Data Coverage and Reconciliation

What portion of a population is covered by a particular data source?

For most populations, no single data source provides comprehensive data for all individuals in the population. Claims provide a comprehensive view of utilization but can have high latency and do not provide clinical concepts such as labs or orders. Data from an EHR may be comprensive but only for a small portion of the population. A given individual’s data may be spread across multiple EHRs and multiple data sources for lab results.

Effective management requires a comprehensive view of clinical status to track progress and model risk factors. Care management requires up-to-the-moment data with clinical detail such as labs.

Key Questions of Interest

• What key utilization such as hospitalization and ED utilization can be detected using each data asset?
• What is the data latency for each data source?
• What events can be detected from multiple data sources? Are these sources being appropriately reconciled?
• How does data coverage vary across time and geographies?

Bulk Data API Requirements

Dashboards that address these key questions can leverage the following tables within the Bulk Data API data model:

• derivedencounter_inpatient
• derivedencounter_ed
• derivedperson
• personattribution

Example Visualizations (Tableau)