Development and Validation of Diabetes Interpreter, a Mobile Application-Based Tool for Point-of-Care Evaluation of Children with Diabetes

Objectives

To develop and validate Diabetes Interpreter, a mobile application-based tool for point-of-care guidance for classifying pediatric and adolescent diabetes.

Methods

The Diabetes Interpreter recommends diagnosis and evaluation based on clinical parameters (age at diagnosis, disease duration, presentation, and insulin requirements). The validation involved comparing the guidance given by the Diabetes Interpreter, two pediatric endocrinologists, one adult endocrinologist, a pediatrician, and a pediatric trainee with the clinical diagnoses of 302 children and adolescents with diabetes treated at authors’ Endocrinology Clinic.

Results

The Diabetes Interpreter agreed highly with clinical diagnoses and guidance (596 out of 604, 98.6%). The concordance rates for Pediatric Endocrinologist I (546; 90.4%) and Pediatric Endocrinologist II (491; 81.3%) exceeded those of the adult endocrinologist (405; 67.1%), pediatrician (287; 47.5%), and pediatric trainee (258; 42.7%). The adult endocrinologist missed the diagnosis of Type 1 diabetes without a suggestion for autoimmune assessment in 15 subjects (6%), while the pediatrician missed it in 16 subjects (6.4%), and the pediatric trainee in 33 (13.2%). Following the guidance of the Diabetes Interpreter could have potentially reduced 31.5%, 51.1%, and 55.9% of the discordances observed in the adult endocrinologist, pediatrician, and pediatric trainee, respectively. The pediatric trainee (250, 100%), the pediatrician (245, 98%), and the adult endocrinologist (124, 49.6%) recommended workups for a more significant proportion of subjects who did not require investigations compared to the Diabetes Interpreter (3, 1.2%).

Conclusions

The high concordance score of the Diabetes Interpreter underscores its importance in point-of-care guidance for assessing children and adolescents with diabetes.

Comments (0)

No login
gif