The process of combating the COVID-19 pandemic has brought to light a number of complicated choices that stakeholders have had to make along the way, often with a lack of clarity on what the right choice is. Protecting livelihoods vs. saving lives, to lockdown or to test faster, which vaccines to equip the population with, which segments to immunise first, and so on. We go through some of these and offer our own hypothesis answers to each, supported by facts and data to suggest their validity. Notwithstanding, each remains only a suggestion and caveats do apply.
Speed of immunisation: without a doubt immunisation should be done as rapidly as possible. For Pfizer-weighted portfolio countries, the safe immunisation threshold to reducing case growth and case fatality rate (CFR %) seems to be ~25 per 100 persons. For others, the results are still not entirely credible. While AZ seems to not reduce infections / case transmissions as effectively, Moderna, on the other hand seems to not reduce CFR % as much.
Choice of vaccine portfolio: while still not fully conclusive, there is some evidence to suggest that a Pfizer-centric portfolio is the most robust to immunise populations. Many European countries have been able to reopen their economies while keeping case growth low and CFR % in gradual decline, each deploying a high proportion of Pfizer in their portfolios. Meanwhile, UK, US, Brazil, for example have high proportion of non-Pfizer vaccines, and while each has reached high immunisation rates in their population, they fail to exhibit equivalent results. The only caveat at this point is that not enough is known at present on the efficacy of any of the vaccine options against the Delta variant.
Prioritising immunisation of ages <18: arguably given the ~100x difference in case fatality rate between >60 years vs <18 younger ages should not be prioritised until older adults are fully immunised. However their potential as carrier vectors of the virus, should lockdowns be lifted and schooling is to resume, means that they are likely to carry equal risk to others, and therefore should be prioritised.
Efficacy of SOPs (i.e. NPIs): our analysis is inconclusive on this matter, but there is reason to suggest that the apparent hysteresis noted in the transmission-mobility correlation could in fact be due to inherent tracking errors in the measurement of mobility - given general low hysteresis % of ~+/-5% and quantised nature to which mobility is measured to the nearest percentage point from baseline of 100%. As such, SOPs beyond those involving physical distancing, for example sanisation of surfaces, mask-wearing, and temperature checks have shown little to no measurable impact in reducing transmission.
Long 2-week lockdowns: viable alternatives make this quite a senseless approach to curtailing infection spread. As background, the two week window is assumed to break the transmission chain given the expected equal amount of time an infected person remains infectious after contracting COVID-19. However mobility data clearly shows that while these lockdowns do reduce aggregate mobility, the method is simply too porous - (i) not only can one ensure robust discrimination towards curtailing mobility of the infected, but (ii) continuous porosity over the period also would allow infections to spread even if an infected person were to only leave their home 1 day out of 14 - imagine this happening rotationally across different infected persons within the 2 weeks. Alternatively, several countries like Austria, Turkey and Ecuador have tried evening and weekend curfews with reasonably good success, while others have noted distinct reduction in transmission during major public holidays. The verdict is: breaking chains even for a short time effectively is better that milder forms of lockdow for a longer period of time. And that is without factoring in the economic impact of locking down.
Improving testing speed and availability: in our opinion faster testing only helps if test results are used to more objectively determine next actions on continued or alternative NPIs. From what we have observed this is not often the case. Nonetheless, keeping test speeds to <10 days and a positivity rate of <5% seem a sensible target to keep to, in order to ensure at least a good handle on the on-going situation. Test speeds should also be observed and contrasted across smaller geographies in order to inform more granular deployment of test resources for example. What is perhaps even more crucial is acknowledging the existence of the test delay (or testing speed in days) in the first place. Decision-makers seem unable to objectively match SOPs and lockdown measures, with they expect visible changes to show in daily case data.
Reopening international borders: given the prevalence of new variants of concern (VOCs) such as Beta and Delta variants, border travel should be non-negotiable and scrutinised with an extremely critical eye. Israel for example had totally locked down their borders during most of their immunisation phase, and so did New Zealand; in fact they are likely to remain closed until 2022. The consequences from the Delta variant entering a country is far too dire from both a livelihood and lives standpoint; observe what happened in India for example. The potential impact from necessary lockdowns, and increased cases and deaths from having to deal with the Delta variant should far outweigh any potential upside from allowing international travellers into the country.
Genomic surveillance: useful but not the only way to monitor for looming dangers. From a clinical standpoint it is useful to understand which strains have made their way into the population, particularly where VOCs are concerned. However monitoring case trajectory, when disassociated from the effects of mobility, provides a reasonably accurate measure of new strain prevalence and their onset by date in a given geography. Comparing dates of strain prevalence as reported by covariants.org or nextstrain.org, with the same determine through empirical fitting of transmission curves, there is reasonably good agreement between the two. The added advantage of the latter is its reliability in countries where seroprevalence testing is not easily accessible or in widespread use.