Skip to main content
Study

Comparison of methods to reduce bias from clinical prediction models of postpartum depression.

Park Y, Hu J, Singh M, et al. Comparison of methods to reduce bias from clinical prediction models of postpartum depression. JAMA Netw Open. 2021;4(4):e213909. doi: 10.1001/jamanetworkopen.2021.3909.

Save

Print

May 12, 2021
Park Y, Hu J, Singh M, et al. JAMA Netw Open. 2021;4(4):e213909.

Machine learning uses data and statistical methods to enhance risk prediction models and it has been promoted as a tool to improve healthcare safety. Using Medicaid claims data for a large cohort of White and Black pregnant females, this study evaluated approaches to reduce bias in clinical prediction algorithms for postpartum depression and mental health service utilization. The researchers found that a reweighing method in machine learning models was associated with a greater reduction in bias than excluding race from the prediction models. The authors suggest further examination of potentially biased data informing clinical prediction models and consideration of other methods to mitigate bias.

Save

Print

Cite
Citation

Park Y, Hu J, Singh M, et al. Comparison of methods to reduce bias from clinical prediction models of postpartum depression. JAMA Netw Open. 2021;4(4):e213909. doi: 10.1001/jamanetworkopen.2021.3909.