"Everything in moderation" – On the effectiveness of early Covid-19 measures
Leonidas Spiliopoulos answers questions about his study published in the journal BMC Public Health
A study conducted at the Max Planck Institute for Human Development examined the effectiveness of early Covid-19 measures, such as lockdowns and other non-pharmaceutical interventions (NPIs). Author Leonidas Spiliopoulos presents the key findings of his study and explains what we can learn for future pandemics with similar characteristics to Covid-19.
What can the study tell us?
Leonidas Spiliopoulos: One of the important aspects of this study is that it provides improved estimates of the effects of non-pharmaceutical interventions (NPIs) across all possible levels of severity and therefore can suggest what the optimal policy mix is. Unlike in other studies, these estimates also account for voluntary behavioral changes made by citizens, as well as government reactions to the severity of a pandemic as it unfolds over time. Using mobility data from Google Maps, it also offers better estimates of how decreasing mobility impacts pandemic variables of interest such as case and death growth rates.
What can’t the study tell us? What are the limitations?
The study can offer broad recommendations for effective NPIs, but not the optimal set of NPIs for a country or regions within a country. These depend on the country or region’s individual characteristics. For example, population density, living conditions, distribution of age groups, and tourism can affect transmission and mortality rates and may vary across regions. The study also cannot provide an accurate measurement of the positive impact of vaccination because it focused on the period prior to the widespread adoption of vaccines. Finally, while the study examined the impact of NPIs on the number of confirmed cases and deaths, it does not capture the impact of post-Covid-19 condition (long Covid).
What does the study tell us about the impact of a single policy?
The study shows that extensive and open testing had half of the impact of the optimal NPI mix, with negligible other social costs. Furthermore, the intermediate goal of significantly reducing social mobility did not lead to the benefits that had been expected. Public information campaigns contributed significantly to educating the public and led to voluntary behavioral changes with a beneficial impact on the pandemic dynamic. Closing down public transportation and imposing stay-at-home restrictions and internal travel restrictions were not found to play an important role in controlling the pandemic. Recommendations (but not necessarily mandates) to work from home, close schools, cancel public events, and restrict events with more than roughly 100 people were helpful in containing the pandemic. International mobility restrictions in the form of quarantining arrivals from high-risk regions were also effective.
This study cannot measure the individual impact of many specific behavioral changes (e.g., mask-wearing, hand-washing, social distancing); either it is difficult or impossible to measure these changes objectively, or not enough data existed in the time period studied to allow for cross-country comparisons. Estimates of these effects can be found in other studies that looked specifically at these individual behaviors. However, this study directly captures the impact of some measurable behavioral changes—for example, people voluntarily reduced their mobility when the growth rate of cases increased. The study also indirectly captures other voluntary behaviors—for example, extensive testing was found to be extremely effective. Testing alone cannot have an impact if people who test positive do not then voluntarily follow-up by self-isolating, avoiding social contact, etc. until they are no longer contagious. It is therefore possible to conclude that people’s voluntary behavioral changes, as a whole, had an important positive impact on the pandemic dynamics.
What can we learn from this study about how to react to a future pandemic?
As is the case with other Covid-19 studies, the conclusions drawn in this study are relevant for pandemics with similar characteristics to Covid-19. The study strongly recommends extensive testing as an efficient mechanism that has a significant impact on the dynamics of a pandemic without imposing high social costs. In a similar future pandemic, resources should be devoted to developing accurate tests as soon as possible, and to making them universally accessible. Severe restrictions were not found to further reduce case and death growth rates compared to the moderate NPIs identified in the study—and can actually lead to a decrease in overall social welfare compared to moderate policies. However, a future pandemic with significantly higher mortality rates may be better handled with more severe restrictions. Another important takeaway of this study is that people made many voluntary behavioral changes that were effective in containing Covid-19, and that public informational campaigns were relatively successful, potentially making mandatory restrictions unnecessary. These findings suggest that an increased focus on information campaigns and communication would be beneficial in similar future instances.
Were the COVID-19 measures and government politics too strict? Were they appropriate?
At the start of any pandemic, there is high uncertainty about its characteristics, particularly transmissibility and mortality rates. Accurate estimates of these variables require a large amount of data, which must be gathered over time. Consequently, it is not unreasonable for the initial reaction to a pandemic to be severe in order to avoid the risk of a high mortality rate. The study argues that as the situation becomes clearer, governments can—and should—re-evaluate the policies and consider the pros and cons of various NPIs so as to avoid imposing unnecessary social hardship. The study does not attempt to identify what the optimal initial policies would have been, and indeed, based on the information that was available at the time, it might not have been possible to do so in the moment. However, by looking at the past with the knowledge that we have now, we will be in a more informed position to face possible future pandemics.
Can we draw conclusions for individual countries?
The advantage of this study is that, in pooling information from so many countries, it can more reliably estimate the average expected effects of different NPIs, since it includes far more data. Country-specific recommendations, however, would be based on much more limited but specific data. Ideally, public health policy needs to be empirically evaluated in studies such as this one that look at average effects and in country-specific studies. This study can be used as a starting point for considering appropriate public policies, but should then be fine-tuned according to country-specific characteristics. For example, countries with higher population density, a higher proportion of elderly citizens, etc. may benefit more from more stringent NPIs than would other countries, which in turn may be able to achieve a better policy mix with fewer social costs.