An algorithm to identify patients aged 0-3 with rare genetic disorders

Area of Science:

• Pediatric rare diseases
• Clinical informatics
• Genomic medicine

Background:

• Identifying rare genetic disorders in children is difficult due to incomplete electronic health records and coding inaccuracies.
• Over 7000 Mendelian disorders exist, presenting diagnostic challenges in early childhood.
• There is a need for improved methods to detect genetic conditions in young children.

Purpose of the Study:

• To develop and validate a digital phenotyping algorithm (PheIndex) for identifying children aged 0-3 at risk for genetic disorders.
• To leverage electronic medical record data for early detection of rare genetic conditions.
• To improve the diagnostic pathway for pediatric patients with potential genetic disorders.

Main Methods:

• Developed the PheIndex algorithm based on 13 criteria derived from expert opinion.
• Utilized electronic medical records to identify children aged 0-3 with potential genetic disorder diagnoses or risks.
• Validated algorithm performance through comprehensive chart review.

Main Results:

• The PheIndex algorithm identified 1,088 children at increased risk for genetic disorders out of 93,154 live births.
• Chart review confirmed the algorithm's high performance: 90% sensitivity, 97% specificity, and 94% accuracy.
• The algorithm successfully flagged children requiring further genetic evaluation.

Conclusions:

• The PheIndex algorithm effectively identifies children who may have a rare genetic disorder.
• This tool can prompt healthcare providers to consider diagnostic genetic testing or referral to a medical geneticist.
• PheIndex enhances the early detection and management of genetic conditions in pediatric populations.