Clinical Research Databasing
For Clinical Research Databasing we offer access to the internationally recognized Research Electronic Data Capture (REDCap) system. REDCap is flexible, easy to use and our instance is hosted here at UTHSC. Please contact us to learn more about REDCap.
Cerner Health Facts:
CBMI is now able to provide UTHSC researchers and affiliates access to the CERNER Health Facts® (HF) database. Since 2000, CERNER Health Facts® database has captured and stored de-identified, longitudinal electronic health record (EHR) patient data, aggregated and organized to facilitate analyses and reporting – it currently contains data on almost 50 million patients and almost 300 million encounters. The database includes data on patient demographics, encounters, diagnoses, prescriptions, procedures, laboratory tests, locations of services/patients (e.g., clinic, ED, ICU, etc.) and hospital information, and billing. HF can be used to track a drug's or device's usage across diagnoses and major procedures, as well as by geographic region and hospital type, or researchers can use it to study practice patterns, treatments, and outcomes.
For cohort discovery and knowledge management we offer controlled access to a series of i2b2-based datamarts. These de-identified data sets are extracted from the medical through our partnership with Methodist LeBonheur Healthcare and represent a major step forward in bringing truly transformative clinical research to our community. To learn more about any of these secure resources contact us directly.
- Pediatric Research Database (PRD) - is funded by the LeBonheur CFRI and contains 466,498 patient encounters for 191,035 distinct patient records.
- Breast Cancer Registry - is funded by the Avon and Komen Foundations and contains screening and diagnostic mammography data augmented with comprehensive tumor staging data. The latest registry update contains over 56,000 patients seen at Methodist and The West Clinic.
- Diabetes Care & Wellness Coalition – is funded jointly by Methodist and the College of Medicine at UTHSC and contains information on diabetic patients treated at Methodist South.
Unity integrated data management system is a fully customizable web-service for data collection linked to an encrypted PostgreSQL database. Unity is accessed via the Internet though secure web-applications. Data can be entered manually into the interface or uploaded in bulk from excel files. For very large datasets and data migrations, the CBMI exploits tools such as Talend and Pentaho Kettle to automate extract-transform-load (ETL) protocols between source database and the Unity PostgreSQL backend.
Per NIH and PCORI expectations, the CBMI makes great effort to maintain semantic and syntactic interoperability between systems. The OMOP-CDM is used as a data standard when possible. For custom fields, the downloadable data dictionary contains rich metadata to facilitate post hoc variable translation. Unity is operating system agnostic and runs on most popular web browsers. Accessing the system requires only a web browser and internet access. Unity’s powerful reporting functions allow complex interrogation of datasets. Data can be bulk-exported or constrained to subsets before export to standard statistical analytics platforms for analysis (SAS, SPSS, Stata etc.).
Security and Provision of data sets
CBMI data security and privacy adheres to HIPAA, Common Rule, and HITECH standards. Access to identified and de-identified data is subject to IRB approval and institutional governance standards at all times and is governed and executed through an honest broker process: trained brokers follow standard operating procedures to assure appropriate approvals and security precautions are in place prior to delivering data sets to researchers.
- Data access is approved through established regulatory and data governance including IRB approval processes and privacy and security settings assuring HIPAA and FISMA compliance.
- Data within the rEDW is available in structured, semi-structured and un-structured formats and delivered into de-identified dimensional models for queries through pre-built data pipelines.