Sunghwan Sohn
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Sunghwan Sohn has a total of 270 co-authors. Here are the top co-authors by number of shared publications:
Hongfang Liu
McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston, Houston, TX, USA
12 shared publicationsMaria Vassilaki
Department of Quantitative Health Sciences, Mayo Clinic, Rochester, USA
9 shared publicationsRonald C. Petersen
Department of Neurology, Mayo Clinic, Rochester, USA
9 shared publications- Sunyang Fu
McWilliams School of Biomedical Informatics, University of Texas Health Science Center at Houston , Houston, Texas
9 shared publications - Hongfang Liu
Department of Quantitative and Systems Health Sciences, University of Texas at Austin Dell Medical School, Austin, TX 78712, USA
8 shared publications Muskan Garg
Department of Artificial Intelligence & Informatics, Mayo Clinic, Rochester, USA
8 shared publications
Latest Publications
- 10 September 2026
- Published by Oxford University Press (OUP) in American Journal of Health-System Pharmacy
Performance evaluation of a large language model for medication management tasks
Purpose: Large language models (LLMs) have proven performance for certain diagnostic tasks; however, limited studies have evaluated their consistency in recommending appropriate medication regimens for a given diagnosis. Medication management is a complex task that requires synthesis of drug formulation and complete order instructions for safe use. Here, the performance of GPT-4o, an LLM available with OpenAI’s ChatGPT, was tested on 3 medication management tasks. Methods: GPT-4o performance was tested on 3 medication tasks: identifying available formulations for a given generic drug name, identifying drug-drug interactions (DDIs) for a given medication regimen, and preparing a medication order for a given generic drug name. For each experiment, the model’s raw text response was captured exactly as returned and evaluated using clinician evaluation in addition to standard LLM metrics, including Term Frequency-Inverse Document Frequency (TF-IDF) vectors, normalized Levenshtein similarity, and Recall-Oriented Understudy for Gisting Evaluation (ROUGE-1/ROUGE-L) F1 score between each response and its reference string. Results: For the first task of drug-formulation matching, GPT-4o had 49% accuracy for generic medications being matched to all available formulations, with an average of 1.23 omissions per medication and 1.14 hallucinations per medication. For the second task of drug-drug interaction identification, the accuracy was 54.7% for identifying the DDI pair. For the third task, GPT-4o generated order sentences containing no medication or abbreviation errors in 65.8% of the cases. Conclusion: Model performance for basic medication tasks was consistently poor. This evaluation highlights the need for domain-specific training through clinician-annotated datasets and a comprehensive evaluation framework for benchmarking performance.
- 10 September 2026
- Published by Oxford University Press (OUP) in American Journal of Health-System Pharmacy
Performance evaluation of a large language model for medication management tasks
Purpose: Large language models (LLMs) have proven performance for certain diagnostic tasks; however, limited studies have evaluated their consistency in recommending appropriate medication regimens for a given diagnosis. Medication management is a complex task that requires synthesis of drug formulation and complete order instructions for safe use. Here, the performance of GPT-4o, an LLM available with OpenAI’s ChatGPT, was tested on 3 medication management tasks. Methods: GPT-4o performance was tested on 3 medication tasks: identifying available formulations for a given generic drug name, identifying drug-drug interactions (DDIs) for a given medication regimen, and preparing a medication order for a given generic drug name. For each experiment, the model’s raw text response was captured exactly as returned and evaluated using clinician evaluation in addition to standard LLM metrics, including Term Frequency-Inverse Document Frequency (TF-IDF) vectors, normalized Levenshtein similarity, and Recall-Oriented Understudy for Gisting Evaluation (ROUGE-1/ROUGE-L) F1 score between each response and its reference string. Results: For the first task of drug-formulation matching, GPT-4o had 49% accuracy for generic medications being matched to all available formulations, with an average of 1.23 omissions per medication and 1.14 hallucinations per medication. For the second task of drug-drug interaction identification, the accuracy was 54.7% for identifying the DDI pair. For the third task, GPT-4o generated order sentences containing no medication or abbreviation errors in 65.8% of the cases. Conclusion: Model performance for basic medication tasks was consistently poor. This evaluation highlights the need for domain-specific training through clinician-annotated datasets and a comprehensive evaluation framework for benchmarking performance.
- Open Access
- 14 May 2026
- Published by Elsevier in Journal of the National Medical Association
- Vol. 118 (4) , 789-816
A lifecycle governance and learning health system framework for trustworthy, generalizable, and sustainable human-ai partnership in clinical practice: Lessons from the asthma-guidance and prediction system (A-GPS)
- 14 May 2026
- Published by Elsevier in Journal of the National Medical Association
- Vol. 118 (4) , 789-816
Open AccessA lifecycle governance and learning health system framework for trustworthy, generalizable, and sustainable human-ai partnership in clinical practice: Lessons from the asthma-guidance and prediction system (A-GPS)
- Open Access
- 17 February 2026
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 14, e80405
Assessing the Impact of Sociodemographic Factors on Artificial Intelligence Models in Predicting Dementia: Retrospective Cohort Study
Background: Artificial intelligence (AI) is increasingly applied to health care, yet concerns about fairness persist, particularly in relation to sociodemographic disparities. Previous studies suggest that socioeconomic status (SES) and sex may influence AI model performance, potentially affecting groups that are historically underserved or understudied. Objective: This study aimed to (1) assess algorithmic bias in AI-driven dementia prediction models based on SES and sex (biological sex), (2) compare the utility of an individual-level SES measure (the Housing-Based Socioeconomic Status [HOUSES] Index) versus an area-level measure (the Area Deprivation Index) for bias detection, and (3) evaluate the effectiveness of an oversampling technique (the Synthetic Minority Oversampling Technique for Nominal and Continuous features) for bias mitigation. Methods: This study used data from two population-based cohorts: the Mayo Clinic Study on Aging (n=3041) and the Rochester Epidemiology Project (n=19,572). Four AI models (random forest, logistic regression, support vector machine, and Naïve Bayes) were trained using a 5-year observation window of structured electronic health record data to predict dementia onset within the subsequent 1-year window. Fairness and model performance were assessed using the balanced error rate (BER) across intersecting SES-sex subgroups. The Synthetic Minority Oversampling Technique for Nominal and Continuous features algorithm was applied to the training data to balance the representation of SES groups. Results: Across both cohorts, individuals with lower SES generally exhibited higher BERs (worse performance) than high SES groups, confirming the presence of bias. In the MCSA cohort, males with high SES, as indicated by the HOUSES Index, consistently exhibited the lowest BERs across all evaluated models. Balancing the training data based on a specific SES measure showed a trend toward reducing the BER disparity when evaluated using that same measure. However, this targeted improvement demonstrated nonuniversal benefits; in some cases, it exacerbated disparities when evaluated using other, unbalanced SES measures. This pattern suggests that fairness interventions are not universally beneficial across different definitions of the protected attribute. While the balancing approach improved fairness in model performance for lower SES groups, it often came at the cost of a slight reduction in overall model performance. However, an exception was observed in the MCSA cohort when balancing based on the HOUSES Index using logistic regression, support vector machine, and Naïve Bayes, where the performances of both the high and low SES groups improved. Conclusions: This research highlights the importance of incorporating sociodemographic context into AI modeling in health care. The choice of SES measure may lead to different assessments of algorithmic bias. The HOUSES Index, as a validated individual-level SES measure, may be more effective for bias mitigation than area-level measures. Future AI development should integrate bias mitigation strategies to ensure models do not reinforce existing disparities in health outcomes.
- 17 February 2026
- Published by JMIR Publications Inc. in JMIR Medical Informatics
- Vol. 14, e80405
Open AccessAssessing the Impact of Sociodemographic Factors on Artificial Intelligence Models in Predicting Dementia: Retrospective Cohort Study
Background: Artificial intelligence (AI) is increasingly applied to health care, yet concerns about fairness persist, particularly in relation to sociodemographic disparities. Previous studies suggest that socioeconomic status (SES) and sex may influence AI model performance, potentially affecting groups that are historically underserved or understudied. Objective: This study aimed to (1) assess algorithmic bias in AI-driven dementia prediction models based on SES and sex (biological sex), (2) compare the utility of an individual-level SES measure (the Housing-Based Socioeconomic Status [HOUSES] Index) versus an area-level measure (the Area Deprivation Index) for bias detection, and (3) evaluate the effectiveness of an oversampling technique (the Synthetic Minority Oversampling Technique for Nominal and Continuous features) for bias mitigation. Methods: This study used data from two population-based cohorts: the Mayo Clinic Study on Aging (n=3041) and the Rochester Epidemiology Project (n=19,572). Four AI models (random forest, logistic regression, support vector machine, and Naïve Bayes) were trained using a 5-year observation window of structured electronic health record data to predict dementia onset within the subsequent 1-year window. Fairness and model performance were assessed using the balanced error rate (BER) across intersecting SES-sex subgroups. The Synthetic Minority Oversampling Technique for Nominal and Continuous features algorithm was applied to the training data to balance the representation of SES groups. Results: Across both cohorts, individuals with lower SES generally exhibited higher BERs (worse performance) than high SES groups, confirming the presence of bias. In the MCSA cohort, males with high SES, as indicated by the HOUSES Index, consistently exhibited the lowest BERs across all evaluated models. Balancing the training data based on a specific SES measure showed a trend toward reducing the BER disparity when evaluated using that same measure. However, this targeted improvement demonstrated nonuniversal benefits; in some cases, it exacerbated disparities when evaluated using other, unbalanced SES measures. This pattern suggests that fairness interventions are not universally beneficial across different definitions of the protected attribute. While the balancing approach improved fairness in model performance for lower SES groups, it often came at the cost of a slight reduction in overall model performance. However, an exception was observed in the MCSA cohort when balancing based on the HOUSES Index using logistic regression, support vector machine, and Naïve Bayes, where the performances of both the high and low SES groups improved. Conclusions: This research highlights the importance of incorporating sociodemographic context into AI modeling in health care. The choice of SES measure may lead to different assessments of algorithmic bias. The HOUSES Index, as a validated individual-level SES measure, may be more effective for bias mitigation than area-level measures. Future AI development should integrate bias mitigation strategies to ensure models do not reinforce existing disparities in health outcomes.
- Open Access
MedMatch: a first step for the automation of large language model performance benchmarking for medication-related tasks
Background: The accuracy and safety of generating medication orders by large language models (LLMs) must be demonstrated. Without standardization, performance evaluation is limited to time and resource-intensive clinician grading. This evaluation aimed to develop a standardized medication format that supports automated performance evaluation (MedMatch). Methods: First, a survey of 40 medication prompts was given to clinicians to assess agreement in medication order communication. Second, a clinician panel developed a standardized medication format (MedMatch) for oral and intravenous medications. Third, a clinician-annotated dataset of medication prompts and standardized answers in the MedMatch format was developed for LLM testing. Finally, LLMs were retested with the same dataset, adjusted to exclude route information, to evaluate the appropriate categorization of medication route. Results: The formal medication orders consistently showed low omission rates and high overlap for all entities, compared to the verbal and brief written communication types. Lexical overlap results demonstrated pattern norms amongst clinicians with entities appearing most commonly in positions 1-5 in the order of drug name, dose, unit, route, and frequency. In the second survey, the formal written group performed the highest with 78.3% of prompts considered appropriate as a computer-generated response. LLM accuracy on MedMatch order standardization was highest in oral solid (64.2-72.5%), intravenous intermittent (72.5-84.3%), and intravenous push (62.7-74.5%) categories. LLMs performed the worst at categorizing medication orders accurately into intravenous push (18-61%) and intravenous intermittent (51-100%) routes. Conclusions: Standardized format for computer-based outputs may support automated performance analysis and enhance the clarity of medication communication.
Open AccessMedMatch: a first step for the automation of large language model performance benchmarking for medication-related tasks
Background: The accuracy and safety of generating medication orders by large language models (LLMs) must be demonstrated. Without standardization, performance evaluation is limited to time and resource-intensive clinician grading. This evaluation aimed to develop a standardized medication format that supports automated performance evaluation (MedMatch). Methods: First, a survey of 40 medication prompts was given to clinicians to assess agreement in medication order communication. Second, a clinician panel developed a standardized medication format (MedMatch) for oral and intravenous medications. Third, a clinician-annotated dataset of medication prompts and standardized answers in the MedMatch format was developed for LLM testing. Finally, LLMs were retested with the same dataset, adjusted to exclude route information, to evaluate the appropriate categorization of medication route. Results: The formal medication orders consistently showed low omission rates and high overlap for all entities, compared to the verbal and brief written communication types. Lexical overlap results demonstrated pattern norms amongst clinicians with entities appearing most commonly in positions 1-5 in the order of drug name, dose, unit, route, and frequency. In the second survey, the formal written group performed the highest with 78.3% of prompts considered appropriate as a computer-generated response. LLM accuracy on MedMatch order standardization was highest in oral solid (64.2-72.5%), intravenous intermittent (72.5-84.3%), and intravenous push (62.7-74.5%) categories. LLMs performed the worst at categorizing medication orders accurately into intravenous push (18-61%) and intravenous intermittent (51-100%) routes. Conclusions: Standardized format for computer-based outputs may support automated performance analysis and enhance the clarity of medication communication.
- Open Access
- 14 January 2026
- Published by Frontiers Media SA in Frontiers in Psychiatry
- Vol. 16, 1726100
Leveraging reddit data for context-enhanced synthetic health data generation to identify low self esteem
Low self-esteem (LoST) is a latent yet critical psychosocial risk factor that predisposes individuals to depressive disorders. Although structured tools exist to assess self-esteem, their limited clinical adoption suggests that relevant indicators of LoST remain buried within unstructured clinical narratives. The scarcity of annotated clinical notes impedes the development of natural language processing (NLP) models for its detection. Manual chart reviews are labor-intensive and large language model (LLM)-driven (weak) labeling raises privacy concerns. Past studies demonstrate that NLP models trained on LLM-generated synthetic clinical notes achieve performance comparable to, and sometimes better than those trained on real notes. This highlights synthetic data’s utility for augmenting scarce clinical corpora while reducing privacy concerns. Prior efforts have leveraged social media data, such as Reddit, to identify linguistic markers of low self-esteem; however, the linguistic and contextual divergence between social media and clinical text limits the generalizability of these models. To address this gap, we present a novel framework that generates context-enhanced synthetic clinical notes from social media narratives and evaluates the utility of small language models for identifying expressions of low self-esteem. Our approach includes a mixed-method evaluation framework: (i) structure analysis, (ii) readability analysis, (iii) linguistic diversity, and (iv) contextual fidelity of LoST cues in source Reddit posts and synthetic notes. This work offers a scalable, privacy-preserving solution for synthetic data generation for early detection of psychosocial risks such as LoST and demonstrates a pathway for translating mental health signals in clinical notes into clinically actionable insights, thereby identifying patients at risk.
- 14 January 2026
- Published by Frontiers Media SA in Frontiers in Psychiatry
- Vol. 16, 1726100
Open AccessLeveraging reddit data for context-enhanced synthetic health data generation to identify low self esteem
Low self-esteem (LoST) is a latent yet critical psychosocial risk factor that predisposes individuals to depressive disorders. Although structured tools exist to assess self-esteem, their limited clinical adoption suggests that relevant indicators of LoST remain buried within unstructured clinical narratives. The scarcity of annotated clinical notes impedes the development of natural language processing (NLP) models for its detection. Manual chart reviews are labor-intensive and large language model (LLM)-driven (weak) labeling raises privacy concerns. Past studies demonstrate that NLP models trained on LLM-generated synthetic clinical notes achieve performance comparable to, and sometimes better than those trained on real notes. This highlights synthetic data’s utility for augmenting scarce clinical corpora while reducing privacy concerns. Prior efforts have leveraged social media data, such as Reddit, to identify linguistic markers of low self-esteem; however, the linguistic and contextual divergence between social media and clinical text limits the generalizability of these models. To address this gap, we present a novel framework that generates context-enhanced synthetic clinical notes from social media narratives and evaluates the utility of small language models for identifying expressions of low self-esteem. Our approach includes a mixed-method evaluation framework: (i) structure analysis, (ii) readability analysis, (iii) linguistic diversity, and (iv) contextual fidelity of LoST cues in source Reddit posts and synthetic notes. This work offers a scalable, privacy-preserving solution for synthetic data generation for early detection of psychosocial risks such as LoST and demonstrates a pathway for translating mental health signals in clinical notes into clinically actionable insights, thereby identifying patients at risk.
- Open Access
Drug or Pokémon? Large language model performance in identification of fabricated medications
Background: Large language models (LLMs) are increasingly used in medication-related tasks despite limited evidence supporting their accuracy and safety. LLMs are vulnerable to adversarial attacks and may confabulate in response to fabricated information in the prompt. Given the substantial risk of harm from medication errors, the purpose of this study was to establish a benchmark task to evaluate LLM performance in distinguishing fabricated from real medications. Methods: Two datasets (brand and generic), each consisting of 250 medication lists were developed, each including 4-6 medications and one fabricated medication (specifically a Pokémon character), with complete dosing information provided. Six LLMs were evaluated (GPT-5-Chat, GPT-4o-mini, DrugGPT, Gemma-3-27B-IT, Llama-3.3-70B-Instruct, and Qwen3-32B). For each medication list, models were queried for dosing information and disease indication across three experimental conditions: standard decoding, standard decoding with mitigation prompting, and deterministic decoding (temperature = 0). Each experiment was performed in triplicate. Outputs were deemed confabulations if the LLM failed to recognize and reject the fabricated medication. The primary outcome was the rate of confabulations. Exact paired-permutation tests were used to compare confabulation rates across LLMs and prompting approaches. Results: Confabulation rates under standard conditions ranged from 2.7% to 99.6%. DrugGPT demonstrated the lowest baseline confabulation rates (2.7-6.4%), whereas other models demonstrated substantially higher rates (47.2-99.6%). Mitigation prompting significantly reduced confabulation rates compared to standard and deterministic decoding across all tasks and datasets (p < 0.001): GPT-5-Chat achieved confabulation rates of 0% with mitigation prompting from baseline rates of 66.4-88.5%. Deterministic decoding did not substantially impact confabulation rates. Conclusions: LLMs performing medication-related tasks are susceptible to confabulation in response to invalid or fabricated inputs, raising important safety concerns for clinical use. While mitigation prompting reduced confabulation rates, this study underscores the importance of robust safeguards when deploying LLMs in medication-related decision-making.
Open AccessDrug or Pokémon? Large language model performance in identification of fabricated medications
Background: Large language models (LLMs) are increasingly used in medication-related tasks despite limited evidence supporting their accuracy and safety. LLMs are vulnerable to adversarial attacks and may confabulate in response to fabricated information in the prompt. Given the substantial risk of harm from medication errors, the purpose of this study was to establish a benchmark task to evaluate LLM performance in distinguishing fabricated from real medications. Methods: Two datasets (brand and generic), each consisting of 250 medication lists were developed, each including 4-6 medications and one fabricated medication (specifically a Pokémon character), with complete dosing information provided. Six LLMs were evaluated (GPT-5-Chat, GPT-4o-mini, DrugGPT, Gemma-3-27B-IT, Llama-3.3-70B-Instruct, and Qwen3-32B). For each medication list, models were queried for dosing information and disease indication across three experimental conditions: standard decoding, standard decoding with mitigation prompting, and deterministic decoding (temperature = 0). Each experiment was performed in triplicate. Outputs were deemed confabulations if the LLM failed to recognize and reject the fabricated medication. The primary outcome was the rate of confabulations. Exact paired-permutation tests were used to compare confabulation rates across LLMs and prompting approaches. Results: Confabulation rates under standard conditions ranged from 2.7% to 99.6%. DrugGPT demonstrated the lowest baseline confabulation rates (2.7-6.4%), whereas other models demonstrated substantially higher rates (47.2-99.6%). Mitigation prompting significantly reduced confabulation rates compared to standard and deterministic decoding across all tasks and datasets (p < 0.001): GPT-5-Chat achieved confabulation rates of 0% with mitigation prompting from baseline rates of 66.4-88.5%. Deterministic decoding did not substantially impact confabulation rates. Conclusions: LLMs performing medication-related tasks are susceptible to confabulation in response to invalid or fabricated inputs, raising important safety concerns for clinical use. While mitigation prompting reduced confabulation rates, this study underscores the importance of robust safeguards when deploying LLMs in medication-related decision-making.
- Open Access
- 16 December 2025
- Published by Springer Nature in Biology of Sex Differences
- Vol. 17 (1)
Sex differences in cognitive decline and impairment: a scoping review in informatics literature
Objectives: A scoping review was conducted to investigate knowledge gaps in the informatics research literature regarding sex differences in cognitive decline and impairment, identifying existing studies and areas requiring further exploration. Methods and materials: Our scoping review follows the Preferred Reporting Items for Systematic reviews and Meta-Analysis Extension for Scoping Reviews (PRISMA – ScR) guidelines. We searched Ovid and other databases (APA PsychInfo, EMB Reviews, and Embase) for studies on sex differences in cognitive decline and impairment, focusing on peer-reviewed informatics journals and conference proceedings from 2000 to 2025. The selected manuscripts were analyzed based on metadata statistics, study attributes, and thematic content. Results: A total of 17 full articles met the inclusion criteria. Most studies were conducted in North America ( n = 7) and the European Union ( n = 5). More than half of the studies were published after 2020 ( n = 10). Our analyses highlight key aspects of selected studies, including bibliometric metadata, study attributes (e.g., study types, methods, and data sources), and thematic findings. Statistical modeling ( n = 8) and machine learning ( n = 4) are the most widely used study methods. Majority ( n = 11) of the publications are single-site studies, while the other multi-site collaborations ( n = 6) have emerged among hospitals, academic institutions, and research institutions. Discussion: Sex-specific disparities in cognitive decline and impairment remain a critical issue in healthcare. Most informatics research has primarily concentrated on identifying generic sex differences in cognitive decline and impairment progression, rather than exploring the complex underlying mechanisms such as observational studies with causal analysis. While these studies are valuable, they lack a holistic approach to understanding sex-specific disparities. Conclusion: There is a significant gap in using informatics to understand how biological, social, and behavioral factors contribute to sex-specific disparities in cognitive decline and impairment. This limitation underscores the need for more comprehensive informatics research that goes beyond mere identification to find the root cause of these disparities in healthcare.
- 16 December 2025
- Published by Springer Nature in Biology of Sex Differences
- Vol. 17 (1)
Open AccessSex differences in cognitive decline and impairment: a scoping review in informatics literature
Objectives: A scoping review was conducted to investigate knowledge gaps in the informatics research literature regarding sex differences in cognitive decline and impairment, identifying existing studies and areas requiring further exploration. Methods and materials: Our scoping review follows the Preferred Reporting Items for Systematic reviews and Meta-Analysis Extension for Scoping Reviews (PRISMA – ScR) guidelines. We searched Ovid and other databases (APA PsychInfo, EMB Reviews, and Embase) for studies on sex differences in cognitive decline and impairment, focusing on peer-reviewed informatics journals and conference proceedings from 2000 to 2025. The selected manuscripts were analyzed based on metadata statistics, study attributes, and thematic content. Results: A total of 17 full articles met the inclusion criteria. Most studies were conducted in North America ( n = 7) and the European Union ( n = 5). More than half of the studies were published after 2020 ( n = 10). Our analyses highlight key aspects of selected studies, including bibliometric metadata, study attributes (e.g., study types, methods, and data sources), and thematic findings. Statistical modeling ( n = 8) and machine learning ( n = 4) are the most widely used study methods. Majority ( n = 11) of the publications are single-site studies, while the other multi-site collaborations ( n = 6) have emerged among hospitals, academic institutions, and research institutions. Discussion: Sex-specific disparities in cognitive decline and impairment remain a critical issue in healthcare. Most informatics research has primarily concentrated on identifying generic sex differences in cognitive decline and impairment progression, rather than exploring the complex underlying mechanisms such as observational studies with causal analysis. While these studies are valuable, they lack a holistic approach to understanding sex-specific disparities. Conclusion: There is a significant gap in using informatics to understand how biological, social, and behavioral factors contribute to sex-specific disparities in cognitive decline and impairment. This limitation underscores the need for more comprehensive informatics research that goes beyond mere identification to find the root cause of these disparities in healthcare.
- Open Access
- 9 December 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 43449
Hypergraph clustering for analyzing chronic disease patterns in mild cognitive impairment reversion and progression
Identifying the sequential patterns of chronic conditions that precede the onset of mild cognitive impairment (MCI) is essential for understanding both the progression and the potential reversal of MCI. This study identifies common sequences of chronic conditions preceding MCI and introduces a novel, network-based clustering framework for characterizing patients with similar progression patterns linked to cognitive trajectories. We used the Mayo Clinic Study of Aging (MCSA) cohort and categorized participants of MCSA into two groups (i) stay at MCI or progressed to dementia, or (ii) reversion to normal within 5 years after the first onset of MCI. We curate the state transition network (patient level) for identifying and introduced hypergraph clustering (patient group level) to characterize participants with similar sequences. We identified generic key indicators (e.g., chronic kidney disease) and highlighted sex-specific potential indicators (e.g., arthritis, hypertension) associated with MCI reversal, opening new research directions to explore potential differences between males and females. There are certain ssequences of chronic conditions (e.g., originating from arthritis) that are more commonly observed in females. However, these observations warrant further validation. The proposed framework – hypergraph clustering – offers a promising method for uncovering similarities in patients through unique trajectories of chronic conditions that precede MCI.
- 9 December 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 43449
Open AccessHypergraph clustering for analyzing chronic disease patterns in mild cognitive impairment reversion and progression
Identifying the sequential patterns of chronic conditions that precede the onset of mild cognitive impairment (MCI) is essential for understanding both the progression and the potential reversal of MCI. This study identifies common sequences of chronic conditions preceding MCI and introduces a novel, network-based clustering framework for characterizing patients with similar progression patterns linked to cognitive trajectories. We used the Mayo Clinic Study of Aging (MCSA) cohort and categorized participants of MCSA into two groups (i) stay at MCI or progressed to dementia, or (ii) reversion to normal within 5 years after the first onset of MCI. We curate the state transition network (patient level) for identifying and introduced hypergraph clustering (patient group level) to characterize participants with similar sequences. We identified generic key indicators (e.g., chronic kidney disease) and highlighted sex-specific potential indicators (e.g., arthritis, hypertension) associated with MCI reversal, opening new research directions to explore potential differences between males and females. There are certain ssequences of chronic conditions (e.g., originating from arthritis) that are more commonly observed in females. However, these observations warrant further validation. The proposed framework – hypergraph clustering – offers a promising method for uncovering similarities in patients through unique trajectories of chronic conditions that precede MCI.
- Open Access
Drug-drug interaction identification using large language models
Background: The purpose of this study was to develop a series of benchmarks specifically for large language model (LLM) performance in the identification of drug-drug interactions (DDIs). Methods: We evaluated six LLMs using a clinician-designed dataset of 750 DDI scenarios. Tasks included: (1) a pointwise two-drug classification task (2) a pairwise three-drug discrimination task and (3) a listwise 4–6 drug selection task. Zero-shot prompting was applied and performance was assessed using precision, recall, F1 score, accuracy, and self-consistency. Results: Model performance varied by task structure and interaction severity. GPT-5-Chat demonstrated the highest recall and F1 score in the pointwise task (80.7% and 80.2%), whereas DrugGPT achieved superior accuracy in the pairwise (93.7%) and listwise task (92.6%). All models had improved performance with more severe compared to less severe DDIs (accuracy 72.0-99.2% vs. 58.3-86.7%). All models tended to have higher self-consistency in the pairwise and listwise experiments compared to the pointwise experiment (68.8-93.6% vs. 31.0-74.2%). No model exhibited uniformly high reliability across all tasks. Conclusions: Current LLMs show inconsistent capabilities in identifying DDIs. Performance degrades as reasoning space expands with limited stability. These findings emphasize the need for multi-format evaluation frameworks and reliability-aware assessment when considering LLMs for medication-safety applications.
Open AccessDrug-drug interaction identification using large language models
Background: The purpose of this study was to develop a series of benchmarks specifically for large language model (LLM) performance in the identification of drug-drug interactions (DDIs). Methods: We evaluated six LLMs using a clinician-designed dataset of 750 DDI scenarios. Tasks included: (1) a pointwise two-drug classification task (2) a pairwise three-drug discrimination task and (3) a listwise 4–6 drug selection task. Zero-shot prompting was applied and performance was assessed using precision, recall, F1 score, accuracy, and self-consistency. Results: Model performance varied by task structure and interaction severity. GPT-5-Chat demonstrated the highest recall and F1 score in the pointwise task (80.7% and 80.2%), whereas DrugGPT achieved superior accuracy in the pairwise (93.7%) and listwise task (92.6%). All models had improved performance with more severe compared to less severe DDIs (accuracy 72.0-99.2% vs. 58.3-86.7%). All models tended to have higher self-consistency in the pairwise and listwise experiments compared to the pointwise experiment (68.8-93.6% vs. 31.0-74.2%). No model exhibited uniformly high reliability across all tasks. Conclusions: Current LLMs show inconsistent capabilities in identifying DDIs. Performance degrades as reasoning space expands with limited stability. These findings emphasize the need for multi-format evaluation frameworks and reliability-aware assessment when considering LLMs for medication-safety applications.