Electronic Medical Records for Genetic Research: Results of the eMERGE Consortium

By Abel N. Kho, J. Pacheco, P. Peissig, L. Rasmussen, K. Newton, Noah Weston, Paul K. Crane, Jyotishman Pathak, Christopher G. Chute, S. Bielinski and 5 more. Science Translational Medicine, 2011

351 citations recorded by the source corpus.

Clinical data captured in electronic medical records accurately identify cases and controls for genome-wide association studies. Where Electronic Records and Genomics Meet There has been a surge of interest in using electronic medical records in hospitals and clinics to capture information about patients that is normally buried in doctors’ handwritten notes. Indeed, the U.S. government has made the implementation of electronic medical records a priority area and has instigated standards for the recording and use of these records. The clinical data captured in electronic medical records including diagnoses, medical tests, and medications provide accurate clinical information that will improve patient care. With the ability to sequence the genomes of individuals faster and cheaper than ever before, it may be possible in the future to include the genome sequences of patients in their electronic medical records. A consortium called the Electronic Medical Records and Genomics Network (eMERGE) has set out to investigate whether clinical data captured in electronic medical records could be used to accurately identify patients with particular diseases for inclusion in genome-wide association studies (GWAS). GWAS scrutinize the genomes of individuals with particular diseases to identify tiny genetic variations that are associated with the risk of developing that disease. Here, the eMERGE consortium reports its study of the electronic medical records from five clinical centers and how accurately it identified patients with one of five diseases: dementia, cataracts, peripheral arterial disease, type 2 diabetes, and cardiac conduction defects. The investigators show that even though the electronic medical records were of different types and did not all use natural language processing to extract information from the records, they were able to obtain robust positive and negative values for identifying patients with these diseases with sufficient accuracy for use in GWAS. They conclude that widespread adoption of electronic medical records will provide real-world clinical data that will be valuable for GWAS and other types of genetic research. Clinical data in electronic medical records (EMRs) are a potential source of longitudinal clinical data for research. The Electronic Medical Records and Genomics Network (eMERGE) investigates whether data captured through routine clinical care using EMRs can identify disease phenotypes with sufficient positive and negative predictive values for use in genome-wide association studies (GWAS). Using data from five different sets of EMRs, we have identified five disease phenotypes with positive predictive values of 73 to 98% and negative predictive values of 98 to 100%. Most EMRs captured key information (diagnoses, medications, laboratory tests) used to define phenotypes in a structured format. We identified natural language processing as an important tool to improve case identification rates. Efforts and incentives to increase the implementation of interoperable EMRs will markedly improve the availability of clinical data for genomics research.

Subfields: Genetics, Molecular Biology

View Paper: https://www.semanticscholar.org/paper/a5d1d23c988c810c6862eec94f6dc06bb9123335

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Molecular Biology

Citations in the corpus

Extracting research-quality phenotypes from electronic health records to support precision medicine (2015, background): “In addition, the inclusion of longitudinal EHR data may lead to more accurate phenotype algorithms [39,49,50].”
Portability of an algorithm to identify rheumatoid arthritis in electronic health records (2012, methodology): “Modern NLP systems can incorporate these features to improve sensitivity and/or PPV of concept identification.(26) The original RA algorithm of Liao et al(7) used HITEx to find relevant disease names, medications, and laboratory results.”
Computational Phenotype Discovery Using Unsupervised Feature Learning over Noisy, Sparse, and Irregular Clinical Data (2013, methodology): “EMRs have previously been used in clinical [25–27] and genomic [28,29] research by leveraging expert domain knowledge to manually engineer phenotype specifications that identify clinical cohorts of interest.”
Rubik: Knowledge Guided Tensor Factorization and Completion for Health Data Analytics (2015, positive): “Successful pheno-typing eﬀorts on EHR data can enable many important applications, such as clinical predictive modeling [12, 27] and EHR-based genomic association studies [12, 17, 27].”
Clustering Longitudinal Clinical Marker Trajectories from Electronic Health Data: Applications to Phenotyping and Endotype Discovery (2015, background): “…to improve the effectiveness of clinical trials (Gundlapalli et al. 2008), drive the design of new genome-wide association studies (Kho et al. 2011; Kohane 2011), and allow medical scientists to view related diseases through a more fine-grained lens that can lead to insights that connect their…”
Developing Predictive Models Using Electronic Medical Records: Challenges and Pitfalls (2013, neutral): “Potential uses of these data include risk stratification or early prediction [1, 2], biomarker discovery [3], cohort detection for clinical trial recruitment [4], and optimizing resource use.”
Mining electronic health records: towards better research applications and clinical care (2012, background): “A high precision 65,66 of the approach has been documented and demonstrated in genome-wide association studies (GWASs) and other types of cohort research 67–74 .”
Genome- and Phenome-Wide Analysis of Cardiac Conduction Identifies Markers of Arrhythmia Risk (2013, background): “7,20 Similarly, the electronic phenotyping experience in eMERGE indicates that recurrent mentions of specific diagnoses or combinations of diverse data types (such as medications plus free text plus diagnostic codes) greatly improve diagnostic accuracy of electronic phenotyping algorithms when asses”