Non-suicidal self-injury at a Canadian paediatric emergency department.

OBJECTIVE: Our primary objective was to determine agreement between non-suicidal self-injury recorded at triage and during subsequent mental health assessment. The secondary objective was to describe patients who reported non-suicidal self-injury. METHODS: This is a health records review of patients aged 12-18 years who had an Emergency Mental Health Triage form on their health record from an ED visit June 1, 2017-May 31, 2018. We excluded patients with diagnoses of autism spectrum disorder or schizophrenia. We abstracted data from the Mental Health Triage form, Emergency Mental Health and Addictions Service Assessment forms and Assessment of Suicide and Risk Inventory. We calculated Cohen's Kappa coefficient, sensitivity, and negative predictive value to describe the extent to which the forms agreed and the performance of triage for identifying non-suicidal self-injury. We compared the cohort who reported non-suicidal self-injury with those who did not, using t-tests, Wilcoxon rank-sum tests, and chi-square tests. RESULTS: We screened 955 ED visits and included 914 ED visits where 558 (58.4%) reported a history of non-suicidal self-injury. There were significantly more females in the group reporting non-suicidal self-injury (82.1%, n = 458) compared to the group not reporting non-suicidal self-injury (45.8%, n = 163). Patients reporting non-suicidal self-injury did so in triage and detailed Mental Health Assessment 64.7% of the time (Cohen's Kappa Coefficient 0.6); triage had sensitivity of 71.5% (95% CI 67.3-75.4) and negative predictive value of 71.2% (95% CI 68.2-74.0). Cutting was the most common method of non-suicidal self-injury (80.3%). CONCLUSION: Screening at triage was moderately effective in identifying non-suicidal self-injury compared to a detailed assessment by a specialised mental health team. More than half of children and adolescents with a mental health-related concern in our ED reported a history of non-suicidal self-injury, most of which were female. This symptom is important for delineating patients' coping strategies.

RéSUMé: OBJECTIFS: Notre objectif principal était de déterminer l'accord entre les blessures non suicidaires enregistrées au triage et lors de l'évaluation subséquente de la santé mentale. L'objectif secondaire était de décrire les patients qui ont déclaré une automutilation non suicidaire. MéTHODES: Il s'agit d'un examen des dossiers de santé de patients âgés de 12 à 18 ans qui avaient un formulaire de triage d'urgence en santé mentale dans leur dossier de santé à la suite d'une visite à l'urgence du 1er juin 2017 au 31 mai 2018. Nous avons exclu les patients présentant un diagnostic de trouble du spectre autistique ou de schizophrénie. Nous avons extrait des données du formulaire de triage en santé mentale, des formulaires d'évaluation des services d'urgence en santé mentale et en toxicomanie et de l'évaluation du suicide et de l'inventaire des risques. Nous avons calculé le coefficient de Kappa de Cohen, la sensibilité et la valeur prédictive négative pour décrire la mesure dans laquelle les formes étaient d'accord et la performance du triage pour identifier l'automutilation non suicidaire. Nous avons comparé la cohorte qui a déclaré une automutilation non suicidaire avec celles qui ne l'ont pas fait, en utilisant des tests t-tests, des tests Wilcoxon rank-sum et des tests chi-carrés. RéSULTATS: Nous avons examiné 955 visites à l'urgence et inclus 914 visites à l'urgence où 558 (58,4 %) ont signalé des antécédents d'automutilation non suicidaire. Il y avait beaucoup plus de femmes dans le groupe déclarant une automutilation non suicidaire (82,1 %, n = 458) que dans le groupe ne déclarant pas une automutilation non suicidaire (45,8 %, n = 163). Les patients ayant déclaré une automutilation non suicidaire l'ont fait dans le cadre du triage et de l'évaluation détaillée de la santé mentale 64,7 % du temps (coefficient de Kappa de Cohen 0,6); le triage avait une sensibilité de 71,5 % (IC à 95 % 67,3­75,4) et une valeur prédictive négative de 71,2 % (IC à 95 % 68,2­74,0). La coupe était la méthode la plus courante d'automutilation non suicidaire (80,3 %). CONCLUSION: Le dépistage au triage a été modérément efficace pour identifier les blessures non suicidaires comparativement à une évaluation détaillée par une équipe spécialisée en santé mentale. Plus de la moitié des enfants et des adolescents ayant un problème de santé mentale à notre DE ont signalé des antécédents d'automutilation non suicidaire, dont la plupart étaient des femmes. Ce symptôme est important pour délimiter les stratégies d'adaptation des patients.

Machine learning to identify attributes that predict patients who leave without being seen in a pediatric emergency department.

PURPOSE: To characterize patients who left without being seen (LWBS) from a Canadian pediatric Emergency Department (ED) and create predictive models using machine learning to identify key attributes associated with LWBS. METHODS: We analyzed administrative ED data from April 1, 2017, to March 31, 2020, from IWK Health ED in Halifax, NS. Variables included: visit disposition; Canadian Triage Acuity Scale (CTAS); triage month, week, day, hour, minute, and day of the week; sex; age; postal code; access to primary care provider; visit payor; referral source; arrival by ambulance; main problem (ICD10); length of stay in minutes; driving distance in minutes; and ED patient load. The data were randomly divided into training (80%) and test datasets (20%). Five supervised machine learning binary classification algorithms were implemented to train models to predict LWBS patients. We balanced the dataset using Synthetic Minority Oversampling Technique (SMOTE) and used grid search for hyperparameter tuning of our models. Model evaluation was made using sensitivity and recall on the test dataset. RESULTS: The dataset included 101,266 ED visits where 2009 (2%) records were excluded and 5800 LWBS (5.7%). The highest-performing machine learning model with 16 patient attributes was XGBoost which was able to identify LWBS patients with 95% recall and 87% sensitivity. The most influential attributes in this model were ED patient load, triage hour, driving minutes from home address to ED, length of stay (minutes since triage), and age. CONCLUSION: Our analysis showed that machine learning models can be used on administrative data to predict patients who LWBS in a Canadian pediatric ED. From 16 variables, we identified the five most influential model attributes. System-level interventions to improve patient flow have shown promise for reducing LWBS in some centres. Predicting patients likely to LWBS raises the possibility of individual patient-level interventions to mitigate LWBS.

RéSUMé: BUT: Caractériser les patients qui sont partis sans être vus (left without being seen LWBS) d'un service d'urgence (SU) pédiatrique canadien et créer des modèles prédictifs utilisant l'apprentissage automatique pour identifier les attributs clés associés au LWBS. MéTHODES: Nous avons analysé les données administratives de SU du 1er avril 2017 au 31 mars 2020 provenant de l'urgence de IWK Health à Halifax, en Nouvelle-Écosse. Les variables comprenaient: disposition de la visite; l'échelle canadienne de triage de la gravité (ETG); mois, semaine, jour, heure, minute et jour de la semaine; sexe; âge; code postal; accès au fournisseur de soins primaires; payeur de la visite; source de l'aiguillage; arrivée par ambulance; principal problème (CIM10); durée du séjour en minutes; distance de conduite en minutes; et la charge de patients de l'urgence. Les données ont été divisées de manière aléatoire en ensembles de données de formation (80%) et de test (20%). Cinq algorithmes de classification binaire d'apprentissage automatique supervisés ont été mis en œuvre pour former des modèles de prévision des patients atteints de LWBS. Nous avons équilibré l'ensemble de données à l'aide de la technique de suréchantillonnage synthétique des minorités (SMOTE) et utilisé la recherche de grille pour le réglage des hyperparamètres de nos modèles. L'évaluation du modèle a été faite en utilisant la sensibilité et le rappel sur l'ensemble de données d'essai. RéSULTATS: L'ensemble de données comprenait 101266 visites aux urgences où les enregistrements de 2009 (2%) ont été exclus et 5800 LWBS (5,7%). Le modèle d'apprentissage automatique le plus performant avec 16 attributs de patient était XGBoost, qui a été en mesure d'identifier les patients LWBS avec 95% de rappel et 87% de sensibilité. Les attributs les plus influents dans ce modèle étaient la charge de patients à l'urgence, l'heure de triage, les minutes de conduite entre l'adresse du domicile et l'urgence, la durée du séjour (minutes depuis le triage) et l'âge. CONCLUSION: Notre analyse a montré que les modèles d'apprentissage automatique peuvent être utilisés sur des données administratives pour prédire les patients qui sont partis sans être vus dans un service d'urgence pédiatrique canadien. À partir de 16 variables, nous avons identifié les cinq attributs de modèle les plus influents. Les interventions au niveau du système visant à améliorer le flux de patients se sont révélées prometteuses pour réduire les LWBS dans certains centres. La prévision des patients susceptibles de LWBS soulève la possibilité d'interventions individuelles au niveau des patients pour atténuer le LWBS.

Predictors of Adherence to Short-Course Probiotics Among Children with Gastroenteritis who are Enrolled in a Clinical Trial.

BACKGROUND: To improve our understanding of adherence to discharge medications in the ED and within research trials, we sought to quantify medication adherence and identify predictors thereof in children with acute gastroenteritis (AGE). METHODS: We conducted a secondary analysis of a randomized trial of twice daily probiotic for 5 days. The population included previously healthy children aged 3-47 months with AGE. The primary outcome was patient-reported adherence to the treatment regimen, defined a priori as having received >70% of the prescribed doses. Secondary outcomes included predictors of treatment adherence and concordance between patient-reported adherence and the returned medication sachet counts. RESULTS: After excluding participants with missing data on adherence, 760 participants were included in this analysis: 383 in the probiotic arm (50.4%); and 377 in the placebo arm (49.6%). Self-reported adherence was similar in both groups (77.0% in probiotic versus 80.3% in placebo). There was good agreement between self-reported adherence and sachet counts (87% within limits of agreement (-2.9 to 3.5 sachets) on the Bland-Altman plots). In the multivariable regression model, covariates associated with adherence were greater number of days of diarrhea post-emergency department visit, and the study site; covariates negatively associated with adherence were age 12-23 months, severe dehydration and greater total number of vomiting and diarrhea episodes after enrolment. CONCLUSIONS: Longer duration of diarrhea and study site were associated with higher probiotic adherence. Age 12-23 months, severe dehydration and greater number of vomiting and diarrhea episodes post enrolment negatively predicted treatment adherence.

Medication-related emergency department visits in pediatrics: a prospective observational study.

BACKGROUND AND OBJECTIVE: There are few data on the rate and characterization of medication-related visits (MRVs) to the emergency department (ED) in pediatric patients. We sought to evaluate the frequency, severity, preventability, and classification of MRVs to the ED in pediatric patients. METHODS: We performed a prospective observational study of pediatric patients presenting to the ED over a 12-month period. A medication-related ED visit was identified by using pharmacist assessment, emergency physician assessment, and an independent adjudication committee. RESULTS: In this study, 2028 patients were enrolled (mean age, 6.1 ± 5.0 years; girls, 47.4%). An MRV was found in 163 patients (8.0%; 95% confidence interval [CI]: 7.0%-9.3%) of which 106 (65.0%; 95% CI: 57.2%-72.3%) were deemed preventable. Severity was classified as mild in 14 cases (8.6%; 95% CI: 4.8%-14.0%), moderate in 140 cases (85.9%; 95% CI: 79.6%-90.8%), and severe in 9 cases (5.5%; 95% CI: 2.6%-10.2%). The most common events were related to adverse drug reactions 26.4% (95% CI: 19.8%-33.8%), subtherapeutic dosage 19.0% (95% CI: 13.3%-25.9%), and nonadherence 17.2% (95% CI: 11.7%-23.9%). The probability of hospital admission was significantly higher among patients with an MRV compared with those without an MRV (odds ratio, 6.5; 95% CI: 4.3-9.6) and, if admitted, the median (interquartile range) length of stay was longer (3.0 [5.0] days vs 1.5 [2.5] days, P = .02). CONCLUSIONS: A medication-related cause was found in ∼1 of every 12 ED visits by pediatric patients, of which two-thirds were deemed preventable. Pediatric patients who present to the ED with an MRV are more likely to be admitted to hospital and when admitted have a longer length of stay.

Medication-related emergency department visits and hospital admissions in pediatric patients: a qualitative systematic review.

OBJECTIVE: To review and describe the current literature pertaining to the incidence, classification, severity, preventability, and impact of medication-related emergency department (ED) and hospital admissions in pediatric patients. STUDY DESIGN: A systematic search of PubMED, Embase, and Web of Science was performed using the following terms: drug toxicity, adverse drug event, medication error, emergency department, ambulatory care, and outpatient clinic. Additional articles were identified by a manual search of cited references. English language, full-reports of pediatric (≤18 years) patients that required an ED visit or hospital admission secondary to an adverse drug event (ADE) were included. RESULTS: We included 11 studies that reported medication-related ED visit or hospital admission in pediatric patients. Incidence of medication-related ED visits and hospital admissions ranged from 0.5%-3.3% and 0.16%-4.3%, respectively, of which 20.3%-66.7% were deemed preventable. Among ED visits, 5.1%-22.1% of patients were admitted to hospital, with a length of stay of 24-72 hours. The majority of ADEs were deemed moderate in severity. Types of ADEs included adverse drug reactions, allergic reactions, overdose, medication use with no indication, wrong drug prescribed, and patient not receiving a drug for an indication. Common causative agents included respiratory drugs, antimicrobials, central nervous system drugs, analgesics, hormones, cardiovascular drugs, and vaccines. CONCLUSION: Medication-related ED visits and hospital admissions are common in pediatric patients, many of which are preventable. These ADEs result in significant healthcare utilization.

Concussion or mild traumatic brain injury: parents appreciate the nuances of nosology.

We explored whether parents of our pediatric patients valued the diagnostic terms "concussion," "minor traumatic brain injury," and "mild traumatic brain injury" as equivalent or nonequivalent. 1734 of 2304 parents attending a regional pediatric emergency department completed a brief questionnaire assessing the equivalence or nonequivalence of the diagnostic terms "concussion," "minor traumatic brain injury," and "mild traumatic brain injury" in a pairwise fashion. Many parents viewed these diagnostic terms as equivalent, when assessed side by side. For those who considered these diagnostic terms nonequivalent, concussion was regarded as considerably "better" (or less "worse") than minor traumatic brain injury (P < 0.001, χ(2) test) or mild traumatic brain injury (P < 0.001, χ(2) test). A moderate degree of variability was evident in parent/guardian responses. As a group, parents reported that concussion or mild/minor traumatic brain injuries are valued equivalently. However, many parents considered them different, with concussion reflecting a "better" (or less "worse") outcome.

"My child can't keep anything down!" Interviewing parents who bring their preschoolers to the emergency department for diarrhea, vomiting, and dehydration.

INTRODUCTION: Viral gastroenteritis with dehydration is one of the most frequent reasons for visits to pediatric emergency departments (ED). Parental intervention before presentation to the ED can make a significant difference in the course of a child's illness. There is a discrepancy between medical knowledge of dehydration and parental fears and understanding. This project is part of a larger program of research developing an educational tool for parents of preschoolers with diarrhea, vomiting, and dehydration. The primary objective was to develop an interview guide. From initial data, the researchers explored parental motivations for bringing their children to the ED. METHODS: Ten families were recruited after their visit to a pediatric ED in the fall of 2007. Included were families of children younger than 4 years who experienced vomiting, diarrhea, and dehydration. Interviews were conducted over the telephone and were transcribed. The interview guide was edited in an iterative process. RESULTS: Thematic analysis focused on parents' decision to take their child to the ED. Making the decision to take a child to the ED is a complex process for parents. This decision involves expectations developed from community-level, family-level, and child factors. Issues of access to care affect parents' decision, including perceived level of urgency, travel time, and modes of transport available. CONCLUSIONS: A framework is proposed, which outlines the most important factors our sample of parents reported when deciding whether to take their ill child to the ED. The interview guide developed will facilitate collection of further information.