Automated Monitoring of Injuries Due to Falls Using the BioSense System

Background: Falls are the leading cause of nonfatal medically attended injuries in the United States. BioSense receives data from 441 non-federal hospitals; falls are one of 11 injury-related concepts tracked. The purposes of this study are to identify and characterize clusters of falls in metropolitan areas during the 2007-08 winter season. Methods: We studied chief complaints of fall in 19 metropolitan areas with >=2 participating EDs during October 1 2007-March 31 2008. The number of fall visits per day was compared with 28-day moving average to calculate statistical significance (clusters with p<0.002 are reported) and excess visits. Among EDs that transmit ED diagnosis data, we searched for final diagnosis ICD-9 codes in the range of 800-829 that indicate fractures. Results: We found 14 clusters of falls in 10 metropolitan areas; 4 areas had 2 clusters and 6 had 1 cluster; 9 clusters lasted 1 day and 5 lasted 2 days. In the clusters, the median number of visits for falls was 52 (range 18 – 324) per day and the median number of excess visits for falls was 40 (range 10 – 233) per day. Among the 2,093 visits for falls, 33% had a diagnosis reported, and 15% of these had an ICD-9 code for fracture; 188 (9%) chief complaints mentioned the word “ice” or “snow.” Nine of the clusters occurred during Dec 6-18, 2007 and 4 on February 12, 2008, periods with winter storm activity. Discussion: We identified several large clusters of ED visits due to falls associated with severe weather. These data are limited in geographic coverage and types of data available. Nevertheless, they illustrate the role that an automated system can play in tracking injuries and potentially assisting with prevention strategies.