Center for Injury Research and Prevention, Children's Hospital of Philadelphia, Philadelphia, Pennsylvania, USA
Division of Emergency Medicine, University of Pennsylvania Perelman School of Medicine, Philadelphia, Pennsylvania, USA.
Inj Prev. 2021 Oct;27(5):472-478. doi: 10.1136/injuryprev-2020-044101. Epub 2021 Mar 8.
Our objective was to describe the development of the New Jersey Safety and Health Outcomes (NJ-SHO) data warehouse-a unique and comprehensive data source that integrates state-wide administrative databases in NJ to enable the field of injury prevention to address critical, high-priority research questions.
We undertook an iterative process to link data from six state-wide administrative databases from NJ for the period of 2004 through 2018: (1) driver licensing histories, (2) traffic-related citations and suspensions, (3) police-reported crashes, (4) birth certificates, (5) death certificates and (6) hospital discharges (emergency department, inpatient and outpatient). We also linked to electronic health records of all NJ patients of the Children's Hospital of Philadelphia network, census tract-level indicators (using geocoded residential addresses) and state-wide Medicaid/Medicare data. We used several metrics to evaluate the quality of the linkage process.
After the linkage process was complete, the NJ-SHO data warehouse included linked records for 22.3 million distinct individuals. Our evaluation of this linkage suggests that the linkage was of high quality: (1) the median match probability-or likelihood of a match being true-among all accepted pairs was 0.9999 (IQR: 0.9999-1.0000); and (2) the false match rate-or proportion of accepted pairs that were false matches-was 0.0063.
The resulting NJ-SHO warehouse is one of the most comprehensive and rich longitudinal sources of injury data to date. The warehouse has already been used to support numerous studies and is primed to support a host of rigorous studies in the field of injury prevention.
我们的目标是描述新泽西安全与健康结果（NJ-SHO）数据仓库的开发情况——这是一个独特而全面的数据源，整合了新泽西州的全州行政数据库，使伤害预防领域能够解决关键的、高优先级的研究问题。
我们采用迭代方法将新泽西州六个全州行政数据库的数据进行链接，时间范围为 2004 年至 2018 年：（1）驾驶员驾照记录，（2）与交通相关的传票和停职，（3）警方报告的撞车事故，（4）出生证明，（5）死亡证明和（6）医院出院记录（急诊、住院和门诊）。我们还链接了费城儿童医院网络的所有新泽西州患者的电子健康记录、普查区一级指标（使用地理编码的居住地址）和全州范围内的医疗补助/医疗保险数据。我们使用了几种指标来评估链接过程的质量。
链接过程完成后，NJ-SHO 数据仓库包括 2230 万条不同个体的链接记录。我们对该链接的评估表明，链接的质量很高：（1）所有接受对中匹配的中位数匹配概率——即匹配真实的可能性——为 0.9999（IQR：0.9999-1.0000）；（2）错误匹配率——即接受对中错误匹配的比例——为 0.0063。
由此产生的 NJ-SHO 仓库是迄今为止最全面、最丰富的伤害数据纵向来源之一。该仓库已经被用于支持许多研究，并为伤害预防领域的大量严格研究做好了准备。
