Provider identity

Medicare PUF (NPI x HCPCS utilization)

vlada_curated.dim_provider_service·live·dim·Medicare Fee-for-Service Utilization & Payments

Rows

14,161,974

Columns

History

N/A

Card

nocolumns only

Typed tools

What it is

1Medicare Physician & Other Practitioners PUF — current deduped provider-service cache used by the ghost-rate filter and provider-service lookups.
2The cache keeps the most recent CMS PUF observation for each (npi, hcpcs_cd, place_of_service) key.

Grain

Key	Grain
N/A	1One row per (npi, hcpcs_cd, place_of_service) — most recent CMS PUF observation (current cache; data_year 2021-2023 reflects per-key recency)

Columns (15)

Column	Type	Meaning
`npi`	string
`hcpcs_cd`	string
`hcpcs_desc`	string
`place_of_service`	string
`provider_type`	string
`provider_state`	string
`provider_zip5`	string
`total_services`	bigint
`total_beneficiaries`	bigint
`avg_submitted_charge`	double
`avg_allowed_amt`	double
`avg_medicare_payment`	double
`avg_standardized_payment`	double
`data_year`	int
`loaded_at`	timestamp

Query it

describe_data_source(source_fqn="vlada_curated.dim_provider_service")     # the card, live
query_substrate(sql="SELECT * FROM vlada_curated.dim_provider_service WHERE <filter> LIMIT 100")

Item	Value
Typed tools that read it	none; raw SQL
Identity key	`npi,hcpcs_cd,place_of_service`
Join	`how_to_join(source_fqn="vlada_curated.dim_provider_service")`
Raw door	Athena: vlada_curated.dim_provider_service
Note	14,161,974 rows; zero duplicate (npi,hcpcs_cd,place_of_service) keys; data_year 2021-2023 reflects per-key recency; single loaded_at batch. Current cache serves the ghost-rate filter and provider-service lookups, not historical trends.

This page as markdown: /docs-md/data/vlada_curated_dim_provider_service.md · every page in one file: /llms-full.txt