Please use this identifier to cite or link to this item:
Title: Development of a Multivariable Risk Prediction Tool to Predict Adverse Outcomes among Children with Type 1 Diabetes: A Pilot Study
Authors: Lieu, Fiona;Martin, Wrivu N;Birt, Stewart ;Mattes, Joerg;McGee, Richard G 
Affliation: Central Coast Local Health District
Gosford Hospital
Issue Date: 20-May-2024
Source: 2024(1), 8335604
Journal title: Pediatric Diabetes
Department: Paediatrics
Abstract: Background. Children and adolescents with type 1 diabetes mellitus (T1DM) are frequently hospitalised for severe hypoglycaemia, hyperglycaemia, and diabetic ketoacidosis (DKA). While several risk factors have been recognised, clinically identifying these children at high risk of acute decompensation remains challenging. Objective. To develop a risk prediction model to accurately estimate the risk of acute healthcare utilisation due to severe hypoglycaemia, hyperglycaemia, and DKA in children and adolescents with T1DM. Materials and Methods. Using a retrospective dataset, baseline demographic and clinical data were collected from patients (<18 years) seen at a regional paediatric diabetes clinic from 1 January 2018 to 1 January 2020. The outcome was the number of emergency department presentations or hospital admissions for severe hypoglycaemia, hyperglycaemia, and DKA across the study period. Variables that were significant in univariate analysis were entered into a multivariable model. Receiver operator characteristic (ROC) curves assessed the model’s discrimination and generated cut-offs for risk group stratification (low, medium, and high). Kaplan–Meier survival analysis measured time to acute healthcare utilisation across the risk groups. Results. Our multivariable risk prediction model consisted of five predictors (continuous glucose monitoring device, previous acute healthcare utilisation, missed appointments, and child welfare services involvement and socioeconomic status). The model exhibited good discrimination (area under the ROC = 0.81), accurately stratified children into low-, medium-, and high-risk groups, and demonstrated significant differences between median time to healthcare utilisation. Conclusion. Our model identified patients at an increased risk of acute healthcare utilisation due to severe hypoglycaemia, hyperglycaemia, and DKA.
Publicaton type: Journal Article
Keywords: Paediatrics
Appears in Collections:Obstetrics / Paediatrics

Show full item record

Page view(s)

checked on Jul 21, 2024

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.