Perneger T, Smith G S
Department of Epidemiology, Johns Hopkins University School of Hygiene and Public Health, Baltimore, MD 21205.
Am J Epidemiol. 1991 Nov 15;134(10):1138-45. doi: 10.1093/oxfordjournals.aje.a116017.
This population-based study examines drivers' characteristics associated with driving errors that resulted in fatal motor vehicle crashes. Routinely collected data from the Fatal Accident Reporting System were used to assess whether a driver initiated the crash (case) or was passively involved (control) in 6,506 two-car collisions (81% of 7,993 eligible events). A paired comparison of cases and controls avoided confounding by environmental factors, exposure to traffic, and differences in case fatality. The strongest predictor of crash initiation is alcohol (odds ratio (OR) = 11.5; 95% confidence interval (CI) 9.57-13.9). Odds ratios are elevated even at the lowest blood alcohol concentration levels and increase dramatically as alcohol levels rise. Drivers aged 40-49 years are least likely to initiate crashes; odds ratios rise in a U-shaped manner to 3.35 in teenagers (95% CI 2.72-4.13) and to 22.1 in drivers over 80 years (95% CI 14.2-34.5). Other risk factors for initiating a fatal crash are the following: not wearing a seat belt (OR = 1.54; 95% CI 1.35-1.75), driving without a valid driver's license (OR = 2.16; 95% CI 1.72-2.73), and having had a crash within the last year (OR = 1.21; 95% CI 1.07-1.38). Driving errors leading to fatal crashes do not occur at random, but are associated with specific driver characteristics. The risk factors for crash initiation among crash-involved drivers are similar to risk factors for crash involvement found in other studies. These findings suggest that driving errors often explain high rates of crash involvement, invite further use of crash initiation in traffic injury research, and underscore the value of population-based registries for analytic epidemiology.
这项基于人群的研究考察了与导致致命机动车碰撞事故的驾驶失误相关的驾驶员特征。利用从致命事故报告系统中定期收集的数据，评估了在6506起两车碰撞事故（占7993起符合条件事件的81%）中，驾驶员是引发事故（案例）还是被动卷入事故（对照）。对案例和对照进行配对比较，避免了环境因素、交通暴露情况以及案例死亡率差异造成的混杂影响。引发碰撞事故的最强预测因素是酒精（优势比（OR）=11.5；95%置信区间（CI）9.57 - 13.9）。即使在最低血液酒精浓度水平时，优势比也会升高，并且随着酒精水平上升而急剧增加。40 - 49岁的驾驶员引发碰撞事故的可能性最小；优势比呈U形上升，青少年为3.35（95%CI 2.72 - 4.13），80岁以上驾驶员为22.1（95%CI 14.2 - 34.5）。引发致命碰撞事故的其他风险因素如下：未系安全带（OR = 1.54；95%CI 1.35 - 1.75）、无有效驾驶执照驾驶（OR = 2.16；95%CI 1.72 - 2.73）以及在过去一年内发生过碰撞事故（OR = 1.21；95%CI 1.07 - 1.38）。导致致命碰撞事故的驾驶失误并非随机发生，而是与特定的驾驶员特征相关。涉及碰撞事故的驾驶员中引发碰撞事故的风险因素与其他研究中发现的碰撞事故卷入风险因素相似。这些发现表明，驾驶失误常常可以解释碰撞事故卷入率高的原因，促使在交通伤害研究中进一步采用引发碰撞事故的情况，并且强调了基于人群的登记系统在分析性流行病学中的价值。
