Publications by authors named "A Bastani"
Article Synopsis
- The study aimed to evaluate the effectiveness of Natural Language Processing (NLP) using large language models (LLMs) like GPT-4 for identifying cirrhosis cohorts compared to traditional diagnosis code classification and manual chart review.
- A review of 3,788 discharge summaries revealed that LLM-based classification provided higher accuracy rates (up to 98.8%) for detecting cirrhosis and its complications, while code-based classification demonstrated lower positive predictive values (PPVs), especially for complex conditions.
- The findings indicate that LLMs could serve as a more efficient and precise alternative to code-based methods, potentially reducing the need for labor-intensive chart reviews in cirrhosis research.
Telerehabilitation is an appealing service delivery option for optimising recovery. Internationally, the equity of telerehabilitation services for people from culturally and linguistically diverse (CALD) backgrounds has been questioned. Using a 31-item survey, our study explored the access, readiness and willingness of 260 patients receiving allied health services from a large tertiary health service located in Sydney, Australia, to use telerehabilitation for adults.
View Article and Find Full Text PDFVaccine hesitancy is a major barrier to vaccination, hindering the success of vaccine efforts and thereby increasing public health risk to viral diseases, including COVID-19. Neurodivergent (ND) individuals, including individuals with an intellectual and/or developmental disability, have demonstrated a heightened risk of hospitalization and death due to COVID-19, highlighting the need for further research specifically on ND communities. We conducted a qualitative analysis using in-depth interviews with medical professionals, non-medical health professionals and communicators, and ND individuals or their caregivers.
View Article and Find Full Text PDF