A framework for employing longitudinally collected multicenter electronic health records to stratify heterogeneous patient populations on disease history

Abstract

Objective To facilitate patient disease subset and risk factor identification by constructing a pipeline which is generalizable, provides easily interpretable results, and allows replication by overcoming electronic health records (EHRs) batch effects. Material and Methods We used 1872 billing codes in EHRs of 102 880 patients from 12 healthcare systems. Using tools borrowed from single-cell omics, we mitigated center-specific batch effects and performed clustering to identify patients with highly similar medical history patterns across the various centers. Our visualization method (PheSpec) depicts the phenotypic profile of clusters, applies a novel filtering of noninformative codes (Ranked Scope Pervasion), and indicates the most distinguishing features. Results We observed 114 clinically meaningful profiles, for example, linking prostate hyperplasia with cancer and diabetes with cardiovascular problems and grouping pediatric developmental disorders. Our framework identified disease subsets, exemplified by 6 “other headache” clusters, where phenotypic profiles suggested different underlying mechanisms: migraine, convulsion, injury, eye problems, joint pain, and pituitary gland disorders. Phenotypic patterns replicated well, with high correlations of ≥0.75 to an average of 6 (2–8) of the 12 different cohorts, demonstrating the consistency with which our method discovers disease history profiles. Discussion Costly clinical research ventures should be based on solid hypotheses. We repurpose methods from single-cell omics to build these hypotheses from observational EHR data, distilling useful information from complex data. Conclusion We establish a generalizable pipeline for the identification and replication of clinically meaningful (sub)phenotypes from widely available high-dimensional billing codes. This approach overcomes datatype problems and produces comprehensive visualizations of validation-ready phenotypes.

Making causal inferences from transactional data: A narrative review of opportunities and challenges when implementing the target trial framework

“…60 Importantly, the use of transactional data for causal inference has facilitated the discovery of disease subgroups and analysis of treatment effect heterogeneity among these groups, a feature which is invaluable to the concept of precision medicine. [61][62][63][64][65] Furthermore, EHR provide a multifaceted trove of longitudinal and temporally rich data. 60 This temporal granularity is of paramount importance for causal inference, as it enables the tracking of individual patients over time, capturing dynamic changes in exposures, interventions, and outcomes.…”

From real-world electronic health record data to real-world results using artificial intelligence

“…clinical trials. [28][29][30] The identification of homogeneous disease subsets and trajectories within these large datasets can support research to disease aetiology and optimise treatment, particularly in the setting of complex heterogeneous diseases. Whether a model is trained in a supervised or unsupervised manner, accurate and generalisable results are important.…”

“…Unsupervised pattern recognition analyses identify subgroups of patient-patient similarity in a high dimensional or graph-based space. In rheumatology, they are most commonly employed for biological studies for instance to differentiate cell types in high-dimensional typing of blood and synovial biopsies, and are increasingly applied to clinical data from observational studies and post-hoc analyses of clinical trials 28–30. The identification of homogeneous disease subsets and trajectories within these large datasets can support research to disease aetiology and optimise treatment, particularly in the setting of complex heterogeneous diseases.…”

“…In rheumatology, they are most commonly employed for biological studies for instance to differentiate cell types in high-dimensional typing of blood and synovial biopsies, and are increasingly applied to clinical data from observational studies and post-hoc analyses of clinical trials. 28–30 The identification of homogeneous disease subsets and trajectories within these large datasets can support research to disease aetiology and optimise treatment, particularly in the setting of complex heterogeneous diseases. Whether a model is trained in a supervised or unsupervised manner, accurate and generalisable results are important.…”

Machine Learning Approaches for Electronic Health Records Phenotyping: A Methodical Review

“…In contrast to the previously discussed ML approaches, unsupervised learning is used for phenotype discovery, including identification of subphenotypes, [39,74,[120][121][122][123][124][125][126][127][128] co-occurring conditions, [69,129] and disease progression patterns. [68,[130][131][132][133][134] Among the 19 articles utilizing unsupervised learning, Latent Dirichlet Allocation (LDA) [69,124,125,127,133] and K-means were the most frequently used methods.…”