Dataset Information

Synthesize high-dimensional longitudinal electronic health records via hierarchical autoregressive language model.

ABSTRACT: Synthetic electronic health records (EHRs) that are both realistic and privacy-preserving offer alternatives to real EHRs for machine learning (ML) and statistical analysis. However, generating high-fidelity EHR data in its original, high-dimensional form poses challenges for existing methods. We propose Hierarchical Autoregressive Language mOdel (HALO) for generating longitudinal, high-dimensional EHR, which preserve the statistical properties of real EHRs and can train accurate ML models without privacy concerns. HALO generates a probability density function over medical codes, clinical visits, and patient records, allowing for generating realistic EHR data without requiring variable selection or aggregation. Extensive experiments demonstrated that HALO can generate high-fidelity data with high-dimensional disease code probabilities closely mirroring (above 0.9 R² correlation) real EHR data. HALO also enhances the accuracy of predictive modeling and enables downstream ML models to attain similar accuracy as models trained on genuine data.

SUBMITTER: Theodorou B

PROVIDER: S-EPMC10471716 | biostudies-literature | 2023 Aug

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Synthesize high-dimensional longitudinal electronic health records via hierarchical autoregressive language model.

Theodorou Brandon B Xiao Cao C Sun Jimeng J

Nature communications 20230831 1

Synthetic electronic health records (EHRs) that are both realistic and privacy-preserving offer alternatives to real EHRs for machine learning (ML) and statistical analysis. However, generating high-fidelity EHR data in its original, high-dimensional form poses challenges for existing methods. We propose Hierarchical Autoregressive Language mOdel (HALO) for generating longitudinal, high-dimensional EHR, which preserve the statistical properties of real EHRs and can train accurate ML models witho ...[more]

PMID: 37652934

Similar Datasets

Project description:Fundamental to elucidating the functional organization of the brain is the assessment of causal interactions between different brain regions. Multivariate autoregressive (MVAR) modeling techniques applied to multisite electrophysiological recordings are a promising avenue for identifying such causal links. They estimate the degree to which past activity in one or more brain regions is predictive of another region's present activity, while simultaneously accounting for the mediating effects of other regions. Including as many mediating variables as possible in the model has the benefit of drastically reducing the odds of detecting spurious causal connectivity. However, effective bounds on the number of MVAR model coefficients that can be estimated reliably from limited data make exploiting the potential of MVAR models challenging for even modest numbers of recording sites. Here, we utilize well-established dimensionality-reduction techniques to fit MVAR models to human intracranial data from ∼100 - 200 recording sites spanning dozens of anatomically and functionally distinct cortical regions. First, we show that high-dimensional MVAR models can be successfully estimated from long segments of data and yield plausible connectivity profiles. Next, we use these models to generate synthetic data with known ground-truth connectivity to explore the utility of applying principal component analysis and group least absolute shrinkage and selection operator (gLASSO) to reduce the number of parameters (connections) during model fitting to shorter data segments. We show that gLASSO is highly effective for recovering ground-truth connectivity in the limited data regime, capturing important features of connectivity for high-dimensional models with as little as 10 seconds of data. The methods presented here have broad applicability to the analysis of high-dimensional time series data in neuroscience, facilitating the elucidation of the neural basis of sensation, cognition, and arousal.

Project description:Current studies regarding the secondary use of electronic health records (EHR) predominantly rely on domain expertise and existing medical knowledge. Though significant efforts have been devoted to investigating the application of machine learning algorithms in the EHR, efficient and powerful representation of patients is needed to unleash the potential of discovering new medical patterns underlying the EHR. Here, we present an unsupervised method for embedding high-dimensional EHR data at the patient level, aimed at characterizing patient heterogeneity in complex diseases and identifying new disease patterns associated with clinical outcome disparities. Inspired by the architecture of modern language models-specifically transformers with attention mechanisms, we use patient diagnosis and procedure codes as vocabularies and treat each patient as a sentence to perform the patient embedding. We applied this approach to 34,851 unique medical codes across 1,046,649 longitudinal patient events, including 102,739 patients from the electronic Medical Records and GEnomics (eMERGE) Network. The resulting patient vectors demonstrated excellent performance in predicting future disease events (median AUROC = 0.87 within one year) and bulk phenotyping (median AUROC = 0.84). We then illustrated the utility of these patient vectors in revealing heterogeneous comorbidity patterns, exemplified by disease subtypes in colorectal cancer and systemic lupus erythematosus, and capturing distinct longitudinal disease trajectories. External validation using EHR data from the University of Washington confirmed robust model performance, with median AUROCs of 0.83 and 0.84 for bulk phenotyping tasks and disease onset prediction, respectively. Importantly, the model reproduced the clustering results of disease subtypes identified in the eMERGE cohort and uncovered variations in overall mortality among these subtypes. Together, these results underscore the potential of representation learning in EHRs to enhance patient characterization and associated clinical outcomes, thereby advancing disease forecasting and facilitating personalized medicine.

Project description:BackgroundPrediction models have demonstrated a range of applications across medicine, including using electronic health record (EHR) data to identify hospital readmission and mortality risk. Large language models (LLMs) can transform unstructured EHR text into structured features, which can then be integrated into statistical prediction models, ensuring that the results are both clinically meaningful and interpretable.ObjectiveThis study aims to compare the classification decisions made by clinical experts with those generated by a state-of-the-art LLM, using terms extracted from a large EHR data set of individuals with mental health disorders seen in emergency departments (EDs).MethodsUsing a dataset from the EHR systems of more than 50 health care provider organizations in the United States from 2016 to 2021, we extracted all clinical terms that appeared in at least 1000 records of individuals admitted to the ED for a mental health-related problem from a source population of over 6 million ED episodes. Two experienced mental health clinicians (one medically trained psychiatrist and one clinical psychologist) reached consensus on the classification of EHR terms and diagnostic codes into categories. We evaluated an LLM's agreement with clinical judgment across three classification tasks as follows: (1) classify terms into "mental health" or "physical health", (2) classify mental health terms into 1 of 42 prespecified categories, and (3) classify physical health terms into 1 of 19 prespecified broad categories.ResultsThere was high agreement between the LLM and clinical experts when categorizing 4553 terms as "mental health" or "physical health" (κ=0.77, 95% CI 0.75-0.80). However, there was still considerable variability in LLM-clinician agreement on the classification of mental health terms (κ=0.62, 95% CI 0.59-0.66) and physical health terms (κ=0.69, 95% CI 0.67-0.70).ConclusionsThe LLM displayed high agreement with clinical experts when classifying EHR terms into certain mental health or physical health term categories. However, agreement with clinical experts varied considerably within both sets of mental and physical health term categories. Importantly, the use of LLMs presents an alternative to manual human coding, presenting great potential to create interpretable features for prediction models.

Dataset Information

Synthesize high-dimensional longitudinal electronic health records via hierarchical autoregressive language model.

Publications

Synthesize high-dimensional longitudinal electronic health records via hierarchical autoregressive language model.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets