Model of Wuhan Trajectory Suggests Longer Social Controls Will Stall a Second COVID-19 Wave

March 26, 2020
Allison Inserro
Allison Inserro

Extending school and workplace closures in Wuhan until April would likely delay a second wave of coronavirus disease 2019 (COVID-19) cases until October, rather than August, and cut new infections by 24%, said a recent report.

An analysis released late Wednesday indicates that keeping tighter measures in place for a longer period of time to control coronavirus disease 2019 (COVID-19) would help relieve pressure on hospitals and healthcare providers and help “flatten the curve.” Using 3 scenarios, the mathematical modeling study estimates the impact of social distancing measures in Wuhan, China, where the novel coronavirus disease began in December 2019.

The study comes as social distancing appears to be easing in Wuhan, as several days have gone by without a new case of infection. It also arrives as President Donald Trump claims he would like COVID-19 business closures to end by April 12—an action not endorsed by public health experts.

Extending school and workplace closures in Wuhan until April would likely delay a second wave of cases until October, rather than August, and cut new infections by 24%, said the report, which was published in The Lancet Public Health.

The model suggests that social distancing has worked in Wuhan, reducing the number of COVID-19 cases. That delayed the peak of the pandemic, giving the health system the chance to expand and respond.

After the emergence of the novel coronavirus (SARS-CoV-2), schools and workplace were closed as part of the Lunar New Year holidays in mid-January 2020. These closures were then extended to reduce person-to-person contact and prevent the spread of infection.

The authors said their model cannot readily be applied to other countries, and the lead author of the study, Kiesha Prem, PhD, of the London School of Hygiene and Tropical Medicine, said in an email to The American Journal of Managed Care® that she could not provide any comments about Trump’s desire to relax social distancing in the United States.

In addition, given the large uncertainties around estimates of the reproduction number—defined as how many people an individual with the virus is likely to infect—and how long a person is infected on average, the actual effect of relaxing physical distancing measures cannot be precisely predicted.

In the model, Wuhan was a “closed system,” meaning that all demographic changes in the population—births, deaths, and aging of the population—are ignored. The study used a short time horizon.

The researchers constructed location-specific contact patterns and layered on Wuhan’s social shutdown. Using estimates of the epidemiological parameters, they simulated the ongoing trajectory of the outbreak in Wuhan using an age-structured SEIR model for several physical distancing measures. A SEIR model estimates the flows of individuals between 4 states: susceptible (S), exposed (E), infected (I), and resistant (R).

The projections show that physical distancing measures were most effective if the staggered return to work was at the beginning of April; this reduced the median number of infections by more than 92% (interquartile range [IQR], 66%-97%) and 24% (IQR, 13%-90%) in mid-2020 and end-2020, respectively.

However, the modelled effects of physical distancing measures vary by the duration of infectiousness and the role school children have in the epidemic.

Using data on the spread of COVID-19 in Wuhan and from the rest of China on the number of contacts per day by age group at school and work, the researchers compared the effects of 3 scenarios, ranging from the hypothetical to what actually happened:

  • No interventions and no holidays (ie, the hypothetical scenario)

  • No physical distancing measures, but holding school winter break and Lunar New Year holidays as normal

  • Intense control measures with school closed and only about 10% of the workforce (eg, healthcare personnel, police, and other essential government staff) —working during the control measures. This started in Wuhan in mid-January.

In addition, they modelled the impact of lifting control measures in a staggered way, and during different stages of the outbreak, in March and April.

The analyses suggest that the normal school winter break and Lunar New Year holidays would have had little impact on the progression of the outbreak. However, putting extreme measures in place to reduce contacts could reduce case numbers and the size of the peak, while also delaying the peak.

Once these interventions are relaxed, case numbers are expected to rise.

"The city now needs to be really careful to avoid prematurely lifting physical distancing measures, because that could lead to an earlier secondary peak in cases. But if they relax the restrictions gradually, this is likely to both delay and flatten the peak,” said Prem in a statement.

The effects of these distancing measures seem to vary by age, with the greatest reductions in new cases among school children and the elderly, and lowest among working-age adults.

“Our results won’t look exactly the same in another country, because the population structure and the way people mix will be different,” said co-author Dr Yang Liu from London School of Hygiene and Tropical Medicine. “But we think one thing probably applies everywhere: physical distancing measures are very useful, and we need to carefully adjust their lifting to avoid subsequent waves of infection when workers and school children return to their normal routine. If those waves come too quickly, that could overwhelm health systems.”

Prem said the team is “in the process of extending this analysis to other countries and contexts where behavioural patterns and demographic structure is different from China.”

The authors noted some limitations to the analysis, including that it assumed no difference in susceptibility between children and that the extreme distancing measures used in Wuhan may have increased the transmission within households. In addition, the model did not capture individual-level differences in contact rates, which could be important in super-spreading events, particularly early on in an epidemic.

Writing in an accompanying commentary, Tim Colbourn, PhD, from University College London, said the study “is crucial for policy makers everywhere, as it indicates the effects of extending or relaxing physical distancing control measures on the [COVID-19] outbreak in Wuhan, China.”

Noting that many countries are in the first phase of cautioning or ordering citizens to shelter-in-place, he said “safe ways out of the situation must be identified.”

“New COVID-19 country-specific models should incorporate testing, contract tracing, and localised quarantine of suspected cases as the main alternative intervention strategy to distancing lockdown measures, either at the start of the epidemic, if it is very small, or after the relaxation of lockdown conditions, if lockdown had to be imposed, to prevent health-care system overload in an already mounting epidemic,” he said.

The study was funded by the Bill & Melinda Gates Foundation, National Institute for Health Research, Wellcome Trust, and Health Data Research UK.

References

1. Prem K, Liu Y, Russell TW, et al.The effect of control strategies to reduce social mixing on outcomes of the COVID-19 epidemic in Wuhan, China: a modelling study [published online March 25, 2020]. Lancet Public Health. doi: 10.1016/S2468-2667(20)30073-6

2. Colbourn T. COVID-19: extending or relaxing distancing control measures [published online March 25, 2020]. Lancet Public Health. doi: 10.1016/S2468-2667(20)30072-4