Lives saved and lost in the first six month of the US COVID-19 pandemic: A retrospective cost-benefit analysis.

In the beginning of the COVID-19 US epidemic in March 2020, sweeping lockdowns and other aggressive measures were put in place and retained in many states until end of August of 2020; the ensuing economic downturn has led many to question the wisdom of the early COVID-19 policy measures in the US. This study's objective was to evaluate the cost and benefit of the US COVID-19-mitigating policy intervention during the first six month of the pandemic in terms of COVID-19 mortality potentially averted, versus mortality potentially attributable to the economic downturn. We conducted a synthesis-based retrospective cost-benefit analysis of the full complex of US federal, state, and local COVID-19-mitigating measures, including lockdowns and all other COVID-19-mitigating measures, against the counterfactual scenario involving no public health intervention. We derived parameter estimates from a rapid review and synthesis of recent epidemiologic studies and economic literature on regulation-attributable mortality. According to our estimates, the policy intervention saved 866,350-1,711,150 lives (4,886,214-9,650,886 quality-adjusted life-years), while mortality attributable to the economic downturn was 57,922-245,055 lives (2,093,811-8,858,444 life-years). We conclude that the number of lives saved by the spring-summer lockdowns and other COVID-19-mitigation was greater than the number of lives potentially lost due to the economic downturn. However, the net impact on quality-adjusted life expectancy is ambiguous.

Introduction

In response to the COVID-19 pandemic in the US, a series of lockdowns mitigating the spread of the pandemic were introduced beginning in March and maintained into July of 2020 when the first wave of the pandemic was largely believed to have passed. With millions of lives and trillions of dollars at stake, the wisdom of imposing lockdowns to control a pandemic is controversial in both the academic literature and the lay press [1-9]. Much of the debate revolves around the tradeoff between lives saved and economic impact, a question routinely addressed through economic evaluations.

Economic evaluations of health-related policies typically compare policy-attributable health benefits and costs in monetary terms by assigning a dollar value to the lives saved or to the resulting life expectancy gained. One problem with this approach is that economic estimates of the value of a human life vary between $8-$11 million per life. Several cost-benefit calculations of the COVID-19 lockdowns estimated the monetary value of lives saved from as low as $1.15 trillion to as high as $65 trillion [7, 9–14], thus only intensifying the national debate about whether the lockdowns were worth the cost. But beyond the challenge of calculating the correct value for cost-benefit calculations, the economic approach of assigning average monetary values to a human life frequently fails to resonate with the public, including individuals worried about themselves and loved ones and with healthcare professionals who took an oath to save lives [3, 15–17].

This study offers an innovative approach to the dollars-per-life-saved conundrum by instead estimating the number of non-COVID-19 deaths potentially attributable to the economic downturn brought on by the lockdowns and other COVID-mitigating measures (regulation-attributable mortality). The COVID-19 lockdowns were not the first time that public safety and wellbeing required restrictions on personal freedoms and economic activity. For example, road safety regulations (e.g. seatbelt laws, lane width and marking codes, and speed limits) save lives but at the same time they are costly to implement and enforce [18]. Many government programs (e.g., occupational safety regulations, homeland security programs) consume taxpayer dollars, therefore diverting them from personal spending on food, housing, and healthcare [19-21]. Since a loss of income is known to negatively impact individuals’ health and wellbeing [22-32], these life-saving regulations, depending on their cost, may end up being counter-productive. For example, a study of regulation-attributable mortality of fire-mitigating programs in Australia found that, at an annual taxpayer cost of $12 billion (in 2020 US Dollars), fire regulations in Australia potentially cause as many deaths from a reduction in people’s incomes as the number of fire fatalities they prevent [33]. Importantly, these analyses of mortality attributable to government regulations do not use the value of statistical life approach—instead, they empirically derive the economic cost that would induce one statistical death based on empirical relationships between income and mortality observed in large national cohorts.

Much like other government safety regulations, COVID-19 lockdowns and other measures were put in place to protect lives, but they also led to a loss of personal income for many. Therefore, the aim of this study was to compare lives saved by the COVID-19-mitigating policy intervention during the spring and summer of 2020 to regulation-attributable mortality potentially caused by the ensuing economic downturn in the US. Our hope is to initiate a broader discussion whether a ‘lives-to-lives’ comparison could allow for assessments of the judiciousness of economic lockdowns without the distraction of normative assessment differences in the value of human life.

Materials and methods

This study is a synthesis-based retrospective cost-benefit evaluation of the US COVID-19-mitigating policy intervention in the US population of all ages during the first six months of the pandemic. Taking a societal perspective, we considered the “benefit” to be COVID-19 mortality potentially averted by the intervention. The “cost” was the economic downturn (loss of national income during the first 6 months of the pandemic) and its attributable mortality. Parameter estimates for our calculations were obtained via rapid review [34, 35] of academic literature and reports by official US government and international bodies.

Intervention

This study evaluated the COVID-19-mitigating public policy intervention as the entire complex of federal, state, and local COVID-19-mitigating measures, including lockdowns and all other measures implemented during the first 6 months of the pandemic (March through August 2020) in the US. We define lockdowns as government-induced mandatory restrictions on private activity, including closures of businesses and public gatherings (e.g., schools, public offices), building capacity regulations, and stay-at-home orders. Other COVID-19-mitigating measures included all non-mandatory public health measures including mask-wearing and hand-washing guidelines, recommendations to refraining from family gatherings, and social distancing. The comparative strategy was ‘no intervention,’ a hypothetical scenario where no COVID-19-mitigating restrictions or measures were implemented in the US during the first six months.

Health outcomes

The primary outcome for evaluating cost and benefit of the COVID-19-public health intervention was cumulative mortality, measured as the number of lives (gained and lost). Our secondary derivative outcome was cumulative life expectancy measured as the number of quality-adjusted life-years (gained and lost).

Time horizon

The time horizon was March 1 through August 31 2020, chosen because the most severe restrictions (sweeping lockdowns, shut-down of non-essential sectors of the economy) were in place during the first six months of the pandemic, and also because epidemiological models of the unmitigated US epidemic predicted it to end by September of 2020 without any mitigating interventions, thus allowing us to attribute the full potential US mortality from COVID-19 under the counterfactual no-intervention scenario to the March 1 through August 31 period [36-39].

Analysis

The number of lives saved by the COVID-19-mitigating policy intervention during the first six months of the US pandemic was calculated in two steps. First, to calculate the potential unmitigated cumulative COVID-19 mortality, we multiplied the US population, 330 million, by the COVID-19 herd immunity threshold (HIT) and by the infection fatality rate (IFR). Then, to calculate the number of lives potentially saved by the intervention, we subtracted out the observed cumulative COVID-19 mortality reported during the same period. Life expectancy gained was calculated as the number of lives potentially saved by the intervention times 5.64 years—the quality-adjusted life expectancy lost among COVID-19 fatalities [40].

Assuming cost-to-death ratio (CDR) estimates and methods from the economic literature on regulation-attributable mortality [17–21, 33, 41, 42] are applicable to evaluating mortality attributable to the COVID-19 policy intervention, the number of lives lost from the economic downturn was calculated by dividing the combined first and second quarter US GDP loss, $2.23 trillion [43], by the CDR. The CDR measures the cost of a government intervention that induces one statistical death. Dividing the full economic cost of a regulation by the CDR yields the total number of regulation-attributable deaths. Quality-adjusted life expectancy lost was calculated as the number of lives lost due to the economic downturn times 36.1 quality-adjusted life years per life lost—the 2019 average US life expectancy (78.7 years) minus the 2019 median age (37.9 years) [44, 45] multiplied by the health-related quality of life score (0.886) [46].

Due to the short-term timespan of our analysis, we did not use discounting.

IFR, HIT, CDR parameters

To obtain the infection fatality ratio (IFR) and the herd immunity threshold (HIT) parameters, we reviewed six projections of COVID-19 cumulative mortality by US and international authorities [37, 38, 47–49], a systematic review and meta-analysis study of the SARS-CoV-2 IFR [50-52], and four recent studies of HIT in the US [53-56]. To obtain the CDR, we searched EconLit, the official reference search tool of the American Economic Association. We used a free-text search using terms “cost-per/to-death,” “regulation-attributable mortality/fatalities,” “mortality/fatality cost,” and “government regulation/intervention.” We included studies that: 1) were published after 1990; 2) used data from the US, Europe, or Australia; 3) empirically estimated the CDR from large, representative cohorts or national databases; and 4) reported adjusted (not only crude) CDR estimates. Studies that used the value of a statistical life instead of the CDR were excluded [17, 21, 42]. Reference lists of the included studies were searched for additional studies meeting the above criteria. We synthesized the evidence based on the analysis method, study population, and type of covariate adjustment. If a study reported several analysis approaches, we extracted the most rigorous CDR estimate using the following criteria: 1) adjusted cross-sectional estimates were preferred to unadjusted (crude) estimates, 2) longitudinal (cohort, panel, time-series) estimates were preferred to cross-sectional estimates, and 3) individual-level analyses were preferred to ecological studies. We converted all CDRs to current 2020 US dollars by using historical data on the Consumer Price Index from the US Bureau of Labor Statistics.

Results

Lives saved

Given the evolving stage of knowledge about COVID-19’s IFR and HIT, we included a broad range of IFR and HIT estimates from the literature and calculated a low and high bounds for the number of lives saved by the intervention. For IFR, we attempted to use estimates obtained from data during the first six months of the pandemic when less was known about the disease and few effective treatment approaches existed. A median IFR of 0.23 was reported by the WHO [52] across 51 locations globally; among the 11 US states included in the report, the IFR ranged from 0.08 in Utah and 0.20 in Missouri, to 1.54 in Connecticut and 1.63 in parts of Louisiana. Overall, locations that had higher number of deaths (most of the US) had a median IFR of 0.57. A recent meta-analysis of 24 IFR studies reported the IFR of 0.68 with the 95% confidence interval between 0.53% and 0.82%, subsequently supported in an age-specific meta-analysis of 111 studies from 33 locations. The HIT is thought to be between 60–70% [53-56].

Applying0.53%-0.82% IFR and 0.60–0.70 HIT to the US population of 330 million, the unmitigated mortality (in the absence of any intervention) was 1,049,400–1,894,200 deaths. After subtracting the 183,050 confirmed COVID-19 deaths on August 31 2020 [57], the number of COVID-19 lives saved by the first six months of the policy intervention was 866,350–1,711,150 lives. Measured in cumulative quality-adjusted life expectancy instead of number of lives saved, life expectancy gained was 4,886,214–9,650,886 quality-adjusted life-years.

Regulation attributable mortality

Seven economic studies of mortality attributable to government regulation satisfied our inclusion and exclusion criteria () [18–20, 22, 23, 33, 41]. Among them, four derived the CDR from cross-sectional differences in annual mortality rates across income groups [18–20, 33], two estimated an individual-level panel regression of annual mortality on personal income [22, 23], and one estimated a time-series regression of annual mortality rate on per-capita national income over several decades [18]. Four of the studies used US data [19, 20, 22, 23], two used Northern European data [18, 41], and one used Australian data [33]. All of the studies adjusted for basic demographic factors (sex, age) and three adjusted for an extensive set of covariates, including family characteristics (e.g. marital status, children) and health (e.g. self-reported health status, disability or functional limitations) [22, 23, 41]. Five studies examined how the CDR may depend on the allocation of the cost burden of a regulation across income groups—proportionally to income (flat rate) or progressively with income (higher income groups affected relatively more than lower income groups) [19, 20, 22, 33, 41], while one study assumed proportional cost allocation [18] and one assumed progressive cost allocation [23]. The CDR ranged $9.1—$15.4 million (in 2020 Dollars) under the proportional cost allocation assumption and $16—$38.5 million under the progressive cost allocation assumption. The two most rigorous studies utilizing a panel-data analysis with adjustments for an extensive set of controls [22, 41] reported the CDR from $10.8 to $12.5 million and $16 to $17 million under the proportional and progressive cost allocations, respectively.

Regulation-attributable annual mortality from the $2.23 trillion economic loss ranged from 144,805–245,055 deaths and 57,922–139,375 deaths under the proportional and the progressive cost allocation assumptions, respectively. The attributable mortality was lower (and the CDR was higher) under progressive cost allocation because higher income individuals are less vulnerable to income reductions. [22, 24, 31]. Taking the minimum and the maximum of these numbers, we estimated regulation-attributable mortality to be 57,922–245,055 deaths. Measured in cumulative life expectancy instead of number of lives, life expectancy lost was 2,093,811–8,858,444 life-years.

Discussion

This study is the first to evaluate the societal costs and benefits of the US public health response during the first six months of the COVID-19 pandemic not in monetary terms but in terms of human lives saved and potentially lost. Our calculations suggest that the number of lives potentially saved by the spring 2020 lockdowns and other mitigating measures impacting the US economy (866,350–1,711,150 lives) far exceeds the number of lives potentially lost during the same time period due to the ensuing $2.23 trillion economic downturn (57,922–245,055 lives). However, because the majority of lives saved are those of older adults with multiple chronic illnesses whose life expectancy is shorter on average, the impact of the intervention on cumulative life expectancy is less clear (4,886,214–9,650,886 quality adjusted live-years saved; 2,093,811–8,858,444 quality-adjusted life-years lost). From an ethical perspective, a potential caveat to the quality-adjusted life-years approach is that, by design, it assigns a smaller value to the lives of older adults compared to younger adults, and persons with disabilities compared to fully-abled persons [58, 59]; racial and ethnic disparities in life expectancy have also been well-documented [60]. To avoid exercising judgement and bias about the value of a human life, the calculation based on the number of lives saved and lost might be the preferred approach.

Lockdowns and other restrictions on private activity have real humanitarian consequences that should not be overlooked. Known as the income gradient in health, the notion that a person’s health and life expectancy is in large part determined by their income is not new [22-32]. The association of income with life expectancy is the strongest among lowest-income individuals who have minimal discretionary spending to cushion the blow [22, 24]. Poverty can limit access to basic needs like transportation, health care, shelter, and clean food and water [22, 24, 31]. Even transient income fluctuations have been linked to reduced essential household expenditures (e.g. food and shelter) and loss of access to health insurance, and to negative health impacts including lower self-reported health [61-65], increased risk of cardiovascular disease, reduced brain health, and increased all-cause mortality [26, 27, 66]. A large amount of literature across disciplines evaluated health and mortality effects of the most recent global Great Recession of 2007–2009 which caused a similar magnitude of economic decline to the current COVID-19 recession [67]. The preponderance of the evidence is consistent with negative health effects on fertility, self-rated health, overall morbidity (including both new-onset and exacerbation of existing chronic illness) and increased mortality for older adults, racial and ethnic minority groups, and individuals in countries with weaker social support structures [67].

It is important to note that the 2020 economic downturn disproportionally impacted the poorest, most marginalized members of our society. According to experts, this recession could be the most unequal in modern U.S. history, as job losses are overwhelmingly affecting minority workers, younger workers, and the less-educated working-poor [68, 69]. Three months after the spring-summer lockdowns had been largely lifted, only one out of three Black Americans who lost their job during the lockdowns regained employment [68]. The disproportional impact of the recession on low-income groups suggests that the indirect mortality attributable to the lockdowns and other COVID-19-mitigating measures may be closer to the more pessimistic projections and higher end of our lives lost estimates, close to 250,000 deaths and 9 million years of lost human lifespan. It is also important to note that while the costs of the lockdowns disproportionally fell on the younger, less educated, low-income workers, the benefits are disproportionally accruing to older adults who are at the highest risk of dying from COVID-19. A better understanding of these distributional effects is needed to inform current and future policy how to anticipate and better manage public sentiment during public health crises requiring a strong government response.

Our calculations are based on the best available evidence at this time. To calculate lives and quality-adjusted life expectancy saved by the intervention, we used IRFs between 0.53% and 0.82% from systematic reviews of data collected during the first 6 months of the pandemic [50, 51]. These estimates are based on observed mortality and serological prevalence data, and therefore incorporate epidemiological and healthcare trends in COVID-19 mortality during that time (e.g., virus becoming less deadly, doctors getting more experience in treating COVID-19, medical supplies becoming more readily available). Our IFR range is considerably lower than the March 2020 estimates based on early data from the Wuhan province in China (IFR between 1.1% and 3.4%) which at the time gave rise to early predictions of potential unmitigated death toll between 2.2 and 2.9 million lives in the US [36-39]. Therefore, our estimate of COVID-19 lives saved, 0.87–1.7 million, reflects our current best understanding of SARS-CoV-2 and potential population-level mortality from COVID-19.

Our results are subject to several qualifications. First, we conducted an “as-is” ex-post evaluation of the public health intervention, including the lockdowns and other COVID-19 mitigating measures (including any voluntary behavioral modification component). This is only a first step, and future research of incremental effectiveness of different public health measures (lockdowns, school closures, travel restrictions) is needed to determine whether a more targeted, nimble, or shorter set of measures may have resulted in a more favorable net balance of lives saved versus lost. Additionally, we only evaluated the initial six months of the US pandemic; therefore, our findings have limited ability to inform decisions regarding any future use of lockdowns and other measures. Specifically, our estimates are based on infection fatality rates during the early stages of the pandemic when less was known about effective treatments and many healthcare facilities experienced shortages of medical supplies (ventilators, personal protective equipment). With effective treatment modalities and vaccines now available, the vulnerable can be protected through targeted measures limiting a need for future blanket restrictions.

We attributed all COVID-19 lives saved (relative to the unmitigated counterfactual) to the public health measures (lockdowns, social distancing recommendations, masking recommendations), even though some voluntary behavioral modifications (e.g., limiting social contacts, trips to the store, or non-essential travel outside the state) would likely have taken place among the public even in the absence of these government interventions. It is not possible to empirically isolate self-induced voluntary behavioral changes that the public would have initiated without any government policies from behavioral changes induced by government regulations through either mandates or public awareness campaigns. Inability to quantify the self-initiated (not induced by government intervention) component of behavioral modification means that our analysis may overstate the impact of the government intervention on COVID-19 lives saved, and, likewise, on the economy. However, because behavioral adaptations (whether they are government induced or not) reduce both sides of the equation (benefits and costs), it is unclear whether and how this limitation may have affected our findings on the net.

Although our study is based on best available evidence, the evidence has several important limitations. First, our COVID-19 mortality estimates are based on IFRs that were observed under a full set of public health measures during the first 6 months of the pandemic; therefore, our analysis likely understated the potential cumulative death toll under the hypothetical unmitigated scenario leading us to understating the number of lives and life expectancy saved by the measures. Second, long-term population-level morbidity and mortality among COVID-19 survivors is yet unknown. Emerging sequelae of a COVID-19 infection include damage to the lungs, heart, and brain, suggesting increased risk of long-term health problems and mortality [70, 71]. Not accounting for these long-term effects further increases the possibility that our study underestimated the cumulative mortality potentially averted by the spring 2020 intervention. Third, a salient limitation of the regulation-attributable mortality literature is that cost-to-death estimates are derived from adjusted cross-sectional comparisons of mortality across individuals with different observed incomes, not clearly distinguishing between long-term income differences and short-term income variability. Theoretically, a short term loss of income, like during the economic recession that followed the COVID-19 lockdowns in the US, may impact individuals less (or differently), compared to a persistent economic disadvantage over the course of a lifetime [72, 73]. Further, the negative impact of the economic downturn on personal incomes was partly ameliorated by the US government’ COVID-19 economic relief spending, providing assistance to workers, families, and businesses. Therefore, our study may have overstated mortality and loss of life expectancy potentially attributable to government regulation, while possibly understating lives saved and life expectancy gained. Lastly, rapid review and synthesis has inherent limitations relative to systematic reviews [34, 35].

Conclusion

Evaluated as a full complex of COVID-19-mitigating restrictions, the number of lives saved by the spring-summer lockdowns and other COVID-19 mitigation was greater than the number of lives potentially lost due to the economic downturn, while the net impact on quality-adjusted life expectancy is less clear owing to the older age and poorer health of individuals at highest risk of mortality from the disease. Moving forward, it is essential that we emerge from the pandemic with the smallest humanitarian cost from the combination of disease impact from the virus and the lost economic opportunity.

14 May 2021

PONE-D-21-09447

Lives Saved and Lost in the First Six Month of the US COVID-19 Pandemic: A Retrospective Cost-Benefit Analysis

PLOS ONE

Dear Dr. Olga Yakusheva ,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by 7 may 2021 due. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see:  http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols . Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at  https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols .

We look forward to receiving your revised manuscript.

Kind regards,

Carlos Alberto Zúniga-González, Ph.D

Academic Editor

PLOS ONE

Additional Editor Comments:

Dear, I consider that you have made a good effort to present this manuscript, but it is very important to make the improvements indicated by the reviewers. I believe that if you include and apply the concept of opportunity cost, it could help you explain the cost-benefit relationship, given the benefit-cost relationship does not clarify the cost as we know it regularly in economics or the reader is used to it, so I suggest you include the cost opportunity

Journal Requirements:

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. In the Methods section of the manuscript please provide additional details regarding the rational for the selection of literature used for the selection of IFR and Hit parameters. Furthermore, please clearly indicate the date range for the literature search conducted. Finally, please provide a justification for the selection of data used from only US, Europe and Australia, as an inclusion criteria during the literature search.

3.  In your Data Availability statement, you have not specified where the minimal data set underlying the results described in your manuscript can be found. PLOS defines a study's minimal data set as the underlying data used to reach the conclusions drawn in the manuscript and any additional data required to replicate the reported study findings in their entirety. All PLOS journals require that the minimal data set be made fully available. For more information about our data policy, please see http://journals.plos.org/plosone/s/data-availability.

Upon re-submitting your revised manuscript, please upload your study’s minimal underlying data set as either Supporting Information files or to a stable, public repository and include the relevant URLs, DOIs, or accession numbers within your revised cover letter. For a list of acceptable repositories, please see http://journals.plos.org/plosone/s/data-availability#loc-recommended-repositories. Any potentially identifying patient information must be fully anonymized.

Important: If there are ethical or legal restrictions to sharing your data publicly, please explain these restrictions in detail. Please see our guidelines for more information on what we consider unacceptable restrictions to publicly sharing data: http://journals.plos.org/plosone/s/data-availability#loc-unacceptable-data-access-restrictions. Note that it is not acceptable for the authors to be the sole named individuals responsible for ensuring data access.

We will update your Data Availability statement to reflect the information you provide in your cover letter.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: N/A

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Report for “Lives Saved and Lost in the First Six Month of the US COVID-19 Pandemic: A Retrospective Cost-Benefit Analysis”

This article evaluates the benefits and costs of U.S. COVID-19 mitigation policy interventions. In terms of benefits, the authors calculate the number of lives saved by subtracting the observed actual number of deaths from the unmitigated potential number of deaths. In terms of costs, the authors estimate the potential number of deaths due to the economic downturn caused by the intervention policies, which is calculated by multiplying the GDP loss in this period with the cost-to-death ratio taken from economic literature.

I have major concerns with respect to the methods used to estimate both the benefits and costs of the intervention policies.

Benefits:

The authors estimate the benefits of intervention policies by subtracting the actual number of deaths from the unmitigated potential number of deaths. By doing so, the authors assume that the decrease in COVID mortality in this time period is all caused by government mitigation policies, which is a questionable assumption in my opinion. First, the virus itself could become less deadly over time. Second, doctors could become more experienced in treating the disease. Third, high death rate in the initial time period could be caused by a shortage of medical equipment supply. Last but not least, people voluntarily adjust their behaviors to reduce the transmission of COVID even without any government policies. I doubt whether the methodology taken by the authors can tease out the impact of voluntary adjustments and get a clean estimate of the policy effects. Thus, the decrease in mortality is caused by multiple factors, including government intervention policies. Attributing all decrease in mortality to intervention policies will overestimate the benefits of these policies.

Costs:

The authors estimate the cost of intervention policies by calculating the potential number of deaths due to the economic downturn caused by the interventions. The key parameter in this calculation is the cost-to-death ratio. However, the authors do not make a clear distinction between short run income shocks and the differences in income levels. Theoretically, the impacts of income shocks can be very different from income levels. If people expect that the income level will return to normal, they may not adjust their behavior. Many estimates cited by the authors estimates the cost-to-death ratio using the differences in income levels. It is questionable whether these estimates can be applied to the COVID-19 pandemic case, which is more appropriately described as an income shock.

Reviewer #2: This study is important because it reveals the impact of the COVID-19 pandemic in the USA in the first six months. The impact of the COVID-19 pandemic has been demonstrated by cost-benefit analysis. The most troubling aspects of the paper are presented under the following heads.

1. On page 5, lines 97-98 indicate that non-Covid-19 deaths are considered as costs. The reason for this has been shown as economic downturn. However, the cost of covid-19 should cover both the losses in the gross domestic product and the health expenses associated with Covid-19. On the other hand, the authors mentioned a different methodology when measuring the cost of Covid-19 under the analysis subheading in the method section (lines 130-149). The explanations in the analysis section are suitable as methods. Therefore, the sentence on lines 97-98 needs to be revised.

2. There are some writing errors in this study (eg Page 6, line 127, suc as COVID-19l). The writing errors should be carefully read and corrected by the authors.

3. I think the conclusion part of this study is not enough. The authors state that the net effect of Covid-19 on quality-adjusted life expectancy is ambiguous. It is suggested that the author explain a little more about this. The authors are expected to evaluate the possible effects of Covid-19 on quality-adjusted life expectancy.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

21 Oct 2021

Reviewer #1:

Report for “Lives Saved and Lost in the First Six Month of the US COVID-19 Pandemic: A Retrospective Cost-Benefit Analysis”

This article evaluates the benefits and costs of U.S. COVID-19 mitigation policy interventions. In terms of benefits, the authors calculate the number of lives saved by subtracting the observed actual number of deaths from the unmitigated potential number of deaths. In terms of costs, the authors estimate the potential number of deaths due to the economic downturn caused by the intervention policies, which is calculated by multiplying the GDP loss in this period with the cost-to-death ratio taken from economic literature.

I have major concerns with respect to the methods used to estimate both the benefits and costs of the intervention policies.

Benefits:

The authors estimate the benefits of intervention policies by subtracting the actual number of deaths from the unmitigated potential number of deaths. By doing so, the authors assume that the decrease in COVID mortality in this time period is all caused by government mitigation policies, which is a questionable assumption in my opinion. First, the virus itself could become less deadly over time. Second, doctors could become more experienced in treating the disease. Third, high death rate in the initial time period could be caused by a shortage of medical equipment supply.

Response: Thank you for the comment. We use a range of infection fatality ratios between 0.53 and 0.82, based on a systematic review of studies that used data from the first 6 months of the pandemic. [1] In updating our literature search for this revision, we included a newly published systematic review of age-specific IFRs that produced evidence consistent with the overall IFR in the 0.53-0.82 IFR range.[2] The 0.53-0.82 range of IFR estimates is based on observed mortality during that time, therefore it incorporates epidemiological (e.g., virus becoming less deadly) and healthcare (e.g., doctors getting more experience in treating COVID, medical supplies becoming more readily available) trends mentioned by the reviewer. Although a lower median IFR of 0.23 was reported by the WHO[3] across 51 locations globally, the same study reported that among the 11 US states included in the report, the IFR ranged from 0.08 in Utah and 0.20 in Missouri, to 1.54 in Connecticut and 1.63 in parts of Louisiana. Overall, locations that had higher number of deaths (500 of more cases per million population), like most of the US, had a median IFR of 0.57. [3] Therefore, we believe that the IFR range of 0.53-0.82 reflects the current best understanding of COVID-19 mortality during the first 6 months of the pandemic.

As context for the magnitude of the IFR estimates used in our study, it may be helpful to note that our IFR range is considerably lower than the one used in the early estimates originally reported by the WHO and CDC in March 2020 based on initial data from the Wuhan province in China (between 1.2 and 3.4) which at the time gave rise to early predictions of the potential unmitigated death toll between 2.2 and 2.9 million lives. [4, 5, 6]

We expanded on these points on pp 9, and 13-14 of the revised manuscript (page numbers here and henceforward are based on the tracked version).

Last but not least, people voluntarily adjust their behaviors to reduce the transmission of COVID even without any government policies. I doubt whether the methodology taken by the authors can tease out the impact of voluntary adjustments and get a clean estimate of the policy effects. Thus, the decrease in mortality is caused by multiple factors, including government intervention policies. Attributing all decrease in mortality to intervention policies will overestimate the benefits of these policies.

Response: The reviewer is correct in pointing out that we attribute all reduction in COVID-19 mortality (relative to the unmitigated counterfactual) to the public health measures (lockdowns, social distancing recommendations, masking recommendations), even though some voluntary behavioral modifications would likely have taken place among the public—even in the absence of these government interventions. It is not possible to empirically isolate self-induced voluntary behavioral changes that the public would have initiated without any government policies from behavioral changes induced by government regulations through either mandates or public awareness campaigns. Inability to quantify the self-initiated (not induced by government intervention) component of behavioral modification means that our analysis may overstate the effect of the government intervention on COVID-19 lives saved. However, self-induced voluntary restrictions on activity also impacted the economy in the same way as government-imposed/induced restrictions. Because behavioral adaptations impact both sides of the cost-benefit equation (lives saved and lives lost to economic shutdown), it is unclear how this may have affected our findings.

We expanded our discussion on pp 14-15 of the revised manuscript.

Costs:

The authors estimate the cost of intervention policies by calculating the potential number of deaths due to the economic downturn caused by the interventions. The key parameter in this calculation is the cost-to-death ratio. However, the authors do not make a clear distinction between short run income shocks and the differences in income levels. Theoretically, the impacts of income shocks can be very different from income levels. If people expect that the income level will return to normal, they may not adjust their behavior. Many estimates cited by the authors estimates the cost-to-death ratio using the differences in income levels. It is questionable whether these estimates can be applied to the COVID-19 pandemic case, which is more appropriately described as an income shock.

Response: We agree with the reviewer. A salient limitation of the regulation-attributable mortality literature is that cost-to-death estimates are derived from adjusted cross-sectional comparisons of mortality across individuals with different observed incomes, not clearly distinguishing between long-term income differences and short-term income variability. Theoretically, a short term loss of income, like during the economic recession that followed the COVID-19 lockdowns in the US, may impact individuals less (or differently), compared to a persistent economic disadvantage over the course of a lifetime. (67, 68) Further, the negative impact of the economic downturn on personal incomes was partly ameliorated by the US government’ COVID-19 economic relief spending providing assistance to workers, families, and businesses. Therefore, our study may have overestimated mortality attributable to government regulation and overstated its cost.

We expanded our discussion on pp 15-16 of the revised manuscript.

Reviewer #2:

This study is important because it reveals the impact of the COVID-19 pandemic in the USA in the first six months. The impact of the COVID-19 pandemic has been demonstrated by cost-benefit analysis. The most troubling aspects of the paper are presented under the following heads.

1. On page 5, lines 97-98 indicate that non-Covid-19 deaths are considered as costs. The reason for this has been shown as economic downturn. However, the cost of covid-19 should cover both the losses in the gross domestic product and the health expenses associated with Covid-19. On the other hand, the authors mentioned a different methodology when measuring the cost of Covid-19 under the analysis subheading in the method section (lines 130-149). The explanations in the analysis section are suitable as methods. Therefore, the sentence on lines 97-98 needs to be revised.

Response: We thank the reviewer for pointing out a lack of clarity in our explanation. Our goal is to evaluate the benefit and cost, of the public health intervention (lockdowns, masking mandates, etc)—in terms of COVID-19 lives saved versus lives potentially lost as the result of the economic downturn. In public health accounting (the societal approach we take in the manuscript), medical treatments represent economic activity (labor, medical supplies, etc) and are accounted for as part of the gross national product accounting on the cost side of the equation. All medical expenditures that occurred during the pandemic are being reflected at the national level. We agree that the sentence on lines 97-98 was poorly written. It is now revised: “The cost was the economic downturn (loss of national income during the first 6 months of the pandemic) and its attributable mortality.”

Please see p 5 (here and henceforward, page numbers reference the tracked version).

2. There are some writing errors in this study (eg Page 6, line 127, suc as COVID-19l). The writing errors should be carefully read and corrected by the authors.

Response: We corrected the typographical error on page 6 and had the manuscript professionally copyedited. These changes are highlighted throughout the manuscript.

3. I think the conclusion part of this study is not enough. The authors state that the net effect of Covid-19 on quality-adjusted life expectancy is ambiguous. It is suggested that the author explain a little more about this. The authors are expected to evaluate the possible effects of Covid-19 on quality-adjusted life expectancy.

Response: Thank you for the comment. We added more context to the first paragraph of the discussion were we report on our findings for lives saved and lost, versus life expectancy gained and lost, as a result of the intervention. We also clarified the sentence in the conclusion, referring to the additional text now included in the discussion.

Please see pp 11-12 and p 17.

References

1. Meyerowitz-Katz G, Merone L. A systematic review and meta-analysis of published research data on COVID-19 infection-fatality rates. Int J Infect Dis. 2020;S1201(20):32180–9.

2. Levin AT, Hanage WP, Owusu-Boaitey N, Cochran KB, Walsh SP, Meyerowitz-Katz G. Assessing the age specificity of infection fatality rates for COVID-19: systematic review, meta-analysis, and public policy implications. Eur J Epidemiol. 2020;35(12):1123-38.

3. Ioannidis J. Infection fatality rate of COVID-19 inferred from seroprevalence data. Bulletin of the World Health Organization. 2021;99(1):19-33F.

4. COVID-19 Pandemic Planning Scenarios: Center for Disease Control and Prevention; 2020 [Available from: https://web.archive.org/web/20200522214936/https://www.cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios-h.pdf.

5. Ferguson NM, Laydon D, Nedjati-Gilani G, Natsuko Imai, Ainslie K, Baguelin M, et al. Report 9: Impact of non-pharmaceutical interventions (NPIs) to reduce COVID-19 mortality and healthcare demand: Imperial College COVID-19 Response Team; Mar 16 2020 [Available from: https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf.

6. Walker PG, Whittake C, Watson O, Baguelin M, Ainslie KEC, Bhatia S. Report 12 - The global impact of COVID-19 and strategies for mitigation and suppression: Imperial College London; March 26 2020 [Available from: https://www.imperial.ac.uk/mrc-global-infectious-disease-analysis/covid-19/report-12-global-impact-covid-19/.

Submitted filename: Responses to Reviewers.docx

Click here for additional data file.

10 Dec 2021

Lives Saved and Lost in the First Six Month of the US COVID-19 Pandemic: A Retrospective Cost-Benefit Analysis

PONE-D-21-09447R1

Dear Dr. Olga Yakusheva ,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Carlos Alberto Zúniga-González, Ph.D

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Dear authors, I have made the decision to accept the manuscript, I consider that the point of the methods used to estimate benefits and costs related to intervention policies is debatable and I think the approach you give it to measure the mortality rate in the first 6 months of a pandemic; It has been a well-discussed manuscript, so I congratulate you for your contribution and effort in this type of research.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: (No Response)

Reviewer #2: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: No

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: N/A

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: I still have my previous concern that the estimation of the decrease in mortality due to intervention policy is not “causal”. The decrease in COVID mortality in the lockdown period is not all caused by intervention policies. The paper lacks a proper methodology to identify the causal effect, for example comparing a state with strict lockdown v.s. another state with less strict lockdown, which is a quite standard methodology in the discipline of economics. Without a causal effect, the value of the cost and benefit analysis will be significantly limited.

Reviewer #2: (No Response)

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

17 Dec 2021

PONE-D-21-09447R1

Lives Saved and Lost in the First Six Month of the US COVID-19 Pandemic: A Retrospective Cost-Benefit Analysis

Dear Dr. Yakusheva:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Prof. Carlos Alberto Zúniga-González

Academic Editor

PLOS ONE

Table 1

Evidence table and regulation-attributable mortality calculations.

Study			Cost-to-death (CDR) estimatea			Regulation-attributable mortalitya,b
			As reported in the study	Inflation adjustment	In current 2020 USD	Lives lost	Comorbidity-adjusted life-years lost
Ashe, B., de Oliveira, F. D., & McAneney, J. (2012)	Fire management to prevent structural and bushfires in Australia	Cross-sectional estimates from (Keeney 1997), age and sex adjusted, rescaled to Australian income and mortality data from the 2006 Australian Census data, Australian adults age 35 and older, 2006	$20; $50 mil 2010 AU	0.77c	$15.4; $38.5 mil	144,805; 57,922	5,908,052; 2,363,221
Chapman, K. S., & Hariharan, G. (1994)	Select health and safety regulations during 1970–1990	Cross-sectional survival analysis of 10-year mortality and personal income; adjusted for age, wealth, employment, family structure, health and disability; 1969–1979 Retirement History Survey merged with Social Security records through 1974, US males 58–62 in 1969	$12.2 mild 1990 US	1.96	$23.9 mil	93,305	3,806,862
Chapman, K. S., & Hariharan, G. (1996)	Occupational safety regulations	Panel survival analysis of annual mortality and personal income; adjusted for age, wealth, employment, education, family structure, number of living parents, and health status; 1966–1990 National Longitudinal Survey of Mature Men, older US men 45–59 in 1966	$6.4; $8.7 mil 1990 US	1.96	$12.5; $17.0 mil	178,400; 131,176	7,278,720; 5,352,000
Gerdtham, U.-G., & Johannesson, M. (2002)	Unspecified government regulation	Panel survival analysis of annual mortality risk and personal disposable income; adjusted for sex, age, wealth, income, disposable income, education, employment, immigrant status, family structure, health status, blood pressure, functional limitations; Swedish adults 20–84, 1980–1986	$6.8; $9.8 mil 1996 US	1.63	$10.8; $16.0 mil	206,481; 139,375	8,424,445; 5,686,500
Elvik, R. (1999)	Road safety regulations	Time series analysis of mortality and income per capita during ten 5-year periods between 1946 and 1995; adjusted for age and sex, data for from Statistics Norway 1996–97	$7.1 mile 1995 USD	1.68	$11.9 mil	187,395	7,645,714
Keeney, R. L. (1990)	Unspecified government regulation	Cross-sectional annual mortality rates across income groups; adjusted for age and sex; 1970 data, white US adults 25–64	$3.14; $7.25 mil 1980 US	3.11	$9.8; $22.5 mil	227,551; 99,111	9,284,082; 4,043,733
Keeney, R. L. (1997)	Unspecified government regulation	Cross-sectional annual mortality risk across income groups, age and sex adjusted, using National Longitudinal Mortality Study 1979–1985 data, 550,000 US adults 25–64	$4.85; $13.33 mil 1991 US	1.88	$9.1; $25.1 mil	245,055; 88,845	9,998,242; 3,624,864

Notes

a Estimates are reported as the estimate under the proportional cost allocation assumption followed by the estimate under the progressive cost allocation assumption, separated by a semicolon.

b Author calculations for $2.23 trillion GDP loss.

c Costs reported in AU$2010, the 0.77 multiplier adjusts for 2020 AU$ to US$ exchange rate 0.65, and US 2010–20 inflation coefficient 1.88.

d Progressive cost allocation is assumed.

e Proportional cost allocation is assumed.