Huai Nang Synthetic Personal Health Record (PHR) Dataset

A fully synthetic benchmark dataset modelling the primary health-care population of Tambon Huai Nang, Amphoe Huai Yot, Changwat Trang, Thailand (census year 2020). The dataset is designed for evaluating Attribute-Based Access Control (ABAC) and blockchain-based audit-chain systems for Personal Health Records (PHR). It is released in two formats: native JSON and FHIR R5 NDJSON. Contents: 4,946 synthetic residents across 5 age buckets and 6 clinical roles 462,800 health records spanning 12 record types (vitals, wearable, app_data, lab_result, diagnosis, case, vaccination, cancer_screening, and more) 10,842 ABAC policies in four policy classes (NORMAL, EMERGENCY, EPIDEMIC, PERSONAL_REQUEST) 5,000 explicit consent grants 200 adversarial test fixtures for 4 attack scenarios FHIR R5 NDJSON export: 471,893 resources across 9 resource types (Patient, Practitioner, RelatedPerson, Observation, Condition, Encounter, Immunization, Procedure, Consent) Blockchain-simulation audit trail per record (tx_hash, prev_hash, version) The dataset is fully deterministic: re-running the generator with seed 20260418 at scale 1.0 produces byte-identical output. SHA-256 checksums for all files are recorded in manifest.json. All data is fully synthetic. No real patient information was used.本数据集为全合成基准数据集，建模对象为泰国董里府（Changwat Trang）怀约县（Amphoe Huai Yot）怀南村（Tambon Huai Nang）的基层医疗服务覆盖人口，数据基于2020年人口普查结果。 本数据集旨在评估面向个人健康记录（Personal Health Records, PHR）的基于属性的访问控制（Attribute-Based Access Control, ABAC）方案，以及基于区块链的审计链系统。该数据集以两种格式发布：原生JSON格式与FHIR R5 NDJSON格式。 数据集内容如下： 1. 覆盖5个年龄分组、6类临床角色的4946名合成居民； 2. 涵盖12类记录类型（生命体征、可穿戴设备数据、应用数据、实验室检查结果、诊断信息、病例、疫苗接种记录、癌症筛查记录等）的462800条健康记录； 3. 4类策略类别（NORMAL、EMERGENCY、EPIDEMIC、PERSONAL_REQUEST）下的10842条ABAC策略； 4. 5000条显性同意授权记录； 5. 针对4种攻击场景的200个对抗测试夹具（adversarial test fixtures）； 6. FHIR R5 NDJSON导出内容：覆盖9类资源类型（Patient、Practitioner、RelatedPerson、Observation、Condition、Encounter、Immunization、Procedure、Consent）的471893条资源； 7. 每条记录对应的区块链模拟审计轨迹（包含交易哈希（tx_hash）、前序哈希（prev_hash）、版本号（version））。 本数据集具备完全确定性：以种子值20260418、缩放比例1.0重新运行生成器，将得到字节完全一致的输出结果。所有文件的SHA-256校验和均已记录于manifest.json文件中。 所有数据均为完全合成生成，未使用任何真实患者信息。