Synthetic version of anonymized Norway Registry data containing prescriptions and hospitalization of the patients

This dataset represents synthetic data derived from anonymized Norwegian Registry Data of pa aged 65 and above from 2011 to 2013. It includes the Norwegian Patient Registry (NPR), which contains hospitalization details, and the Norwegian Prescription Database (NorPD), which contains prescription details. The NPR and NorPD datasets are combined into a single CSV file. This real dataset was part of a project to study medication use in the elderly and its association with hospitalization. The project has ethical approval from the Regional Committees for Medical and Health Research Ethics in Norway (REK-Nord number: 2014/2182). The dataset was anonymized to ensure that the synthetic version could not reasonably be identical to any real-life individuals. The anonymization process was done as follows: first, only relevant information was kept from the original data set. Second, individuals' birth year and gender were replaced with randomly generated values within a plausible range of values. And last, all dates were replaced with randomly generated dates. This dataset was sufficiently scrambled to generate a synthetic dataset and was only used for the current study. The dataset has details related to Patient, Prescriber, Hospitalization, Diagnosis, Location, Medications, Prescriptions, and Prescriptions dispatched. A publication using this data to create a machine learning model for predicting hospitalization risk is under review.

该数据集为基于2011至2013年65岁及以上人群匿名化挪威注册数据生成的合成数据，包含挪威患者注册库（Norwegian Patient Registry, NPR）与挪威处方数据库（Norwegian Prescription Database, NorPD）两部分内容——前者记录住院详情，后者记录处方详情。NPR与NorPD数据集已合并为单个CSV文件。此真实数据集曾用于一项研究老年人用药情况及其与住院关联的项目，该项目已获得挪威医学与健康研究伦理区域委员会（Regional Committees for Medical and Health Research Ethics in Norway, REK-Nord）的伦理批准（编号：2014/2182）。为确保合成版本无法与任何真实个体合理匹配，数据集已完成匿名化处理，具体步骤如下：首先，仅保留原始数据中的相关信息；其次，将个体出生年份与性别替换为合理范围内的随机生成值；最后，将所有日期替换为随机生成日期。该数据集经充分混淆处理以生成合成数据集，且仅用于本次研究。数据集涵盖患者、处方者、住院、诊断、位置、药物、处方及处方分发等相关细节。利用该数据构建住院风险预测机器学习模型的论文目前正在评审中。