On May 3, 1999, powerful tornadoes, including a category F5 tornado, swept through Oklahoma. The authors examined all tornado-related deaths, hospital admissions, and emergency department visits to identify important risk factors. Data on deaths and injuries directly related to the tornadoes and information obtained from a survey of residents in the damage path of the F5 tornado were used in a case-control analysis. The direct force of the tornadoes caused 40 deaths, 133 hospital admissions, and 265 emergency department outpatient visits. The risk of death from the F5 tornado was greater for persons who were in mobile homes (odds ratio (OR) = 35.3, 95% confidence interval (CI): 7.8, 175.6) or outdoors (OR = 141.2, 95% CI: 15.9, 6,379.8) when the tornado struck than for those in permanently anchored houses. Risk of severe injury was also greater for persons in mobile homes (OR = 11.8, 95% CI: 3.4, 51.7) or outdoors (OR = 34.3, 95% CI: 4.4, 1,526.2). However, the risk of death (OR = 0.0, 95% CI: 0.0, 9.9), severe injury (OR = 0.0, 95% CI: 0.0, 2.0), or minor injury (OR = 0.8, 95% CI: 0.1, 3.1) was not greater among persons in motor vehicles than among those in houses. The risk of death (OR = 0.6, 95% CI: 0.1, 1.7), severe injury (OR = 0.2, 95% CI: 0.1, 0.6), or minor injury (OR = 0.3, 95% CI: 0.2, 0.7) was lower among those fleeing their homes in motor vehicles than among those remaining. Recommendations involving the relative safety of motor vehicles during a tornado should be evaluated using experience from recent tornado events.