Biomedical text readability after hypernym substitution with fine-tuned large language models.

The advent of patient access to complex medical information online has highlighted the need for simplification of biomedical text to improve patient understanding and engagement in taking ownership of their health. However, comprehension of biomedical text remains a difficult task due to the need for domain-specific expertise. We aimed to study the simplification of biomedical text via large language models (LLMs) commonly used for general natural language processing tasks involve text comprehension, summarization, generation, and prediction of new text from prompts. Specifically, we finetuned three variants of large language models to perform substitutions of complex words and word phrases in biomedical text with a related hypernym. The output of the text substitution process using LLMs was evaluated by comparing the pre- and post-substitution texts using four readability metrics and two measures of sentence complexity. A sample of 1,000 biomedical definitions in the National Library of Medicine's Unified Medical Language System (UMLS) was processed with three LLM approaches, and each showed an improvement in readability and sentence complexity after hypernym substitution. Readability scores were translated from a pre-processed collegiate reading level to a post-processed US high-school level. Comparison between the three LLMs showed that the GPT-J-6b approach had the best improvement in measures of sentence complexity. This study demonstrates the merit of hypernym substitution to improve readability of complex biomedical text for the public and highlights the use case for fine-tuning open-access large language models for biomedical natural language processing.

Effect of Recent Abortion Legislation on Twitter User Engagement, Sentiment, and Expressions of Trust in Clinicians and Privacy of Health Information: Content Analysis.

BACKGROUND: The Supreme Court ruling in Dobbs v Jackson Women's Health Organization (Dobbs) overrules precedents established by Roe v Wade and Planned Parenthood v Casey and allows states to individually regulate access to abortion care services. While many states have passed laws to protect access to abortion services since the ruling, the ruling has also triggered the enforcement of existing laws and the creation of new ones that ban or restrict abortion. In addition to denying patients the full spectrum of reproductive health care, one major concern in the medical community is how the ruling will undermine trust in the patient-clinician relationship by influencing perceptions of the privacy of patient health information. OBJECTIVE: This study aimed to study the effect of recent abortion legislation on Twitter user engagement, sentiment, expressions of trust in clinicians, and privacy of health information. METHODS: We scraped tweets containing keywords of interest between January 1, 2020, and October 17, 2022, to capture tweets posted before and after the leak of the Supreme Court decision. We then trained a Latent Dirichlet Allocation model to select tweets pertinent to the topic of interest and performed a sentiment analysis using Robustly Optimized Bidirectional Encoder Representations from Transformers Pre-training Approach model and a causal impact time series analysis to examine engagement and sentiment. In addition, we used a Word2Vec model to study the terms of interest against a latent trust dimension to capture how expressions of trust for our terms of interest changed over time and used term frequency, inverse-document frequency to measure the volume of tweets before and after the decision with respect to the negative and positive sentiments that map to our terms of interest. RESULTS: Our study revealed (1) a transient increase in the number of daily users by 576.86% (95% CI 545.34%-607.92%; P<.001), tweeting about abortion, health care, and privacy of health information postdecision leak; (2) a sustained and statistically significant decrease in the average daily sentiment on these topics by 19.81% (95% CI -22.98% to -16.59%; P=.001) postdecision leak; (3) a decrease in the association of the latent dimension of trust across most clinician-related and health information-related terms of interest; (4) an increased frequency of tweets with these clinician-related and health information-related terms and concomitant negative sentiment in the postdecision leak period. CONCLUSIONS: The study suggests that the Dobbs ruling has consequences for health systems and reproductive health care that extend beyond denying patients access to the full spectrum of reproductive health services. The finding of a decrease in the expression of trust in clinicians and health information-related terms provides evidence to support advocacy and initiatives that proactively address concerns of trust in health systems and services.

Website Usability Analysis of U.S. Military Residency Programs.

INTRODUCTION: The Military Match is the residency matching system for medical students attending the Uniformed Services University of Health Sciences, and the students were funded by the Health Professions Scholarship Program through the U.S. Army, Air Force, and Navy. To evaluate and compare military residency programs, students use residency program websites. Often, the residency program's website serves as a key source, or the only point of reference, when considering residency options, especially during times when face-to-face interactions are limited.This report aims to provide a systematic evaluation of military residency programs and their websites. MATERIALS AND METHODS: Utilizing a previously published website usability scoring system, military residency programs were categorized to objectively and quantitatively analyze their websites. Usability was divided into four categories for quantifiable analysis: accessibility, marketing, content quality, and technology. The methodology for this analysis was replicated from published reports that have examined healthcare website usability. Each website was analyzed and scored in four categories: accessibility, content quality, marketing, and technology. A "General Usability" score was calculated for each website using a composite of the key factors within the four categories. An overall score was generated utilizing the weighted percentage across all four categories. To address deficiencies of the original methodology, a secondary analysis was performed on the listed websites utilizing an automated methodology for website usability. RESULTS: A comprehensive list of 125 Accreditation Council for Graduate Medical Education U.S. Military residency program websites was compiled. Of these, 96 programs and 106 websites were evaluated. The primary analysis employing usability methodology identified technology as the highest ranked category with a score of 0.749 (SD ± 0.039) (SE 0.005) (P < .05). Marketing and content quality were the lowest scoring categories with mean scores of 0.414 (SD ± 0.054) (SE 0.006) and 0.428 (SD ± 0.229) (SE 0.027), respectively (P < .05). There was no significant difference in overall usability rankings or scores among the 96 residency program websites across the three branches (P < .05).Secondary analysis with the new usability methodology demonstrated military residency websites to exhibit more external backlinking compared to internal backlinking (P < 0.05) and no social media backlinking to any of the 106 analyzed websites. When comparing the three services, the Army had significantly lower external backlinking ranking 43.4 (P < .05) and overall backlinking ranking 56.4 (P < 0.05) when compared to the Navy (mean 48.8 and 71.7, and 43.4). There were no other differences in backlinking rankings across the three branches. CONCLUSIONS: Residency websites have become a primary way to communicate information to applicants. By assessing the overall usability of the various military residency websites, we determined the effectiveness of these websites to relay information to prospective students interested in applying for military residency. We predict that by improving website accessibility, residency programs increase their effectiveness at communicating information to potential applicants and increase interest in military residency programs.

Risk Adjustment Is Necessary in Value-Based Payment Models for Arthroplasty for Oncology Patients.

BACKGROUND: Value-based payment models such as bundled payments have been introduced to reduce costs following total hip arthroplasty (THA). Concerns exist, however, about access to care for patients who utilize more resources. The purpose of this study is thus to compare resource utilization and outcomes of patients undergoing THA for malignancy with those undergoing THA for fracture or osteoarthritis. METHODS: We queried the American College of Surgeons National Surgical Quality Improvement Program database to identify all hip arthroplasties performed from 2013 to 2016 for a primary diagnosis of malignancy (n = 296), osteoarthritis (n = 96,480), and fracture (n = 13,406). The rates of readmissions, reoperations, comorbidities, mortality, and surgical characteristics were compared between the 3 cohorts. To control for confounding variables, a multivariate analysis was performed to identify independent risk factors for resource utilization and outcomes following THA. RESULTS: Patients undergoing THA for malignancy had a longer mean operative time (155.7 vs 82.9 vs 91.0 minutes, P < .001), longer length of stay (9.0 vs 7.2 vs 2.6 days, P < .001), and were more likely to be discharged to a rehabilitation facility (42.1% vs 61.8% vs 20.2%, P < .001) than patients with fracture or osteoarthritis. When controlling for demographics and comorbidities, patients undergoing THA for malignancy had a higher rate of readmission (adjusted odds ratio 3.39, P < .001) and reoperation (adjusted odds ratio 3.71, P < .001). CONCLUSION: Patients undergoing THA for malignancy utilize more resources in an episode-of-care and have worse outcomes. Risk adjustment is necessary for oncology patients in order to prevent access to care problems for these high-risk patients.

Age-driven modulation of tRNA-derived fragments in Drosophila and their potential targets.

BACKGROUND: Development of sequencing technologies and supporting computation enable discovery of small RNA molecules that previously escaped detection or were ignored due to low count numbers. While the focus in the analysis of small RNA libraries has been primarily on microRNAs (miRNAs), recent studies have reported findings of fragments of transfer RNAs (tRFs) across a range of organisms. RESULTS: Here we describe Drosophila melanogaster tRFs, which appear to have a number of structural and functional features similar to those of miRNAs but are less abundant. As is the case with miRNAs, (i) tRFs seem to have distinct isoforms preferentially originating from 5' or 3' end of a precursor molecule (in this case, tRNA), (ii) ends of tRFs appear to contain short "seed" sequences matching conserved regions across 12 Drosophila genomes, preferentially in 3' UTRs but also in introns and exons; (iii) tRFs display specific isoform loading into Ago1 and Ago2 and thus likely function in RISC complexes; (iii) levels of loading in Ago1 and Ago2 differ considerably; and (iv) both tRF expression and loading appear to be age-dependent, indicating potential regulatory changes from young to adult organisms. CONCLUSIONS: We found that Drosophila tRF reads mapped to both nuclear and mitochondrial tRNA genes for all 20 amino acids, while previous studies have usually reported fragments from only a few tRNAs. These tRFs show a number of similarities with miRNAs, including seed sequences. Based on complementarity with conserved Drosophila regions we identified such seed sequences and their possible targets with matches in the 3'UTR regions. Strikingly, the potential target genes of the most abundant tRFs show significant Gene Ontology enrichment in development and neuronal function. The latter suggests that involvement of tRFs in the RNA interfering pathway may play a role in brain activity or brain changes with age.