Characterizing pediatric dermatological presentations in an outer metropolitan emergency department: a single-center Western Australian study

Article information

Pediatr Emerg Med J. 2025;12(1):20-29

Publication date (electronic) : 2025 January 1

doi : https://doi.org/10.22470/pemj.2024.01151

Tristen Tze Wei Ng^,1,2,3,*, Darren Zhi-Yang Low^4,*, Amelia Ye Chiung Tang¹, Mabel Zhi Qi Foo⁵, Dale Wesley Edgar^2,6,7, Paul Anthony Hill¹

¹Department of Emergency Medicine, Armadale Health Service, Mount Nasura, WA, Australia

²The University of Western Australia Medical School, Crawley, WA, Australia

³Department of Dermatology, Royal Perth Hospital, Perth, WA, Australia

⁴Department of Emergency Medicine, Singapore General Hospital, Singapore

⁵Department of Infection Prevention and Epidemiology, Singapore General Hospital, Singapore

⁶Safety and Quality Unit, Armadale Health Service, Mount Nasura, WA, Australia

⁷The Institute for Health Research, The University of Notre Dame Australia, Fremantle, WA, Australia

Corresponding author Tristen Tze Wei Ng Department of Dermatology, Royal Perth Hospital, Victoria Square, Perth, WA 6000, Australia Tel: +61-8-9224 2244 Fax: +82-31-780-8409 E-mail: tristen.ng@uwa.edu.au

*These authors contributed equally to this work as co-first authors.

Received 2024 November 8; Revised 2024 December 11; Accepted 2024 December 12.

Abstract

Introduction

Pediatric dermatological conditions are primarily managed in general practice settings in Australia, Singapore, the United States, Italy, and Germany (1-3). However, many children present to the emergency department (ED) due to factors such as symptom onset outside regular clinic hours, heightened parental concern, or difficulty accessing a general practitioner (1). Globally, pediatric dermatological conditions comprise 6%-14% of patients presenting to the ED, highlighting their importance in emergency medicine (2-11). Most dermatological presentations do not require emergency treatment, are not life-threatening, and usually result in sameday discharge or referral back to general practitioners for ongoing management (5,12). However, to the authors’knowledge, no studies in Australia have specifically examined the proportion of pediatric skin presentations to an ED leading to hospitalization (13,14). This gap in research underscores the need for further investigation.

This Western Australian study was primarily aimed at characterizing the clinical features and proportion of pediatric dermatological presentations that necessitated hospitalization, thereby contributing to the growing body of knowledge in this field. The study was secondarily aimed at tracking the ED length of stay (EDLOS) and characterizing the proportion of outpatient referrals from the ED to a tertiary dermatology clinic. The time to access the outpatient dermatology clinic in a tertiary hospital was also analyzed.

Methods

We conducted a retrospective, cross-sectional study from December 1, 2022 through to November 30, 2023 using data from the ED Information System (EDIS), a digital medical records program of the Armadale Health Service (AHS). AHS is a secondary outer metropolitan hospital with the ED that provides care for more than 67,000 presentations annually, 25% of which involve pediatric patients (15). It is located approximately 35 km away from Perth’s Central Business District and over 40 km from the Perth Children’s Hospital, a tertiary pediatric hospital. AHS also serves a culturally and linguistically diverse population with socioeconomic disadvantage (16). This project was exempted under provisions of the Armadale-Kalamunda Group Safety and Quality Unit (IRB no. GEKO 51013). The East Metropolitan Health Service Research Ethics Committee waived ethics approval to publish the results from de-identified data collected retrospectively.

The inclusion criteria were any patient aged 0-16 years who presented to the ED with a primary dermatological complaint resulting in a dermatologic diagnosis during the study period. Multiple presentations by the same patient during the defined period were included as separate encounters. The exclusion criteria were thermal burns except sunburns and non-accidental injuries.

We analyzed age, sex, dermatologic diagnosis, arrival time in the ED, EDLOS, acuity based on the Australasian Triage Scale (ATS), and ED disposition. The age was categorized into infants (≤ 1 year), preschoolers (> 1 to ≤ 6 years), schoolers (> 6 to ≤ 12 years), and adolescents (> 12 to ≤ 16 years). The disposition status on the EDIS was streamed into 4 categories: “admitted,”“transferred to a tertiary hospital,”“discharged from ED,”and “discharged against medical advice or self-discharged (left before disposition was finalized).”We defined hospitalization as an inpatient admission to the Department of Pediatrics at AHS or discharge within 24 hours after the ED stay under plans for admission.

The ATS is a nationwide clinical tool for categorizing the maximal waiting time for medical review and management of patients presenting to EDs, and it uses 5 categories. Category 1 presentations are the most life-threatening conditions requiring immediate assessment and management. In contrast, Category 5 presentations are minor conditions (e.g., warts) or “clinico-administrative”issues (e.g., dressing changes and prescription issuances) (17). A triage nurse had assigned an ATS category before the patient was seen by medical staff. The category was rarely changed unless the patient deteriorated in the waiting room.

Patients were identified using 2 preset umbrella terms (“+ALLERGY AND IMMUNE”and “+SKIN”) and subsequently cross-referenced against the diagnostic codes (D17, D18, D22, D23, L00-L008, L10- L14*, L20-L30, L40-L45*, L50-54*, L60-75, and L80-L99*) of the International Statistical Classification of Diseases, 10th revision, on EDIS. Relevant dermatological conditions were manually selected and added to this list of patients.

Patients who met the inclusion criteria had their electronic and paper records interrogated to confirm the final dermatologic diagnoses. The diagnostic codes obtained from EDIS were grouped into broad categories of the diagnoses to stratify the data into meaningful clusters: “anaphylaxis and angioneurotic edema (AAE),”“allergy-related and urticarial dermatitis (AUD),”“eczema and other dermatitis (EOD),” “infective dermatoses,”and “not elsewhere classified.” The first category contained patients who met the criteria for anaphylaxis or anaphylactoid reactions, while the second contained those with allergic conditions other than such reactions. Eczema and other nonspecific dermatoses were combined into the same category of EOD. In contrast, other dermatologic diagnoses that did not fit into the first 4 categories were consolidated in “not elsewhere classified.”Further details are found in Appendix 1 (https://doi.org/10.22470/pemj.2024.01151).

All presentations within the study period were interrogated to identify outpatient dermatology clinic referrals from the ED presentation, either by the ED clinicians or Pediatrics team. The time from the ED referral to the outpatient dermatology appointment was also traced. Data were extracted from EDIS, imported into Microsoft Excel ver. 2406, and processed in REDCap^TM, a research electronic data capture platform manufactured by Vanderbilt University (18).

Descriptive statistics were used to summarize deidentified patient demographics and clinical characteristics. Categorical variables were reported as frequencies and percentages, while the median values with interquartile ranges were calculated for EDLOS. Associations of categorical variables (e.g., age group) with the primary dermatologic diagnosis were assessed using Pearson’s chi-square tests. Significant overall associations (P < 0.05) were further evaluated with post-hoc tests using Bonferroni corrections to identify pairwise differences for categorical variables with more than 2 groups, such as the age group. The Kruskal-Wallis tests were used to compare median values of EDLOS across diagnostic and ATS categories, and the Mann-Whitney U tests to examine differences in EDLOS between hospitalized and discharged patients.

Univariable logistic regression was conducted to assess potential factors associated with hospitalization from the ED. Variables with a P < 0.05 in the univariable analyses were considered significant. No a priori variables were selected for the analysis; all variables were evaluated through the univariable analysis. Multivariable logistic regression was not performed, as only 1 variable, dermatologic diagnoses, had a significant association with hospitalization in the univariable analysis. In all analyses, we used Stata ver. 18 (StataCorp).

Results

Of the 540 pediatric patients who presented to the ED with a primary dermatological complaint, 41 patients (7.6%) were hospitalized while 4 (0.7%) were transferred to tertiary hospitals (Table 1). Patients with AAE had a higher hospitalization rate (15.2%) compared to the other dermatologic diagnoses. The dermatologic diagnoses consisted of AUD (34.3%), infective dermatoses (29.3%), EOD (23.3%), AAE (8.5%), and not elsewhere classified (4.6%). Only 9 patients were referred for dermatology follow-up at Perth Children’s Hospital, with a median waiting time of 45 days (interquartile range, 23.5-85.0) (Appendix 1). No dermatologic emergencies, such as Stevens-Johnson syndrome, were identified during the study period (Appendix 2, https://doi.org/10.22470/pemj.2024.01151).

Table 1.

Summary of pediatric dermatological presentations and patient demographics at ED of Armadale Health Service

The age group consisted of 100 infants (18.5%), 222 preschoolers (41.1%), 144 schoolers (26.7%), and 74 adolescents (13.7%) (Table 1). There was a significant association between the age group and dermatologic diagnoses. A higher proportion of infants had EOD (33.3%) compared to other dermatological conditions such as AAE (8.7%), infective dermatoses (7.6%), AUD (21.1%). Among the 158 patients with infective dermatoses, 41.1% and 32.9% were preschoolers and schoolers, respectively. Pairwise comparisons showed differences in the diagnoses between the age groups, such as the infants vs. preschoolers and the schoolers vs. adolescents for EOD, as well as infective dermatoses (Appendix 3, https://doi.org/10.22470/pemj.2024.01151). In addition, ED disposition varied significantly by the diagnoses, with hospitalized patients differing from those discharged or transferred to tertiary hospitals (Appendix 4, https://doi.org/10.22470/pemj.2024.01151). Significant differences were observed for dispositions related to anaphylaxis, cutaneous drug eruptions, eczema, and other dermatitis, and diagnoses categorized as not elsewhere classified (P < 0.001) (Table 1). Median values of EDLOS varied significantly across the dermatologic diagnoses. The patients with AAE had the longest median EDLOS, while those under infective dermatoses had the shortest value (3.8 vs. 2.0 hours) (Table 1).

The proportion of hospitalized patients was significantly higher among those diagnosed with AAE, EOD, and infective dermatoses than those discharged (Table 2). The ATS indicated a trend towards significance, with a higher hospitalization rate in ATS 1-3 (P = 0.059). Additionally, the hospitalized patients had a longer median EDLOS than those discharged (5.0 vs. 2.1 hours).

Table 2.

Characteristics according to the hospitalization or discharge from the emergency department (N = 534^*)

Univariable logistic regression showed a significant association between dermatologic diagnosis and hospitalization (Table 3). The patients with AUD were less likely to be hospitalized (odds ratio, 0.06; 95% confidence interval, 0.12-0.30; compared with AAE). No association was found between the age group or ATS and hospitalization. The median EDLOS was longest in the ATS 1 cases and shortest in the ATS 5 cases (Table 4).

Table 3.

Univariable logistic regression analysis of factors associated with hospitalization from the emergency department

Table 4.

Distribution of pediatric emergency presentations rated by the Australasian Triage Scale at the emergency department of Armadale Health Service with the example diagnoses and EDLOS

Discussion

This study confirmed that most pediatric dermatological presentations to the ED (91.3%) were discharged regardless of the clinical urgency. This finding is consistent with similar studies on pediatric dermatology patterns and hospitalization rates in Switzerland, the U.S., Portugal, and Saudi Arabia (4,9,12,19). Interestingly, our study showed that AUD (34.3%) was the most typical ED presentation, unlike the infective dermatoses in France, Italy, Thailand, and India (5-8,20). Although more patients presented with AUD, those with infective dermatoses had a higher hospitalization rate (1.1% vs. 11.4%).

The association between the age group and dermatologic diagnoses likely reflects the environmental changes in the child’s development. For example, the highest proportion of pediatric patients who presented with EOD were preschoolers (Table 1), concordant with 85% of childhood eczema diagnosed before 5 years of age (21). In addition, the lower proportion of infective dermatoses among infants compared to their total representation (7.6% vs. 18.5%; Table 1) might reflect a reduced microbial exposure in this age group, due to limited mobility and close parental supervision (22). Conversely, a higher proportion of infective dermatoses among the adolescents (18.4% vs. their total representation of 13.7%; Table 1) could be attributed to the increased exposure through social activities and riskier behaviors, such as shared equipment or limited attention to personal hygiene (23).

The skin microbiome has also been implicated in the pathogenesis of eczema, with Staphylococcus aureus identified as a potential trigger. Skin flora changes are seen at baseline and during flares in pediatric eczema, with diminished bacterial diversity and overgrowth of S. aureus upregulating the inflammatory response (21,24,25).

Most dermatological (99.8%) presentations in our study were managed without direct dermatology input. This may be due to patient and hospital factors. Life-threatening dermatological emergencies tend to present directly to a tertiary pediatric center rather than a secondary outer-metropolitan hospital. Given our hospital’s absence of on-site dermatology service, we relied on on-call dermatology support from tertiary centers, available only from 08:30-16:30 on weekdays. In addition, most cases were classified as an ATS category 3-4 and could be effectively managed by emergency physicians or pediatricians without dermatologists’input. For conditions assigned to ATS 1, such as anaphylaxis, emergency physicians or pediatricians may be more appropriate clinicians in the acute setting than dermatologists. The limited dermatology access and the predominantly non-urgent nature of case presentations likely explain this study’s low dermatology referral rates.

Although the patients were hospitalized for further investigation and management, no patients were confirmed to have any dermatologic emergencies listed in Appendix 2. The absence of such emergencies was possible because the cases tend to present in tertiary pediatric centers. Many non-urgent and occasional cases of benign presentations could be managed in primary care settings. Although infective dermatoses appear to warrant an extended ED stay for either topical or intravenous therapy, the patients having such conditions are likely hemodynamically stable and could be managed in the ambulatory or community care settings. Thus, improving access to community care through outpatient dermatology clinics or after-hours primary care clinics may help reduce the strain on the ED with non-urgent presentations.

Teledermatology may be a solution to manage non-urgent dermatological presentations to EDs. General practitioners, particularly in rural areas, currently use it to contact dermatologists with non-urgent inquiries (26,27). This could be expanded to metropolitan general practitioners to reduce ED presentations. For emergency physicians, teledermatology may be used in diagnostically complex or life-threatening dermatologic emergencies where visualization of the patient’s skin lesions is critical.

There were limitations to our study. Firstly, the retrospective design and limited diagnostic options available in EDIS might have introduced misclassification bias and inaccurate diagnosis in the ED. The final dermatological diagnosis ascribed may also be incorrect, as non-dermatologists made diagnoses. Secondly, we might have missed some pediatric emergencies with no final dermatologic diagnoses under the umbrella terms “+ALLERGY AND IMMUNE”and “+SKIN.”The independence of each patient also could not be guaranteed given that multiple presentations by the same patient were included as separate encounters. Lastly, we could not access primary health records to verify whether general practitioners independently made dermatology referrals after an ED visit, potentially underestimating the number of referrals. The absence of a standardized protocol for initiating dermatology referrals, whether by emergency physicians during the visit or by general practitioners following discharge, might explain the low referral rate.

Our study provides contemporary insight into the types of pediatric dermatological presentations in an outer metropolitan Australian ED. Most presentations were benign or low acuity, underscoring the need for improved triage and management to efficiently allocate the resources for dermatology practice. Emergency physicians or pediatricians should remain vigilant for rare but critical dermatologic emergencies, and ensure prompt stabilization and specialist consultation if necessary. The low rate of dermatology referrals from the ED suggests an underuse of outpatient services. Rapid-access dermatology clinics, increased availability of after-hours care, and teledermatology may help reduce the number of non-emergent dermatology referrals, alleviate pressure on the ED, and ensure timely specialist-led follow-up.

Article information Continued