All politics is local. Infections are local too. In the Covid-19 epidemic it makes sense to consider some sort of management of social behaviour, with attention to the local character of quarantine.
- Use information about disease status and risk.
- Insulate the vulnerable (elderly and diseased) by stricter rules.
- Allow a controlled gradual re-introduction of the virus as its own vaccine, so that the economy can recover.
Update 2020-04-07: (1) Sake de Vlas and Luc Coffeng had the same idea of gradually re-introducing the virus, see their preprint of April 1 2020. They use 10 regions of each 1 million people, with transport of ICU patients between them. I am thinking of hospitals and their own smaller service areas, and quarantined patches within those, so that transport is normal. There will be similar numbers of re-infected nationwide, and a same period of 16 months. De Vlas & Coffeng do not yet mention the death count of the strategy. (2) Paul Romer suggests daily testing of randomly 7% of the population, when those tests are available, and also considers daily randomly allocating 50% of the population to lockdown when there are no tests yet (and self-quarantine if you get symptoms). The latter are ways at containment / suppression of a homogeneous population while I would tend to look at isolation / immunity and ways to scale up ICU capacity for a heterogeneous population.
Update 2020-04-19: RIVM has reported more than 3,500 deaths now, as officially attributed to Covid-19, but also a rise of untested deaths overall. Official deaths below the age of 50 are 23, which corroborates below numerical setup. Table 3 is slightly adapted to better fit with table 2.
The idea of colour-coding for the status of infection is rather natural. Fruits and flowers already use the gimmick.
China now has this app: (1) green = travel relatively freely, (2) yellow = home isolation, (3) red = confirmed Covid-19 patient who must be in quarantine. (Guardian) I had suggested this type of coding in 2004 for HPV that had no vaccine at that time yet – see this working paper – though this particular virus soon got a vaccine.
The last weeks I have been reconsidering the coding scheme. There are more groups to deal with. It is also better to use red for the quarantine borders between the different groups. My present suggestion is this overall scheme:
The reasoning behind the code is to use the natural combination of R, G and B as much as possible. This is the overall scheme.
The following is an example map. It contains two hospitals: one for the vulnerable and uninfected and one for the infected and infectables. The example has a geographic layout, but one would start out with a functional allocation, and then see how it would work out geographically. One would suppose that the homes of the elderly, or at least the entries and exits, are surrounded by red barriers, and guards checking who crosses the boundary. Basically one would try to identify larger regions that have the same code, so that people can move freely within their region. Indeed, villages that are unaffected would set up village border patrols again. Make visits by appointment: not only the doctor and hairdresser but also schools and cafes. Below calculation shows that a cohort of less-vulnerable (young and non-diseased) persons can be of size of 270,000 persons, such that if the virus goes rampant in that cohort then their allocated hospital would still have enough ICU beds to serve the severe cases amongst those less-vulnerables. It is an option to deliberately infect such cohorts, since for the less-vulnerables the virus has little effect and is basically a vaccine for itself (though it can hurt). For Holland, such a scenario would take 16 months. See the discussion below.
Why would we consider such colour coding of the disease status ?
Findings by the Imperial College Covid-19 team
The Imperial College Covid-19 team estimates that R0 =3.87 from a study on European nations, with an Infection Fatality Factor (IFF) 0.66 (that they call a “rate”). For the UK they calculate an IFF 0.9. Their underlying estimates using data from China and the Chinese conventions at hospitalisation are not at their website but originally at the medical archive and now (shorter) in The Lancet. Their “impact paper” by Neil Ferguson et al. (March 16 2020) sums up the current message on Covid-19:
- Best is suppression of all infections down to the capacity of the health system, and there-after start with surveillance with tracing and isolating infections, for the whole period till we have a vaccine, which may take 12-18 months. This scenario fits with keeping R[t] < 1. Taiwan likely gives the best practice for containment / suppression of Covid-19. It will be quite a challenge for the health care system to set up such a surveillance system, e.g. requiring at least five investigators per suspected infection.
- Worse is the idea of mitigation, i.e. to try a combination with trying to get herd immunity. This plays with the idea that R[t] > 1 so that eventually a large majority of the population builds up immunity. This would come with a huge risk of overburdening the health care system. However, the Imperial College team has not yet considered an approach of structured quarantines, see below.
Perhaps not all arguments have been mentioned why Covid-19 will be with us for at least a year, and why it is best to plan for at least two years. The reasons are rather standard from an introductory course in infectious disease:
- getting a vaccine takes time … and then give it to seven billion people on the planet
- health care is under strain, will perform less well, thus allowing breaches and ever newer infections
- the virus already shows mutations and is likely to continue to do so: every new mutation would require a swift reaction for new containment – but the health care system already is under strain.
We can expect waves of new infections, like with the flu, but then 10 times more infectious / deadlier than the flu, with the risk of shorter intervals because of faster mutations. The Northern hemisphere now benefits from the upcoming Summer, but in Autumn the reduced health because of the common cold and flu will combine with Covid-19, causing increased joint mortality. This upcoming Summer should rather not be wasted.
Hammer and dance
What the Imperial College proposes as “Containment” is called “Hammer & Dance” by Tomas Pueyo. Pueyo employs a slightly different terminology. He compares “Do nothing” with “Mitigation” (towards herd immunity) and “Containment / suppression“. He advises the latter for the USA, and presents his variant as a “hammer and dance” approach, which is still the Imperial College proposal:
- first contain / suppress by lockdown to the level of health care capacity (likely not in a constant state of emergency),
- and then follow the Taiwan and South Korea model of slow release but suppression by trace and quarantine of new infections.
In Pueyo’s graph, observe the distinction between the horizontal axis and the capacity of the health system just above it (the height of the green curve in the “ongoing” phase). Pueyo also points to the epidemic calculator by Gabriel Goh.
Containment and Hammer & Dance are not enough
Given the expectations above, the Imperial College “suppression” and Pueyo’s “Dance” are dubious. There is no way how we can currently reduce the number of infections down to the manageable level except by lockdown. Perhaps the USA and Europe with more resources can try to copy the Taiwanese example but can they really, and what about other regions ? Given that the world is not Taiwan, the world now has only the option of lockdown, continued hammering, and this has nasty effects:
- In lockdown, there are the “collateral” deaths of persons who would normally receive care but who remain untreated. The Dutch Volkskrant newspaper reports that 40% of normal hospital care has been cancelled. (a) People with a disease are vulnerable to Covid-19 and may fear a greater risk of infection within the health system itself. (b) Health care resources are reallocated from normal care to Covid-19 related cases. The latter is rational, given that untreated infections are a risk for the whole population. Not treating infected cases (by hospitalisation or self-quarantine with supervision) creates such risk. We must compare the collateral deaths to the “avoided deaths by treating those with infections“. PM. See my earlier weblog about the value of life.
- In lockdown, the economy is severely affected. Bankruptcies would strain the legal system. The government currently tries monetary, financial and tax arrangements, but real production would collapse, and more money chasing fewer goods will mean rising inflation and the need for price controls. Thus, we are getting a war time economy. For Holland, production already goes from expected positive growth of 1.7% to -1.2%, a loss of at least 2.9% of GDP, or some EUR 24 bn of EUR 800 bn – see the CPB-scenario’s of 2020-03-26. Each death has come along together with a loss of EUR 30 million. See also Richard Baldwin at VoxEU on preparing for the second wave.
The Imperial College team and Pueyo present the “mitigation / controlled infection” scenario as too risky with too many deaths. One tends to agree with them, except for above nasty effects that they don’t actually discuss.
If “mitigation” comes with a degree of control then there are some aspects that are worth considering. When Covid-19 is relatively harmless for a large low-risk group, it can work as its own vaccine. The objective of this present weblog is to show a way how to enhance control: by better identifying and handling of the various quarantine categories.
Dutch data about Covid-19: using the 70-79 group as the Canary in the Mine
The following uses data of April 2.
The Imperial College estimates give problems for the Dutch data. With 121 deaths in the Dutch 60-69 age group, the London age-specific IFF gives 5500 infected in the population while their “symptomatic cases per hospitalised” gives 7663 symptomatic cases in the population, which is too much since we are assuming that the flu season is over. Holland has 29 hospitalised children of age 0-9, and the London symptoms / hospital ratio for this group gives 29000 symptomatic children in the Dutch population, which would create panic if true. Looking the issue over, I cannot find a match. It must be remarked that the Dutch “reported number of cases” is rather useless, because of the lack of tests, and their preferred application to medical personel rather than patients. Also the death count is understated since non-hospitalised deaths are not tested. See Table 1 below.
However, the Dutch 70-79 age group may be used as canary in the mine. The number 2951 of “reported cases” will be accurate for this group, since they do not belong to medical personel. These patients will have some symptoms (like “feeling really sick”) and not be tested for nought. The reported number of 2951 means only 0.19% of the whole age group. The Imperial College IFF for this group gives an estimate that 8137 would be infected, or a share of 0.005346 or 0.5%. We arrive at the problem that we are not in the steady state. Either these elderly “infected but non-patients-yet” have a stronger immune system or they are due to arrive at the hospital at a later moment. With lack of other information, we can still presume that this is the overall prevalence of infection (haves and have-beens) in Holland. When we apply this prevalence to the whole population, then we get age-group specific ratios of hospitalisation and IFF that show the same pattern as in China and the London research group. Especially relevant is the “hospitalised per infected ratio” (H/I). See Table 2 below. NB. This uses IFR and CFR, namely as “rates” while it actually are factors IFF and sCFF (symptomatic case fatality factor).
The intermediate conclusions are:
- On April 5, Holland has about 92.391 infected persons, or 0.5% of the population. The reported number of cases by RIVM is 16% or 1/6 of the true number. The current IFF for Holland is 1.4% because of the high share of elderly people (with comorbidity). If you are younger than 60 then your IFF is 0.04% (4 basispoints) and for 60+ it is 5.5%.
- If 100% of the population would get the virus then there will be 20,945 deaths younger than 60 years and 197,896 deaths in the 60+ group. The normal deaths in 2019 were 151,737 persons. Life expectancy would roughly reduce by 1% – till there is the vaccine.
A scenario with a Dance with Managed Quarantine
Table 2 in the last column (at the bottom RHS) has the option of using the health care capacity for the coming two years:
- The group of 60+ is put under quarantine, so effective that at most 1% of them gets infected. This would cause the death of 1979 persons in that group. (Actually, it will be wise to also include the younger diseased in this vulnerable group, see below.)
- The group younger than 60 years is put under quarantine, but also: step by step exposed to the virus, in cohorts of size 271,360, using the virus as its own vaccine, so that eventually herd immunity at 75% of this group is attained (using R0 = 4 and 75% = 1 – 1/R0). This would cause the death of 15,709 persons in this group. (But there will be less deaths if we shift the younger diseased.)
- It would take 16 months to achieve this. By that time, there ought to be a vaccine, and the vulnerable people in the population can be vaccinated, while the less-vulnerable people already will have achieved herd immunity for their section.
More detailed calculations are in Table 3 and this excel workbook. I suppose that the population still will grow a bit. Blue letters and figures are parameters that can be adjusted. The other colour coding is taken from above. For this calculation, the vulnerable group consists of 60+ and the younger diseased, so that the protected group counts 5,000,000 people. Implications are:
- A vulnerable person who gets infected anyway (slips through quarantine – the 1%) needs 3 weeks of ICU time, while a less-vulnerable person takes 1.5 weeks of ICU time. Given the required loads of service, 328 covid-beds serve the vulnerable and 1272 covid-beds serve the less-vulnerable. There are still 800 beds for non-covid ICU cases, allocated to the different quarantine areas. The parameters can be adjusted to different values, and then the scheme might take a different number of months.
- In this scheme, the number of deaths is reduced from above 218,841 to 19,661 (a bit different from above rougher 17,688) over a 16 months period. The normal death count is about 150,000 per year, so this rises with 10% per year over a period of 16 months. The causes for the much lower death toll are: (a) the strict protection to the vulnerable group, (b) the shift of the group of younger vulnerables from the younger group to the vulnerable group.
- While the less-vulnerable group would be the “economicially productive” group of society, they could restart their business in two manners: (1) first under the quarantine of “unsymptomatic and untested” (Cyan) group (with restricted number of contacts), then a pause for the phase of cohort infection and recovery (Blue, for some Magenta and some death (black, not shown)), then (2) secondly as recovered and likely immune and no longer a carrier (Green) (with restricted number of contacts even when no carrier, when it is not clear what the recent contacts have been).
- For an evaluation, we need an estimate of how many “collateral lives” would be lost, if we do not restore some normality.
For people and goods crossing borders, testing is important (not only whether one carries the virus but also whether one once did). Such tests are now in short supply, and when they come in supply then the priority is for the health system. Overall, they would be important for the “dance” phase. However, with this scenario, they will also be important for the checking of the quarantine boundaries and the management of the deliberate infection of the cohorts.
(Table 3 not shown anymore since the one in the next weblog entry is better.)
Overall, it seems possible to start up the economy again in the beginning of May, without the risk of another wave of infections, provided that society finds a way to manage and control the states of quarantine.
NB. For an evaluation, we need an estimate of how many “collateral lives” will be lost, if we do not restore some normality.
PM. See the excel workbook for details and references to authors who inspired this kind of calculation. The CBS Statistics Netherlands StatLine tables do not provide five-year age groups of January 1 2020 yet, and beware that there are different subgroups of 95+.
Disclaimer. Limited earlier experience in research on infectious diseases
In 2002-2004 I collaborated at Erasmus Medical Center on the modeling of the Human Papilloma Virus (HPV) as the cause of cervical cancer. My background in modeling and also logistics was relevant because diseases may look like a Markov logistics process with stages and transition probabilities. There can be the same issues of test reliability, criteria of lives-saved or life-years-gained, and cost-effectiveness of screening and treatment. I also followed the discussion about the SARS epidemic of 2003. My period at Erasmus MC was too short to allow for publishing peer reviewed papers but let me mention two working papers.
- Working paper 2004: Modifying behaviour with a passport. At that time there was no HPV-vaccine yet. An option was to manage human behaviour. The status of infection can be recorded in the medical dossier: free (green) or carrier (red). While children can gets warts, an assumption might be that children start out uninfected by the harmful HPV variants (status green). When couples meet and want to get into a serious relationship – in the sense of sharing their germs – then they can show each other their status of infection in their medical dossier and discuss the implications. From this working paper, we may take the idea of recording the status of infection, and using colour coding for clear communication. For Covid-19, it is better to use “red” (alarm, or hungry in Chinese restaurants) for the barrier between zones and groups.
- Working paper 2003: On the value of life. This compares the lives-saved and life-years-gained measures, and develops a compromise: a “unit-square-root” measure, that regards each life as 100% and takes the square root of the relative gain. This is discussed in the former weblog entry.