Polygenic scores and social determinants of health: Their correlations and potential biases

• Published in: HGG advances 2024-11, Vol.6 (1), p.100389, Article 100389
• Main Authors: Schaid, Daniel J., McDonnell, Shannon K., Akhtari, Farida S., Sinnwell, Jason P., Batzler, Anthony, Cobran, Ewan K., Motsinger-Reif, Alison
• Format: Article
• Language: English
• Subjects: biased risk; genetic risk scores; health disparities; social determinants of health
• ISSN: 2666-2477; 2666-2477
• DOI: 10.1016/j.xhgg.2024.100389

The use of polygenic scores (PGS) for personalized medicine has gained momentum, along with caution to avoid accentuating health disparities. Greater ancestral diversity in genetic studies is needed, as well as close attention to the social determinants of health (SDoH).We measured the correlations between 3,030 PGS from the PGS Catalog and SDoH among participants in the Personalized Environment and Genes Study (PEGS). Correlations mainly ranged from −0.05 to 0.05, yet there was a heterogeneity of correlations across SDoH themes, with the largest amount of heterogeneity for PGS predicting body measures and smoking, as well as some common diseases. We also quantify the expected bias of PGS effect size on disease risk when strong predictors, such as SDoH, are omitted from models, emphasizing the importance of including SDoH with PGS to avoid biased estimates of PGS risk and to achieve equitable precision medicine. Analysis of 3,030 polygenic scores (PGS) in PEGS participants showed a range of correlations with social determinants of health (SDoH), with accentuated correlation heterogeneity for traits like body measures and smoking. Omitting strong predictors can bias the effects of included variables, emphasizing inclusion of SDoH to ensure equity in precision medicine.