%). There were no repeat ED visits within 5 days resulting in admission, nor were there any deaths in the entire study cohort. (Table 2)

Table 2:

Indication for an enrolled child transported to the emergency department (ED) by Emergency Medical Services (EMS) being excluded from the low acuity group

Variable	N	%
EMS Interventions
Administration of any medication	160	18.1
Procedure completeda	104	11.7
Patient vital signs
Hypoxia (SpO2 < 96%)	147	14.8
Systolic hypotension (<5th percentile for age)	27	2.7
Vital signs triggered ED “Sepsis Alert”	14	1.4
ED Interventions
Radiographic testing	357	35.8
Blood or CSF testing	294	29.5
Administration of intravenous medications	266	26.7
Procedure completedb	98	9.8
Patient Disposition
Admission to hospital (general inpatient unit or ICU)	230	23.1
Discharged with repeat ED visit within 5 days leading to admission	0	0.0
Death in the ED	0	0.0
TOTAL (all patients with 1 or more indicators for being excluded from low acuity group) c	668	67.1%

^aEMS Procedures included airway interventions, placement of a peripheral intravenous catheter, application of a splint or cervical collar.

^bED procedures included laceration repair, fracture reduction, and procedural sedation.

^cA patient could be excluded for multiple reasons (e.g., hypoxia and admission to hospital); thus, the total of each individual exclusion factor is greater than 668.

In bivariable analysis, low acuity status was strongly associated with ESI triage level, with 88.0% of ESI 5 and 52.4% of ESI 4 patients classified as low acuity. Young age was significantly associated with low acuity status, with children < 1 year and 1 – 3 years more likely to be low acuity compared to the referent group of adolescents aged 13 – 18 years. Children involved in a motor vehicle crash were more likely to be low acuity. Patient encounters requiring the use of an interpreter were less likely to be low acuity. There was no association between low acuity and patient gender, race/ethnicity, insurance status, median household income by zip code, or arrival time. (Table 3)

Table 3:

Factors associated with low acuity status for enrolled children transported to the emergency department (ED) by Emergency Medical Services (EMS) in univariable and multivariable regression modeling

Variable	Low-acuity (%)	OR (95% CI)	aOR (95% CI)a
Age (years)
< 1	50.4	3.2 (2.0, 5.1)	3.1 (1.9, 5.1)
1 – 3	41.1	2.2 (1.5, 3.2)	2.1 (1.4, 3.2)
4 – 6	27.8	1.2 (0.7, 2.0)	1.2 (0.7, 2.0)
7 – 12	23.9	1.0 (0.6, 2.0)	0.9 (0.6, 1.5)
13 – 18	24.2	Reference	Reference
Sex
Male	29.3	Reference	Reference
Female	36.9	1.4 (1.1, 1.8)	1.4 (1.1, 1.9)
Race
Black/African-American, NH	36.3	1.5 (0.9, 2.4)	1.3 (0.8, 2.3)
Caucasian, NH	27.0	Reference	Reference
Hispanic	28.2	1.1 (0.6, 1.8)	1.0 (0.5, 1.8)
Other, NH	23.0	0.8 (0.4, 1.6)	0.8 (0.4, 1.6)
Unknown	42.9	2.0 (0.6, 6.3)	1.6 (0.5, 5.4)
Interpreter Use
Yes	18.8	0.4 (0.3, 0.8)	0.4 (0.2, 0.8)
No	34.2	Reference	Reference
Insurance status
Private insurance	28.7	Reference	Reference
Medicaid	36.8	1.4 (1.0, 2.1)	1.2 (0.8, 1.9)
No insurance	36.2	1.4 (0.8, 2.6)	1.4 (0.7, 2.7)
Unknown	21.4	0.7 (0.4, 1.1)	0.5 (0.2, 1.1)
Median household income by zip code
Fourth quartile	30.8	Reference	Reference
Third quartile	33.0	1.1 (0.7, 1.6)	1.0 (0.7, 1.5)
Second quartile	30.6	1.0 (0.7, 1.5)	0.9 (0.6, 1.4)
First quartile	39.9	1.5 (1.0, 2.2)	1.2 (0.8, 1.8)
Missing	25.3	0.8 (0.4, 1.3)	1.0 (0.6, 1.9)
Date/time arrival
Office Hours (Mon-Fri, 8 am – 5 pm)	35.3	Reference	Reference
Out of Office Hours	29.3	0.8 (0.6, 1.0)	0.8 (0.6, 1.1)
Chief Complaint
Behavioral/psychiatric	39.1	1.0 (0.5, 1.9)	2.1 (1.0, 4.1)
CNS/Neurologic	20.3	1.2 (0.6, 2.2)	0.5 (0.3, 0.7)
Gastrointestinal	24.3	0.6 (0.3, 1.1)	0.5 (0.3, 0.9)
Global/general	44.4	1.5 (1.0, 2.1)	1.4 (1.0, 2.1)
Musculoskeletal/skin	35.1	Reference	Reference
Pulmonary	27.9	0.7 (0.5, 1.1)	0.5 (0.3, 0.8)
Other	35.3	1.0 (0.5, 1.9)	1.2 (0.6, 2.2)
Injury
Yes	35.0	1.2 (0.9, 1.5)	1.2 (0.9, 1.7)
No	31.6	Reference	Reference
Motor vehicle crash victim
Yes	49.3	2.1 (1.3, 3.4)	2.4 (1.4, 4.0)
No	31.8	Reference	Reference
ESI triage level
2	8.3	Reference	Reference
3	24.6	3.6 (1.9, 6.9)	3.6 (1.9, 7.0)
4	52.4	12.1 (6.3, 23.3)	11.0 (5.7, 21.5)
5	88.0	80.7 (20.8, 312.7)	75.3 (18.9, 330.8)

^aMultivariable regression model is adjusted for patient age, race/ethnicity, insurance status & household zip code income quartile.

The multivariable regression model (Table 3) was adjusted for age, race/ethnicity, insurance status, and household zip code income quartile. The findings were not significantly different than in the bivariable analyses. Young patient age and mechanism of motor vehicle crash were more likely to be low acuity, and patients who used an interpreter were less likely to be low acuity.

DISCUSSION

Almost one-third of pediatric EMS transports to the pediatric ED in the study region are for low acuity complaints. These children had stable clinical vital signs, no emergent resource needs from either EMS or the ED, and an outcome disposition of discharged. These patients may still have time-sensitive medical and social needs, but these could be addressed safely through an alternative EMS disposition program. Young age and being a motor vehicle crash victim were associated with low acuity patient status. Patients who required an interpreter were less likely to be low acuity. Most other factors, including other patient demographics, chief complaint, and arrival in the ED out of office hours, were not associated with low acuity.

The percentage of low acuity patients in the study, while substantial, is lower than most previous studies, which have reported low acuity rates of up to 61%.^7,9,23 This may be explained by the different definitions these studies used for low acuity. This study intentionally used a conservative definition to try and capture only those children who clearly had no emergent or urgent resource needs from EMS or the ED. For example, a child who needed a laceration repair or radiograph for an ankle injury would have been excluded from the low acuity group but may have been classified as low acuity in other studies. It is also possible that the study sample was biased, as some EMS agencies in the study transport a small number of lower acuity cases to other community hospitals. The proportion of pediatric patients taken to other facilities for EMS agencies in the region varies from 9 – 30%. The finding that young age is associated with low acuity status is consistent with previous studies.¹¹ The finding that patient race/ethnicity, insurance status, and household income by zip code were not associated with low acuity status differs from previous studies.^8,9,11,24. The patient population in this study may not have sufficient variability in these variables to capture significant differences. In this study, patients who required an interpreter were less likely to have a low acuity complaint. This is consistent with previous studies,^25,26, and suggests that non-English speaking patients may not feel comfortable accessing EMS. Finally, the strong association between ESI triage level and low acuity in this study provides evidence for the face validity of the novel outcome measure for low acuity. The ESI triage levels were designed for use in the ED to sort patients on the basis of acuity and resource needs. These findings suggest this may be of use in the prehospital setting, but studies have found only moderate inter-rater agreement when EMS clinicians and ED nurses assign ESI triage levels to patients.²⁷

This study confirms that when using a composite definition for low acuity that includes both triage criteria and outcome measures, a substantial proportion of pediatric EMS transports are for low acuity complaints. Previous studies show that another one-third of all pediatric EMS calls are left at the scene,^28–31 with most of these children being classified as low acuity by EMS. When taken together, this suggests that probably more than half of the children assessed by EMS in the study region could be candidates for alternative EMS disposition programs. These children might be able to receive appropriate care at clinic sites, be safely transported by private vehicle, or not need any additional medical care beyond the initial EMS assessment. Given that children involved in a motor vehicle crash, and those aged less than 1 year old were more likely to be low acuity in this study, these might be specific populations for initial pilot testing of alternative EMS disposition programs.

The use of EMS for low acuity complaints places significant stress on EMS systems, leading to prolonged wait times.³² Furthermore, because EMS is traditionally only reimbursed for transporting patients to the ED, this also reinforces the behavior of seeking care in EDs for minor illnesses, which contributes to ED crowding. ED crowding is associated with higher rates of leaving without treatment,³³ return visits, medication errors,³⁴ and procedure complications.³⁵ The high prevalence of non-emergent EMS calls thus has an adverse impact on the care of all patients. The use of the EMS system for low acuity complaints is also inefficient, costing the federal government at least $560 million annually.³

Given the high prevalence of low acuity pediatric EMS calls and the associated harms, there is an urgent need to develop alternatives for this patient population. This could include the development of alternatives to calling 9–1-1(such as the 9–8-8 National Suicide Prevention Lifeline³⁶), and more robust and accessible out of hours primary care advice lines.³⁷ The evidence for these types of interventions reducing EMS and ED utilization for non-emergent complaints, however, is mixed. Another solution to this problem would therefore be to develop alternative disposition programs for EMS. It will be important to determine whether EMS clinicians can accurately predict children without emergent resource needs. A meta-analysis of United States (US) paramedic determinations of medical necessity found that the negative predictive value (NPV) for paramedic determinations of medical necessity of ambulance transport was 0.62 – 0.99, and the NPV for the necessity of ED evaluation was 0.45 – 0.98.³⁸ Only one study in this review included significant numbers of children.³⁹ In addition, the reference standard for most studies was problematic, using either physician judgment or hospital admission. Knowing how often EMS clinicians accurately predict a low acuity pediatric EMS patient will be critical. It will guide how much detail needs to be incorporated into a triage protocol and how much education will be needed with protocol implementation. Caregiver perceptions of acuity will also need to be better understood prior to implementing pediatric alternative EMS disposition programs. Previous studies suggest that patients are broadly supportive of alternative EMS disposition programs,^40,41 but have included limited numbers of pediatric caregivers and non-white respondents.

This study has several limitations. First, this was a single-center study undertaken in an urban area. Caution should be taken, extrapolating these findings to other populations. Second, this was a convenience sample of patients enrolled when research staff were available in the ED. It is reassuring that the patient sample had similar characteristics to overall ED patient data, and the analysis did not find an association between arrival out of office hours and low acuity status. Third, by excluding patients triaged as ESI 1 the study will have over-estimated the proportion of low acuity patients by up to 9.4%. Fourth, the study was conducted during the novel Coronavirus-19 (COVID-19) pandemic. The pandemic has resulted in changes in overall EMS call volumes and patient characteristics, with studies finding patient acuity during the early phases of the pandemic increased.^20,21 This may reflect a reluctance of patients to utilize 9–1-1 for a less acute complaint, due to fear of possible exposure to COVID-19 in the ED²². Therefore, it is likely that this data reflect an underestimate of low-acuity transport rates during non-pandemic times. Finally, the outcome of low acuity is subject to potential ascertainment bias. The study was only able to assess children transported to the ED by EMS. The outcomes of non-transported children are not known. This likely represents additional low acuity patients not ascertained in the study sample.

CONCLUSION

In summary, almost one-third of EMS transports to the pediatric ED in this urban region are low acuity encounters, with no emergent resource needs. These findings suggest an unmet need for alternative disposition programs for low acuity pediatric patients. Further research is needed to determine whether EMS clinicians can identify these children accurately and whether children can be safely included in alternative EMS disposition programs.