3.Closure of schools
4.Closure of nonessential businesses
5.Travel restrictions
6.Orders to stay at home or shelter in place
Interestingly, data from the 1918 influenza pandemic also show that, when social distancing measures were relaxed after the first wave of infections subsided, the second wave of infections led to higher peak death rates in cities that had imposed social distancing earlier. This was presumably because of a greater proportion of susceptible people in those cities. One important difference between 1918 and the present times is modern medicine. There was little medical treatment that could be provided to sick patients in 1918. During the COVID-19 pandemic, physicians quickly optimized treatment strategies to reduce fatalities. So, after relaxation of social distancing measures and the expected increase in cases, optimized clinical treatment can reduce fatalities even if there is still a larger number of susceptible people in the population.
Later in the chapter, we will return to how mathematical models can be used to suggest ways to mitigate the effects of waves of the epidemic. But first let us ask whether social distancing can completely extinguish an epidemic.
Can Social Distancing “Crush the Curve”?
If strict social distancing is imposed for a long enough time, the value of Reff would drop below 1, and the epidemic would be extinguished. Assume that public health measures were such that no one could leave their homes for any reason. Food would be delivered to the door by government workers who would not have any contact with you, and the same would be true for any drugs needed for chronic conditions. Imagine that at the start of imposing such measures, 5,000 people were infected. With these measures in place, at most they would infect people who shared a home with them, but the cohabitants thus infected would not infect anyone new. So, over time, the Reff would drop below 1, and ultimately the epidemic would wane. This is the situation that is implied by the phrase “crush the curve.” Is it possible to achieve this?
Very strictly enforced and intense social distancing measures are said to have crushed the curve in Shenzen and Wuhan in China. After Reff falls below 1, the number of infections starts to fall, but it takes a while for the virus to be extinguished. Crushing the curve requires enforcing strict social distancing measures for a long time and forbidding travel in and out of the region. It seems that crushing the curve is a strategy that can work, but might not be practical in all settings.
Weathering the Storm
Social distancing measures can flatten the curve to save lives during a spreading epidemic and, imposed in extreme form for a sufficiently long time, can extinguish an epidemic. But social distancing measures also have many adverse consequences. The most obvious cost is the economic downturn caused by shutting down businesses and the concomitant loss of jobs. For example, during the first months of the COVID-19 pandemic, 40 percent of US households earning less than $40,000 per year lost their jobs. Such economic carnage and the loss of citizens’ livelihoods can have many ripple effects that include depression and addiction.
Given these serious concerns, and the fact that we do not know exactly how efficacious each of the social distancing strategies noted above are, some have suggested that a better strategy may be to just weather the storm and let the wave of infections pass through. A related strategy, “shield,” temporarily shuts down society to allow the isolation of those most at risk for serious illness and death, such as the elderly and the immunocompromised. Restrictions are then relaxed, and the less vulnerable weather the storm. The hope is that people would naturally react to a pandemic by being vigilant, thus not allowing the number of infections and deaths to grow so fast as to overwhelm the healthcare system. Ultimately, the number of infections would die out and things would return to normal. Indeed, before the advent of modern medicine and therapeutics, all infectious disease epidemics ended this way. But why does an epidemic end by weathering the storm?
As the number of people who recover from disease and become immune to infection increases, the proportion of the population that is susceptible to infection drops. To illustrate how this works, let us consider a specific location where the social network is such that an average person interacts with 100 people during an infectious period of 10 days. In the beginning of the epidemic, all 100 people are susceptible to infection. As people become infected and recover, a proportion of the population becomes immune and is no longer susceptible to infection. Suppose we reach a point where 95 percent of the population is immune. Now, only five out of the 100 people an infected person interacts with during the infection could get infected. If the value of R0 for a particular viral infection at this location is 2, instead of infecting two people, an infected person would only infect 2 × (5/100) people. The Reff in this situation would be far less than 1. So, once a sufficiently large proportion of the population becomes immune, the epidemic ends over time because an infected person is very unlikely to infect anyone else. Even a susceptible person has a very low chance of being infected. When this situation prevails, the population is said to have acquired “herd immunity.”
Knowing the proportion of the population that must be immune so that herd immunity is acquired is important. If this proportion is very large, acquiring herd immunity takes a long time because many people have to become infected. If a significant proportion of people infected with the pandemic-causing virus require hospitalization and die, many would die during this time. Let us estimate the proportion of the population that must become immune for a population to acquire herd immunity.
Consider a viral infection characterized by a particular value of R0, and let us denote the proportion of the population that is immune as pI. For example, if 50 percent of the population is immune, pI = 0.5. If a person interacts with 100 others during the infectious period, as in the example we considered above, only 100 × (1 − pI) people that a person encounters are susceptible (50 people). So, instead of infecting R0 people during the infectious period, an infected individual would infect a proportionately smaller number, which is equal to
This is the value of Reff now. So, if 50 percent of the population is immune (pI = 0.5), then Reff is half of R0 (Reff = R0 × 0.5). Herd immunity is acquired when Reff falls below 1. So, Reff (R0 × (1 − pI)) must be less than 1, or, equivalently, pI must be more than 1 – 1/R0. When the proportion of the population who have recovered from the disease exceeds 1 – 1/R0, the population acquires herd immunity. So, if R0 is 2, pI must be more than one half; that is, herd immunity is acquired when more than 50 percent of the population has recovered from disease.
Because the proportion of the population that must be immune for herd immunity to be established depends on R0, what it takes to acquire herd immunity depends on the virus and on local conditions. If the R0 is high, such as 18 for measles, the magic number for herd immunity to be established is 1 – 1/18, which is about 0.94. So, roughly 94 percent of the population has to be immune. Highly infectious viruses, like measles, are very difficult to control by hoping that natural infection will lead to herd immunity because it requires essentially everyone in the population to have been infected. Vaccines play a key role in such cases by immunizing a large number of people such that herd immunity is achieved without natural infection with the virus.
The MMR vaccine, given to children, protects against measles, mumps, and rubella, all viruses with high values of R0. Almost all states in the United States require this vaccine for preschool and elementary school admission because without herd immunity many children would be infected and some would die. Measles is not a benign virus. It has a mortality rate of about 0.2 percent in the United States, similar to influenza, and infection can also cause permanent neurological damage. Herd immunity acquired by vaccination programs is an important reason for the sharp decline in childhood mortality in the twentieth century. Maintaining herd immunity requires vigilance as outbreaks will occur if the proportion of vaccinated children falls below that required for herd immunity. This is why
, for example, measles outbreaks occur frequently in communities where people do not vaccinate their children. When this happens, vulnerable children with compromised immune systems (e.g., due to chemotherapy), who are protected from infection by herd immunity, are at risk again.
If the value of R0 is not too large, herd immunity can be established by natural infection. In 2016, the Zika virus began to spread from Brazil to North and South America. Infection caused a fever, joint pain, and a skin rash, but up to 80 percent of infections were asymptomatic. Sadly, many babies born to women who were infected during pregnancy had a brain defect called microcephaly. The Zika epidemic was declared a public health emergency by the WHO. Steps were taken by public health officials in many countries to mitigate the spread of Zika, including the development of a vaccine. However, as 2016 drew to a close, new infections started to wane and the epidemic died out. Testing for Zika antibodies showed that in parts of Brazil and El Salvador, hot spots of the epidemic, more than 60 percent of the population was infected. The estimated R0 for Zika is between 2.1 and 2.5, so herd immunity is acquired when 52 to 60 percent of the population is immune. So, naturally acquired herd immunity may have controlled the epidemic.
For SARS-CoV-2, estimates of the value of R0 vary between 2 and 3. If the value is 2, herd immunity would be acquired when half of the population is immune. For an R0 of 3, herd immunity would be acquired when two thirds of the population is immune. During the COVID-19 pandemic, the United Kingdom initially seemed to be on the verge of adopting a “weather the storm” policy. But when mathematical models predicted a scale of infections that would completely overwhelm the healthcare system and cause many deaths, the government reversed course. Sweden, however, decided to essentially weather the storm. They did not lock down the country, nor close schools for younger children, and recommended only personal social distancing measures. By September 2020, the percentage of deaths in the Swedish population was higher than in neighboring Scandinavian countries, about the same as the overall rate of deaths in the United States, and substantially less than in New York City. Antibody testing suggests that Sweden has not attained herd immunity. Perhaps this is not surprising. It took a few years for the plague to pass through Europe, and more than two years for the 1918 flu to extinguish itself, largely without mitigation. It remains unclear how high the death toll in Sweden will be before herd immunity is acquired naturally or by vaccination. Comparing different strategies that have been used across the world will provide valuable knowledge for mitigating future pandemics.
Many factors, including cultural differences resulting in better compliance with public health measures and self-imposed social distancing, which cannot be anticipated by epidemiological models, may explain why Sweden was able to control infections better than anticipated without a lockdown. Similar factors may have been important for the way that Japan controlled the pandemic. This is also an opportune point to emphasize again that parameters used in models for one location may not be appropriate for another.
During the COVID-19 pandemic, in many countries, such as the United States, social distancing imposed during the first months of the pandemic flattened, but did not crush, the curve. In the United States, as expected, infections did increase as social distancing rules were subsequently relaxed. While, as of September 2020, there is no evidence that herd immunity had been acquired in any country, it may have been achieved in some neighborhoods in New York City. This suggests that in some parts of the world where infection rates were high, the spread of the virus is slowing.
Is a “weather the storm” strategy the only alternative when social distancing measures are relaxed? This strategy could be dangerous. Data from the 1918 influenza pandemic showed that cities that imposed stronger social distancing measures initially experienced a bigger second wave. After the first phase of an epidemic, how does one try to balance the needs of keeping the economy and some semblance of normal life going without overwhelming the healthcare system, which can result in avoidable deaths? Epidemiological models can explore various scenarios and make qualitative comparisons that can guide public officials in this regard.
Acquiring Herd Immunity by Intermittent Social Distancing
One strategy that could be considered for mitigating subsequent waves of infection if the curve was not crushed and herd immunity was not acquired is intermittent imposition of social distancing measures. This strategy aims to keep the economy and normal life going as long as possible without overwhelming the healthcare system. For epidemiological models to explore the consequences of such a strategy and how it should be implemented, the values of various parameters that go into the SEIR models discussed earlier must be known or assumed. For example, one would have to know or estimate the extent to which different social distancing measures influence Reff. The difference in values of Reff with and without imposition of social distancing measures is related to the difference in the parameters that determine the rate at which susceptible people are exposed to the virus (S + I → E) and subsequently become infected (E → I). Many other variables are also important. For example, for how long are people who have recovered from the disease immune? What is the seasonal variation in Reff? Is it larger in the winter, as it is for influenza? With clearly stated assumptions, which are important to know, epidemiological models can offer some useful insights.
As just one example of such an insight, we describe projections made in April 2020 by Grad, Lipsitch, and colleagues (epidemiologists at Harvard University) about how intermittent social distancing might affect the COVID-19 pandemic under different scenarios. These epidemiologists assumed that Reff cycled between higher values in the winter and smaller values in the summer. The trigger for imposing social distancing measures should be when new cases start to rise and exceed a threshold value. This threshold value would be set to prevent overwhelming the capacity of the healthcare system. When the number of cases declines below this threshold, social distancing measures would be relaxed. As time ensued, more and more people would be infected and become immune, and so the population would progress toward herd immunity. An interesting effect of healthcare capacity was predicted. If hospital capacity was increased, the threshold value of cases when social distancing measures are turned on could be higher. This is because the healthcare system would be less easily overwhelmed. Thus, periods with no social distancing measures could be longer. This would enable the population to acquire herd immunity faster, and keep normal life and the economy going longer. As more people became immune, the duration between imposition of social distancing measures would get longer as Reff would be smaller and fewer people would get infected. Ultimately, when the number of new cases becomes small enough, testing, isolation, and contact tracing will be sufficient to control the epidemic.
These types of projections are very useful for guiding public policy. But without real-world data, the predictions cannot be quantitatively accurate. For example, to make models like this predictive, we need extensive testing to know what proportion of the population is immune. We need immunological studies to determine whether immunity declines and how long it lasts. Many other parameters are also needed, some of which can be difficult to measure.
Massive amounts of data are being acquired across the world about the COVID-19 pandemic. At the same time, we have enormous computational resources today. Modern machine-learning approaches could mine the data being collected to obtain an understanding of pandemic control that far exceeds anything we have had before. Perhaps, in the future, this understanding will allow us to develop models that can help find public policies that optimally control hospitalizations and deaths to acceptable levels while not causing other kinds of damage to society. Thus, we can hope to be more resilient when the next pandemic comes.
The quickest way to end a pandemic that is difficult to control is to have an effective drug that cures the disease, or a vaccine that provides herd immunity. We now turn to these two topics.
6 Antiviral Therapies
&nb
sp; In 1981, five young gay men in the United States developed pneumonia caused by a fungus called Pneumocystis jirovecii. Previously, pneumocystis pneumonia was an extremely rare disease that developed mainly in immunocompromised people. The cluster of cases affecting otherwise healthy young men suggested that their immune systems were compromised, thus reducing their ability to fight off infections like pneumocystis pneumonia. This was because a new virus, HIV, that infected immune cells had entered the human population.
Almost 40 years later, we still cannot cure HIV, but therapeutics have been developed that allow HIV-infected people to live relatively normal lives. In 1995, a combination of antiviral drugs was developed that can keep the amount of virus in an infected person at very low levels. Before these drugs became available, a positive test for HIV was essentially a death sentence.
The approaches used to develop antiviral treatments for HIV forged a path that has led to successful development of treatments for other viral infections. In this chapter, we will explore the strategies that are used to develop drugs that can treat viral infections.
Identifying the Virus
The first step toward developing therapies that can treat a disease caused by a new virus is to identify the causative virus. This is because antiviral drugs interfere with the virus’s ability to replicate in humans and cause disease. We have to know what type of virus it is in order to know how it replicates, and so how to interfere with these mechanisms. Identifying a virus was a formidable task until the middle of the twentieth century because they are so tiny. In 1949, John Enders, Frederick Robbins, and Thomas Weller transformed the study of viruses when they figured out how to make poliovirus multiply and grow in animal cells in test tubes. Enders and colleagues used these methods to grow, identify, and study viruses, including those that cause measles and mumps. This work directly led to vaccines that protect against the childhood diseases caused by these viruses. Enders, Robbins, and Weller would be awarded a Nobel Prize for their achievements. Others quickly adopted their approach to growing a virus in the laboratory making it much easier to identify and study viruses.
Viruses, Pandemics, and Immunity Page 11
