Abstract
AbstractObjectivesEvaluating methods for building data frameworks for application of AI in large scale datasets for women’s health studies.MethodsWe created methods for transforming raw data to a data framework for applying machine learning (ML) and natural language processing (NLP) techniques for predicting falls and fractures.ResultsPrediction of falls was higher in women compared to men. Information extracted from radiology reports was converted to a matrix for applying machine learning. For fractures, by applying specialized algorithms, we extracted snippets from dual x-ray absorptiometry (DXA) scans for meaningful terms usable for predicting fracture risk.DiscussionLife cycle of data from raw to analytic form includes data governance, cleaning, management, and analysis. For applying AI, data must be prepared optimally to reduce algorithmic bias.ConclusionAlgorithmic bias is harmful for research using AI methods. Building AI ready data frameworks that improve efficiency can be especially valuable for women’s health.Lay SummaryWomen’s health studies are rare in large cohorts of women. The department of Veterans affairs (VA) has data for a large number of women in care. Prediction of falls and fractures are important areas of study related to women’s health. Artificial Intelligence (AI) methods have been developed at the VA for predicting falls and fractures. In this paper we discuss data preparation for applying these AI methods. We discuss how data preparation can affect bias and reproducibility in AI outcomes.
Publisher
Cold Spring Harbor Laboratory