Redelmeier Donald A, Zipursky Jonathan S
Evaluative Clinical Sciences, Sunnybrook Research Institute, Canada.
Department of Medicine, University of Toronto, Canada.
Am J Lifestyle Med. 2021 Nov 26;17(2):276-279. doi: 10.1177/15598276211058378. eCollection 2023 Mar-Apr.
The COVID pandemic provides a natural experiment examining how a 50-60% reduction in pedestrian activity might lead to a reduction in pedestrian deaths. We assessed whether the reduction in pedestrian deaths was proportional to a one-to-one matching presumed in statistics correlating mobility with fatality. The primary analysis examined New York (largest city in US), and the validation analysis examined Toronto (largest city in Canada). We identified pedestrian activity in each location from the Apple Mobility database, normalized to the baseline in January 2020. We calculated monthly pedestrian deaths from the Vision Zero database in each city with baseline data from 3 prior years. We found a large initial reduction in pedestrian deaths during the lockdown in New York that was transient and not statistically significant during the summer and autumn despite sustained reductions in pedestrian activity. Similarly, we found a large initial reduction in pedestrian deaths during the lockdown in Toronto that was transient and not sustained. Together, these data suggest the substantial reductions in pedestrian activity during the COVID pandemic have no simple correlation with pedestrian fatality counts in the same locations. An awareness of this finding emphasizes the role of unmeasured modifiable individual factors beyond pedestrian infrastructure or other structural contributors.
新冠疫情提供了一个自然实验，用以研究行人活动减少50%-60%会如何导致行人死亡人数的下降。我们评估了行人死亡人数的减少是否与将出行与死亡关联起来的统计数据中假定的一对一匹配成比例。主要分析考察了纽约（美国最大城市），验证分析考察了多伦多（加拿大最大城市）。我们从苹果移动数据库中确定了每个地点的行人活动情况，并将其与2020年1月的基线进行了归一化处理。我们根据每个城市“零交通事故愿景”数据库计算了每月的行人死亡人数，并采用了前三年的基线数据。我们发现，纽约在封锁期间行人死亡人数最初大幅下降，但这种下降是短暂的，尽管行人活动持续减少，但在夏季和秋季并没有统计学意义。同样，我们发现多伦多在封锁期间行人死亡人数最初也大幅下降，但这种下降也是短暂的，没有持续下去。总体而言，这些数据表明，新冠疫情期间行人活动的大幅减少与同一地点的行人死亡人数之间没有简单的关联。认识到这一发现凸显了除行人基础设施或其他结构性因素之外未被测量的可改变个体因素的作用。
