Skip to content

Nigeria NDPA 2023 identifier checklist

Decision support only - not legal advice. This checklist describes the technical defaults in OpenMed's ng_ndpa de-identification profile. It does not determine whether processing is lawful, establish a section 30 ground, certify compliance with the Nigeria Data Protection Act 2023 (NDPA), or replace advice from Nigerian counsel or a data protection officer.

The profile is a strict, irreversible de-identification posture: it requires the safety sweep, does not retain a replacement mapping, and assigns a non-keep action to every canonical label. The NDPA's interpretation provision in section 65 defines sensitive personal data to include genetic and biometric data used for unique identification, race or ethnic origin, religious or similar beliefs, health status, sex life, political opinions or affiliations, trade union memberships, and further information prescribed by the Nigeria Data Protection Commission under section 30(2). Section 30(1) restricts processing of sensitive personal data unless one of its stated grounds applies. The full statutory text is available from the Nigeria Data Protection Commission.

Section 30 sensitive-data category reference

Every label below is a member of CANONICAL_LABELS. replace means an irreversible synthetic replacement under this profile; mask means a typed placeholder. Both are non-keep actions. These mappings are conservative technical anchors, not findings that every occurrence of a label is sensitive personal data under the NDPA.

NDPA class key Statutory category Canonical label anchors Profile action
GENETIC_AND_BIOMETRIC_DATA Sections 65 and 30: genetic and biometric data used to uniquely identify a natural person ID_NUM, EYE_COLOR, HEIGHT, GENE_SYMBOL, VARIANT_DESCRIPTOR, PROTEIN_CHANGE, ZYGOSITY, CLINICAL_SIGNIFICANCE mask
RACE_OR_ETHNIC_ORIGIN Sections 65 and 30: race or ethnic origin ETHNICITY, OTHER mask
RELIGIOUS_OR_SIMILAR_BELIEFS Sections 65 and 30: religious or similar beliefs, including conscience or philosophy OTHER mask
HEALTH_STATUS Sections 65 and 30: health status, including health-record identifiers and clinical content under this profile ID_NUM, MICROORGANISM, ANTIBIOTIC, SUSCEPTIBILITY, CONDITION, MEDICATION, LAB_TEST, PROCEDURE, BODY_SITE, DEVICE, ANESTHESIA_TYPE, ANESTHETIC_AGENT, AIRWAY_MANAGEMENT, ASA_CLASS, DIET_TYPE, NUTRITION_TARGET, FEEDING_ROUTE, NUTRITIONAL_STATUS, VACCINE_NAME, DOSE_NUMBER, ADMINISTRATION_ROUTE, VACCINE_LOT, VACCINE_SERIES, INTAKE_OUTPUT, LINE_DRAIN_TUBE, NURSING_RISK_SCORE, CARE_INTERVENTION, GENE_SYMBOL, VARIANT_DESCRIPTOR, PROTEIN_CHANGE, ZYGOSITY, CLINICAL_SIGNIFICANCE, GLYCEMIC_MEASURE, THYROID_MEASURE, HORMONE_LEVEL, INSULIN_REGIMEN, ENDOSCOPIC_FINDING, GI_SYMPTOM, GI_SCORE, POLYP_DESCRIPTOR, CKD_STAGE, DIALYSIS_MODALITY, RENAL_FUNCTION_MEASURE, URINE_FINDING, SPIROMETRY_MEASURE, OXYGEN_SUPPORT, RESPIRATORY_FINDING, DYSPNEA_GRADE, GROWTH_PARAMETER, GROWTH_PERCENTILE, DEVELOPMENTAL_MILESTONE mask
SEX_LIFE Sections 65 and 30: sex life GENDER, OTHER mask
POLITICAL_OPINIONS_OR_AFFILIATIONS Sections 65 and 30: political opinions or affiliations ORGANIZATION, OTHER mask
TRADE_UNION_MEMBERSHIPS Sections 65 and 30: trade union memberships ORGANIZATION, JOB_DEPARTMENT, OTHER mask
OTHER_COMMISSION_PRESCRIBED_DATA Sections 30(2) and 65: other information prescribed as sensitive personal data by the Commission OTHER mask

ETHNICITY represents explicitly contextualized race or ethnic-affiliation detections; OTHER remains the conservative fallback for sensitive attributes without a dedicated canonical label. The profile action protects a correctly normalised detection; it does not itself discover belief, sex-life, political, or trade-union terms. Sites should validate detector coverage on synthetic or otherwise approved local data and add approved custom recognizers where necessary. EYE_COLOR and HEIGHT are treated only as biometric hints; the table does not claim that they are biometric data in every context.

Linked identifier controls

The strict profile also removes identifiers that can link sensitive clinical content to a person. NIN-style and BVN-style values are technical examples in this table; their inclusion is a conservative de-identification default, not a claim that the NDPA categorically defines every NIN or BVN as sensitive personal data.

Technical identifier class Canonical labels Profile action
Names and user handles PERSON, FIRST_NAME, LAST_NAME, MIDDLE_NAME, PREFIX, USERNAME replace
Contact details EMAIL, PHONE, URL replace
Lagos, Abuja, and other location details LOCATION, STREET_ADDRESS, BUILDING_NUMBER, ZIPCODE, GPS_COORDINATES, ORDINAL_DIRECTION mask
Dates and demographic quasi-identifiers DATE, DATE_OF_BIRTH, TIME, AGE, GENDER, EYE_COLOR, HEIGHT mask
NIN-style national IDs, health-record IDs, and biometric-template IDs ID_NUM, SSN mask
BVN-style and other account or financial identifiers ACCOUNT_NUMBER, CREDIT_CARD, CVV, IBAN, BIC, BITCOIN_ADDRESS, ETHEREUM_ADDRESS, LITECOIN_ADDRESS, MASKED_NUMBER mask
Authentication, network, device, and vehicle identifiers PASSWORD, PIN, API_KEY, IP_ADDRESS, MAC_ADDRESS, USER_AGENT, VIN, VEHICLE_REGISTRATION, IMEI mask

Sections 41-42 and local-first processing

Section 41 restricts transfers of personal data from Nigeria to another country unless the recipient is covered by an adequate protection mechanism or a section 43 condition applies. It also requires the transfer basis and section 42 adequacy to be recorded. Section 42 describes adequacy and the factors used to assess it.

When OpenMed runs entirely on the user's device and sends neither input text nor de-identified output to a recipient in another country, the de-identification operation itself is not a cross-border transfer. A fully local workflow therefore avoids the sections 41-42 transfer-basis and adequacy question for that operation. Later cloud storage, telemetry, support access, model calls, exports, or other disclosures can create a separate transfer and must be assessed independently. The profile does not evaluate adequacy, section 43 conditions, consent, contracts, or any other lawful basis.

Implementation notes

  • NDPA_SENSITIVE_CLASS_LABELS is a many-to-many statutory cross-map because broad fallback labels such as OTHER can anchor more than one category.
  • ng_ndpa contains no jurisdiction-specific engine branch. It uses the common policy loader, compiler, action resolver, mandatory safety sweep, and risk budget machinery.
  • All committed fixtures for this profile are synthetic and must not be replaced with real patient, NIN, BVN, address, phone, or biometric data.