An Iterative Process for Identifying Pediatric Patients With Type 1 Diabetes: Retrospective Observational Study

Area of Science:

• Pediatric Endocrinology
• Computational Health Informatics
• Diabetes Research

Background:

• Rising incidence of type 1 diabetes (T1DM) and type 2 diabetes (T2DM) in pediatric populations.
• Current methods for diagnosing pediatric diabetes are labor-intensive and costly.

Purpose of the Study:

• To develop a computable phenotype for precise and efficient identification of diabetes in children.
• To differentiate between T1DM and T2DM in pediatric patients.

Main Methods:

• Retrospective study using electronic health records from the University of Florida Health Integrated Data Repository.
• Development and validation of a computable phenotype to identify pediatric diabetes cases.
• Manual review of medical records by endocrinology specialists to confirm diagnoses.

Main Results:

• The computable phenotype demonstrated high accuracy with a positive predictive value of 94.7% and sensitivity of 96.9%.
• Specificity was 95.8% and negative predictive value was 97.6%.
• Successfully identified 128 cases of T1DM and 35 cases of T2DM among 295 reviewed records.

Conclusions:

• A validated computable phenotype enables accurate and efficient identification of T1DM in pediatric patients.
• This tool offers a cost-effective solution for researchers to identify patient cohorts for intervention studies.
• Facilitates improved ease and accuracy in selecting participants for future diabetes research.