Case Study: OMOP common data model
Overview
Long-term overall survival is often estimated using clinical trial information. However, these estimates can involve considerable uncertainty because they rely on future projections based only on observed trial data. Real-world evidence, including hospital and medical records, can help reduce this uncertainty and improve understanding of treatment patterns and patient outcomes.
Comparing treatment and outcomes between hospitals is an important way of ensuring good standards of care. However, undertaking studies that compare outcomes between hospitals, particularly across countries, presents challenges because of data privacy concerns.
To address this challenge, the Observational Health Data Sciences and Informatics (OHDSI) network developed the Observational Medical Outcomes Partnership (OMOP) Common Data Model. The OMOP model allows hospitals to standardise their data locally. Once hospitals have formatted their data to this common structure, the same analysis can be run at each site and the resulting summaries can be shared and compared. In this approach, patient-level data does not leave the hospitals and only aggregated results are shared after appropriate disclosure checks.
Edinburgh Cancer Informatics participates in a partnership of universities and research organisations across Europe that are applying the OMOP standard to examine cancer survival rates from the date of diagnosis. Using this approach, researchers can evaluate the incidence and prevalence of several cancers and assess survival outcomes across multiple healthcare systems.
The project analyses the occurrence and survival rates of specific cancers to assess how effectively mathematical models can forecast cancer survival at 1, 5, and 10 years following initial diagnosis using real-world data. Researchers are studying mortality from all causes across several cancer types including breast, colon, lung, liver, prostate, head and neck, pancreatic, oesophagus and stomach cancers.
Participation in this work also allows NHS Lothian to test its recently mapped breast cancer OMOP dataset and local processes needed to support federated analysis within a secure environment.
Who is involved?
The project is led from the UK by the National Institute for Health and Care Excellence (NICE).
Partners include the University of Oxford, the University of Edinburgh, Edinburgh Cancer Informatics, IDIAP Jordi Gol, SIDiAP, the Netherlands Comprehensive Cancer Organisation, the Cancer Registry of Norway, Helsinki University Hospital, Hospital Universitario Virgen Macarena, Université de Genève, Institut d’Assistència Sanitària, ULSM, ULS Região de Aveiro, ULSGE, PCi, University Medical Center Rotterdam and the University of Tartu.
The work is supported through the European Health Data and Evidence Network (EHDEN) and the Oxford NIHR Biomedical Research Centre.
DataLoch provides the secure research environment that supports NHS Lothian’s participation in the analysis.
More information
Standardized Data: The OMOP Common Data Model
Image credit: Neuronal synapse, artwork. Stephen Magrath. Source: Wellcome Collection.