Linking Genomic and Clinical Data for Discovery and Personalized Care

Electronic Health Records (EHRs) are a powerful tool to improve health care quality while reducing its costs. As a longitudinal repository of patient diagnoses, treatments, and responses to treatment, EHRs are also being increasingly recognized as an important tool for research as well as clinical care. By coupling EHRs with DNA biobanks, EHRs can also provide clinical phenotypes for genomic studies. This chapter summarizes the role and needed methods to use EHR data for genomic discovery and its implementation into clinical care. There are a number of challenges to accurate interpretation and repurposing of EHR data for clinical and genomic research. Typically, investigators employ multimodal “phenotype algorithms” to find accurate cases and controls. Such algorithms integrate billing codes, medication records, laboratory and test result data, and clinical notes to achieve necessary recall and precision. Since much of the content in the clinical record is in unstructured (narrative) clinical documentation, use of natural language processing is often required. Despite these challenges, researchers have been successful in replicating known genetic associations and making new discoveries using EHR data. Early demonstration projects show that the EHR, coupled with advanced genome-enabled decision support, may be the ideal repository to operationalize genomic data for clinical use.