Chuan Hong

Has grant

Duke University
Country flagUnited States
ApplyKite TwinLearns as you apply

Make your next move.

Prepare your application

Review drafts before you send them.

Explore related searches

Recent Grants

Grant: Close

Evaluation of Cardiovascular Risk Prediction Equations across Diverse Sociodemographic Subgroups

Open Date: 2024-04-01

Close Date: 2025-03-31

Articles (8)

BACKGROUND Electronic health record (EHR) based phenotyping algorithms are typically trained and/or validated using a small set of gold-standard labels manually annotated via medical chart review by domain experts. To reduce the time and labor cost, silver-standard labels such as self-reported outcomes that are highly predictive of the true disease statuses have been used as a proxy of gold standard labels, which are unfortunately subject to misclassification due to insufficient documentation or human error. OBJECTIVE Ignoring such labeling errors may lead to biases in both the estimated classification model and its accuracy for predicting the true underlying phenotype status. The objective of this study is to develop a calibration algorithm for both phenotyping classification and downstream association analysis using noisy silver-standard labels, where the true gold-standard disease status is not observable. METHODS In this paper, we propose an imperfectly supervised calibrated algorithm for phenotyping and regression (SCAPER) that simultaneously produces calibrated phenotyping classifications and association regression modeling by utilizing a small number of noisy silver-standard labels and a large set of unlabeled observations on predictive features and biomarkers (possibly genetic data). The proposed approach yields an improved predicted probability of phenotype for each patient, a threshold for classifying participants with phenotype yes/no, and bias corrected regression coefficient for predictors in the downstream association study. The algorithm was validated by both phenotyping and genetic association studies for rheumatoid arthritis from a large tertiary care center, and that for type II diabetes from two large tertiary care centers. RESULTS When validating against the gold-standard labels for phenotyping performance, the proposed algorithm achieved higher AUC compared to the International Classification of Diseases (ICD) codes and existing unsupervised phenotyping algorithms. The downstream association study suggests that the proposed approach detected previously validated associations with higher power when compared to the standard association studies based on ICD codes. CONCLUSIONS The proposed unsupervised calibration increased the accuracy of phenotype definition and corrected the biases in the downstream association studies such as genetic association studies.

Year:

2025

Background Integrating data is essential for advancing clinical and epidemiological research. However, because datasets often describe variables (eg, demographic and health conditions) in diverse ways, the process of integrating and harmonizing variables from research studies remains a major bottleneck. Objective The objective was to assess a natural language processing–based method to automate variable harmonization to achieve a scalable approach to integration of multiple datasets. Methods We developed a fully connected neural network (FCN) method, enhanced with contrastive learning, using domain-specific embeddings from the Bidirectional Encoder Representations from Transformers for Biomedical Text Mining language representation model, using 3 cardiovascular datasets: the Atherosclerosis Risk in Communities study, the Framingham Heart Study, and the Multi-Ethnic Study of Atherosclerosis. We used metadata variable descriptions and curated harmonized concepts as ground truth. We framed the problem as a paired sentence classification task. The accuracy of this method was compared with a logistic regression baseline method. To assess the generalizability of the trained models, we also evaluated their performance by separating the 3 datasets when preparing the training and validation sets. Results The newly developed FCN achieved a top-5 accuracy of 98.95% (95% CI 98.31%‐99.47%) and an area under the receiver operating characteristic (AUC) of 0.99 (95% CI 0.98‐0.99), outperforming the standard logistic regression model, which exhibited a top-5 accuracy of 22.23% (95% CI 19.91%‐24.87%) and an AUC of 0.82 (95% CI 0.81‐0.83). The contrastive learning enhancement also outperformed the logistic regression model, although slightly below the base FCN model, exhibiting a top-5 accuracy of 89.88% (95% CI 87.88%‐91.68%) and an AUC of 0.98 (95% CI 0.97‐0.98). Conclusions This novel approach provides a scalable solution for harmonizing metadata across large-scale cohort studies. The proposed method significantly enhances the performance over the baseline method by using learned representations to categorize harmonized concepts more accurately for cohorts in cardiovascular disease and stroke.

Year:

2025

BACKGROUND Integrating data is essential for advancing clinical and epidemiological research. However, the process of integrating and harmonizing variables from research studies remains a major bottleneck. OBJECTIVE The objective was to assess a natural language processing (NLP)-based method to automate variable harmonization to achieve a scalable approach to integration of multiple datasets. METHODS We developed a fully connected neural network method, enhanced with contrastive learning, using domain-specific embeddings from BioBERT using three cardiovascular datasets: Atherosclerosis Risk in Communities (ARIC) study, the Framingham Heart Study (FHS) and the Multi-Ethnic Study of Atherosclersosis (MESA) . We used meta-data variable descriptions and curated harmonized concepts as ground truth. The accuracy of this method was compared to a logistic regression baseline method. RESULTS The newly developed fully connected neural network achieved a top-5 accuracy of 98.95% (95% CI: 98.31%-99.47%) and an AUC of 0.990 (95% CI: 0.988-0.991), outperforming the accuracy of to the standard logistic regression model. CONCLUSIONS This novel approach provides a scalable solution for harmonizing meta-data across large-scale cohort studies.

Year:

2025

BACKGROUND The current landscape of Emergency Care (EC) is marked by high demand leading to issues such as Emergency Department boarding, overcrowding and subsequent delays that impact the quality and safety of patient care. Integrating data science into EC can enhance decision-making with predictive, preventative, personalized, and participatory approaches. However, gaps in adherence to fairness, accountability, interpretability, and responsibility are evident, particularly due to barriers in data-sharing, which often result in a lack of transparency and robust oversight in these applications. OBJECTIVE The Fair, Accountable, Interpretable and Responsible (FAIR)-EC collaboration adapts the existing FAIR principles to address emerging challenges as data science integrates with EC. This initiative aims to transform EC by establishing ethical artificial intelligence (AI) standards specifically tailored for this integration. By bridging the gap between EC professionals, data scientists and other stakeholders, the collaboration promotes international cooperation that leverages advanced data science techniques to enhance EC outcomes across different care settings. METHODS We propose a federated research design that enables analyses of extensive datasets from various global institutions without compromising patient privacy. This approach transforms epidemiological research with advanced data science techniques, emphasizing the harmonization of data for comprehensive analyses across different healthcare systems. RESULTS The FAIR-EC initiative has facilitated the collection and analysis of datasets from diverse geographical regions, enabling the examination of regional variations in EC practices. Initial projects have demonstrated promising outcomes, including the successful development of a federated scoring system and the adaptation of association studies and predictive models across various regions. These efforts highlight the feasibility of leveraging advanced data science techniques to address the complexities of EC while preserving patient privacy. CONCLUSIONS FAIR-EC integrates data science ethically and effectively into EC, addressing challenges like fragmented data, real-time handoffs, and public health crises. Its federated design harmonizes diverse data streams while preserving privacy, and its emphasis on ethical AI aligns with the dynamic nature of EC. Despite challenges in data variability and system complexity, FAIR-EC establishes a strong foundation for innovation in global EC.

Year:

2025

Importance Early identification of fever or hypothermia is crucial for diagnosing sepsis. Despite their increased use across healthcare systems, concerns have been raised about the accuracy of temporal thermometers among Black patients. Objective To study the performance of temporal thermometry across race and ethnicity, and its impact on the initiation of the sepsis management bundle (SEP-1). Design In this retrospective cohort study, records from 2008–2024 in 123 U.S. hospitals were analyzed, including electronic health records from Duke, MIMIC-IV, eICU-CRD-1, and eICU-CRD-2. Patients were included if they had a temporal measurement within one hour of an oral/core measurement (esophageal, bladder, rectal). Main Outcomes and Measures Hidden Fever was defined when the temporal thermometer read ≤ 38°C but oral/core measurement > 38°C and Hidden Hypothermia as temporal temperature ≥ 36°C but oral/core < 36°C. The primary outcome was Hidden Hyper/Hypothermia (HHH) when either Hidden Fever or Hidden Hypothermia happened. Observed Hyper / Hypothermia (OHH) corresponded to an agreement between both measurements, used as a control. To study the impact of temporal thermometry underperformance on SEP-1 bundle initiation, we assessed time to culture ordering, antibiotics administration, and lactate measurement. A composite of these three, whichever happened first, was defined as our secondary outcome. Results 6,921 paired temperature measurements were studied for the primary outcome analysis, corresponding to 4,248 hospitalizations (2.6% Asian; 12.4% Black; 2.6% Hispanic/Latino; 83.4% White; 44.9% female; 7.7% with HHH) from 115 hospitals. After adjusting for confounders, Black [OR (95% CI): 1.760 (1.219, 2.541), p =0.003] and Hispanic/Latino [OR (95% CI): 2.183 (1.226, 3.888), p= 0.008] patients were significantly more likely to present with HHH than White patients. For the study of the secondary outcome, 434 patients had either OHH or HHH. Compared to patients with OHH, patients with HHH had significantly delayed secondary outcomes, with log-rank p-value of 0.002. Conclusions and Relevance Solely relying on temporal thermometry can lead to missed hyper/hypothermia events, especially among Asian, Black, and Hispanic patients. This has been associated with delays in the initiation of the SEP-1 bundle, which may hamper health systems’ compliance with reimbursement programs. Key Points Question Are there racial differences in temporal thermometry that result in missed hyper/hypothermia, and are these associated with delays in sepsis bundle care (SEP-1) initiation? Findings In this multicenter retrospective cohort study of 4,248 hospitalizations, Asian, Black, and Hispanic patients were more likely to have hidden hyper/hypothermia when using temporal thermometers, as compared to oral/core measurements. These missed events were associated with delayed SEP-1 interventions (three out of four were studied: culture orders, antibiotic administration, and lactate measurements). Meaning Reliance on temporal thermometry could exacerbate health disparities and hamper hospital systems’ compliance with reimbursement programs. Temporal thermometers should therefore be routinely validated with contact thermometers.

Year:

2025

BACKGROUND Developing computable phenotypes (CP) based on electronic health records (EHR) data requires "gold-standard" labels of patient charts obtained from clinicians. Charts are most often sampled randomly, but random sampling may fail to capture the diversity of a given patient population, which may lead to bias of the CP. OBJECTIVE We proposed an unsupervised sampling approach designed to better capture a diverse patient cohort and improve the information coverage of chart review samples. METHODS Our coverage sampling method utilizes clustering and stratified sampling to ensure diverse representation in chart review samples. We use simulations and a real-world data example to compare the performance of our method with random sampling. The performance of the samples was evaluated based on the information coverage and area under the receiver operator characteristic curve (AUROC). RESULTS Our simulation studies demonstrate that our unsupervised approach provided better coverage of patient populations and equal or improved CP performance compared to random samples, especially in scenarios where minority sub-groups were present. In the real-world application, the method also outperformed random sampling, yielding more representative samples and enhancing CP performance. CONCLUSIONS The proposed coverage sampling method enhances the coverage of chart review samples, leading to the development of CPs that can capture outcomes of interest in a diverse patient population. This approach is particularly beneficial in cohorts with complex or minority sub-groups, providing a robust alternative to random sampling in EHR-based research.

Year:

2025

BACKGROUND Medical texts present significant domain-specific challenges, and manually curating these texts is a time-consuming and labor-intensive process. To address this, natural language processing (NLP) algorithms have been developed to automate text processing. In the biomedical field, various toolkits for text processing exist, which have greatly improved the efficiency of handling unstructured text. However, these existing toolkits tend to emphasize different perspectives, and none of them offer generation capabilities, leaving a significant gap in the current offerings. OBJECTIVE This study aims to describe the development and preliminary evaluation of Ascle. Ascle is tailored for biomedical researchers and clinical staff with an easy-to-use, all-in-one solution that requires minimal programming expertise. For the first time, Ascle provides 4 advanced and challenging generative functions: question-answering, text summarization, text simplification, and machine translation. In addition, Ascle integrates 12 essential NLP functions, along with query and search capabilities for clinical databases. METHODS We fine-tuned 32 domain-specific language models and evaluated them thoroughly on 27 established benchmarks. In addition, for the question-answering task, we developed a retrieval-augmented generation (RAG) framework for large language models that incorporated a medical knowledge graph with ranking techniques to enhance the reliability of generated answers. Additionally, we conducted a physician validation to assess the quality of generated content beyond automated metrics. RESULTS The fine-tuned models and RAG framework consistently enhanced text generation tasks. For example, the fine-tuned models improved the machine translation task by 20.27 in terms of BLEU score. In the question-answering task, the RAG framework raised the ROUGE-L score by 18% over the vanilla models. Physician validation of generated answers showed high scores for readability (4.95/5) and relevancy (4.43/5), with a lower score for accuracy (3.90/5) and completeness (3.31/5). CONCLUSIONS This study introduces the development and evaluation of Ascle, a user-friendly NLP toolkit designed for medical text generation. All code is publicly available through the Ascle GitHub repository. All fine-tuned language models can be accessed through Hugging Face.

Year:

2024

Background Post–COVID-19 condition (colloquially known as “long COVID-19”) characterized as postacute sequelae of SARS-CoV-2 has no universal clinical case definition. Recent efforts have focused on understanding long COVID-19 symptoms, and electronic health record (EHR) data provide a unique resource for understanding this condition. The introduction of the International Classification of Diseases, Tenth Revision (ICD-10) code U09.9 for “Post COVID-19 condition, unspecified” to identify patients with long COVID-19 has provided a method of evaluating this condition in EHRs; however, the accuracy of this code is unclear. Objective This study aimed to characterize the utility and accuracy of the U09.9 code across 3 health care systems—the Veterans Health Administration, the Beth Israel Deaconess Medical Center, and the University of Pittsburgh Medical Center—against patients identified with long COVID-19 via a chart review by operationalizing the World Health Organization (WHO) and Centers for Disease Control and Prevention (CDC) definitions. Methods Patients who were COVID-19 positive with either a U07.1 ICD-10 code or positive polymerase chain reaction test within these health care systems were identified for chart review. Among this cohort, we sampled patients based on two approaches: (1) with a U09.9 code and (2) without a U09.9 code but with a new onset long COVID-19–related ICD-10 code, which allows us to assess the sensitivity of the U09.9 code. To operationalize the long COVID-19 definition based on health agency guidelines, symptoms were grouped into a “core” cluster of 11 commonly reported symptoms among patients with long COVID-19 and an extended cluster that captured all other symptoms by disease domain. Patients having ≥2 symptoms persisting for ≥60 days that were new onset after their COVID-19 infection, with ≥1 symptom in the core cluster, were labeled as having long COVID-19 per chart review. The code’s performance was compared across 3 health care systems and across different time periods of the pandemic. Results Overall, 900 patient charts were reviewed across 3 health care systems. The prevalence of long COVID-19 among the cohort with the U09.9 ICD-10 code based on the operationalized WHO definition was between 23.2% and 62.4% across these health care systems. We also evaluated a less stringent version of the WHO definition and the CDC definition and observed an increase in the prevalence of long COVID-19 at all 3 health care systems. Conclusions This is one of the first studies to evaluate the U09.9 code against a clinical case definition for long COVID-19, as well as the first to apply this definition to EHR data using a chart review approach on a nationwide cohort across multiple health care systems. This chart review approach can be implemented at other EHR systems to further evaluate the utility and performance of the U09.9 code.

Year:

2024

Collaborators (21)

Katherine Liao

Associate Professor

Harvard Medical School

UNITED STATES
View details

Mahmoud Alwakeel

Duke University

UNITED STATES
View details

Tiarnan Keenan

National Institutes of Health

UNITED STATES
View details

Mengling Feng

National University of Singapore (NUS)

SINGAPORE
View details

Yu Rang Park

Yonsei University College of Medicine

SOUTH KOREA
View details

Irene Li

University of Tokyo

JAPAN
View details

Shyam Visweswaran

Chair & Professor

University of Pittsburgh

UNITED STATES
View details

Malarkodi Jebathilagam Samayamuthu

Senior Research Scientist

University of Pittsburgh

UNITED STATES
View details

Matthew Engelhard

Assistant Professor

Duke University

UNITED STATES
View details

Qingyu Chen

Assistant Professor

National Institutes of Health

UNITED STATES
View details

Emily Chew

NIH Distinguished Investigator

-

UNITED STATES
View details

Hua Xu

Yale University

UNITED STATES
View details

Zongqi Xia

Principal Investigator

University of Pittsburgh

UNITED STATES
View details

Bibhas Chakraborty

National University of Singapore (NUS)

SINGAPORE
View details

Michael J. Pencina

Duke University School of Medicine

UNITED STATES
View details

Dragomir Radev

Professor

Yale University

UNITED STATES
View details

Nan Liu

Associate Professor

National University of Singapore (NUS)

SINGAPORE
View details

Ricardo Henao

Associate Professor

Duke University

UNITED STATES
View details

Tingting Zhu

Associate Professor in AI for Digital Health

University of Oxford

UNITED KINGDOM
View details

Yilin Ning

Assistant Professor

Duke-NUS Medical School

SINGAPORE
View details

Kendall Ho

The University of British Columbia

CANADA
View details