no string id rule implementation

[sshenzha]

*_source_value columns contain the raw, original codes from the source system before ETL — things like "I10" (ICD-10 for hypertension) or "99213" (CPT for an office visit). These are institution-specific, encoding-specific strings. They are not a stable or portable way to identify clinical concepts.

OMOP provides the mapped equivalent for every source_value: *_source_concept_id (the source code mapped to an OMOP concept) and *_concept_id (the standard concept).

-- BAD: identifying hypertension by its raw source string
WHERE condition_source_value = 'I10'

-- GOOD: using the concept_id that the ETL already resolved for you
WHERE condition_source_concept_id = 44331282
-- or the standard concept
WHERE condition_concept_id = 316866

[sshenzha]

The rule defines an explicit whitelist of every *_source_value column across all clinical tables in OMOP CDM, grouped by table:

Table	Columns watched
condition_occurrence	condition_source_value
drug_exposure	drug_source_value, route_source_value, dose_unit_source_value
procedure_occurrence	procedure_source_value, modifier_source_value
measurement	measurement_source_value, unit_source_value, value_source_value
observation	observation_source_value, unit_source_value, qualifier_source_value
device_exposure	device_source_value
visit_occurrence	visit_source_value, admitted_from_source_value, discharged_to_source_value
visit_detail	visit_detail_source_value, admitted_from_source_value, discharged_to_source_value
person	gender_source_value, race_source_value, ethnicity_source_value
death	cause_source_value
specimen	specimen_source_value, unit_source_value, anatomic_site_source_value, disease_status_source_value
episode	episode_source_value
note	note_source_value
payer_plan_period	payer_source_value, plan_source_value, sponsor_source_value, stop_reason_source_value

Any time you filter on a *_source_value column using a string.

-- Equality
WHERE drug_source_value = 'metformin'

-- IN list
WHERE condition_source_value IN ('I10', 'I11', 'I12')

-- Pattern matching
WHERE drug_source_value LIKE '%aspirin%'

All three are flagged.

EXAMPLE:

-- Instead of this:
WHERE drug_source_value = 'metformin'

-- Do this:
WHERE drug_concept_id IN (
    SELECT concept_id FROM concept
    WHERE concept_code = '372897005'
      AND vocabulary_id = 'RxNorm'
)