Linking Emergency Medical Services Data to Hospital Discharge and Emergency Department Data for the Surveillance of Non-Fatal Work-Related Injuries in New Jersey, 2011

BACKGROUND:   Though work-related injuries and illnesses are preventable, thousands of workers are injured every day in the US. In New Jersey there have been over one million non-fatal work-related injuries and illnesses reported by employers since 2000. Joining New Jersey Emergency Medical Services (EMS) data with hospital and emergency department discharge data (HD&ED) may enhance the classification and capture of work-related non-fatal injuries. Primarily we will determine whether EMS electronic data can be linked to HD&ED data. We will also assess whether we have properly classified a work-related injury and whether the linkage of the two data sources provides useful additional information for the surveillance of work-related non-fatal injuries.

METHODS:   Records were abstracted from 2011 EMS data if worker’s compensation was the payer or the incident location type was industrial premises or mine or quarry for individuals 16 to 65 years of age. Two sets of record abstractions from HD&ED data were performed each for 2011 for those 16 to 65 years of age. The first abstraction was restricted to those with worker’s compensation as payer. The second abstraction was not restricted to worker’s compensation. An open source software program, Fine-Grained Record Integration and Linkage Tool (FRIL), was used to match records from EMS data with HD&ED data. The linkage algorithm depended on patient name, date of birth, admission date and gender.

RESULTS:   EMS abstraction resulted in 3,027 records. Of the EMS records, 89% were matched to the HD&ED records while 11% remained unmatched. Among the matched records, less than 1% were incorrectly matched by FRIL. Among the matched records 51% had worker’s compensation as payer in the HD&ED data while 66% had worker’s compensation as payer in the EMS. Among the unmatched EMS records 49% had worker’s compensation as payer.

CONCLUSIONS: It is possible to link EMS electronic data with HD&ED data using name, date of birth, admission date and gender. Future objectives of this project are to address whether EMS abstraction criteria is sufficient for capturing work-related injuries and to determine whether the linkage of these two datasets is useful in the surveillance of work-related injuries.