Medicare Advantage, Part D, Medicaid

Vlada Plan Value, county basis (formerly BenVAT) with per-benefit values and real county enrollment

medicare_semantic.vw_plan_value_county·live·semantic_view

Rows

445,771

Columns

History

N/A

Card

nocolumns only

Typed tools

What it is

1Vlada Plan Value county basis (run d06cf1bf7, signed 2026-08-27; typed rebuild 2026-09-17 verified cell-for-cell against the signed CSVs): the plan value re-priced and re-utilized per county (v_b) against the national value (v_national), decomposed into price_effect, util_effect and interaction, with the reference and plan PMPM, MOOP relief, the rake diagnostics and the 35 per-benefit-code value columns (v_by_code_*).
2Adds real county enrollment from the CMS February file with its suppression flag and bounds, the count of a plan's segments in the county, and `snp_type` as labels.
3Modeled; never a CMS figure.
4Owner source: register w3 vlada_w3_benvat (door RAW, locked by Ben 2026-09-03).
5Supersedes medicare_semantic.benvat_county_v1, which read CSV positionally and so mis-named 26 columns for 2023 and 2024.

Grain

Key	Grain
N/A	1One row per benefit_design_year x contract x plan x segment x county_ssa (measured 2026-09-17 in Athena, 445,771 rows = 445,771 distinct keys, Athena efc66d87; 2844 contracts, CY2023-CY2026). 3Real types (double / int / boolean / varchar); no cast is needed or wanted.

Columns (76)

Column	Type	Meaning
`contract`	string
`plan`	string
`segment`	string
`benefit_design_year`	int
`organization`	string
`plan_name`	string
`plan_type`	string
`snp_type`	string
`parent_organization`	string
`county_ssa`	string
`fips`	string
`county_name`	string
`state`	string
`partial_county`	string
`eghp`	string
`v_b`	double
`v_b_raw_f_sensitivity`	double
`v_price_b`	double
`v_util_b`	double
`v_cplus_additive`	double
`accumulator_effect_b_minus_cplus`	double
`v_price_first_order`	double
`v_national`	double
`price_effect`	double
`util_effect`	double
`interaction`	double
`ref_pmpm_b`	double
`plan_pmpm_b`	double
`v_moop_relief_pmpm`	double
`rake_converged`	boolean
`rake_max_rel_gap`	double
`ma_penetration_warned`	boolean
`plan_member_year_refusals`	double
`identity_gap`	double
`refusal_reason`	string
`v_by_code_10a1`	double
`v_by_code_10a2`	double
`v_by_code_11a`	double
`v_by_code_11b`	double
`v_by_code_11bs`	double
`v_by_code_12`	double
`v_by_code_14e2`	double
`v_by_code_15c`	double
`v_by_code_15m`	double
`v_by_code_16a`	double
`v_by_code_1a`	double
`v_by_code_1b`	double
`v_by_code_2`	double
`v_by_code_3c`	double
`v_by_code_3p`	double
`v_by_code_4a`	double
`v_by_code_4b`	double
`v_by_code_6`	double
`v_by_code_7a`	double
`v_by_code_7b`	double
`v_by_code_7c`	double
`v_by_code_7d`	double
`v_by_code_7e`	double
`v_by_code_7f`	double
`v_by_code_7h`	double
`v_by_code_7i`	double
`v_by_code_7k`	double
`v_by_code_8a1`	double
`v_by_code_8a2`	double
`v_by_code_8b1`	double
`v_by_code_8b2`	double
`v_by_code_8b3`	double
`v_by_code_9a1`	double
`v_by_code_9a2`	double
`v_by_code_9b`	double
`plan_enrollment_feb_per_segment_proxy`	double
`county_enrollment_feb`	int
`county_enrollment_suppressed`	boolean
`county_enrollment_lower_bound`	int
`county_enrollment_upper_bound`	int
`segments_of_plan_in_county`	bigint

Query it

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

Item	Value
Typed tools that read it	none; raw SQL
Identity key	`contract,plan,segment,county_ssa,benefit_design_year`
Join	`how_to_join(source_fqn="medicare_semantic.vw_plan_value_county")`
Raw door	Athena (read-only via query_substrate): medicare_semantic.vw_plan_value_county
Company names	`parent_organization` resolve to a permanent Vlada ID: `resolve_entity(text=<name>)`; the company pages hold owners and holdings
Note	Name: Ben chose "Vlada Plan Value" on 2026-09-17. "BenVAT" survives in the source id, the code and the retired table names until a rename lane changes them.

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