Skip to Content

Research platform optimizes use of electronic health data

A visualization created using the Knowledge Extraction via Sparse Embedding Regression Network, or KESER, tool available on CIPHER.
A sample knowledge extraction visualization created using the Knowledge Extraction via Sparse Embedding Regression Network, or KESER, tool available on CIPHER at https://phenomics.va.ornl.gov/web/cipher/vistools. Visualizations like this can help researchers understand, learn from, and utilize electronic health data.
By Ashley Galloway, associate director for strategic partnerships and outreach, CIPHER

Launched in June 2023, the public-facing website for VA’s Centralized Interactive Phenomics Resource, known as CIPHER, provides an online knowledge-sharing platform designed to improve the use of electronic health records, or EHRs, for research and clinical care.

“CIPHER has revolutionized the way researchers use EHR data in research by providing a standardized and streamlined platform to look up phenotypes, get a head start in health research, and help with portability across health systems,” said Dr. Sumitra Muralidhar, director of the VA Million Veteran Program, and lead sponsor for CIPHER.

The CIPHER library contains definitions of computable phenotypes – clinical conditions, diseases, or characteristics derived from health data – data mappings, programming code, and tools for data visualization. Open to both VA and non-VA researchers, the library can allow research findings to be shared more broadly within the scientific community. The goal of this universal access is to help accelerate research and facilitate the use of complex health data, leading to improved care for Veterans and others.

The platform serves as an online searchable library of EHR-based phenotypes which are de-identified to maintain patient privacy – VA has strong safeguards in place to protect Veteran privacy. Data sources can include health registries, claims data, physician notes and patient surveys.

VA is the largest integrated health care system in the United States, with more than 170 medical centers. its EHR system contains billions of records that include laboratory tests, diagnoses and medical procedures. EHR-based phenotypes can help standardize health definitions across different groups, research projects and hospital operations.

EHR-based phenotypes are used in research, quality improvement initiatives, and clinical care for patients. However, creating phenotypes can be time intensive due to variation in data sources and health systems. The process also requires expertise in the use of large, complex data sets. CIPHER helps overcome those challenges by optimizing health data sources and introducing standards for collecting phenotype definitions. To date, the CIPHER platform has cataloged more than 6,000 phenotypes using a phenotype collection standard that was developed in house.

CIPHER was established through a VA partnership with the U.S. Department of Energy’s Oak Ridge National Laboratory. The collaboration leverages the lab’s computational resources and expertise in large, complex datasets.

“We expect the CIPHER knowledgebase to improve the rigor and efficiency of research that uses code-based phenotypes. Contributing to CIPHER will also extend the impact of our work by making our phenotypes more readily available to other researchers,” said Dr. Maureen Dubreuil, VA Boston Healthcare System rheumatologist and associate professor of medicine at Boston University Chobanian & Avedisian School of Medicine.

CIPHER is accessible to the public at phenomics.va.ornl.gov. Details on the resources and standards that make CIPHER possible are in the Journal of the American Medical Informatics Association (CIPHER Resources, PMID: 38481028; CIPHER Metadata Standard, PMID: 36882092)