The value of a life in the time of Covid-19 – A moral and economic dilemma

By Hagop Kantarjian, M.D.
Nonresident Fellow in Health Policy, Baker Institute

 

Introducing the dilemma through the lens of Covid-19

The course of Covid-19 is morphing from an acute pandemic (as happened in Spain, Italy and New York) into a simmering slow-motion “endemic” threat that may wax and wane over the course of months or years. It will continue until there are effective and safe vaccines or antiviral drugs, or until herd immunity develops, through exposure or vaccination, in a large proportion of the population. We might also benefit from “black swan” events that reduce the spread of Covid-19 or its severity: warm and humid weather; a more rapid than expected exposure rate that reaches herd immunity; or subsequent waves with weaker forms of the virus.

Europe and the U.S. have been afflicted with large numbers of confirmed cases. But Covid-19 is, or may already be, spreading undetected in poorer countries because of inadequate health care conditions, lack of testing and underreporting. Once it starts surging in these vulnerable geographies (Africa, India, Brazil, Mexico, emerging nations), it might once again become an overwhelming event due to their dense populations in capitals and cities, and, in many cases, rudimentary health care facilities, crowded multigenerational households, poor hygienic conditions, and circumstances where the need to work and make a living supersedes the need for anti-Covid-19 measures.

As Covid-19 continues to claim lives and economies without a foreseeable end, it raises an important question: How does a society balance the need for social restrictions and precautions, in an effort to save as many lives as possible, while protecting civil liberties and economies? This can be framed by two emotional slogans at opposite ends of the spectrum: “every life is priceless,” and, quoting Patrick Henry, a founding father of the United States, “I know not what course others may take, but as for me, give me liberty or give me death.”

A thoughtful discussion of this dilemma requires considering several factors: 1) How many lives will be potentially lost to Covid-19 with or without social preventions? 2) How will the imposition of extreme measures to reduce the viral spread affect short and longer term economies, and what are possible unanticipated long-term damages? The fundamental question is: What is the value of a life saved or lost in comparison to the financial and other costs of earlier or delayed reopening of economies?

The impact of subtle variations in wording the discussion

Words matter. The choice of particular words and how we frame controversies may unite or divide. Let us analyze the emotional impact of different words that describe the same conditions. The “value of a life saved” implies a willful, at times heroic, intervention. This is what drives physicians to consider every life as “priceless.” The “value of a life lost” implies perhaps an accidental or unavoidable loss. Both are viewed through the lens of the individual. Using the “value of a statistical life” changes the discussion from personal to societal.

Do the circumstances surrounding a loss of life influence the debate?

Is the life of a 3-year-old child more valuable than that of a 70-year-old person? If a mother who ignores Covid-19 preventive measures to pursue her work and livelihood —  or to rally for a right to liberty — is infected and spreads Covid-19 to a stranger or to a family member, who then dies from the virus, might she assess her actions or the deaths through different tragic lenses? Is the life of a poor person worth the same as that of a wealthier or more “important” one? Using the gross domestic product (GDP) of different nations as a reference point, is the death of a person in the U.S. or Europe more valuable than that of one in a poor nation? The U.S. paid about $2,000-$5,000 for each accidental civilian life lost in Afghanistan and Iraq. Libya paid $8 million for every life lost in the Lockerbie terrorist attack.

The concept of the “value of a statistical life”

Various government agencies (Environmental Protection Agency, Food and Drug Administration, others) have proposed the concept of the “value of a statistical life” (VSL) to estimate how much people are willing to pay to reduce their risk of death. Using this measure also eliminates emotional elements related to age, wealth, risk of a profession (soldier, health care worker, firefighter, pilot), and individual versus societal considerations. This may seem like a cold and impersonal description, but it allows the assignment of an equal value to a life in a particular geography. With many caveats and assumptions, experts have placed the VSL in the U.S. at about $10 million. Simply stated, to prevent one death in a million Americans, every one of these 1 million Americans is willing to spend $10, or a total of $10 million. An important question is whether this American standard is applicable in geographies with lower GDP per capita. The value of a statistical life is about $4 million in Australia and New Zealand, $72,000 (range $40,000-$2 million) in Russia, and $500,000 in Turkey.

The value of a life saved through the lens of physicians

Health care workers and physicians may have a unique vision of the value of a life saved. The Hippocratic Oath mandates that physicians protect their patients “from harm and injustice” at any cost. “First do no harm” is forged into the medical lexicon so deeply that one might forget that treatments come with a potential risk of harm. This risk may be infinitesimal, say an allergic reaction to a life-saving antibiotic necessary to treat a pneumonia, or higher, say a stem cell transplant that may cure a patient with leukemia but that comes with a possible mortality of 10% to 30%. Hence, the need for consent forms that explain that the aim of a treatment is perhaps, more accurately, to “first do no net harm.”

Encoded into doctors’ DNA throughout their training, and perpetuated through their careers, is how precious life is. This immediate reflex to attempt to save the life of every patient is sacrosanct. Life is priceless. They are rarely, if ever, faced with moral dilemmas, as is the case during wars, mass shootings, natural disasters (floods, hurricanes, fires), or in the recent situation of medical infrastructures overwhelmed with Covid-19 patients (who gets a ventilator if not enough are available). For example, should a doctor, in the process of evacuation in a hospital during Hurricane Katrina, have allegedly euthanized the most desperately ill patients on ventilators who could not be evacuated, once electricity ran out, hoping to alleviate unnecessary suffering?

The ethos of the Hippocratic Oath is what drives, in the Covid-19 discussions, the vast majority of infectious disease experts to favor recommendations that will minimize mortality, while at the same time realizing that these measures — enhanced hygiene, social distancing, lockdowns of nonessential commerce when indicated, widespread testing, contact tracing and isolation, and cautious reopening only when PCR nasal swab and serology testing are broadly available — are the  same ones that will maximize economic losses.

Value of a statistical life through the lens of economists

Unlike physicians, economists evaluate two potentially conflicting mandates: saving lives and saving economies. While physicians’ concerns are more immediate and relate to the patients in front of them, economists’ views are more long-term and consider that destroying economies may lead to a cascade of unforeseeable consequences, particularly in poorer countries. These may include hopelessness and despair, suicides, extreme poverty, famine, social unrest/upheaval, mass rebellions and revolutions, armed conflicts, breakdown of civil society, widespread crimes, and others. All these could result in more deaths over the next decade than the number of lives saved by Covid-19 early interventions. But these potential long-term consequences are influenced by numerous variables, many dependent on local geographies, national gross domestic products, geopolitical structures, and others. In order to measure the impact of early versus later reopening of economies, a set of modeling parameters needs to be agreed upon, in order to calculate the ultimate toll of Covid-19.

Estimating the lives lost from Covid-19 and assumptions

A word of caution: Forecasting the future with Covid-19 is akin to predicting the outcome of a chess game that lasts 40 moves, when we are still in the fifth move. There are still too many changing variables: timing of availability of effective vaccines and antiviral drugs; infectivity rate with precautions; effect of warm and humid weather; recurrence of the virus and whether it will be less or more virulent; current true exposure and mortality rates; and whether exposed individuals acquire short- or long-term immunity.

For example, using a mortality rate of 0.5% or 1% could double or halve the total number of deaths or dollars saved or lost. Also, implementing some forms of protection (hygiene, PPE) without an economic lock-down could change the overall estimates.

This is what leads to different recommendations by experts along the geopolitical spectrum.

Early Covid-19 models dealt with multiple unknowns, assumed wide ranges of variables, and borrowed information from previous pandemics. The 1918-1919 Spanish influenza infected 500 million worldwide (one-third of the 1.5 billion world population) in three waves, the second one the most virulent, and caused the death of about 50 million people — a mortality rate of 10% among exposed people and of 3% of the world population. We do not know whether herd immunity was triggered at an exposure rate of 35% or whether the virus was eradicated by warm seasons and by mutations. Also, there were no vaccines, antiviral medications, antibiotics or modern medical infrastructures to reduce mortality.

Covid-19 appears to be highly infectious. Every infected individual can infect 5-7 others without precautions, and 2-3 with precautions. This is termed R0 (or R naught). With this level of infectivity, it is estimated that herd immunity will happen once 60% of a given population is exposed or vaccinated. Currently, 3.8 million people worldwide are known to be clinically infected (0.05%) and 260,000 have died (mortality rate 6-7%). These numbers are rising rapidly and can be extremely misleading. Recent studies suggest that as many as 80% of people infected with Covid-19 are asymptomatic. Widespread testing in New York and Sweden indicated that up to 25% of people had been exposed. In regions with wide testing, the estimated mortality rates are 0.5% to 1%, much lower than in the U.S. and worldwide. If correct, then many more people (5 to 10 times more than estimated) may have been already exposed and are asymptomatic, which will require a revision of all projections, including projecting an estimated mortality rate of only 0.5% to 1% from Covid-19 (still 5 to 10 times higher than that of seasonal influenza).

Early models to predict the course of Covid-19, by the CDC, assumed an infection rate of 1%-20% and a mortality rate of 0.25%-10%. With a world population of 8 billion, using these parameters and without precautions, the total number of infected (not exposed) people would range from as few as 80 million to as many as 1.6 billion, and the number of deaths would range from 200,000 to 160 million. Therefore, the early models were not useful in developing long-term strategies.

With our knowledge today: 1) herd immunity triggers at 60% exposure; 2) the mortality rate is 0.5% to 1% ; and 3)  the estimated exposure rate  would be 20% to 40% by the end of 2020 (or after August 2020) without serious precautions, and 10% with serious precautions.

The unique approach to Covid-19 in Sweden

Early on, Britain and Sweden adopted a different approach to Covid-19, allowing the viral spread in order to accelerate herd immunity. Facing mounting casualties, Britain soon revised its course. Sweden then adopted a hybrid approach of protecting  vulnerable and older people, but allowing a natural spread of the disease among the younger population, in whom the mortality rate is estimated to be very low. The reasoning there is that preventive measures to “flatten the curve” will not reduce overall mortality, but will spread the same number of deaths over a longer period. Thus, acquiring herd immunity earlier will protect the economy of the nation. Currently, the mortality from Covid-19 is 2-10 times higher in Sweden than in other Nordic countries, mostly among the migrant population (25% of Swedes; possibly due to housing densities). But Swedish experts reported that about 25% of Swedes have already been exposed, and anticipate the 60% herd immunity to be attained within weeks.

Using the logic of the Swedish model, the total number of deaths in the long run is the same with or without preventive measures, and so the damage caused by a long-term economic shutdown is not worthwhile, because it does not save lives. This argument may not be valid, since preventive measures may allow enough time for the discovery of effective vaccines and/or antiviral medications, and for the possibility that Covid-19 may abate in the summer and come back in a less virulent form, or not at all. To clarify this: If vaccines or drugs are discovered before the year’s end, or if Covid-19 is permanently suppressed by August (effect of warmer weather), then many deaths incurred so far in Sweden would have been avoided. Conversely, most of the anticipated 2 million lives lost in the U.S. in the next two years would be saved.

The Swedish model may ultimately replay as the natural scenario in several geographies. This is already unfolding in Brazil, and may happen in poorer areas (Africa, India, Mexico, others). Brazil has not implemented protection or Covid-19 testing measures and is rapidly becoming a new Covid-19 epicenter. Some poorer countries may not be able to afford the necessary Covid-19 precautions due to population density, overcrowding, poor sanitation and hygiene, and rudimentary health care. They may become theaters for an unmitigated course of Covid-19. This would be similar to the Swedish experiment, with additional positive and negative assumptions: There are no protections for vulnerable and older individuals (higher mortality), but these nations have a lower average age (lower mortality); and hygiene and medical measures are sparse (higher mortality). In these geographies, the consequences on lives and livelihoods would then be contrasted with geographies that implemented protective measures.

Let us now put forward some estimates to guide us

Herd immunity is estimated to require that 60% of a population be exposed. This may take a shorter period of time without precautions (Sweden), and up to two years without vaccines or drugs. If vaccines and drugs are available by the end of the year, Covid-19 exposure by then may be 20% to 40%. An average mortality rate of 0.75% will be used in all calculations.

Let us assume a cautiously optimistic scenario in the U.S.: 1) The number of those exposed is about 5% to 10%, much higher than currently estimated; therefore, some herd immunity is already slowing the viral spread. 2) With human ingenuity and massive investments in scientific infrastructures and research, effective vaccines and treatments may become available before the end of 2020. Remdesivir, an anti-Covid-19 drug, was approved by the FDA within a record time of only two months; and, of about 100 vaccines under investigation, five or more are already in clinical trials (one is expected to be available by September 2020). 3) Covid-19 may be suppressed in warm, humid weather and either may not recur, or may recur in a less virulent and contagious form.

With the maintenance of reasonably serious protective measures tailored to the pattern of cases in particular regions, the optimistic view is that, by the end of 2020, 10% will be exposed, and 0.75% of them will die. In this view, the total number exposed worldwide will be 800 million, and the total number of deaths will be 6 million. In the U.S. (330 million), the number of exposed will be 33 million, and the number of deaths 247,500.

As of today, close to 1.3 million Americans have been “clinically infected” (infection rate 0.4%; with minimal testing) and 73,000 have died (mortality rate 5.6%). Both figures may be significantly different (exposure rate 5 to 10 times higher; mortality rate 5 to 10 times lower). Assuming the current rate of infections per unit time and a mortality rate of 0.75%, the overall number of exposed people by July-August (current doubling time of 15 days) would be about 12 million to 16 million, and the number of deaths 90,000-120,000+ (could be as high as 200,000+ with accurate reporting of all cases). A February 2020 report from the Imperial College estimated the number of deaths in the U.S. to be about 2.2 million without precautions. Therefore, preventive measures are already working. However, the Washington University reputable modeling reference, used by many experts, just doubled (May 4, 2020) its estimated total number of deaths in the U.S. from about 72,500 to 134,000 by August 4, 2020. This may still be an underestimate, since there is frequent speculation about underreporting, and the number of deaths in the U.S. is already 70,000.

A pessimistic scenario would assume that no vaccines or drugs are developed in two years, and that Covid-19 might spread continuously until herd immunity develops, with an exposure rate of 60% worldwide. This would be a total of 4.8 billion people exposed. Even with a very conservative mortality rate estimate of 0.5%, the total death toll from Covid-19 would be 24 million (36 million if we use a mortality rate of 0.75%). This is still lower than the estimated total deaths from the Spanish influenza epidemic (50 million), but in our modern times, it could be a devastating loss of lives and economies that would change the world as we have known it.

Some consequences may be unforeseeable. For example, the Russian defeat in the Russo-Japanese war, together with the 1917 cholera epidemic may have contributed to the 1917 Bolshevik revolution in Russia. The bubonic plague (Black Death; Yersinia pestis) of 541-750 AD may have precipitated the downfall of the Byzantine Roman Empire.

Mercifully, the Covid-19 pandemic, together with lower oil prices, may have forced the quieting of conflicts in Yemen and Syria, both proxy battlegrounds for regional powers (Iran, Saudi Arabia, Russia, Turkey, and the U.S.). How will the combination of Covid-19 and low oil prices affect the longer-term geopolitical structures of these regions? Will they cause social unrest, mass upheavals and changes in these societies, or will they allow consolidation of existing powers? How will this impact the geopolitical situation in Brazil? Will it result in social unrest, more poverty, famine and death, the emergence of a military government or a socialist one, or both in sequence? How will Covid-19 impact the longer-term situations in the two most populous nations, China and India? Time will tell who will come out as winners and losers in this 40-move chess game, which is still in its 5th move.

Applying the assumptions to the value of statistical life

Here, we will apply the information outlined above (acknowledging that worse scenarios would justify longer economic shutdowns) and translate assumptions into numbers.

Let us start with the U.S., and use a 60% infection rate (until August 2020) and a mortality rate of 0.75%, without Covid-19 precautions. The total number of exposed people would be 198 million, and the number of deaths 1.49 million.

Reducing the number of deaths to 247,500 with precautions (330 million x 10% exposure x 0.75% mortality) would save about 1.32 million lives. This would be worth spending $13.2 trillion (using the $10 million value of a statistical life for the U.S.).

But to compare apples to apples, we should assume the availability of vaccines and drugs in both scenarios (no protection versus full protection) before the end of the year. With acceptable protections (hygiene, social distancing) but without economic shut-downs, the natural exposure rate in the U.S. would be about 20% to 40%, not 60%. Therefore, the number of Americans exposed to Covid-19 would be 66 million to 132 million, and the number of deaths 500,000 to 1 million. The number of estimated lives saved with severe protections would be 500,000 to 1 million, which would be worth a $5 trillion to $10 trillion.

Using the same apple-to-apple comparison for the world (exposure rate 20% to 40% by the end of the year, rather than 60%, and a mortality rate of 0.75%), reducing the number of infected people worldwide from 1.6-3.2 billion (20% to 40% x 8 billion) if Covid-19 is left relatively unchecked, to 800 million (with social restrictions), would result in reducing the total number of deaths from 12 million-24 million to 6 million, saving a range of 6 million to 18 million lives. If we apply the U.S. $10 million “value of statistical life,” this would be worth $60 trillion to $180 trillion.

One might argue that a life in a particular geography should be valued based on its GDP. Let us use the hypothetical value of statistical life translated to the average world GDP per capita, about $14,000. This would translate into a value of statistical life one-fifth of that in the U.S., or $2 million. So worldwide, saving lives from Covid-19 would  then be worth $12 trillion to $36 trillion.

As the U.S. government and states mull over the reopening of economies, the domestic costs of continued mitigation efforts have begun to mount. A preliminary analysis conducted by the Congressional Budget Office (CBO) suggests that second-quarter 2020 real GDP will decline by roughly 12%, an annualized rate of about 40%, and the unemployment rate may reach 14% in the near term, with continued job losses and a prolonged recovery expected to last years. The macroeconomic impact will extend beyond those values projected by the CBO. One conservative estimate places the domestic price tag at around $7 trillion per year. Even that calculation does not account for the far-reaching extent of the damage, including business closures and the economic losses related to educational disruptions.

Given the tradeoffs between loss of life and economic loss related to pandemic mitigation, some economists and researchers have recently incorporated epidemiology models into standard macroeconomic models to evaluate this tradeoff and recommend a course of action. For example, Glover et al. (2020) use this general framework to show that partial mitigation should extend through July 2020. Farboodi et al. (2020) use a similar framework to show the social distancing measures never need to be extremely restrictive, though, absent a vaccine, the mitigation efforts should be in place for an extended duration — possibly years. Both studies claim that optimal mitigation efforts imposed by the government can begin easing in the near term, but suggest different timing for a full return to normal activity.

Summary

Even when assumptions are made as to the value of a statistical life, the calculations result in a good value for continuing reasonable preventive measures (not necessarily economic lockdowns) until more optimal Covid-19 conditions prevail. These conditions include 1) optimal infrastructures for Covid-19 testing, contact tracing and isolation, and 2) a reduction in the number of daily new infections in large geographies (1+ million) to low-double digits or a low hundreds number (depending on one’s political convictions), over at least a seven-day period.

Finally, it is worth noting that even subtle variations in the estimates of exposure or mortality rates, and of regional estimates of the GDP per capita and of the value of a life saved, can sway the argument for restrictive or more relaxed measures, depending on individual levels of comfort and political-economic convictions. This is what is already happening in different states in the U.S., and in different countries, depending on the geopolitical inclinations of their leadership.

 

Hagop Kantarjian, M.D., is a medical oncologist and a nonresident fellow in health policy at the Baker Institute. His opinions do not reflect those of his institution affiliation.