Stanley Sciortino, PhD and Michael L. Radetsky, MA, MPH. Community Health Education Section, San Francisco Department of Public Health, 30 Van Ness Ave, Suite 2300, San Francisco, CA 94102, 415 581-2419, stanley.sciortino@sfdph.org

Background: San Francisco has high rates of pedestrian injuries and fatalities related to specific locations. Our project assessed the strengths and limitations of injury surveillance along with Geographic Information Systems (GIS) mapping techniques to better target injury interventions.

Methods: We linked pedestrians (n=1911) injured by motor vehicles from the Statewide Integrated Traffic Reporting System (SWITRS) with visits to San Francisco General Hospital’s Trauma Center (n=1321) in 2000 and 2001. We conducted a multivariate logistic regression analysis of several outcomes and mapped collision locations.

Results: We found that SWITRS and police collision reports underestimated the total number of injuries in San Francisco by at least 21% (531/2442). Fifty six percent (299/531) of the injured pedestrians who sought medical treatment and who were not identified by police reports arrived at the hospital in an ambulance or were admitted directly to the hospital. African Americans were less likely than whites (OR=0.53, 95% CI=0.38 to0.75) and females were more likely than males (OR=1.46, 95% CI=1.15 to 1.86) to have a SWITRS report. Police designated fatal or severely injured pedestrians were more likely to be hospitalized (90% vs. 25%) and had higher Injury Severity Scores than pedestrians with minor injuries identified by the police.

Conclusion: Though police collision reports show a reporting bias, they are easily mapped and provide useful, though crude, severity coding to help target street-level environmental interventions. Hospital records lacked accessible collision location information. The ability to map injury collisions, especially the most severe, is crucial to injury countermeasure choice and implementation.

Learning Objectives: