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

Journal of the American Medical Informatics Association, vol. 29, pp. 761–769

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.

Authors 14

1. Marc P. Maurits corresponding Aachen Department of Rheumatology, Leiden University Medical Center Leiden Computational Biology Center

   Leiden University Medical Center

   Affiliation as printed

   Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands

   Leiden Computational Biology Center, Leiden University Medical Center, Leiden, The Netherlands

2. Ilya Korsunsky

   Brigham and Women's Hospital · Harvard University

   Affiliation as printed

   Center for Data Sciences, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA

3. Soumya Raychaudhuri

   Brigham and Women's Hospital · Harvard University

   Affiliation as printed

   Center for Data Sciences, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA

4. Shawn N. Murphy

   Mass General Brigham

   Affiliation as printed

   Research Information Science and Computing, Mass General Brigham, Boston, MA, USA

5. Jordan W. Smoller

   Massachusetts General Hospital

   Affiliation as printed

   Center for Precision Psychiatry, Department of Psychiatry, Massachusetts General Hospital, Boston, MA, USA

   Psychiatric and Neurodevelopmental Genetics Unit, Center for Genomic Medicine, Massachusetts General Hospital, Boston, MA, USA

6. Scott T. Weiss

   Brigham and Women's Hospital · Harvard University

   Affiliation as printed

   Channing Division of Network Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA

7. Lynn Petukhova

   Columbia University Irving Medical Center

   Affiliation as printed

   Lynn M. Petukhova, Department of Dermatology at NewYork-Presbyterian/Columbia University Medical Center (CUMC)

8. Chunhua Weng

   Columbia University

   Affiliation as printed

   Chunhua Weng, Biomedical Informatics - Columbia University

9. Wei‐Qi Wei

   Vanderbilt University

   Affiliation as printed

   Wei-Qi Wei, Biomedical Informatics in the School of Medicine at Vanderbilt University Wei

10. Tom W. J. Huizinga Aachen Department of Rheumatology, Leiden University Medical Center

    Leiden University Medical Center

    Affiliation as printed

    Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands

11. Marcel J. T. Reinders Aachen Leiden Computational Biology Center

    Leiden University Medical Center · Delft University of Technology

    Affiliation as printed

    Leiden Computational Biology Center, Leiden University Medical Center, Leiden, The Netherlands

    The Delft Bioinformatics Lab, Delft University of Technology, Delft, The Netherlands

12. Elizabeth W. Karlson

    Brigham and Women's Hospital · Harvard University

    Affiliation as printed

    Division of Rheumatology, Inflammation and Immunity, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA

13. Erik B. van den Akker Aachen Leiden Computational Biology Center Section of Molecular Epidemiology

    Leiden University Medical Center

    Affiliation as printed

    Leiden Computational Biology Center, Leiden University Medical Center, Leiden, The Netherlands

    Section of Molecular Epidemiology, Leiden University Medical Center, Leiden, The Netherlands

14. Rachel Knevel Aachen Department of Rheumatology, Leiden University Medical Center

    Brigham and Women's Hospital · Harvard University · Leiden University Medical Center

    Affiliation as printed

    Department of Rheumatology, Leiden University Medical Center, Leiden, The Netherlands

    Division of Rheumatology, Inflammation and Immunity, Brigham and Women's Hospital, Harvard Medical School, Boston, MA, USA