logo
ResearchBunny Logo
A fair individualized polysocial risk score for identifying increased social risk in type 2 diabetes

Medicine and Health

A fair individualized polysocial risk score for identifying increased social risk in type 2 diabetes

Y. Huang, J. Guo, et al.

Discover how researchers from the University of Florida developed an innovative machine learning pipeline to create an individualized polysocial risk score for type 2 diabetes patients. This groundbreaking study addresses the challenges faced by racial and ethnic minorities, showcasing an effective tool for predicting hospitalization risks with a focus on social determinants of health.

00:00
00:00
Playback language: English
Abstract
Racial and ethnic minorities disproportionately experience type 2 diabetes (T2D) and its complications, influenced heavily by social determinants of health (SDOH). This study proposes a machine learning pipeline to create an individualized polysocial risk score (iPsRS) for identifying T2D patients at high social risk of hospitalization. Using electronic health records (EHR) data, the iPsRS incorporates both contextual and person-level SDOH, with fairness optimization across racial and ethnic groups. After optimization, the iPsRS achieved a C-statistic of 0.71 in predicting 1-year hospitalization, demonstrating its ability to fairly and accurately screen for increased social risk in T2D patients.
Publisher
Nature Communications
Published On
Oct 05, 2024
Authors
Yu Huang, Jingchuan Guo, William T. Donahoo, Yao An Lee, Zhengkang Fan, Ying Lu, Wei-Han Chen, Huilin Tang, Lori Bilello, Aaron A. Saguil, Eric Rosenberg, Elizabeth A. Shenkman, Jiang Bian
Tags
type 2 diabetes
social determinants of health
machine learning
polysocial risk score
hospitalization prediction
racial and ethnic minorities
fairness optimization
Listen, Learn & Level Up
Over 10,000 hours of research content in 25+ fields, available in 12+ languages.
No more digging through PDFs, just hit play and absorb the world's latest research in your language, on your time.
listen to research audio papers with researchbunny