Agreement between caregivers’ assessment of their children’s urgency of illness and the Emergency Severity Index in an urban pediatric emergency department

Article information

Pediatr Emerg Med J. 2026;13(3):85-92

Publication date (electronic) : 2026 May 7

doi : https://doi.org/10.22470/pemj.2025.01515

Motoki Yasuda^,1, Jesus Jaile¹, Jinal Shah², David Mortel¹

¹Department of Pediatrics Harlem Hospital Center, Columbia University, New York, NY, USA

²Department of Internal Medicine, Harlem Hospital Center, Columbia University, New York, NY, USA

Corresponding author: Motoki Yasuda Department of Pediatrics, Harlem Hospital Center, Columbia University, 506 Lenox Ave, New York, NY 10037, USA Tel: +1-212-939-2777; E-mail: motokiyasuda@gmail.com

Received 2025 December 7; Revised 2026 April 3; Accepted 2026 April 7.

Abstract

Introduction

Pediatric emergency department (PED) visits account for approximately 20% of all emergency department (ED) visits in the United States (1), with over 30,000,000 visits annually among patients younger than 18 years (2). Given limited ED resources and increasing patient volumes, accurate triage is essential to prioritize care and allocate resources appropriately. The Emergency Severity Index (ESI) is a widely used 5-level triage system in U.S. PEDs that prioritizes patients based on acuity and anticipated resource needs (3), and has demonstrated reliability and validity in pediatric populations (4-6). ESI has been shown to have good interrater reliability and validity between physicians and nurses as shown in several pediatric studies (5), and its version 4 has been found a valid predictor for hospitalization for pediatric patients (6). The ESI triage algorithm is detailed in Appendix 1 (https://doi.org/10.22470/pemj.2025.01515).

Nonurgent ED visits are a controversial issue. Despite the lack of consensus on definition of a nonurgent ED visit, Afilalo et al. (7) defined it as “patients who could have waited 2 hours before being seen by a physician.” A literature review reported wide variability in the proportion of nonurgent ED visits (5%–90%; median, 32%) (8). This wide variation leads to crowding and delays in treatment of severely ill patients (8,9). Moreover, there is a gap between patients’ and healthcare team’s perception of the level of urgency. There is lack of correlation between patients’ and physicians’ perception in adult population (10-18). However, to our knowledge, there have been few pediatric studies. Given the central role of caregivers in decision-making for their children, the urgency of a children’s condition should be considered from the caregiver’s perspective.

The primary objective of this study was to evaluate the agreement between caregivers’ assessments of their children’s urgency of illness and ESI score assigned at triage in an urban PED. A secondary objective was to identify characteristics of caregivers or their visits associated with higher or lower agreement.

Methods

Results

1. Study population

A total of 338 caregivers completed the surveys. The mean patient age was 6.4 ± 5.5 years (Table 1). Most patients were black (75.1%), from English-speaking households (78.7%), and described by their caregivers as being in excellent health (69.5%). Nearly all patients had their PCPs (94.7%), although only 10.6% contacted the PCPs before visiting the PED. More than half of the patients had neither prior ED visits in the last 6 months (68.6%) nor trauma.

Table 1.

Baseline characteristics of the study population (N = 338)

2. Agreement between caregiver’s assessments and ESI

Among the patients, the most common ESI level assigned by triage nurses was level 4 (46.1%), followed by levels 3, 5, and 2 (Table 2). Caregivers most frequently rated the urgency as less than 1 hour (37.3%), followed by less than 15 minutes, 1–2 hours, and more than 2 hours. These assessments agreed exactly with the ESI level in 121 patients (35.8%), while over- and under-triage were observed in 54.4% and 9.8%, respectively.

Table 2.

Agreement between caregivers’ assessment of urgency level and Emergency Severity Index level

Table 3 shows the agreement between caregiver’s assessment and ESI score. Overall, there is fair agreement (ICC = 0.53 [95% confidence interval, 0.14–0.72]), in particular among caregivers of children with ≥4 prior ED visits, non-native English speakers, and visits in the daytime.

Table 3.

Interrater agreement between caregivers’ assessment and Emergency Severity Index level

3. Predictors of caregiver perceptions

Table 4 presents the findings of the regression model examining factors associated with caregivers’ assessments of urgency level. In an adjusted model, visits in the evenings (adjusted odds ratio, 0.57 [95% confidence interval, 0.37–0.88]) or nights (0.32 [0.15–0.64]) and trauma-related visits (0.52 [0.30–0.90]) were associated with a higher caregivers’ urgency level.

Table 4.

Ordinal logistic regression with outcome of caregivers’ assessments

Discussion

This study shows an overall fair agreement between caregivers’ assessment and ESI classification. Although this level of agreement indicates that caregivers’ perceptions do not fully align with formal triage outcomes, it may be higher than that reported in adult EDs where previous studies have demonstrated poor agreement between patient-perceived urgency and clinician-assigned triage levels (11-13). In this comparison, the level of agreement observed in our pediatric population suggests that the caregivers may have a somewhat better ability to recognize the urgency of their children’s conditions than that observed in adult patients.

Caregivers who are non-native English speakers have better agreement than native speakers, which is different from Ruud et al.’s study (11). This could be due to that the non-native speakers tend to wait until they truly feel their children needs emergency care for fear of language barrier. Unlike Ghasem-Sam et al.’s study (14), caregivers of patients with more frequent ED visits also have better agreement, which might be because they have learned from prior ED experiences where emergency care providers would provide feedback and return precautions. Caregivers who came during daytime also have better agreement, which might be because more knowledgeable caregivers could wait until clinic office time if they think their children’s conditions are not too urgent.

Notably, 90% of patients were seen by the emergency physicians less than 10 minutes after the triage. This is mainly because of low patient volume during COVID-19 pandemic. The number of patients during our study period decreased by 55% compared to the same season in a pre-pandemic period. Due to the consequently short waiting time, it is less likely for the caregivers to be biased to overstate urgency in the hope of being seen earlier.

Although no hospitalizations occurred among the patients categorized by their caregivers as nonurgent (Appendix 4), this finding should be interpreted cautiously given the limited sample size. Importantly, the agreement in the patients hospitalized in the intensive care unit suggests that caregivers generally recognized severe illnesses (see “4. PED disposition” in the results).

However, we still observed discrepancies between caregivers’ perception and formal triage classification. The former assessment is inherently subjective and may be influenced by parental anxiety, previous illness-related experiences, or individual perceptions of disease severity. Therefore, caregivers’ perceptions should not replace structured triage systems, although they may provide complementary contextual information when evaluating pediatric patients. Our findings suggest the importance of educating caregivers, particularly those who come to PEDs after clinic hours, and give both reassurance and appropriate return precautions to the worried well.

The ordinal logistic regression showed that visits in the evening or night were associated with higher perceived urgency compared with visits in the daytime, which is similar to Ruud et al.’s study (11). This might stem from that regardless of the patients’ actual urgency, the caregivers believed their sick children were unable to wait until next morning to visit their PCPs’ offices. We believe that this represents an opportunity to educate caregivers who come to PEDs at night with low acuity conditions on recognition of warning signs of a true medical emergency. It is also worth considering that the caregivers who bring their children with seemingly nonurgent conditions in the night may be unable to leave work during the daytime to see their PCPs, and thus represent a more vulnerable population. Caregivers of injured patients may feel that their cases were more urgent, as traumatic injuries are often unpredictable and can be frightening for the caregivers, prompting more immediate PED visits. In contrast, another variable, such as age, sex, or prior PED visits, was not associated with the caregivers’ urgency perception, suggesting that such perception was influenced more by context of the visits than by patient characteristics per se.

The observed differences in hospitalization rates across ESI levels and caregiver urgency ratings suggest that both ESI and caregiver assessment are associated with clinical disposition. Notably, higher caregiver-perceived urgency was associated with an increased likelihood of hospitalization, whereas patients perceived as nonurgent were not hospitalized. This pattern suggests that caregivers may have some ability to recognize severe illness, although discrepancies with formal triage remain.

Since March 2020, PED volumes have declined due to COVID-19 pandemic. Instead, telehealth provided an option for patients to receive information and medical treatment safely and conveniently (20). During telehealth visits, clinicians must determine the need for further in-person assessment in PEDs or offices based largely on medical history obtainable over the phone or video connection (21-24). Since our study showed caregivers’ assessments are more reliable than previous adult studies, parental input may be helpful in reaching a decision on whether children should go to PEDs.

This study has several limitations. First, it was conducted in the single urban PED using a convenience patient sample, by the same physicians who cared for them. This study design might have introduced potential bias, limiting the generalizability of the findings. Second, agreement in the most critically ill population could not be adequately evaluated because patients with ESI level 1 were excluded and the number of patients hospitalized to the intensive care unit was small. Third, because the reason of visit was classified only as trauma or not, we were unable to collect more detailed chief complaint data. Finally, this study was conducted during COVID-19 pandemic when PED visits markedly declined. There may be a difference between caregivers’ decision-making or patterns of PED visits during the pandemic and during non-pandemic periods.

In conclusion, the caregivers’ assessments of urgency of the children showed fair agreement with ESI triage classification in this PED. While the caregivers generally recognized severe illness, we observed discrepancies between the caregivers’ perception and formal triage. These findings suggest that caregivers’ input may provide additional contextual information but should not replace formal structured triage systems.

Article information Continued