Ensemble pretrained language models to extract biomedical knowledge from literature.

OBJECTIVES: The rapid expansion of biomedical literature necessitates automated techniques to discern relationships between biomedical concepts from extensive free text. Such techniques facilitate the development of detailed knowledge bases and highlight research deficiencies. The LitCoin Natural Language Processing (NLP) challenge, organized by the National Center for Advancing Translational Science, aims to evaluate such potential and provides a manually annotated corpus for methodology development and benchmarking. MATERIALS AND METHODS: For the named entity recognition (NER) task, we utilized ensemble learning to merge predictions from three domain-specific models, namely BioBERT, PubMedBERT, and BioM-ELECTRA, devised a rule-driven detection method for cell line and taxonomy names and annotated 70 more abstracts as additional corpus. We further finetuned the T0pp model, with 11 billion parameters, to boost the performance on relation extraction and leveraged entites' location information (eg, title, background) to enhance novelty prediction performance in relation extraction (RE). RESULTS: Our pioneering NLP system designed for this challenge secured first place in Phase I-NER and second place in Phase II-relation extraction and novelty prediction, outpacing over 200 teams. We tested OpenAI ChatGPT 3.5 and ChatGPT 4 in a Zero-Shot setting using the same test set, revealing that our finetuned model considerably surpasses these broad-spectrum large language models. DISCUSSION AND CONCLUSION: Our outcomes depict a robust NLP system excelling in NER and RE across various biomedical entities, emphasizing that task-specific models remain superior to generic large ones. Such insights are valuable for endeavors like knowledge graph development and hypothesis formulation in biomedical research.

The IMPACT framework and implementation for accessible in silico clinical phenotyping in the digital era.

Clinical phenotyping is often a foundational requirement for obtaining datasets necessary for the development of digital health applications. Traditionally done via manual abstraction, this task is often a bottleneck in development due to time and cost requirements, therefore raising significant interest in accomplishing this task via in-silico means. Nevertheless, current in-silico phenotyping development tends to be focused on a single phenotyping task resulting in a dearth of reusable tools supporting cross-task generalizable in-silico phenotyping. In addition, in-silico phenotyping remains largely inaccessible for a substantial portion of potentially interested users. Here, we highlight the barriers to the usage of in-silico phenotyping and potential solutions in the form of a framework of several desiderata as observed during our implementation of such tasks. In addition, we introduce an example implementation of said framework as a software application, with a focus on ease of adoption, cross-task reusability, and facilitating the clinical phenotyping algorithm development process.

Telehealth Impact on Primary Care Related Ambulance Transports.

Introduction: Telehealth has been used nominally for trauma, neurological, and cardiovascular incidents in prehospital emergency medical services (EMS). Yet, much less is known about the use of telehealth for low-acuity primary care. We examine the development of one telehealth program and its impact on unnecessary ambulance transports. Objective: The objective of this study is to describe the development and impact of a large-scale telehealth program on ambulance transports. Methods: We describe the patient characteristics and results from a cohort of patients in Houston, Texas who received a prehospital telehealth consultation from an emergency medicine physician. Inclusion criteria were adults and pediatric patients with complaints considered to be non-urgent, primary care related. Data were analyzed for 36 months, from January 2015 through December 2017. Our primary dependent variable was the percentage of patients transported by ambulance. We used descriptive statistics to describe patient demographics, chi-square to examine differences between groups, and logistic regression to explore the effects with multivariate controls including age, gender, race, and chief complaint. Results: A total of 15,067 patients were enrolled (53% female; average age 44 years ± 19 years) over the three-year period. The 3 primary chief complaints were based on abdominal pains (13% of cases), nausea/vomiting/diarrhea (NVD) (9.4%), and back pain (9.3%). Ambulance transports represented 11.2% of all transports in the program, while alternative taxi transportation was used in 75.6%, and the remainder were self- or no-transports. Taxi transportation to an alternate, affiliated clinic (versus ED) was utilized in 5% of incidents. After multivariate controls, older age patients presenting with low-risk, non-acute chest pain, shortness of breath, and dizziness were much more likely to use ambulance transport. Race and gender were not significant predictors of ambulance transport. Conclusions: We found telehealth offers a technology strategy to address potentially unnecessary ambulance transports. Based on prior cost-effectiveness analyses, the reduction of unnecessary ambulance transports translates to an overall reduction in EMS agency costs. Telehealth programs offer a viable solution to support alternate destination and alternate transport programs.