Feature article: Toward an exit strategy [2] – Effective flows of people and confirmation of vaccination history –

At present, the fifth wave has been brought under control in Japan and the infection situation is relatively stable. On the other hand, it is still necessary to consider measures that allow economic activity to recover and prevent infections spreading again in the medium to long term.

In particular, it will be essential to take preventive measures against the spread of infections in situations where the flows of people in the community are expected to increase, so consideration of those measures, their scale and the timing of their implementation is required.¹.

On this AI simulation project, multiple researchers have been using various methods and models to analyze scenarios with similar assumptions in order to consider the situation from multiple perspectives. (* Link about the COVID-19 AI simulation project) As a common result of these analyses, it is suggested that a vaccine-only measure may require a vaccination rate of 85% of the whole population to prevent infections from spreading again, and if children aged up to 12 years are not vaccinated, a vaccination rate in excess of 90% for people aged 12 years and older may be required. On the other hand, these figures cannot be achieved easily given that age limits have been established for vaccinations and the fact that some people cannot be vaccinated for various reasons1. In addition, as vaccine infection prevention effects, etc., decrease over time2, booster vaccinations are required to maintain those effects even supposing that 85% of the whole population is vaccinated. So apart from vaccination, what kinds of measures are possible? We grasped the present correctly and considered the measures required.

※It should be noted that the following are the results of calculations and simulations by researchers, and not the official opinion of the national government.

Why did the fifth wave of infections come under control so quickly even though flows of people did not decrease that much? The concept of “effective flows of people” exists as a candidate answer to that question, which sounds like a paradox.

With past spreads of infections (third, fourth and fifth waves), it was possible to confirm the phenomenon of the effective reproduction number rising in association with increased flows of people. On the other hand, a different trend could be seen in the latter half of the fifth wave (from September), when the vaccination rate increased, and although flows of people tended to increase, the effective reproduction number did not. It is thought that the main reason for this is that while the proportion of people with IgG antibodies due to vaccines increased, on the other hand, the flow of people without antibodies taking actions with a high risk of infection was suppressed effectively by things like media reports of pressure on the medical system.

It is important to sort out what you are referring to when you say “flow of people” like here. For example, the flow of people at a major station from morning to evening is a flow of people that occurs in association with business activities such as commuting (here referred to as “commuter flow”), and this does not decrease except on weekends and major consecutive holidays (Year-end and New Year holidays, Obon holidays, Golden Week, etc.) unless measures like telework expand dramatically. However, when the actual data is looked at, it can be understood that commuter flow decreased uniformly to a certain degree in 2021 compared to the second half of 2020. At Shinagawa Station, for example, commuter flow in the latter half of 2020 was 50% of the level prior to COVID-19, but was about 60% lower in 2021 and so about 20% lower than the latter half of 2020. In addition, a decrease of about 10% to 20% was observed during the transition from the prevention of the spread of infections to the declaration of a state of emergency.

On the other hand, flows of people in downtown areas at nighttime (here referred to here as “nighttime flow”) are highly likely to involve high-risk behavior because it is assumed to correlate with meals, karaoke, and going in and out of facilities subject to the Act on Control and Improvement of Amusement Business, etc. However, it needs to be kept in mind that the high-risk behaviors and locations are not the same with the more infectious Delta strain and earlier variants.

After the declaration of the state of emergency in response to the fifth wave, nighttime flow decreased gradually, by 20 to 30% from before the declaration and by about 70% compared to before COVID-19. Before the fifth wave, there was a sharp and significant decrease in flows of people before and after a declaration of a state of emergency came into effect, but there was no sudden decrease after the declaration in response to the fifth wave and flows decreased gradually down to the Obon holidays in mid-August.

Figure 1: Flows of people around the main stations and downtown areas
From, Office of the Governor for Policy Planning, Tokyo Metropolitan Government, “State of crowding in downtown Tokyo and changes in the number of people present”
(/action/survey/covid19-ai.jp/ja-jp/presentation/2021_rq3_countermeasures_simulation/articles/article140/)

The same changes as commuter flow were apparent in the decreases in the latter half of 2020 and 2021 and during the transition from preventing the spread of infections to the declaration of an emergency, with the size of the decrease tending to be slightly larger than commuter flow. This is more evident when looking at data on people present by time zone (Figure 2). For example, if you look at the change from 6pm to 8pm in contrast to the change in daytime population from 12 noon to 6pm, it can be seen that the number of people present was almost the same in the latter half of 2020, but in 2021, the decrease in people present from 6pm to 8pm was large. This is marked in particular during the period of the declaration of a state of emergency.

Analysis of the relationship between these changes in flows of people and the spread and decrease of infections is shown in Figure 3. In this analysis, the correlation between changes in flows of people at stations during the third, fourth and fifth waves and the effective reproduction number is visualized. From this, a certain correlation can be seen during the infection spread phase of the third and fourth waves between flows of people and the effective reproduction number. However, in the decreasing phase, there is an inverse correlation between flows of people and the effective reproduction number. It is thought this may be due to flows of people returning while continuing to see a decrease in infections. If analyzing the relationship between flows of people and the spread of infections without distinguishing between the spread phase and the control phase, this kind of correlation trend may not be found. In order to analyze the relationship between “flows of people” and the state of infections, detailed analysis is essential, including classification of whether it is during a period of spread or control, and whether downtown population is during the day or at night. The fact is that infections do not occur randomly and the risk of infection increases if people take high-risk actions. Therefore, it is important to ascertain not simply how many people are moving, but how many people are moving in places and at times of high-risk behavior. Moreover, we have found that in the Kinki region, flows of people and the spread of infections correlate strongly in the north and south of Osaka and around the major stations, but in Tokyo, this correlation is relatively weak. This raises the possibility that the places at risk of infection are more diverse and dispersed in Tokyo than in other areas. In addition, a change has also been seen from the period when frequently occurring cases of infections clusters were mainly during eating and drinking, and at hospitals and facilities for the elderly, etc., to places where children and students in the age groups that have not been vaccinated gather (childcare centers and schools). In other words, ascertaining the qualitative aspects of flows of people is important.

As for increases in flows of people during periods when infections come under control, in the past, it was thought that such flows of people would lead to the next spread of infection and that is actually what happened. During the fifth wave too, a correlation between the flow of people at major stations and the effective reproduction number was observed during the period of spreading infections following the expansion of flows of people from the control phase of the fourth wave, but an increase in the effective reproduction number not correlated with a change in the flow of people was observed during the period the Olympics were held. This is thought to have been the result of specific changes in the quality of behavior during this period.

In addition, since July, vaccinations have started for working people aged up to 65 and the proportion of people at high risk of infection for reasons such as being unvaccinated among people going out has decreased rapidly. In order for infections to spread, the presence of people who can be infected is required and the flows of such people are known as “effective flows of people.” Because the number of people with high IgG levels after vaccination whose infection prevention effect is being maintained is increasing, and the quality of flows of people has changed greatly, it is necessary to focus on the “effective flow of people” (the flow of people made up of people with a risk of infection). In other words, the presence of people who can be infected in places where the risk of infection is high is the effective flow of people.

In Japan, vaccination of working generations began in earnest from the beginning of July and was the fastest in the world at a peak of 1.7 million vaccinations a day. When these effects are considered, it is thought that the effective figure for high risk nighttime flow decreased almost the same as it did under the first declaration of a state of emergency (Figure 4). In other words, the observation that “although flows of people are not decreasing, infections are decreasing” in the phase of the fifth wave when infections were decreasing is superficial, and it is thought that nighttime effective flow, which appears to be related strongly to infections, was dramatically reduced.

The introduction of confirmation of vaccination history has advanced in various foreign countries as a means of controlling “effective flows of people.” In the UK and Singapore, which introduced confirmation of vaccination history earlier than other countries, the phenomenon of decreasing effective reproduction while the overall volume of human

In the case of London (Figure 5), the effective reproduction number was kept below one while maintaining flows of people due to the introduction of confirmation of vaccination history for travel and bars, etc., which were likely to be the origin of the spread of infections. This tendency was also observed in Singapore, and the effectiveness of the measure is recognized.
It is thought that confirmation of vaccination history, which can encourage changes in the quality of flows of people like this, is effective as a complementary measure that prevents the spread of infections while allowing economic activity to recover in a situation where vaccinations are not possible for various reasons, sufficient effects are not obtained after vaccination and the effects of the vaccine decrease over time.

A technology demonstration using the vaccine-test package, etc., was started in Japan from October 2021, allowing the relaxation of restrictions due to infection control measures by confirming either vaccination history or test results. How could the introduction of the vaccine-test package change our lives?

In terms of the spread of infections, complete infection protection is not possible even after vaccination. It has become clear that with the Pfizer + BioNTech and Moderna vaccines, the ones mainly used in Japan, very high infection-prevention effects of 90% or more are maintained from about two weeks to three months after the second dose, but they then decrease after that. Because of that, there is also a view that tests should be carried out at the same time as confirmation of vaccination rather than just the latter.

The focus of discussion here is mainly to protect people who cannot be vaccinated, people in whom adequate immunity has not been established even after vaccination, and people in whom the effects of the vaccine have weakened. It can be assumed that the onset or worsening of disease will be suppressed effectively in people who have been vaccinated and established adequate immunity, but it is necessary to control infection itself to protect people who cannot be vaccinated and people in whom adequate immunity cannot be established despite vaccination4. In addition, if people in whom immunity has been established as a result of vaccination are infected, known as breakthrough infection, it will be necessary to wait for future investigation of rates of increased severity of illness and sequelae at the stage when vaccine effects begin to decrease in particular. Because of this kind of situation, it will be necessary for the time being also to control the number of infections, rather than just aim to control the frequency of increased severity.

The following analysis (Figure 6) is a simulation (assuming the delta strain) assuming that a state of emergency will not be declared even if the number of infections increases and that flows of people have returned to intermediate level between the pre-COVID-19 and crowd level of last year. It has been pointed out that if vaccine-test packages are not

Figure 6: Tokyo: Number of new infections in Tokyo in the case of 0% (top) and 50% (bottom) self-restraint rate among unvaccinated people (analysis by Professor Kurihara, Keio University)
Related research: RQ4: “Proposal of human behavior model and infection simulation based on SNS and news report data,”
/action/survey/covid19-ai.jp/ja-jp/presentation/2021_rq3_countermeasures_simulation/articles/article169/

In addition, effectiveness is also shown in the results of simulation if the declaration of a state of emergency was lifted on October 1 with a maximum vaccination rate of 80%. In this simulation, the effects of the vaccine-test package were analyzed assuming a 25% increase in the number of people present in downtown areas at 7pm after lifting of the declaration. In view of the increase in flows of people during the year-end and New Year holidays, there are concerns that infections may peak again in January, but it is estimated that the peak can be kept to about half by the introduction of a vaccine-test package.

Figure 7: Analysis of the effect of booster vaccinations (booster shots) and a vaccine-test package (certification) (analysis by Professor Kurahashi, University of Tsukuba)
Related research: RQ4: “Estimates of effects of the date of lifting of declaration, increase in flows of people, vaccine effect, booster vaccinations, vaccination rates, and vaccination certificate restrictions (Tokyo),”
/action/survey/covid19-ai.jp/ja-jp/presentation/2021_rq3_countermeasures_simulation/articles/article142/

In this article, we have used simulations to consider how to grasp the “flows of people” that have an impact on infections and countermeasures against them. We added consideration of the possibility of a vaccine-test package to control “effective flows of people” and the quality of flows of people for economic recovery and infection control.

For more precise simulation, it will be necessary to introduce factors such as attenuation of infection prevention effects dependent on the period after vaccination and the probability of false negative due to the test method, and this is a future issue.

At the same time, consideration of the period of validity of vaccination certificates will be required together with the accumulation of scientific and medical data. For example, in Singapore, the period of validity is one year from 14 days after the second dose, but this will be reviewed along with the accumulation of data5. Our knowledge of COVID-19