Skip to main content Skip to secondary navigation

STARR supports a number of data models including Epic Clarity data model, a Stanford data model, OMOP and PEDSNet Common Data Models. Fitness of the data model depends on what you are trying to do.

Data models

Main content start

In the context of health data, the data model is the schema representing the patient data in a clinical data warehouse.  STARR supports a number of data models as mentioned below.

  • Epic Clarity data model: This is our raw data from the two hospitals. There are two separate Clarity data models - they are similar but not identical. The data model is complex and there are over 10,000 tables. There is lack of strong type e.g., many numerical data fields are captured as text fields. Nearly a 1000 tables have data that is frequently used for research. You can access Clarity data via data consultation
  • Stanford in-house data model: When STARR Tools (fka STRIDE) first launched in 2008, Common Data Models didn't exist. This first generation research Clinical Data Warehouse (r-CDW) developed an in-house data model. It is specific to Stanford, has been refined over the years to meet the needs of Stanford community.  The typical use case for research using the STRIDE data model is chart review, and prospective studies. You can access this data via STARR Tools self-service.
  • OMOP Common Data Model (CDM): In 2019, STARR launched its next generation r-CDW using Observational Medical Outcomes Partnership (OMOP) CDM from OHDSI consortium. This data model is particularly suitable for advanced electronic phenotyping due to richness of the vocabulary and high quality of Stanford mapping to standard concepts. Our OMOP database is used for CTSA CLIC data quality reporting. The OHDSI analytical tools also support collaborative and reproducible research, allowing Stanford to participate in a large number of network studies. Furthermore, STARR-OMOP is specifically suited to researchers interested in Big Data analytics and method development such as Natural Language Processing. You can access a fully de-identified OMOP r-CDW via self-service or request eProtocol approved data via data consultation
  • PEDSNet Data Model: The latest data model in the STARR portfolio is a PEDSNet. Similar in spirit to OMOP, the PEDSNet is optimized for pediatric research. This data has limited access right now. Please request a data consultation if you want to learn more.