Primary Healthcare Centres Clinical Notes, Bahrain (2023): De-identified Encounters with ICD-10 Annotations

This dataset contains de-identified clinical notes and associated ICD-10 diagnosis labels drawn from nine public primary healthcare centers in the Kingdom of Bahrain for calendar year 2023. Each record corresponds to a single encounter and includes the following variables: encounter_id patient_id sex_at_birth age_at_encounter encounter_month care_service provider_role diagnosis_icd10 clinical_note Provenance. Source data originate from the national health information system. Extraction was performed under the authority of the Primary Healthcare Centers (PHC). Post-extraction processing included schema standardization, UTF-8 normalization, validation of ICD-10 codes, coercion of month-level dates, top-coding of age at 90+, and generation of pseudonymous identifiers. De-identification combined deterministic rules (regular expressions, dictionaries, gazetteers) with PHI-aware NER; dates in text were normalized to month-level precision. A stratified manual audit of [n] notes confirmed removal of direct identifiers; flagged content was redacted. Ethical/Legal. Secondary research was approved by the PHC Research Committee (November 21, 2024) with a waiver of informed consent. Use must comply with the Kingdom of Bahrain’s Personal Data Protection Law (PDPL, Law No. 30 of 2018). Re-identification attempts are prohibited. If any content could reasonably identify a person, notify the maintainer at [e-mail] so that a corrected release can be issued and prior versions deaccessioned by repository policy. License/Access. Recommended license: CC0 1.0 with Terms of Access specifying PDPL compliance and no re-identification. Older public versions may be deaccessioned when updates are released, preserving provenance via a tombstone record.