Public health mitigation of COVID-19

Speed and scale are key to mitigation of COVID-19, due to the fat-tailed nature of pandemic risk and the exponential growth of COVID-19 infections.[1] For mitigation to be effective, (a) chains of transmission must be broken as quickly as possible through screening and containment, (b) health care must be available to provide for the needs of those infected, and (c) contingencies must be in place to allow for effective rollout of (a) and (b).

Goals of mitigation include delaying and reducing peak burden on healthcare (flattening the curve) and lessening overall cases and health impact.[2][3] Moreover, progressively greater increases in healthcare capacity (raising the line) such as by increasing bed count, personnel, and equipment, help to meet increased demand.[4]

Mitigation attempts that are inadequate in strictness or duration—such as premature relaxation of distancing rules or stay-at-home orders—can allow a resurgence after the initial surge and mitigation.[2][5]

Part of managing an infectious disease outbreak is trying to delay and decrease the epidemic peak, known as flattening the epidemic curve.[2] This decreases the risk of health services being overwhelmed and provides more time for vaccines and treatments to be developed.[2] Non-pharmaceutical interventions that may manage the outbreak include personal preventive measures such as hand hygiene, wearing face masks, and self-quarantine; community measures aimed at physical distancing such as closing schools and cancelling mass gathering events; community engagement to encourage acceptance and participation in such interventions; as well as environmental measures such surface cleaning.[6]

More drastic actions aimed at containing the outbreak were taken in China once the severity of the outbreak became apparent, such as quarantining entire cities and imposing strict travel bans.[7] Other countries also adopted a variety of measures aimed at limiting the spread of the virus. South Korea introduced the mass screening and localised quarantines and issued alerts on the movements of infected individuals. Singapore provided financial support for those infected who quarantined themselves and imposed large fines for those who failed to do so. Taiwan increased face mask production and penalised hoarding of medical supplies.[8]

Simulations for Great Britain and the United States show that mitigation (slowing but not stopping epidemic spread) and suppression (reversing epidemic growth) have major challenges. Optimal mitigation policies might reduce peak healthcare demand by two-thirds and deaths by half, but still result in hundreds of thousands of deaths and overwhelmed health systems. Suppression can be preferred but needs to be maintained for as long as the virus is circulating in the human population (or until a vaccine becomes available), as transmission otherwise quickly rebounds when measures are relaxed. Until now, the evidence for public health (nonpharmaceutical) interventions such as social distancing, school closure, and case isolation comes mainly from epidemiological compartmental models and, in particular, agent-based models (ABMs).[9] Such models have been criticized for being based on simplifying and unrealistic assumptions.[10][11] Still, they can be useful in informing decisions regarding mitigation and suppression measures in cases when ABMs are accurately calibrated.[12] An Argentinian modelling study asserted that complete lockdowns and healthcare system overextension could be avoided if 45 percent of asymptomatic patients were detected and isolated.[13] Long-term intervention to suppress the pandemic has considerable social and economic costs.[14]

In August 2020, a working paper by the National Bureau of Economic Research (NBER) questioned major effects of many mitigation and suppression measures. The authors compared the development of casualties connected to SARS-CoV-2 until July 2020, in 25 US states and 23 countries that had counted more than 1.000 overall deaths each. From the date a state passed a threshold of 25 deaths, the statistical study observed a largely uniform development, independently from type and time frame of governmental interactions. Thus, the growth rate of casualties dropped to zero within 20–30 days, and the variability between regions was low, except at the beginning of the epidemics. The authors computed the effective reproduction number R_eff with the aid of different models like the SIR model, and found it hovering around one everywhere after the first 30 days of the epidemic. Hence, they did not find evidence for an influence of lockdowns, travel restrictions or quarantines on virus transmission.[15] For contradicting studies, they assume an omitted variable bias. Candidates for ignored effects could be voluntary social distancing, the structure of social interaction networks (some people contact more networks faster than others), and a natural tendency of an epidemics to spread quickly at first and slow down, which has been observed in former Influenza pandemics, but not yet completely understood. The reviewer Stephen C. Miller concludes “that human interaction does not conform to simple epidemiological models”.[16][15]