Home

Phenoflow is an electronic health record phenotype library that stores definitions under a standard model, and allows computable forms of those definitions to be downloaded. These can then be executed against local datasets.

Although predominantly designed to support research into next-generation phenotype libraries, Phenoflow holds real definitions and is thus also designed to support both researchers and clinicians in analysing patient datasets.

Specifically, Phenoflow offers the ability to:

• Browse a collection of Phenotypes parsed from popular data sources (e.g. the HDR UK phenotype library), or authored directly, and view them in a standard form.
• Immediately download phenotypes as computable workflows, which can be executed with minimal dependencies against local data sources.
• Customise these workflows prior to download
  • Choose a connector based on the target data source standard (e.g., OMOP CDM, FHIR), allowing the same logic to be executed against a variety of data sources.
  • Author a new connector, for proprietary data standards.
  • Customise the realisation of the phenotype logic itself, selecting languages that are most suitable for further development.
• Upload specifications for new definitions, to be hosted on the site.
• Create persistent identifiers for phenotypes, allowing them to be referenced in published works.

Phenoflow aims to ensure phenotype definitions are easy to execute (portable), capture the logic intended by the definition author(s) accurately (reproducible) and are validated.

The Phenoflow library is realised by a set of microservices, which perform a number of different functions, including serving the library web application itself. The source code for these services is contained within this repository.