Large Language Models for Efficient Medical Information Extraction.

Extracting valuable insights from unstructured clinical narrative reports is a challenging yet crucial task in the healthcare domain as it allows healthcare workers to treat patients more efficiently and improves the overall standard of care. We employ ChatGPT, a Large language model (LLM), and compare its performance to manual reviewers. The review focuses on four key conditions: family history of heart disease, depression, heavy smoking, and cancer. The evaluation of a diverse sample of History and Physical (H&P) Notes, demonstrates ChatGPT's remarkable capabilities. Notably, it exhibits exemplary results in sensitivity for depression and heavy smokers and specificity for cancer. We identify areas for improvement as well, particularly in capturing nuanced semantic information related to family history of heart disease and cancer. With further investigation, ChatGPT holds substantial potential for advancements in medical information extraction.

HIV/STI/HCV Risk Clusters and Hierarchies Experienced by Women Recently Released from Incarceration.

This study examines cross-sectional clusters and longitudinal predictions using an expanded SAVA syndemic conceptual framework-SAVA MH + H (substance use, intimate partner violence, mental health, and homelessness leading to HIV/STI/HCV risks)-among women recently released from incarceration (WRRI) (n = 206) participating in the WORTH Transitions (WT) intervention. WT combines two evidence-based interventions: the Women on the Road to Health HIV intervention, and Transitions Clinic. Cluster analytic and logistic regression methods were utilized. For the cluster analyses, baseline SAVA MH + H variables were categorized into presence/absence. For logistic regression, baseline SAVA MH + H variables were examined on a composite HIV/STI/HCV outcome collected at 6-month follow-up, controlling for lifetime trauma and sociodemographic characteristics. Three SAVA MH + H clusters were identified, the first of which had women with the highest overall levels of SAVA MH + H variables, 47% of whom were unhoused. Hard drug use (HDU) was the only significant predictor of HIV/STI/HCV risks in the regression analyses. HDUs had 4.32-fold higher odds of HIV/STI/HCV outcomes than non-HDUs (p = 0.002). Interventions such as WORTH Transitions must differently target identified SAVA MH + H syndemic risk clusters and HDU to prevent HIV/HCV/STI outcomes among WRRI.