To assess the effectiveness of containment policies such as lockdowns or travel restrictions, it is crucial to remember some basic facts on the COVID-19 epidemic.

Several studies on the COVID-19 infection suggests that the incubation period range between 2 and 14 days, with an average incubation period of 5 days. Concerning the deaths, the median delay between illness and deaths is on average 13 days. Thus, a lockdown policy, if respected, should show some effects on reported cases one week after its implementation. The impact on the number of death should begin, on average, two weeks after its implementation.

In this post we provide a graphical analysis of the potential impact of two policies: stay-at-home requirements and workplace closures. We start with a simple correlation analysis, so for the moment, no causality can be inferred from our results (the analysis of causal impacts are still in progress!). We use information from 115 countries between January 15 and May 6 of 2020. For each country, we identify the date in which each policy was implemented and compare trends in the growth of cumulative cases and deaths before and after the implementation. The metric we use is the change in the logarithm of the cumulative number of cases and deaths, a measure that aims to capture if the epidemiological curve is flattening.

We focus on cases in which each policy was mandated, not only recommended. Countries that implemented stay-at-home restrictions also tended to implement workplace closures (correlation of 0.68) at about the same, so in many cases both policies are active.

Looking at these four figures, both policies appear to produce immediate effects on the growth rate of the number of reported cases. As expected, the effect on the growth rate of the number of reported deaths has a longer delay, of around two weeks, although it must be said that there is no obvious trend break. The figures are consistent with the epidemiological characteristics of the virus and suggest that the policies had a significant effect on the rate of growth of the number of cases and deaths.

However, the analysis suffers from several drawbacks. As we already have mentioned, this is a simple correlational analysis that does not take into account heterogeneity across countries. For example, not all countries were facing the same challenges and the same cost/benefit ratio when they implemented the policies. Moreover, some countries had time to learn from other experiences when dealing with the epidemic (A specific focus on the learning time and the quickness of policy reaction is addressed here.). Moreover, countries also vary substantially in dimensions such as population density, median age, and the strength of the health-care system, etc.

In an upcoming post, we will tackle those issues using statistical models better suited to deal with this heterogeneity. More to come ...

References

Linton, N. M., Kobayashi, T., Yang, Y., Hayashi, K., Akhmetzhanov, A. R.,Jung, S.-m., Yuan, B., Kinoshita, R., & Nishiura, H. (2020). Incubation period and other epidemiological characteristics of 2019 novel coronavirus infections with right truncation: A statistical analysis of publicly available case data.Journal of clinical medicine,9(2), 538.