ROID — Rate Object ID

A ROID uniquely identifies a pricing question. It's a 24-character hex string (first 12 bytes of SHA256) derived from 9 fields.

The 9 ROID Components

Field	Example	Source	Notes
provider_type	'Hospital'	provider_spine	Determines which code plausibility rules apply
payer_id	'643'	network_spine	Partition key for all downstream chunking
network_id	'-4821937561028374920'	network_spine	Deterministic xxhash64 of payer_id || network_name
provider_id	'abc123'	provider_spine	Clear Rates internal provider identifier
bill_type	'Inpatient'	code_spine	Inpatient / Outpatient / Professional
billing_code	'470'	code_spine	The procedure or DRG being priced
billing_code_type	'MS-DRG'	code_spine	HCPCS, MS-DRG, APR-DRG
facility	'' (empty)	code_spine	'true'/'false' for PG codes, '' for all others
billing_code_modifier	''	code_spine	e.g. '26' (professional component), '' if none

Hash Formula

ROID hash — Trino SQL

-- rate_object_space.sql
to_hex(
  substr(
      sha256(
          json_format(
              CAST(
                  ARRAY[
                      provider_type,
                      payer_id,
                      CAST(network_id AS VARCHAR),
                      provider_id,
                      bill_type,
                      billing_code,
                      billing_code_type,
                      facility,
                      billing_code_modifier
                  ] AS JSON
              )
          )
      ),
      1, 12
  )
) AS roid

Worked Example

ROID generation — UHC Choice Plus PPO + Mass General + MS-DRG 470

payer_id: '643' (UHC)

network_id: '-4821937561028374920' — xxhash64 of '643' || 'UHC Choice Plus PPO'

provider_type: 'Hospital'

provider_id: 'mgh_boston_001' (Mass General's Clear Rates internal ID)

bill_type: 'Inpatient'

billing_code: '470'

billing_code_type: 'MS-DRG'

facility: '' (empty — Hospital provider type)

billing_code_modifier: '' (no modifier for MS-DRG)

9 fields assembled → JSON array → SHA256 → first 12 bytes → 24-char hex ROID. This ROID is inserted into tmp_rate_object_space. Every downstream phase (raw data, transformations, imputations, accuracy, rate selection) joins to this ROID to add columns.

warning

ROID depends on network_id, not network_name. network_id is the xxhash64 of (payer_id || network_name), so renaming a network changes its network_id, which changes all ROIDs for that network. Historical rates from prior Clear Rates versions won't join to the new ROIDs — they will appear as gaps rather than carrying forward.