We note that long-standing land use and transportation policy are critical factors in creating traffic conditions and will have to play a role in reducing injuries. We present the historical progression of events that created current traffic conditions in Boston Chinatown and an analysis of traffic-related injuries at the community level for the years 1996-1998. Injuries were found to be as likely on weekends as on weekdays and frequently occurred late at night. Nighttime occupant injuries were found to be more likely on Friday, Saturday, and Sunday nights (relative risk = 2.26; confidence interval = 1.35-3.78, P =.0014). Injuries varied significantly by location for occupant (P = <.001) and for pedestrian injuries (P =.039). There were no peaks of injuries at traditionally defined commuter hours, which have been the standard time for assessing "worst case" traffic impacts by developers and government agencies. There was, however, a strong association between injuries and vehicle volume at 9 intersections with simple configurations for AM and PM commuter hours (R(2) = 0.589, P =.010), resulting in a calculated increase of 3-5 injuries per year for each increase of 1,000 vehicles. There was no such association at 10 intersections with complex configurations (R(2) = 0.104, P =.397). The 24-hour weekend patterns of vehicle volumes showed that traffic abated only between 3 and 7 AM, and the patterns appeared qualitatively to mirror the 24-hour pattern of injuries, suggesting that they were also indicative of injury risk. We suggest that there is a need for both long-term changes in policy and more immediate interventions. We also conclude that researchers should be cautious about assuming that traffic patterns conform to naive expectations such as rush hour peaks.