New York – Around the world, people are itching to end lockdowns. Traffic is rising in the United Kingdom, and garden centers and hardware shops have reopened. In a handful of U.S. states, hair stylists are back at work and restaurants and movie theaters have opened. Germans are shopping again, although many remain wary of the risks.
That caution is well-advised. The example of China, as well as epidemiologic models attuned to the longer-term aspects of the pandemic, suggests that most nations are many months away from anything like normal, even with continued social distancing. In the United States, realistic scenarios include the chance of a new winter epidemic, as well as sporadic outbreaks further on.
It is a little depressing, stuck at home, to read scientists estimating what may happen with the epidemic in 2021, 2022 and even 2024. But there are good reasons to question the more optimistic forecasts for when it may be safe to end the lockdowns. Models based on simulations of the infection’s dynamics foresee huge uncertainties and risks ahead.
Among the optimistic projections are those from the Institute for Health Metrics and Evaluation, cited even by the White House. These suggest safe dates for various U.S. states to begin reopening: May 28 for New York, for example, and June 27 for Georgia. These are based on statistical curve-fitting to data and project when the level of infections may drop to one per 1 million people. Graphs with reassuring downward sloping lines accompany the predictions.
Unfortunately, this picture is almost certainly too optimistic, because the basic model assumes that infection numbers will fall rapidly once reaching their peak. This doesn't come from the data. The optimistic downward projection is an artifact of the mathematical framework used, which isn't really designed for projection, as the researchers acknowledge.
There is a similar but improved statistical model that more accurately reflects the great uncertainties in these projections. But it still just assumes a rapid decay in numbers following the peak, and the researchers behind it stress it shouldn't be used to forecast beyond two to three weeks. Other models based on fitting data with actual simulations of disease dynamics aren't so optimistic. For example, one suggests that the rate of infections in Georgia may in fact double by early August following the state's partial reopening and relaxation of distancing rules.
Even without complex models, data from China offer a crude reality check on the likely timescale required to get the epidemic under temporary control. During the initial exponential growth phase, it took roughly 3½ weeks to reach the peak rate of daily deaths. It then took roughly three times as long, or about three months, for the numbers to fall back to only a few new deaths each day. Most nations have shown something similar: The rise in infections was steeper and quicker than the fall seen so far. Assuming a roughly similar pattern for the U.S. — where the epidemic is only now peaking — gives an expectation of another six months to really control the virus, not just a month or two.
And remember: China achieved its success with thorough testing and contact tracing, as well as maintenance of strict social distancing measures. Steps toward relaxing such measures, already happening in several states and elsewhere around the world, will mean a longer wait to get the virus under control.
Some appreciation of the more likely scenarios extending into the future come from a study by Marc Lipsitch of Harvard University and colleagues. Using an epidemiological model, they projected a number of plausible possibilities, acknowledging that much depends on prevailing unknowns about the virus and on what people and authorities do. Changing seasons could affect viral transmission, as could the degree of immunity attained by recovered individuals, which could be permanent or only temporary.
Overall, they found that even if we continue with strict social distancing, there's still a significant chance of a major new outbreak this winter. If it turns out that recovered individuals only have temporary immunity, the virus will probably enter into regular circulation, causing annual outbreaks. If there are strong seasonal effects on transmission, the wintertime outbreaks may very well be worse than the initial ones.
And social distancing, either continuous or intermittent, might be required into 2022, with ongoing surveillance through 2024. It's not at all realistic to think the virus will soon recede and disappear, letting us all go back to normal.
Mark Buchanan is a physicist and science writer.
In a time of both misinformation and too much information, quality journalism is more crucial than ever.
By subscribing, you can help us get the story right.