The vice dean for data science at Duke University School of Medicine spoke about the major disparities observed in models that assess stroke risk. [WATCH TIME: 4 minutes]
WATCH TIME: 4 minutes
“There is growing evidence of knowledge that the ability for machine learning to contribute something meaningful using data that's less complex, is generally not high. Our results confirm that.”
Using systems such as artificial intelligence (AI) has become a popular tool in neurology as it has the ability to incorporate deep learning and machine learning algorithms. As stroke continues to remain a major public health concern—being one of the leading causes of mortality worldwide—incorporating AI technology may have potential in improving care for these patients.
A recently published study in the Journal of American Medical Association compared performance of stroke-specific algorithms with pooled cohort equations between different subgroups and assessed the value of machine learning.1 Michael Pencina, PhD, corresponding author of the study, and colleagues, noted from the results that prediction models assessing for risk and the algorithm technique did not significantly improve discriminative accuracy for new-onset stroke with cohort studies.
In a recent interview with NeurologyLive®,Pencina discussed the generalizability of the results based on cohort data from the United States. He shared how the findings compared with the original hypothesis of the study. Pencina, professor in the Department of Biostatistics and Bioinformatics and director of AI Health at Duke University School of Medicine, discussed potential future research on the health inequalities in America and how it can relate to his findings.