Early in the Omicron wave in South Africa some local doctors were commenting that they were seeing many “mild” cases of Omicron, quite different from what they had seen with Delta in the previous wave in the country.

Soon there were two competing takes on Omicron in the public discourse, one that concludes that Omicron is more like a bad cold and another that the true virulence of Omicron, even if it was somewhat lower, would matter less than the larger number of cases from its rapid spread. “A small percentage of a very large number can also be a large number,” is how many of the experts cautioned us.

The picture was further clouded by the fact that Omicron was spreading in parallel with Delta: Many of the countries that were seeing Omicron waves were already in the midst of a Delta wave connected to the onset of winter in the northern hemisphere. For example, Japan had cases growing about 20 percent week on week for several weeks before it saw significant numbers of confirmed Omicron cases. Many of the deaths seen while Omicron cases were already growing rapidly were still Delta cases.

Death numbers always lag a few weeks behind case numbers. For example, in Tokyo the average gap between Covid diagnosis and death is about 15 days. Then it takes another 16 days on average for the death to be reported in the evening news. This means, the death count reported in the media was from people who on average died 16 days ago after having a positive PCR test 31 days ago. The positive result of the PCR test would have been included in the daily count on average about 2-3 days after the test was taken.

There is some evidence that Omicron is less virulent than Delta. A study on hamsters showed that the variant seems less capable of infecting lung cells and seems to be more focussed on the nose, throat and windpipe, which may also contribute to its easier spread. It’s not clear how these animal study results translate to humans. In December, different numbers were reported for how much less virulent Omicron was overall or specifically for unvaccinated people, with risk reductions mentioned anywhere from 15 to 70 percent.

In any case, milder than Delta does not mean mild: Delta was worse than Alpha which was worse than the D614G strain that devastated Italy and New York city in March 2020, which was worse than the original Wuhan strain. Omicron may still be as bad as one of its pre-Delta ancestors, which were deadly enough.

Finally, we have some numbers coming in.

During the Delta wave the UK 7-day trailing count of deaths had reached a peak of around 1,200 in early November 2021, when almost 100% of cases were still Delta. On January 14, 2022, when most UK cases were Omicron, this had increased to about 1,800 deaths in the last 7 days, about 50 percent more deaths per week than at the peak of Delta.

The picture is similar in Canada, where 7-day deaths peaked at 344 in October with Delta. On January 15, 2022 the same metric reached 606, almost twice as deadly, with Omicron.

While the Case Fatality Rate (CFR), which is defined as the number of deaths divided by confirmed cases, has declined with Omicron relative to Delta, this is primarily because of the extraordinarily rapid growth of cases, which grew faster than deaths climbed. The CFR went down even as deaths increase. Thus Omicron can appear to be milder and not so mild after all, as it ends up killing more people per week than Delta did. If cases go up 5-fold but deaths “only” double, the CFR will fall by 60 percent.

The only saving grace may be that Omicron will most likely burn through the infectable population more quickly than Delta ever did. Cases have peaked in South Africa, the UK and most of the US. By February they will probably also peak in Japan (before I and most others can get a booster here).

Wear a good mask, if you meet people do it outdoors, ventilate, get vaccinated and get a booster if you can.

Links:

• WHO data for new cases and deaths by country (CSV)

In December, 470,000 medical workers, 280,000 seniors (65 and above) and 90,000 others in Tokyo were qualified to receive the 3rd dose because of the time interval since their second dose and their status. Only 63,292 (7.6%) of them actually received their dose. 1,829,880 booster doses were supplied by the central government. As for January, out of 620,000 qualified individuals, 29,645 (4.8%) have received their dose so far. The scheduled supply for this month is a further 1,702,800 doses for Tokyo, for an accumulated total of 3,532,680. The number of people eligible for boosters in Tokyo by the end of February is only 2,100,000.

Also, according to the Cabinet Office (kenbetsu-vaccination_data2.xlsx file), Tokyo had received 16,586,310 million doses of Pfizer for 1st/2nd shots in 2021, of which they have used 15,358,949 doses (92.60%). Therefore, about 1,227,361 Pfizer doses should still be have been available before any of the above mentioned supplies were delivered.

Nationwide, about 10 million doses of Pfizer have been supplied but not used (165,148,620 vs. 155,989,956). Of the 50,000,000 doses of Moderna that the central government had contracted for, 34,251,400 had been distributed and 31,938,353 actually used. That leaves about 18 million doses unused, which translates to 36 million booster shots, if you add 18 million sets of syringes (Moderna boosters are half doses of regular 1st/2nd doses).

It’s not a supply problem. Based on these numbers, it’s going to be something like a mere 120,000 people with a booster in Tokyo by the beginning of February, when literally millions could have received them. Meanwhile, new cases numbers (most of them Omicron) have been increasing 4 to 5-fold week on week. We will soon hit unprecedented numbers of cases.

Without a 3rd dose, two doses of Pfizer received 5 months ago offer little protection against symptomatic infection with Omicron, though they still reduce the risk of hospitalizations and other severe outcomes. This is because of the immunity escape from Omicron. Even two weeks after the second shot, efficacy is only a little over 60% with Omicron, significantly lower than with Delta. A third shot boosts efficacy to a slightly higher level than after the 2nd shot, winding the clock back by 6 months or more. Accelerating the booster campaign by using left-over vaccine doses as soon as possible should be a high priority.

Another important point is to improve the messaging on masks: Many people are still using simple masks that cover the face but do not fit particularly well. Mask policy in Japan is still not based on the recognition that Covid-19 is airborne disease that spreads via aerosols. One person can infect another without them being in the room at the same time. High grade masks like N95, KN95, KF94, FFP2 offer much better filtration because they reduce the unfiltered side stream. They offer the best first line of defense against the explosive spread of Omicron. A rapid roll-out of boosters should be the second line.

Numbers listed for all 47 prefectures in this MHLW document:
追加接種対象者数、接種回数及びワクチンの供給量 (“Number of subjects to be additionally vaccinated, number of times vaccinated, and quantity of vaccine supplied”, mhlw.go.jp)

Vaccination data by prefecture from Cabinet Office:
kenbetsu-vaccination_data2.xlsx (kantei.go.jp)

Compared to over 100,000 new cases a day in countries like the UK and France (each with roughly half the population of Japan), 500+ Covid cases in Japan may seem almost trivial. Japan has recently experienced something of a post-5th wave honeymoon after months of falling cases, but that is set to end very soon.

As of today (December 31), the weekly average of new cases in Tokyo has been higher than the weekly average one week earlier for 23 consecutive days. In the entire month of December the 7-day average only decreased week on week on 4 days, all of them towards the beginning of the month.

What’s more, the rate of increase is going up. Before Dec 19, the highest week on week increase was 1.26. Since then it has been 1.40 or higher. Today’s weekly comparison reached 1.68, the highest since August 4. Only six days in the 5th wave had a higher weekly increase, all between July 30 and August 4, when Delta cases outgrew Alpha cases.

Is this because the more infectious Omicron strain is outgrowing Delta now? Most likely not yet! The number of confirmed Omicron cases from community spread is still small compared to Delta. That means the cause is likely to be social: People are feeling relatively safe because of the much lower case count compared to last December’s 3rd wave and also they want to do things that they will most likely not be able to do in a couple of weeks, once Omicron takes over. Sadly, that’s exactly how Omicron will take over. We’ve entered super-exponential growth, even if it’s from a low basis.

What’s the Japanese response to this? So far, no new restrictions have been imposed yet. Opportunities for free testing, regardless of symptoms, have been expanded. However, the plan for 3rd vaccine doses has been left virtually unchanged. The booster program was originally created when it was found that immunity to Delta dropped after a number of months. Japan then decided on an 8 month interval between 2nd and 3rd doses. However, it takes a much higher level of antibodies to neutralize the heavily mutated Omicron strain, even if you’re still far from 8 months. A third shot of Pfizer/BioNTech will raise vaccine efficiency against symptomatic infection from under 30 percent to over 70 percent, close to what it was for Delta with two doses. Japan has not responded to this new reality yet. The official policy on the Cabinet Office website still states that 3rd doses are due “in principle” after 8 months. That means, people who were vaccinated in August and September are not due for booster shots until April or May. There is no question that Omicron will blow up into big number in January and February already.

It may be some comfort that the number of cases in Tokyo on New Year’s Eve 2021 was only 6 percent of the cases of a year before. On the other hand, it took South Africa only 14 days to go from a weekly count of ~2,000 cases to ~33,000 cases when Omicron arrived. Do the math!

From December 2 (the first day of the booster campaign) to December 28, the latest date for which data is available, Japan has deployed 531,296 third shots to healthcare workers. That’s just 11% of the healthcare workers who received 2 doses and about 0.4% of the entire population. That means before the New Year starts, 8 out of 9 healthcare workers who will soon have to be dealing with an onslaught of Omicron cases have not yet had their immunity level doubled.

In many countries in Europe, between 20 and 50% of the adults with 2 doses have received their third dose already and yet they are experiencing tens of thousands of cases. For example, Portugal, which has only 70% of the population of Tokyo but a higher rate of two vaccine doses in its population and where a quarter of the population have had 3rd shots, counted over 17,000 new cases on December 29. Tokyo counted a mere 76 cases.

While two doses will still offer good protection from hospitalization and death, they will not prevent a rapid increase in cases. On present plans, the third shots will come too late for many healthcare workers and senior citizens.

The other lines of defense should be masks and ventilation. Most people in Japan do not yet use high grade masks such as N95/KN95/FFP2/KF94/JN95 and I am seeing little effort to promote their use over regular surgical or other masks. Likewise, there is little emphasis on preventing airborne spreading, such as installing HEPA filters, upgraded ventilation, etc.

Three members of a family hospitalized with Covid-19 symptoms in Osaka have tested positive for the Omicron variant. They are the first cases reported in Japan without a travel history or linked to someone with travel history. All other cases were either detected at immigration inspections or in quarantine or linked to someone who had recently arrived from abroad.

It must be assumed however that there are many other cases that remained undetected and spreading under the radar screen. As the Washington Post reported on December 15, the Philippines had quarantined a man on December 1 who had tested positive on arrival from Japan and who was later confirmed to have been infected with Omicron. It therefore sounds likely that there was already some community spreading in Japan three weeks ago, soon after the variant was isolated by scientists in South Africa.

During the Delta-driven fifth wave in July/August, the 7-day average of cases peaked on August 24, with the health care system in crisis. At that point the reproduction rate fell below 1. For several weeks the fall was so steep, cases almost halved every week. The most likely explanation for that was increasing immunity in the population from a successful vaccination campaign in combination with continued wearing of masks and reduced mobility (e.g. working from home, less eating out in restaurants). After two months the reproduction rate slowly increased again but as long as it stayed below 1, it still resulted in a further reduction of cases. Weekly cases reached a temporary minimum on November 9. For about a week, cases slightly increased before sliding again at a somewhat slower rate than before. Since then the reproduction rate has been gradually drifting up. Since the minimum in late November/early, December Tokyo cases have been rising from a very small basis of about 100 cases per 7 days, 99.7 percent below the 33,000+ cases per 7 days in late August. They are now around 200 cases per 7 days, a doubling in three weeks.

For the last 4 days the 7-day average increase from a week earlier has been over 40 percent, which is equivalent to a doubling every two weeks. This is not really surprising. Once cases had dropped to less than one percent of the peak, people felt safe to resume activities that they had not been able to do for a very long time, such as enjoying year end parties or having wedding receptions with guests. They felt the risk was low enough.

However, Delta has not gone away. It is still around and as people’s level of antibodies from vaccines gradually wanes, they become more susceptible to infection again, even if they remain largely protected against hospitalization or death. This is compounded by greater risk-taking when cases are perceived to be “rare enough.”

We knew that Omicron would eventually spread in the community. When it does, it will not immediately replace Delta. For a while, both will be increasing but at different rates. A level of immunity and non-medical interventions (such as mask-wearing) that is only just about adequate to control Delta will be unable to halt the spread of much more infectious Omicron. Under the current conditions where even Delta cases are increasing, we would continue seeing Delta grow relatively slowly while Omicron cases would explode. Eventually, especially as hospitals start to fill, people will modify their behaviour to reduce exposure, for example by avoiding restaurants or by using KN95/KF94 masks instead of surgical masks or cloth masks. This would then push the reproduction rate of less infectious Delta below 1 and its cases would fall while cases of more infectious Omicron would still increase but at a slightly reduced rate. That is the point from which Omicron will start to replace Delta rather than spread in addition to it.

Third doses can reduce the exposure to Omicron, especially for the older generations who were first to get vaccinated this spring and summer but Japan has not taken any measures yet to speed up the additional shots. From the beginning of December to December 21, third shots averaged a mere 10,000 per day which is less than one percent of the rate achieved and maintained for several months in the Summer.

Japan is still running its booster campaign as if its only purpose was to prevent a resurgence of Delta (like in Israel last summer) rather than to protect the population as well as possible against the immunity escape that comes with the new Omicron variant. This has to change.

I can only hope that the first official case of community transmission of Omicron will galvanize the government into action for dramatically speeding up the distribution of third shots, especially to senior citizens and people with medical conditions that make them more vulnerable.

My home state of Bavaria (population: 13.1 million) in Germany has had 30,117 new Covid-19 cases in the past 7 days, a 7-day incidence rate of 229 per 100,000. Meanwhile, Tokyo (population: 14.0 million) has had 142, a 7-day incidence rate of 1 per 100,000. The difference in numbers is simply staggering. Given that Germany started vaccinating its citizens months before Japan, it had a headstart on the road to immunity but it has since given up this advantage. A larger share of Japanese residents is fully vaccinated in every age group than in Germany. The growth in the vaccination rate slowed to a crawl in Germany months ago, while it’s still continuing at a healthy clip in Japan.

About two months ago, Tokyo’s Covid incidence (147 per 100,000 in 7 days on 2021-09-03) was quite similar to Bavaria’s current rate. But while the Covid incidence rate has been falling week after week in Japan since early September, it has almost doubled in Germany. The Japanese drop in cases has been amazingly consistent. When cases fall by half every 8 days and this continues for 8 weeks then cases will fall over 100-fold overall. That’s what an exponential decrease looks like. It happens when the reproduction rate of the virus drops below 1.

According to a numerical model created by Kris Popendorf, the two main contributors to the swings in the reproduction rate of the virus are mobility and immunity (you can read more details about his model here). One pushes the number up, the other pushes it down. The combined changes of the two either push the number above 1 (case numbers grow) or below 1 (cases drop). There are other factors, such as adherence to mask-wearing in indoor situations but in Japan’s case this has barely budged even when cases have been falling. Even when the risk of encountering infected individuals has decreased, few Japanese have stopped wearing masks as they still strive to comply with social norms.

With the end of the state of emergency, mobility has increased. People have started to go out and travel again and companies are reducing remote work. On the other hand, while rates of immunity will increase further, this increase is slowing down as the vaccination campaign will be nearing the saturation point over the next month or so. This will shift the balance of the two factors driving the reproduction rate of the virus towards the factor that drives an increase. We can therefore expect the rate of drop to slow and eventually rebound when mobility reaches a level that outweighs the level of immunity in the population. To achieve herd immunity regardless of mobility, the vaccination rate would have to reach a level estimated to be as high as 85-90 percent, which even Japan is unlikely to reach.

Case numbers in Tokyo are now at a level about 1/200 of the peak in August. Therefore even when the numbers start rebounding and grow again, there remains some time for public messaging to prevent a return to a caseload that would overwhelm the healthcare system.

I am much more concerned about the situation in Germany, which has a significant population of people reluctant to protect themselves and others by getting vaccinated. While not as large as in Russia, Romania, Bulgaria or the USA, it makes it very difficult to get numbers under control. In Bavaria, the Covid-19 incidence amongst unvaccinated people is 9 times higher than amongst fully vaccinated people (451.5 vs. 50.9). This means that even though about 65 percent of the population are fully vaccinated, the vast majority of cases are of unvaccinated people. Only 39 percent of teenagers are fully vaccinated there while more than two thirds of teenagers in Tokyo will soon have both shots. Even for age 65+ the rate is 80 percent in Bavaria vs. 91 percent in Tokyo, which means there are proportionally more than twice as many unvaccinated seniors in Bavaria as in Japan.

Unfortunately it will be very difficult to change the attitudes behind the vaccine resistance in Germany and other countries, as it is an issue of trust. Many of the people reluctant to get the shots trust neither politicians nor mass media nor medical professionals nor science in general. They will therefore be difficult to reach.

Restrictions on public activities, such as eating out or travel that are becoming more convenient again for low-risk vaccinated people will gradually erode the non-vaccinated population share but that will take time.

On September 24, 2021 Japan’s first vaccine dose rate passed Germany’s rate (67.8 % vs 67.7 %) but with a steeper trajectory still: While Germany only has a gap of 3.8 percentage points between first and second doses, in Japan it’s still 12.0 percentage points, indicating much faster recent growth. The 1st/2nd dose gap is the equivalent of the number of additional first doses over the past 3+ weeks, since 2nd doses are given 3 weeks after first doses with Pfizer/BioNTech, 4 weeks in case of Moderna and 4-12 weeks in case of AstraZeneca. Both in Germany and in Japan the vast majority of doses used (~80%) have been Pfizer/BioNTech.

This means Japan has not only matched Germany, it is still vaccinating at three times the rate of Germany and will therefore be left with far fewer unvaccinated people. Reluctance to get vaccinated is a much bigger problem in Germany than it is in Japan. Especially in eastern states vaccination rates are much lower than average. Saxony, Saxony-Anhalt, Thuringia, Brandenburg have the lowest vaccination rates, followed by the southern states of Bavaria and Baden-Württemberg and the northernmost of eastern states, Mecklenburg-Vorpommern. Not coincidentally, the states with the lowest vaccination rates also have the highest support for the far right AfD. According to opinion polls, about 60% of AfD-supporters had no intention of getting vaccinated.

Links:

• Cabinet Office vaccine numbers
• Impfdashboard (Vaccination dashboard, English version)
• Impfbereitschaft nach Parteien-Anhänger

My wife and I have received both of our shots, three weeks after the first shot. In not quite two weeks we will be fully vaccinated.

The first shot triggers an immune system reaction that produces antibodies against the spike protein of the virus. After about two weeks, the immune system also produces Memory T cells, which are its way of remembering how to make more antibodies should they be needed again in the future. After three or more weeks, when the second dose is given, it puts those newly formed T cells through their paces, boosting the antibody levels.

According to the latest published numbers for Tokyo, the more infectious Delta variant already causes more new cases than all other strains of SARS-CoV-2 combined and is most likely to completely displace them soon. This is part of the reason why the reproduction rate of the virus has been increasing for several weeks, with weekly averages of new cases growth accelerating from under 20 percent a week to over 50 percent recently.

Now it is crucially important that more people get shots as soon as possible. Not all prefectures and municipalities in Japan have made the same use of the vaccine they have received from the central government.

Tokyo has the questionable distinction as the prefecture with the largest portion of unused vaccine doses, even though people are desperate for vaccine appointments. By Sunday, July 25 Tokyo had received 15,227,660 doses. According to the Vaccination Recording System (VRS), 4,081,931 of these doses had been used as first doses and another 2,380,397 as second doses, for a total of 6,462,328 – just over 42 percent of the total. That means about 8.8 million doses either have not yet been reported after use or they’re still sitting in freezers, earmarked for shots that are weeks away.

The basic supply of future vaccines are going to be about 1.3 million doses every two weeks distributed to the municipalities in Tokyo by population by the central government and another 320,000 or so allocated to the prefecture to give to the cities that need extra supplies the most because they have the smallest stocks.

In the past week (July 19 through July 26), 8,620,790 doses have been distributed for general vaccinations in all of Japan. During the same period, 3,521,414 doses were used as first shots and 3,921,705 doses as second shots, for a total of 7,443,119 shots. The same picture shows in Tokyo, with delivered stocks growing by 2,100,180 million (13,127,480 on 7/19 to 15,227,660 on 7/26) while only 895,327 doses were used (502,489 first doses and 392,838 second doses).

This would actually mean that unused stocks in freezers grew, rather than being shrunk to vaccinate as many people as early as possible to protect them against the Delta variant. Vials in freezers do not protect against illness, only shots in arms do.

However, it is also possible that the recent 4-day long weekend (Thursday, 7/22 through Sunday, 7/25) caused bigger than usual delays in the VRS reports, in which case the actual results of vaccine use could be a little better than these numbers appear to show.

Daily vaccination statistics published by the Cabinet Office of the Japanese prime Minister suffer from lag problem: They are not reported back to a central database the day the vaccinations take place. This has the curious effect of making the published chart of daily first and second shots always trend downwards for the most recent 7-120 days or so.

There is a relatively simple way to compensate for the under-count: Its ratio is consistent between first and second doses. In Japan first and second doses for Pfizer are almost religiously spaced 21 days apart. Thus by looking at the first dose count 21 days earlier one can get a close approximation of the real second dose count for a given date. The ratio between the published incomplete second dose number and the approximation derived from the more accurate first dose count can then be applied to the 1st dose Pfizer count for the date you’re looking at. Those two numbers plus the Moderna numbers (which are not subject to the same delays because of the way the SDF sites — the main consumers of Moderna doses — are operating) then give you the real up-to-date General Vaccination count 🙂

From the current numbers we can see that first doses peaked on June 8 at around 650,000 doses day. However, second doses were still steeply climbing then and the total of first second doses exceeded 1,000,000 doses a day on June 15. On June 29, exactly 21 days after the first dose peak, second doses peaked at about 625,000 doses. For about 5 days, second doses clearly outnumbered first doses (June 28-July 2).

After July 3, first doses pulled ahead again, as many municipalities had sent vaccination tickets to people with medical conditions or younger then 65 years. The overall total seems to have largely stabilised at 1.1-1.2 million doses a day.

During July, August and September, the government will distribute a basic allocation of 8,000 boxes (9.36 million doses) every two weeks plus an “adjustment” of 2,000 boxes (2.34 million doses) that will go to municipalities with the best progress in using their allocated doses. This is pushing municipalities to update data on VRS more quickly, as this will be the metric used by the government to allocate the “adjustment” doses. The basic allocation is based on the number of residents age 12-64 in each municipality (the amounts for age 65 and above have already been fully distributed).

One potential problem with this is the fact that some local clinics have been vaccinating people who did not have a vaccination ticket with bar code yet, making this portion of the vaccination total hard to track. The doses have been allocated to the municipalities but may appear unused in VRS.