Study develops COVID-19 pandemic reopening strategies in California

By Dr. Liji Thomas, MDApr 30 2021

The coronavirus disease 2019 (COVID-19) pandemic has brought devastation to most of the earth, not only by causing death and disease, but also hampering vital social and economic interactions. The primary weapon against the virus spread has been the implementation of non-pharmaceutical interventions (NPIs), including social distancing, mask-wearing and school/business closures.

However, with all closures, the harm is proportional to the duration and extent of shut-down. The time has now come to seriously consider viable strategies to open up states and countries, even while herd immunity remains a goal and not a reality.

Study: Reopening California: Seeking Robust, Non-Dominated COVID-19 Exit Strategies. Image Credit: Alexander Lukatskiy / Shutterstock

This news article was a review of a preliminary scientific report that had not undergone peer-review at the time of publication. Since its initial publication, the scientific report has now been peer reviewed and accepted for publication in a Scientific Journal. Links to the preliminary and peer-reviewed reports are available in the Sources section at the bottom of this article. View Sources

Reopening strategies – different approaches

A working paper from researchers at the RAND corporation explores the various options to come up with a proposed robust plan for reopening up regions where vaccines have not yet been deployed across the entire population. The paper has been released on the medRxiv* preprint server.

The successive waves of the current pandemic have made this a doubly difficult task, with some regions recognizing the need to shut down again after a brief reopening, or setting restrictions on previously permitted activities. Both uncertainties, as well as trade-offs, must be considered to ensure a durable plan.

Rather than a centralized lockdown, many local bodies have taken to the adaptive use of NPIs based on the neighboring districts or counties. A clear list of NPIs to be implemented under a given set of criteria is easy to recommend, but fails to consider different sources of variation, such as viral biology (including the transmissibility and virulence of different viruses), human behavior (both before and after vaccination), and technological advancement (including vaccine development and efficacy).

The current paper assesses multiple options to adjust the levels of NPIs, from those that operate on predetermined thresholds to adaptive thresholds. Secondly, the researchers optimize the trade-offs between health protection and economic recovery. And finally, it reports the use of Robust Decision Making (RDM), which allows different NPI strategies to be compared, keeping in view the profound uncertainties of the situation.

RDM allows deep uncertainty

RDM has already been used in other areas of uncertainty to help shape policies optimally, avoiding what are called pareto-dominated strategies – strategies that are always worse than others in the same comparison group, in all respects, and therefore unworthy of adoption.

Currently, California has a plan called Blueprint for a Safer Economy, which sets a threshold of daily cases below 7 per 100,000, and positivity rates below 8%, before the highest level of NPIs can be removed. However, a reopening plan also comprises decisions on which businesses can operate at different levels of risk, the capacity at which they can operate, and how to adapt to pandemic conditions.

The current paper does not go into these decisions but examines how to shift from a higher risk level to a lower one as the situation changes. The policies here concern large-scale blunt NPIs such as business and school closures.

The level of restriction in these areas can be relaxed or tightened in accordance with how individual low-cost NPIs like physical distancing or mask use in public are complied with.

The ideal and the real

In an ideal situation, high population compliance and stringent initial closures would have ensured a good start on containing viral transmission. If followed by limited movement, with a test-and-trace policy, along with strict isolation and quarantine of cases and contacts, it may have been possible to rid a region of the virus, in theory, which would allow complete reopening.

However, this has rarely been the case, even in the most authoritarian of countries. Most regions thus face the reality of having to balance economic demands with the requirements of health policy.

The model

The current paper uses a model used to construct an earlier RAND COVID-19 State Policy Tool. It envisages six levels of NPI stringency, with an estimated economic loss at each level.

There are five strata of the population, determined by age, chronic illness, and occupation. The level of mixing between these strata is also described. The current analysis ran this model for the state of California, for the period from December 25, 2020, through January 31, 2021.

Using parameters for the level of caution before a level of NPI is mandated, as well as alternative reopening strategies, the study estimates the loss of life and number of cases with each approach. In addition to the short-term economic savings of early reopening, the total loss of welfare over the long term with prolonged NPIs is another important input.

The findings

The results underline the need for a high baseline level of caution in order to minimize deaths. Using a fixed level of caution apparently balances the number of deaths and social welfare costs.

However, these are pareto-dominated by many other strategies based on vaccination and time-based parameters, unless the highest or lowest levels of intervention are considered. This indicates that adaptive strategies are superior to fixed-level-of-caution strategies.

The better strategies would begin with a high level of caution, followed by relaxation with vaccination and advancing time. This results in lower levels of social welfare costs as well as deaths.

New Zealand is likely to use such a strategy, beginning with a very stringent lockdown, with relaxation planned as immunity becomes widespread.

As the immunization coverage increases, beginning with the groups at the highest risk, the infection fatality rate will probably decrease, which indicates that an adaptive strategy would be superior to one with a fixed level of caution.

The eventual predicted outcome of the pandemic is the subsidence of the outbreak into an endemic state. NPIs will no longer be of any benefit, and strategies that keep this in mind will dominate fixed-level-of-caution strategies.

The speed of this transition depends on many different factors, including the rate of increase of immunity, differences in IFR between population groups, and vaccination strategies. As vaccination becomes more widely available, stringent policies may have to give way to others based on temporal trends or vaccine coverage.

If the baseline level of caution is too low, it is possible that the number of deaths will be too high for the level of economic benefit gained, as well as an increased risk that new and more infectious strains will emerge, thus giving the pandemic another shot.

What are the conclusions?

Calling decisions on reopening policies “arguably the most important of 2021”, the researchers conclude, “Adaptive reopening strategies with fixed thresholds can be dominated by alternatives that are more stringent but change their stringency over time.”

The paper indicates that the expansion of vaccination programs will be the single most important determinant of the reopening strategy, but the scientists also comment that a vaccination-based policy will have to be implemented as a time-based one for practical purposes.

This raises some difficulties in that vaccination rates to differ considerably across different counties and regions. Even so, such schedules will still be superior to fixed-threshold reopening strategies, the paper says.

The trade-offs are most distinct when it comes to the most vulnerable population – the poor and minority communities who have neither savings nor the ability to work from home – who are both exposed to higher levels of infection and to economic losses.

How NPIs are managed in the next several months will determine the outcomes of the COVID-19 pandemic and will shape the trade-offs that these populations face.”

Further analyses should focus on behavioral variations such as greater social mixing after vaccination, increased transmission, and vaccine unwillingness; the result in the form of new surges; and the lower severity and IFR of reinfections following immunization.

Other long-term effects of the prolonged hiatus in in-person education, long covid, non-availability of medical care for other health conditions, as well as of financial stress, should also be considered in future iterations.

Overall, the RDM proves to be among the tools that offer the opportunity to perform a stress test of a spectrum of reopening strategies even under deep uncertainties. It is similar to other decision-making under deep uncertainty (DMDU) approaches in its goals, but takes a different route.

A survey of these differences and the contribution of DMDU approaches to models of infectious disease would help build on existing knowledge and improve decision-making both now and in future pandemic threats.

This news article was a review of a preliminary scientific report that had not undergone peer-review at the time of publication. Since its initial publication, the scientific report has now been peer reviewed and accepted for publication in a Scientific Journal. Links to the preliminary and peer-reviewed reports are available in the Sources section at the bottom of this article. View Sources