Epic Clarity holds detailed reporting records extracted nightly from Chronicles, while Caboodle reshapes Clarity and other data into a star-schema warehouse for reporting and SlicerDicer. Choose Caboodle when an implemented subject area already supports your measure. Choose Clarity for a new definition, detailed records or information your Caboodle model does not contain.
A leader asks for readmissions by service line. One analyst opens an existing Caboodle report, while another starts a Clarity query. Their reports can start with Epic records and still count different encounters or apply different readmission rules.
Find out which records and definitions each report used before comparing the numbers. Check the source, the completed load and the rules applied to the encounters.
Epic data moves through three databases
Chronicles supports the live clinical record, while Clarity and Caboodle provide reporting structures downstream. Analysts use extracted records in those reporting databases to study the care and activity recorded in Chronicles.
Clarity receives a nightly extract from Chronicles. It places detailed operational information into a relational structure that analysts can query. The result reflects the completed extract, so a change entered after that refresh may be absent from the report.
Caboodle then brings Clarity records and data from other systems into a star schema. Its model brings related business concepts together for analytical queries. UC Davis Health describes both extracts as nightly, and the local refresh schedule determines when each reporting database has completed its load.
When discussing source access, name the reporting database the application will read. Clarity and the implemented Caboodle subject areas supply extracted records for analysis without a direct analytical query against Chronicles.
Clarity keeps the full detail of Epic activity
A research request can call for records that the existing reports have never joined together. An analyst can use Clarity to assemble them. UC Davis Health describes its Clarity records as detailed and closely tied to the activity recorded in the EHR.
That detail comes through a complicated relational structure. A report writer needs to understand how the records relate and which statuses belong in the requested population. The selected fields, joins and local status rules together determine which records the query counts.
For a research cohort, the analyst needs to explain why each record belongs in the population. Clarity lets that review start with the detailed records and the rules used to select them. Those rules also depend on the institution's Epic build. A query from another hospital needs review against the local mappings before anyone reuses it.
Clarity schema knowledge is often concentrated among specialists. A detailed request therefore needs both a clear definition and someone who can trace that definition through the reporting records. Keep the SQL and its assumptions with the report so the next analyst can repeat the review.
Caboodle reshapes Clarity data for faster reporting
A report writer can start with the relationships already prepared in a Caboodle subject area. Its facts and dimensions bring together the records needed for that subject, provided the local model covers the question being asked.
UC Davis Health describes Caboodle as a star-schema warehouse for general operational analysis. Its value depends on whether the prepared model includes the population, attributes and measure the report needs. For a standard measure, the analyst can use the existing structure and spend the review on the records behind a change.
SlicerDicer uses Caboodle for population exploration. Vanderbilt also describes a SQL interface for Caboodle and links to external data within its implementation. A researcher exploring a population and an analyst writing a recurring report can therefore use the same subject area through different interfaces.
Caboodle can include records from systems outside Epic. Look for the external records and relationships already included in your implementation. Their presence determines whether the warehouse can answer a question spanning several systems.
Clarity and Caboodle differ in model and users
Check how much of the required measure has already been modelled before choosing the query route. Clarity gives the report writer detailed records to assemble. Caboodle offers prepared analytical structures that can serve a standard measure when the institution has implemented them.
The table describes the reporting pattern documented by UC Davis Health and Vanderbilt. Local access and subject-area coverage still determine what an analyst can use. A database name alone does not tell you whether a particular report definition is available.
Use the refresh row when reconciling results. Two reports can use the same definition and still show different records if one completed its load later. Compare the reporting period and completed load before investigating differences in the calculation.
| Attribute | Clarity | Caboodle |
|---|---|---|
| Loaded from | Chronicles | Clarity and other systems |
| Data model | Detailed relational structure | Star schema with facts and dimensions |
| Refresh | Nightly extract | Nightly load following source refresh |
| Coverage | Records extracted from Chronicles | Subject areas implemented locally |
| Main users | SQL report writers and analysts | Analysts, report builders and SlicerDicer users |
| Best for | Detailed operations, new measures and research cohorts | Standard measures and general analysis |
Caboodle coverage depends on your build
An institution can only report from Caboodle subject areas it has implemented. Before promising a result, check that the model contains the required records and that its existing definition matches the proposed measure.
Vanderbilt describes Caboodle as holding much of the data available in its Clarity environment, together with links to external sources. That describes Vanderbilt's implementation. Another institution may have prepared different subject areas or linked different systems.
A subject area can also be available without answering the exact question. A general operational measure may omit the detail or exclusions needed for a research cohort. Inspect the definition with the person who owns the measure before treating the existing report as the answer.
Neither nightly reporting database supplies live clinical data. Reporting Workbench serves operational lists using current entries inside Epic. Use it for requests that need records entered since the last completed reporting extract.
Choose the database by the question
Start with Caboodle when its implemented subject area and measure match the requested analysis. Start with Clarity when the question requires a new definition, detailed records or information absent from that model.
For a recurring trend, use the prepared Caboodle relationships when they express the calculation your team has agreed. Confirm the reporting period, included encounters and local rules, then use the existing model consistently across the report's views.
For a new cohort or audit, Clarity gives a report writer a way to examine the records behind the requested definition. The analyst can assemble the required relationships and inspect which records meet each rule. This is useful when the purpose of the report includes explaining why a particular record was included.
When both databases contain the information, compare the definitions before choosing. Caboodle may provide the reporting route, with Clarity used to examine the underlying detail. A mismatch between results should lead to a review of scope, refresh and calculation. It does not establish that one database is the wrong choice for every report.
Warehouse projects usually start from Clarity
A warehouse project needs a reliable extract and definitions that survive the move into a new analytical environment. Clarity usually supplies the detailed reporting records, while implemented Caboodle subject areas can add prepared relationships where they help.
Keep the raw loads and their timing available alongside the prepared data. That gives analysts a way to compare a later measure with the records extracted for that period. Moving the records alone does not settle the local status mappings or the definitions used in finance, operations and research.
Databasin's Epic connector was created at WashU Medicine. Our team extracts Clarity records first and uses Caboodle subject areas where present, then prepares the local mappings and measures. The result can land in Snowflake, Databricks Unity Catalog or an open Iceberg lakehouse in your Azure tenant.
The Epic integration describes the source connection. The guide to Clarity in Snowflake or Databricks follows the warehouse project further. Decide which questions the destination must answer, then retain the extracted records analysts need to review its measures.
Example: one readmissions question in both databases
In this illustrative case, a team needs last quarter's 30-day readmission rate by service line and considers both reporting databases.
| Step | In Caboodle | In Clarity |
|---|---|---|
| Find the encounters | Use the inpatient subject area, where implemented | Join local encounter, admission and discharge records |
| Apply the definition | Review the modelled measure against local rules | Define the 30-day window and exclusions |
| Data as of | Last completed nightly load | Last completed nightly extract |
| Who can run it | An analyst or an authorised SlicerDicer user | A report writer familiar with the local schema |
| Best use | A recurring trend using the existing measure | A new cohort or a detailed audit |
An implemented Caboodle subject area can reduce the query preparation needed for the recurring trend. Clarity provides the detailed records for reviewing the definition and investigating differences.
Five checks before you pick a source
- List your Caboodle subject areasCheck which populations and measures the local implementation already supports before choosing Caboodle for a proposed report or research question.
- Find who writes Clarity SQLIdentify the analyst who can trace local records and statuses, particularly when the request introduces a new definition or cohort.
- Write down shared definitionsAgree the population, period and exclusions with the measure owner so reports from either database can be compared using the same rules.
- Check when each database refreshesRecord the completed extract and load times, then make the data's reporting period clear to the people using the result.
- Decide the warehouse sourceSelect Clarity detail and available Caboodle models according to the destination's questions, preserving the source records needed to review prepared measures.
Sources
UC Davis Health, Clinical Data Sources, 2026 · health.ucdavis.edu/data/sources.html
Vanderbilt Clinical Informatics Center, Epic Data Resources, 2026 · vumc.org/vclic/epic-data-resources
Get answers from your Epic data
We can discuss the questions your team is waiting to answer and which reporting source can support them. A demo explores a working example, the local definitions behind it and how our team would prepare the data.