In the early stages of a pandemic, non-pharmaceutical interventions (NPIs) that encourage physical distancing and reduce contact can decrease and delay disease transmission. Although NPIs have been implemented globally during the COVID-19 pandemic, their intensity and timing have varied widely. This paper analyzed the country-level determinants and effects of NPIs during the early stages of the pandemic (January 1st to April 29th, 2020). We examined countries that had implemented NPIs within 30 or 45 days since first case detection, as well as countries in which 30 or 45 days had passed since first case detection. The health and socioeconomic factors associated with delay in implementation of three NPIs—national school closure, national lockdown, and global travel ban—were analyzed using fractional logit and probit models, and beta regression models. The probability of implementation of national school closure, national lockdown, and strict national lockdown by a country was analyzed using a probit model. The effects of these three interventions on mobility changes were analyzed with propensity score matching methods using Google’s social mobility reports. Countries with larger populations and better health preparedness measures had greater delays in implementation. Countries with greater population density, higher income, more democratic political systems, and later arrival of first cases were more likely to implement NPIs within 30 or 45 days of first case detection. Implementation of lockdowns significantly reduced physical mobility. Mobility was further reduced when lockdowns were enforced with curfews or fines, or when they were more strictly defined. National school closures did not significantly change mobility.

在大流行的早期阶段，鼓励保持身体距离和减少接触的非药物干预 (NPI) 可以减少和延缓疾病传播。尽管 NPI 在 COVID-19 大流行期间已在全球范围内实施，但其强度和时间差异很大。本文分析了大流行初期（2020 年 1 月 1 日至 4 月 29 日）NPI 的国家层面决定因素和影响。我们检查了在发现首例病例后 30 或 45 天内实施 NPI 的国家，以及自发现首例病例后 30 或 45 天后实施 NPI 的国家。使用分数 logit 和 probit 模型分析了与延迟实施三个 NPI 相关的健康和社会经济因素——国家学校关闭、国家封锁和全球旅行禁令，和 beta 回归模型。使用概率模型分析了一个国家实施全国学校关闭、全国封锁和严格的全国封锁的概率。使用 Google 的社会流动性报告，通过倾向得分匹配方法分析了这三种干预措施对流动性变化的影响。人口较多且卫生准备措施较好的国家在实施方面的延迟较大。人口密度更高、收入更高、政治制度更民主以及首例病例较晚到达的国家更有可能在首例病例发现后 30 或 45 天内实施 NPI。实施封锁大大降低了身体流动性。当通过宵禁或罚款实施封锁时，或者当它们的定义更加严格时，流动性进一步降低。