Distributed Network and Common Data Model
Distributed Network
The BEST Initiative infrastructure is constructed as a distributed network of data providers and scientific collaborators. BEST utilizes multiple data sources, such as administrative claims data collected by private insurance companies for billing purposes, electronic health records (EHRs) collected by health care facilities for both clinical care and billing purposes, and linked claims-EHR databases generated for research and surveillance activities. Multiple organizations contribute their data to the BEST Initiative for surveillance activities. However, to protect the data and increase security and privacy, the data providers retain their data behind their local firewalls without FDA direct access to the data. When surveillance activities are performed, FDA receives aggregated results with individual identifiers removed.
BEST scientific collaborators also come from a large network of academia, health care, health insurance, and pharmaceutical industries, and other government agencies throughout the country.
Common Data Model
Since the BEST Initiative infrastructure is assembled as a distributed network, different data providers offer their data in different structures and formats comprised of different data elements. Therefore, all the databases must be harmonized by converting them into the same format and using the same vocabulary and nomenclature, namely transformed into a common data model (CDM).
| Administrative Claims Data Source* | Number of Patients Covered (Millions) |
|---|---|
| CMS- Medicare | 105 |
| Optum - Pre adjudicated | 22 |
| MarketScan Commercial and Medicare Supplemental | 254 |
| CVS Health | 26 |
| Blue Health Intelligence | 33.6 |
| MarketScan Medicaid | 48 |
| OneFlorida Clinical Research Consortium - Medicaid | 6.7 |
| Optum - Adjudicated | 66 |
| HealthCore | 76 |
| EHR Data Source* | Number of Patients Covered (Millions) |
|---|---|
| OneFlorida Clinical Research Consortium - EHR | 5.6 |
| Optum EHR | 102 |
| MedStar Health Research Institute | 6 |
| PEDSnet | 6.2 |
| Linked Claims-EHR Data Source* | Number of Patients Covered (Millions) |
|---|---|
| IBM CED | 5.4 |
| Optum Integrated Claims - EHR | 25 |
| OneFlorida Clinical Research Consortium – Linked EHR Claims | 1.5 |
*Data lag varies for different databases, and it is approximately 3 months.