Association of driver air bags with driver fatality: a matched cohort study.

OBJECTIVE To estimate the association of driver air bag presence with driver fatality in road traffic crashes. DESIGN Matched pair cohort study. SETTING All passenger vehicle crashes in the United States during 1990-2000 inclusive. SUBJECTS 51 031 driver-passenger pairs in the same vehicle. MAIN OUTCOME MEASURES Relative risk of death within 30 days of a crash. RESULTS Drivers with an air bag were less likely to die than drivers without an air bag (adjusted relative risk 0.92 (95% confidence interval 0.88 to 0.96)). This estimate was nearly the same whether drivers wore a seat belt (adjusted relative risk 0.93) or not (0.91). Air bags were associated with more protection for women (0.88 (0.82 to 0.93)), than for men (0.94 (0.90 to 0.99)). Drivers wearing a seat belt were less likely to die than unbelted drivers (0.35 (0.33 to 0.36)). Belted drivers with an air bag were less likely to die than unbelted drivers without an air bag (0.32 (0.30 to 0.34)). CONCLUSIONS If the associations are causal the average risk of driver death was reduced 8% (95% confidence interval 4% to 12%) by an air bag. Benefit was similar for belted and unbelted drivers and was slightly greater for women. However, seat belts offered much more protection than air bags.