Overview
Four registry entities describe who is contracting and for which population.
The entities don’t reference each other directly. They are linked through canonical observations.
Payer
Payers come from two tiers:- Global payers are shared by every organization. The list is built from the Stedi payer network, NAIC-registered insurers, government programs and self-pay.
- Organization payers exist only for your organization, such as a local TPA. If one has the same name as a global payer, yours takes precedence.
Global payers also carry matching attributes:
id_int64 ranges
A small number of adopted IDs also sit at 9,000,000 and up. IDs are unique and never reused, but don’t rely on the range to tell global and organization payers apart; use
is_global.
Payer Group
Groups are seeded from the parent payers of the payers you contract with. A payer has no fixed group column. The payer-to-group link is recorded on each observation, so the same payer can roll up differently in different contexts.
Line of Business
Organizations seeded with PriceMedic’s standard list get these values. Organizations whose lines of business come from their own data, for example during mastering, have their own list, and
bucket may be empty:
Plan type alone doesn’t set the LOB. A Medicare HMO is Medicare Advantage, not Medicare.
Practice
name, ein and group_npi are each unique within your organization. Incoming data that would give two practices the same TIN or NPI is flagged for review, not merged.
How the IDs relate
Each record from your source data is resolved into a canonical observation: one row holding the IDs of the entities that appeared together. This is the only place the entities are linked.
For example, one observation can say that Practice A contracts with Aetna, under the Aetna (CVS Group) payer group, for Medicare Advantage.
Raw values are matched to entities in this order: a payer already seen for the same practice, exact name, a prior resolution, fuzzy match, AI-assisted match, then human review. Each match records the method used.
Data products
The entities are also published as tables in your organization’sregistry database:
The
*_id columns in fact_canonical_observation_dp join to id in the matching dim_*_dp table. dim_payer_dp includes your payers and all global payers, so every payer_id joins. org_slug identifies your organization and synced_at is the last refresh time.