Health

    Prevention & Personalised Health

Mécénat des Mutuelles

France

Evidence-Based Growth Monitoring (EBGM)

Growth monitoring aims to identify severe pediatric health conditions early and improve their prognosis. Substantial empirical evidence shows that current growth monitoring practices are suboptimal, with very long diagnostic delays (on average 2.3 to 5.2 years) for some severe conditions, and a large number of futile referrals of children (on average 95%) with non-pathological variants of growth. This may be due to a lack of validated, personalized, and effective tools for identifying abnormal growth in children.

The EBGM project, led by Pauline Scherdel, aims to externally evaluate and refine an artificial intelligence (AI) algorithm for detecting abnormal growth in children. Initially, we developed and internally evaluated an algorithm for children aged between 1 and 12 years for detecting two target and priority conditions: Turner syndrome and growth hormone deficiency. The AI algorithm shows good diagnostic performance, achieving a sensitivity of over 90%, a specificity of around 95%, and a median theoretical reduction in diagnostic time of 2 to 4 years.  The aim is to externally evaluate this AI algorithm using data of diseased children from the AP-HP clinical data warehouse and apparently healthy children included in the ELFE national mother-child cohort.

Implementing a personalized and validated AI algorithm into primary care physicians' software will optimize growth monitoring by improving the prognosis of severe health conditions and reducing futile referrals for specialist consultations.  

Pauline
SCHERDEL

Institution

INSERM

Country

France

Nationality

French

ORCID Open Researcher and Contributor ID, a unique and persistent identifier to researchers