Search

Large language models (LLMs) continue to exhibit noteworthy capabilities across a spectrum of areas, including emerging proficiencies across the health care continuum. Successful LLM implementation and adoption depend on digital readiness, modern infrastructure, a trained workforce, privacy, and an ethical regulatory landscape. These factors can vary significantly across health care ecosystems, dictating the choice of a particular LLM implementation pathway. This perspective discusses 3 LLM implementation pathways-training from scratch pathway (TSP), fine-tuned pathway (FTP), and out-of-the-box pathway (OBP)-as potential onboarding points for health systems while facilitating equitable adoption. The choice of a particular pathway is governed by needs as well as affordability. Therefore, the risks, benefits, and economics of these pathways across 4 major cloud service providers (Amazon, Microsoft, Google, and Oracle) are presented. While cost comparisons, such as on-demand and spot pricing across the cloud service providers for the 3 pathways, are presented for completeness, the usefulness of managed services and cloud enterprise tools is elucidated. Managed services can complement the traditional workforce and expertise, while enterprise tools, such as federated learning, can overcome sample size challenges when implementing LLMs using health care data. Of the 3 pathways, TSP is expected to be the most resource-intensive regarding infrastructure and workforce while providing maximum customization, enhanced transparency, and performance. Because TSP trains the LLM using enterprise health care data, it is expected to harness the digital signatures of the population served by the health care system with the potential to impact outcomes. The use of pretrained models in FTP is a limitation. It may impact its performance because the training data used in the pretrained model may have hidden bias and may not necessarily be health care-related. However, FTP provides a balance between customization, cost, and performance. While OBP can be rapidly deployed, it provides minimal customization and transparency without guaranteeing long-term availability. OBP may also present challenges in interfacing seamlessly with downstream applications in health care settings with variations in pricing and use over time. Lack of customization in OBP can significantly limit its ability to impact outcomes. Finally, potential applications of LLMs in health care, including conversational artificial intelligence, chatbots, summarization, and machine translation, are highlighted. While the 3 implementation pathways discussed in this perspective have the potential to facilitate equitable adoption and democratization of LLMs, transitions between them may be necessary as the needs of health systems evolve. Understanding the economics and trade-offs of these onboarding pathways can guide their strategic adoption and demonstrate value while impacting health care outcomes favorably., (©Radha Nagarajan, Midori Kondo, Franz Salas, Emre Sezgin, Yuan Yao, Vanessa Klotzman, Sandip A Godambe, Naqi Khan, Alfonso Limon, Graham Stephenson, Sharief Taraman, Nephi Walton, Louis Ehwerhemuepha, Jay Pandit, Deepti Pandita, Michael Weiss, Charles Golden, Adam Gold, John Henderson, Angela Shippy, Leo Anthony Celi, William R Hogan, Eric K Oermann, Terence Sanger, Steven Martel. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 14.11.2024.)

Molecular diagnosis for pediatric epilepsy patients can impact treatment and health supervision recommendations. However, there is little known about factors affecting the time to receive a diagnosis. Our objective was to characterize factors affecting the time from first seizure to molecular diagnosis in children with epilepsy. A retrospective, population-based review was used to analyze data from pediatric patients with a genetic etiology for epilepsy over a 5 year period. A subgroup of patients with seizure onset after 2016 was evaluated for recent trends. We identified 119 patients in the main cohort and 62 in a more recent (contemporaneous) subgroup. Sex, race, and ethnicity were not significantly associated with time to molecular diagnosis. A greater number of hospitalizations was associated with a shorter time to diagnosis (p < 0.001). Developmental delay was associated with a longer time to diagnosis (p = 0.002). We found no association for time to diagnosis with a diagnosis of autism, utilization of free genetic testing, or epilepsy type. In the recent subgroup analysis, commercial insurance was associated with decreased time to diagnosis (p = 0.02). Developmental delay, public insurance, or patients in the outpatient setting had longer times to molecular diagnosis. These findings suggest that there may be opportunities to implement interventions aimed at accelerating the provision of genetic testing in pediatric epilepsy. PLAIN LANGUAGE SUMMARY: Genetic diagnosis for pediatric epilepsy patients can impact treatment and care. This study looked at factors that affect how long it takes a pediatric epilepsy patient to receive a genetic diagnosis. We found that sex, race and ethnicity, epilepsy type, and whether the patient had autism did not affect how long it took the patient to receive a diagnosis. However, we found that patients with developmental delay, fewer hospitalizations, and public insurance took a longer time to receive a diagnosis. Our findings suggest potential strategies for reducing the time to receive a genetic diagnosis., (© 2024 The Author(s). Epilepsia Open published by Wiley Periodicals LLC on behalf of International League Against Epilepsy.)

The HerediGene Population Study is a large research study focused on identifying new genetic biomarkers for disease prevention, diagnosis, prognosis, and development of new therapeutics. A substantial IT infrastructure evolved to reach enrollment targets and return results to participants. More than 170,000 participants have been enrolled in the study to date, with 5.87% of those whole genome sequenced and 0.46% of those genotyped harboring pathogenic variants. Among other purposes, this infrastructure supports: (1) identifying candidates from clinical criteria, (2) monitoring for qualifying clinical events (e.g., blood draw), (3) contacting candidates, (4) obtaining consent electronically, (5) initiating lab orders, (6) integrating consent and lab orders into clinical workflow, (7) de-identifying samples and clinical data, (8) shipping/transmitting samples and clinical data, (9) genotyping/sequencing samples, (10) and re-identifying and returning results for participants where applicable. This study may serve as a model for similar genomic research and precision public health initiatives., (©2023 AMIA - All rights reserved.)

Objective: High BMI is associated with many comorbidities and mortality. This study aimed to elucidate the overall clinical risk of obesity using a genome- and phenome-wide approach., Methods: This study performed a phenome-wide association study of BMI using a clinical cohort of 736,726 adults. This was followed by genetic association studies using two separate cohorts: one consisting of 65,174 adults in the Electronic Medical Records and Genomics (eMERGE) Network and another with 405,432 participants in the UK Biobank., Results: Class 3 obesity was associated with 433 phenotypes, representing 59.3% of all billing codes in individuals with severe obesity. A genome-wide polygenic risk score for BMI, accounting for 7.5% of variance in BMI, was associated with 296 clinical diseases, including strong associations with type 2 diabetes, sleep apnea, hypertension, and chronic liver disease. In all three cohorts, 199 phenotypes were associated with class 3 obesity and polygenic risk for obesity, including novel associations such as increased risk of renal failure, venous insufficiency, and gastroesophageal reflux., Conclusions: This combined genomic and phenomic systematic approach demonstrated that obesity has a strong genetic predisposition and is associated with a considerable burden of disease across all disease classes., (© 2022 The Obesity Society. This article has been contributed to by U.S. Government employees and their work is in the public domain in the USA.)

Precision health's more individualized molecular approach will enrich our understanding of disease etiology and patient outcomes. Universal implementation of precision health will not be feasible, however, until there is much greater automation of processes related to genomic data transmission, transformation, and interpretation. In this paper, we describe a framework for genomic data flow developed by the Clinical Informatics Work Group of the NIH National Human Genome Research Institute (NHGRI) IGNITE Network consortium. We subsequently report the results of a genomic data flow survey administered to sites funded by NIH-NHGRI for large scale genomic medicine implementations. Finally, we discuss insights and challenges identified through these survey results as they relate to both the current and a desirable future state of genomic data flow., (©2019 AMIA - All rights reserved.)

Developing a forecasting tool for patient census allows for improved staffing, better resource utilization and mobilization, and improved timing of educational campaigns around the disease control process. Using a neural network approach we evaluated several different models and variables for predicting patient census prospectively. These initial studies enabled selection of a subset of predictor variables and show that different network models, and variables must be used based on the season.