Publications by authors named "D Mowery"
Background: The increasing use of social media to share lived and living experiences of substance use presents a unique opportunity to obtain information on side effects, use patterns, and opinions on novel psychoactive substances. However, due to the large volume of data, obtaining useful insights through natural language processing technologies such as large language models is challenging.
Objective: This paper aims to develop a retrieval-augmented generation (RAG) architecture for medical question answering pertaining to clinicians' queries on emerging issues associated with health-related topics, using user-generated medical information on social media.
Background: Stopping or reducing risky or unneeded medications ("deprescribing") could improve older adults' health. Electronic health data can support observational and intervention studies of deprescribing, but there are no standardized measures for key variables, and healthcare systems have differing data types and availability. We developed definitions for chronic medication use and discontinuation based on electronic health data and applied them in a case study of benzodiazepines and Z-drugs in five diverse US healthcare systems.
View Article and Find Full Text PDFArticle Synopsis
- * A comprehensive review of SDoH in EHRs revealed approaches for screening, data collection, and using natural language processing (NLP) to extract data, but highlighted inconsistencies across methods and outcomes.
- * There is a pressing need for the development of standardized measures and coordinated interventions to effectively integrate SDoH data into clinical practice, which is crucial for improving health equity.
Article Synopsis
- Patients recovering from COVID-19 often experience lingering symptoms known as Long COVID, which can manifest weeks or months after their initial infection, but the prevalence of this condition is not well understood.
- To address this, a collaborative initiative called the Long COVID Computational Challenge (L3C) was launched to develop effective risk prediction tools for identifying individuals at risk of Long COVID using extensive healthcare data from over 75 institutions in the U.S.
- The challenge resulted in 74 teams creating 35 predictive models, with the top models achieving high accuracy scores, demonstrating the potential for machine learning to enhance the identification of patients at risk for Long COVID.
Few studies examining the patient outcomes of concurrent neurological manifestations during acute COVID-19 leveraged multinational cohorts of adults and children or distinguished between central and peripheral nervous system (CNS vs. PNS) involvement. Using a federated multinational network in which local clinicians and informatics experts curated the electronic health records data, we evaluated the risk of prolonged hospitalization and mortality in hospitalized COVID-19 patients from 21 healthcare systems across 7 countries.
View Article and Find Full Text PDF