Pediatric and Adult Interfacility Transports: A Descriptive Comparison Open Access

Interfacility transport (IFT) is an important component in the health care system of pediatric patients.^1–4 When presenting with a complex, resource-intensive, or life-threatening issue, these patients are often referred to tertiary pediatric centers. Previous work demonstrated that more than 30% of the pediatric population lives over 50 miles from a pediatric capable hospital.^5,6 On a national level, recent data showed an increase in pediatric interfacility transports in the last 5 years.⁷ Specialization and regionalization of pediatric services over the past 2 decades have increased the distance and, therefore, time to definitive care.^8–11 

Given the increasing number of pediatric patients needing interfacility transport, clinicians, teams, and facilities who primarily care for adults will inevitably be involved in their care. The intricacies of pediatric care derive from the anatomic, physiologic, and developmental differences that set them apart from adults. Teams may have familiarity with adult IFTs but not pediatric IFTs.³ Although the field of transport medicine continues to mature, the literature on pediatric IFTs is overall lacking. Rosenthal et al provided a cross-sectional analysis through the Kids Inpatient Database using data from 2012 and found that interfacility transfers are relatively common and had a higher association with higher illness severity, principal diagnosis, insurance status, and race.¹² There are few other examples of large-scale descriptive studies of pediatric IFTs.^13,14 The remaining pediatric IFT literature is focused on trauma care,^15,16 use of specialized transport teams,^15,17–21 modality decisions,²² communication^23–25 and access to care.^26–29 There are no known studies comparing pediatric with adult IFTs or how knowledge of adult IFTs may impact pediatric IFTs.

The objective of this study is to bridge that gap by describing the key demographic, clinical, and transport characteristics of pediatric as compared with adult interfacility transports from a large recent national database. We hope to better inform transport teams, first-point-of-contact clinicians, and training resources such that better, more effective en route care can be delivered to this important patient population.

This is a large national cross-sectional study using the National Emergency Medical Services (EMS) Information System (NEMSIS) research dataset.³⁰ The NEMSIS is a federally funded standardized national database used to collect and store agency, provider, and patient EMS data. Participating state EMS agencies are mandated to report their data after each patient/transport encounter. To avoid any confounding effects of the COVID-19 pandemic, the 2019 NEMSIS dataset was selected. The use of the NEMSIS dataset has been described in other studies.³¹ After acquiring the data in American Standard Code for Information Interchange comma-delimited flat files, they were reconstituted in a relational database management system (Microsoft SQL Server). The data were cleaned and validated, and relationships were established between corresponding tables. The inclusion criteria included completed interfacility transports whereby the disposition was either the “Emergency Department” or an “Inpatient” destination. Records were excluded if they did not contain or had incomplete age, sex, geographic, or timestamp data because these were deemed critical for analysis. Transport records not “transported by this EMS” were removed to limit duplicated patient encounters and exclude those patients not transported. Finally, the data were split into adult (>17 years old) and pediatric encounters.

Given the vast data available in the NEMSIS research dataset, focus was applied to relevant variables with preference given to mandated fields to reduce missingness. Variables used “as-is” and included demographics, day of the week, season, initial and final acuity, primary organ system involved, US census division of the origin facility, population setting of the origin facility, transport modality, initial destination, final disposition, the presence of injury, and the occurrence of arrest. For time of transport, the clock was divided into 0600 to 1400, 1400 to 2200, and 2200 to 0600 to represent early to midday, afternoon to early evening, and late night to early morning, corresponding to common shift hours. Length of transport was calculated from the duration between timestamps corresponding to time when the EMS crew left the origin facility until arrival at the destination facility, divided into 30-minute segments. The NEMSIS dataset can include multiple sets of vital signs for each encounter. For this analysis, an age-appropriate cutoff was derived for each patient using American Heart Association, Advanced Cardiac Life Support, and Pediatric Advanced Life Support guidelines. The calculated percentage reflects the number of unique encounters with the labeled vital sign abnormality (eg, 2 readings of tachycardia for a single transport were only tallied once). Both the symptoms and impressions fields were derived from entered International Statistical Classification of Diseases, Tenth Revision codes and then sorted and ranked by highest occurrence. Because these are ranked lists, a direct statistical comparison was not done.

Record sets were exported to R version 4.0.5 (R Foundation for Statistical Computing, Vienna, Austria), and descriptive statistics were performed. We used 2-proportion Z-testing for intracategorical proportion comparisons and χ² testing to compare categorical groups; both used an α of 0.05.

This study was certified as exempt from institutional review board review given its secondary use of retrospective data by the Colorado Multiple Institutional Review Board (COMIRB #20-2715). Data use agreements were obtained for the NEMSIS research dataset. There was no source of funding for this project and, therefore, no disclosures to include.

This study analyzed 190 410 pediatric and 1 535 018 adult interfacility transports after inclusion/exclusion criteria (Figure 1). Table 1 illustrates the primary demographics between those groups, including age, sex, race, insurance status, and geographic information. Across both pediatric and adult patients, all age groups were well represented. By sex, there was a noticeable female predominance in pediatric transports (53.1% female vs 46.9% male). The NEMSIS database allows multiple selections for the race field. To account for this, Table 1 shows those transports wherein the patient was identified as at least that race; therefore, the sum of the percent column will exceed 100%. Because of this, cross-categorical comparison was not done. There was no statistical difference in the proportion of pediatric vs adult transports with “white” or “Hispanic or Latino” race selected. Secondary analysis of all transports showed 77 233 (4.5%) marked 2 races and 5550 (0.3%) marked 3 or more. Insurance data were unavailable for 46% of transports. Of the data available, there was a statistically significant difference in the proportions of those who selected Medicare (1% vs 15%), Medicaid (15% vs 5%), private insurance (24% vs 18%), and self-pay (3% vs 3.3%) between pediatric and adult transports, respectively (P < .01 across all insurance options). Finally, Table 1 also provides masked geographic information of the originating facility, including population setting (urbanicity) and US census division. There was a greater percentage of pediatric urban transports compared with adults (77.9% vs 69.6%; P < .01). Census division proportions were similar between the 2 groups but retained overall statistical difference (P < .01).

Table 2 examines the transport characteristics of pediatric vs adult IFTs. Ground transportation was the prominent modality (89.2% pediatric vs 91.4% adult IFTs) followed by rotor wing (6.9% vs 5.9%; P < .01) and fixed wing (2% vs 1%; P < .01). Although initial disposition was an inclusion criterion and, therefore, only 2 options were allowed, the difference in proportions between pediatric and adult IFTs dispositions appeared similar but retained statistical difference (P < .01). Inpatient destinations were the majority (53.7% vs 55.7%) over the emergency department (ED) (46.3% vs 44.3%). Final disposition was initially examined, but, because of missingness being more than 75%, it was excluded.

Temporal information is also shown in Table 2. Pediatric transports more commonly occurred during the late-night/early-morning block (2200–0600; 30.5% vs 22.5%; P < .01), and adult patients were transported more so during the early/midday (0600–1400; 25.4% vs 30.8%; P < .01). Day of the week analysis indicated a greater proportion of weekend pediatric IFTs (26.5% vs 23.9%; P < .01) and a greater proportion of weekday adult IFTs (73.5% vs 76.1%; P < .01). Pediatric transports occurred most often in the fall (26.1% in pediatrics vs 24.3% in adults; P < .01), whereas adults were transported most often in the summer (23.9% vs 26.3%; P < .01). When looking at the length of transport, there was a statistically significant difference (P < .01) between pediatric and adult transports. Adult IFTs accomplished transport 79% of the time in less than 1 hour, whereby only 46% of pediatric interfacility transports finished under the 60-minute mark. Over 45% of pediatric interfacility transports lasted longer than 60 minutes, nearly 7% of which were over the 120-minute mark. In contrast, 19% of adult transports lasted over 60 minutes.

Clinical characteristics were outlined in Table 3. Injury played a larger role in pediatric interfacility transports (10% vs 7.42%; P < .01). Incidents of documented cardiopulmonary arrests were infrequent but greater in adults (0.1% vs 0.17%; P < .01). At least 1 vital sign was abnormal for age in 40.6% of pediatric transports, as compared with 25.8% of adult transports (P < .01). Pediatric patients tended to have more tachycardia (25.4% vs 5.3%; P < .01), hypotension (4.9% vs 3.3%; P < .01), and tachypnea (13.3% vs 9.4%; P < .01). The initial and final acuity fields are often missing data (26.3%–32.7% recorded not available [“N/A”]), but of those recorded, most patients were labeled as “lower acuity” (46.7% vs 49.3%; P < .01). The proportion of those patients with a final acuity of “dead” was not statistically different between pediatric and adult transports (0.03%; P = .455). Primary organ system involved was also queried. Pulmonary concerns were the most common in pediatric IFTs (10.2% vs 5%; P < .01). Cardiovascular involvement was the highest in adult IFTs (1.4% vs 10.1%; P < .01). There were also nearly double the number of pediatric IFTs involving the “behavioral” organ system (7.3% vs 3.8%; P < .01).

Figure 2 shows the top 10 primary symptoms and top 10 primary impressions. Both groups reported generalized pain, need for observation, and suicidal ideation as top primary symptoms. Unique to pediatric IFTs was fever with chills. Unique to adult IFTs were chest pain and changes in mental status. Pediatric transports documented specific respiratory symptoms (shortness of breath, dyspnea, or abnormalities of breathing) 8.2% of the time. Overall, there were 842 unique pediatric symptoms and 1359 unique adult symptoms, with the top 10 symptoms composing 36.2% of pediatric transports and 40.2% of the adult transports.

Top 10 primary impressions seen in Figure 2 showed variation between pediatric and adult IFTs, with 2 of the pediatric impressions being respiratory (acute respiratory distress syndrome, respiratory disorder) and 3 of the top adult impressions being cardiovascular in nature (cardiac arrythmia, chest pain, disorder of circulatory system). Fever with chills is unique to pediatrics (2.5%), whereas concern for stroke was present for 2.1% of adult transports. Abdominal pain takes a position on each impression list at equivalent proportions (4.5%). Both groups also have injury, mental illness, and mental status changes in the top impressions. These top 10 primary impressions accounted for 29.7% of pediatric transports and 30.5% of adult transports. There were 1804 unique impressions for pediatric IFTs and 3681 unique impressions for adult IFTs.

This work represents the largest known national cross-sectional study comparing pediatric and adult interfacility transports. Our goal was to leverage recent retrospective data to fill a sizable knowledge gap in this field. We closely examined the key differences in demographic, transport, and clinical characteristics in over 1.6 million interfacility transports. In this study, 11% of transports were pediatric IFTs, similar to the 10% cited in other descriptions of pediatric EMS encounters.³² 

Because of the large sample sizes involved, virtually all comparisons resulted in statistical significance, even when the proportional differences were minimal (1%–3%). We aim to interpret our results in the context of both statistical significance and practical applicability.

Aside from age, demographics were overall proportionally similar between pediatric and adult IFTs. These transports included a wide distribution of patients, including neonates, school-age children, and adolescents, suggesting that transport crews should be familiar with all ages, both sexes, and patients from both urban and rural environments. Proportionally, the pediatric and adult IFT US census divisions were similar, often within 2%. However, when compared with the US national census data from 2020,³³ our US Census Division distribution indicated a higher proportion from the West South Central division (22.3% pediatric vs 21.5% adult; national distribution 12.3%) and relatively lower from the New England, Middle Atlantic, and the Pacific divisions. This suggests our data are generally representative of typical demographic measures, but our geographic representation may be skewed. It is unclear why the West South Central division states (Arkansas, Louisiana, Oklahoma, and Texas) were overrepresented. Influence by access to health care, geography, and natural disasters may be to blame. It is challenging to interpret our primary insurance data given the large degree of missingness (56%–58%), but pediatric patients tended to use private insurance more than adults, likely reflecting coverage by a parent’s insurance plan. Medicaid and Medicare use differed as expected between children and adults.

Transport characteristics including modality, destination, and use timing were generally similar. Although this study selected only those IFTs going to an ED or inpatient unit, the proportions for each group did not appear different between pediatric and adult IFTs. This draws into question indications for transfer and the capabilities of the sending facility. Given little difference between the pediatric and adult IFT destinations, it may suggest that similar levels of discomfort might be happening, with adult patients in low-acuity or resource-constrained originating facilities who still need further primary workup in an ED. There was a small but noted increase in the use of air transports for pediatric IFTs. Other literature has commented on this trend and may be the result of overtriaging.³⁴ Both pediatric and adult IFTs occurred most over the 1400 to 2200 timeframe, which corresponds to literature suggesting a circadian pattern to most medical emergencies plus the time needed to triage the subsequent IFT.³⁵ Pediatric IFTs also happened in slightly increased frequency on the weekend—similar to what is reported by Rosenthal et al.¹² Pediatric IFTs tended to occur least during the summer; the opposite is true in adult IFTs. This supports illness exposure patterns during the school year and adult trauma patterns during the summer.

Most significantly, we noted differences in transport duration, with only 47% of pediatric IFTs completed during the first 60 minutes compared with 79% of adult IFTs. Because we can assume similar modes of transport would operate at similar speeds, this discrepancy likely reflects the distance over which pediatric IFTs travel. This would be consistent with the literature describing regionalization and distribution of pediatric specialty care centers.^8–11 

Finally, when looking at clinical characteristics, expected differences were noted. There were more respiratory, behavioral, and infectious concerns in pediatric patients and more cardiovascular impressions and symptoms in adults. Diggs et al demonstrated similar pediatric results with mental health, pain, and respiratory complaints dominating top symptoms and impressions.³² Rosenthal et al also demonstrated prevalent behavioral health concerns.¹² We suspect this is multifactorial and in part due to the low availability of pediatric mental health resources and fewer treatment options in the setting of increased nationwide prevalence of pediatric mental health concerns.³⁶ 

This study recognized a greater proportion of pediatric IFTs with age-adjusted abnormal vital signs during transport as compared with adult IFTs. Despite acuity and disposition being proportionally similar between the 2 groups, this may suggest pediatrics IFTs are more ill, or recognition was poor. Drayna et al also found a high percentage of abnormal vital signs in their study of metropolitan pediatric EMS encounters but suggested possible underrecognition among EMS crews.³⁷ Whether vital signs differences in our study are simply underrecognized or have implications for outcomes remains unknown.

There are several limitations to this study. As a retrospective cross-sectional study, it cannot establish causal inference, observe trends, or calculate incidence. It retains susceptibility to biases including recall, misclassification, and selection. Specifically, the NEMSIS data are still considered a convenience sample and are, at times, limited by missing or inaccurate responses. Furthermore, since many fields in the NEMSIS database require selection from a multiple-choice list, EMS crews may oversimplify or incorrectly assign a “default” or “filler” value such as “Not available” when faced by real-life events or other time pressures. We estimate a moderate degree of misclassification bias and a smaller degree of selection and recall bias. This is a widely recognized challenge.^30–32 Future NEMSIS datasets have improved this issue given greater national uptake among state EMS systems and better data definitions.³⁸ We anticipate these weaknesses are at least partially mitigated by the large national sample size. When compared with other NEMSIS studies, our dataset appears to be representative even after removing incomplete records. Additionally, the NEMSIS database poorly captures outcomes because it merely represents the time EMS cared for the patient until handoff to the next care team. For interfacility transports, EMS teams are unlikely to remain at the destination facility long enough to determine the final disposition or outcome.

Despite the aforementioned limitations, generalizability is a strength of this study. Our data source is a large nationally representative database, with contributions from 47 states and 10 062 EMS agencies. Our variables of interest are key factors in transports, and our intention is that these results can aid quality improvement initiatives, training, and future research endeavors.

Pediatric IFTs exhibit distinct demographic, transport, and clinical characteristics compared with adult IFTs. Notably, pediatric patients more frequently present with age-specific abnormal vital signs. Transport characteristics indicate increased use of rotor- and fixed-wing operations, along with significantly longer transport times for pediatric patients. Although the implications of these differences remain unclear, our study provides valuable insights for transport crews, providers, and medical control teams, informing training, simulation, clinical guideline development, and future research directions. The next step involves linking this NEMSIS data with relevant clinical outcomes to better understand the impact of these transport differences in addition to a more detailed temporal analysis to elucidate the significantly longer transports.

ED

emergency department

EMS

Emergency Medical Services

IFT

interfacility transport

NEMSIS

National EMS Information System