Agricultural Injury Surveillance Using Multiple Existing Data Sources

Objective: Due to limited OSHA oversight, most agricultural injuries go unreported regardless of severity. The main objective of this research was to develop an effective agricultural injury surveillance system using existing data sources.   Methods: We identified farm injuries cases through “farm” location check boxes and narrative text review on ambulance reports (AR).  We compared cases (number and type) from AR vs. news sources and interview reports from emergency services officials – “ active community surveillance” (ACS), and vs. hospital (inpatient and emergency department) data.   Results: Since January, 2007, we reviewed over 275,000 AR in 10 NY counties for farm injury cases.   From these, we identified 277 AR cases.  Approximately half the AR cases were missing the “farm” designation, and were only identified through the free text fields.  Review of AR identified 2.4 times as many cases as ACS.  However ACS was better for identifying fatal cases.  Among the three methods, the largest number of injury events by far was obtained from the hospital data (261; 39 inpatient; 222 emergency department).  This was followed by the AR method (107) and then the ACS (44).  We noted minimal overlap in cases discovered by the three methods. Conclusion: These results suggest that a fully electronic agricultural injury surveillance system combining a number of different data sources has the potential to yield more complete data than has yet been assembled on fatal and non-fatal agricultural injury.  This will be further explored by researchers at NEC over the next five year period.