[论文解读] Mask mandates and COVID-19: A Re-analysis of the Boston school mask study

Tracy Beth Høeg, Ambarish Chandra+3 more|arXiv (Cornell University)|Jul 22, 2023

COVID-19 and healthcare impactsMedicine被引用 3

一句话总结

本再分析挑战了马萨诸塞州在2021–2022学年期间学校口罩令减少新冠肺炎病例的主张。利用美国疾病控制与预防中心和马萨诸塞州教育部的州级数据，作者发现，与解除口罩令的学区相比，维持口罩令的学区在病例率上并无显著下降，病例率甚至在这些学区中略有上升；先前的社区免疫状况可解释高达66%的病例率差异。

ABSTRACT

Background: A recent epidemiological analysis of staggered policy implementation reported a 29.4% reduction in COVID-19 cases by maintaining school mask mandates in the greater Boston area during the first half of 2022. The robustness of their results and the appropriateness of methodology are explored. Methods: Using data from the Massachusetts Department of Elementary and Secondary Education and the Centers for Disease Control and Prevention, we re-analyze differences in COVID-19 incidence in school districts that did and did not lift mask mandates using the same districts as the original study and expanded the analysis to the entire state of Massachusetts. We present changes in case rates and differences in prior immunity in areas with different mask lifting policies. Results: The Boston and Chelsea districts, which maintained mask mandates, were outliers in terms of size, demographics, and testing. We failed to find a notable change in student cases in mask mandate districts compared with the 70 districts in the original study (-0.08/1000; p=0.98) and found a slight increase compared with a statewide control group +3.63/1000 (p=0.291). Results were similar for students and staff combined. Districts that dropped mask mandates first experienced the largest decreases in cases (22% drop vs 12% in the masked districts). There was a moderate to strong relationship (R2 = 0.35-0.66; p-values <0.001) between prior community infection burden and district case rates in Spring 2022, with prior immunity alone explaining as much as two-thirds of the variation in case rates in Spring of 2022. Conclusions: We fail to find any consistent notable negative relationship between school mask mandates and infection rates in the Greater Boston Area or state of Massachusetts during the 2021-2022 academic year.

研究动机与目标

• 重新评估先前一项声称2022年初波士顿学校口罩令导致新冠肺炎病例减少29.4%的研究的有效性与方法论。
• 评估马萨诸塞州学区在2021–2022学年期间口罩令是否与较低的感染率相关。
• 考察先前社区免疫力在解释2022年春季学区病例率差异中的作用。
• 通过本地和全州控制组，比较维持口罩令与解除口罩令学区的病例趋势。
• 确定波士顿和切尔西学区在人口统计、规模或检测频率上的差异是否影响了原研究的结论。

提出的方法

• 重新分析了马萨诸塞州教育部和美国疾病控制与预防中心在2021–2022学年间的学校层面新冠肺炎病例数据。
• 使用与原研究相同的70个学区，但将分析范围扩大至马萨诸塞州所有学区。
• 对维持口罩令与解除口罩令的学区之间应用双重差分法和时间序列比较。
• 通过决定系数R²和p值量化先前社区感染负担（作为免疫力的代理指标）与学校后续病例率之间的关系。
• 对师生合计的子群进行分析，并使用p值评估统计显著性。
• 使用补充图和表验证结果并评估模型的稳健性。

实验结果

研究问题

• RQ1 马萨诸塞州的学校口罩令是否在2022年春季导致学生和教职员工新冠肺炎病例率显著下降？
• RQ2 维持口罩令的学区与解除口罩令的学区在本地和全州范围内的病例率如何比较？
• RQ3 先前社区免疫力在多大程度上解释了2022年春季学区病例率的差异？
• RQ4 波士顿和切尔西学区（维持口罩令）是否代表了马萨诸塞州的总体趋势，还是属于异常值？
• RQ5 在有口罩令和无口罩令学区之间观察到的病例率差异具有多高的统计显著性？

主要发现

• 与原研究的70个学区相比，维持口罩令的学区在学生病例率上未观察到显著下降（变化：-0.08/1000；p=0.98）。
• 与全州控制组相比，维持口罩令的学区病例率出现轻微上升（+3.63/1000；p=0.291）。
• 率先解除口罩令的学区病例率下降了22%，而维持口罩令的学区病例率仅下降12%。
• 先前社区感染负担与学校病例率之间存在中等至较强的相关性（R² = 0.35–0.66；p < 0.001）。
• 仅先前免疫力就可解释2022年春季各学区病例率差异的最高66%。
• 波士顿和切尔西学区被识别为在规模、人口统计和检测频率方面存在统计异常，表明原研究结论可能存在偏差。