The United States has just passed the dark milestone of 200,000 COVID-19 deaths, with many experts warning of the possibility of hundreds of thousands more dead and gravely ill in the months ahead. The country has also experienced enormous economic and social upheaval, with millions forced into unemployment, hunger, and homelessness. In the absence of a vaccine, the interventions policies that are known to stop the spread of the coronavirus—like stay-at-home orders and school closures—have also upended the course of normal life. How to strike the proper balance between the economic and social costs on one hand and the health costs on the other has been at the core of a prolonged, and sometimes bitter, national debate—one that has largely been conducted in the absence of hard data. New research by Harvard Kennedy School’s Soroush Saghafian, associate professor of public policy, seeks to provide that important context. Saghafian and his coauthor (his former PhD student who is now an assistant professor of statistics and family medicine at Michigan State University), painstakingly collected health information—including infection rates, hospitalizations, residents’ mobility obtained from cell phone data, and the use of intensive-care beds and ventilators—and provided the first detailed, national and state-by-state measure of economic costs and health gains. The results can help policymakers better make a cost-benefit analysis of policies, including the timing and duration of interventions. We asked Saghafian to walk us through his findings.
Q: How were you able to measure the tradeoffs between the potential health gains versus the economic burdens flowing from pandemic interventions?
Some intervention policies, like stay-at-home orders, are known to be effective in controlling the spread of the novel COVID-19. However, across the United States, with authorities worried about the economic burden of those policies, restrictions were relaxed relatively rapidly as states moved towards reopening.
But decision making in most states has been challenging, especially because of a dearth of quantitative evidence on the health gains versus economic burdens of different intervention policies. Our goal was to assist decision makers by providing them with that quantitative evidence. This enables them to better understand which policies are most effective, whether they should impose stricter polices, and when to move towards reopening.
We used data from all 50 states and the District of Columbia on various factors, including the number of tests, positive and negative results, hospitalizations, ICU beds and ventilators used, residents’ mobility (obtained from cell phone data), and deaths. We also looked at both direct and indirect costs to society. For instance, when people are restricted from going to work, economic growth slows down, people lose their income, and so on. But in addition, as the rate of infection increases, we end up using more ICU beds and ventilators, among many other hospital resources. And so healthcare expenditures increase as well. The model we developed with this data allowed us to observe consequences of adopting different policies (even policies that weren’t implemented) and examine whether authorities should have adopted a different set of strategies.
Q: What did you find?
Our results show that, compared to a hypothetical no intervention during March-June 2020, the policies undertaken across the country on average saved each person living in the United States up to 4.04 days-worth of Quality-Adjusted Life Years (QALY)—a widely used measure of health gains that assigns a value of a full year lived in perfect health—and ended up costing $3,285 per person. We also found that stricter policies could have increased the average QALY gain per person to 6 days and raised the cost to $4,954. Remember that, to see the overall impact, these numbers have to be multiplied by the total population—more than 320 million.
In addition to quantifying the health and economic impacts of intervention policies, our results allow federal and state authorities to avoid following a “one-size-fits-all” strategy, and instead enact policies that are better suited for each state. This is because we provide insights for each state separately and show what policies would have worked in that particular state. For example, when comparing the cost of the policies actually implemented to no intervention at all, we found South Dakota’s per capita cost was $217 per day of QALY, while that of neighboring Nebraska was $6,346.
Another interesting aspect of our study is that we use cell phone data to gauge the mobility of residents in each state. Imposing stricter policies cannot help unless residents abide by them. For example, even during the stay-in-home orders, cell phone data shows that residents were still mobile, with levels that varied across the states. Roughly speaking, states that were able to restrict the mobility of their residents were better able to control the spread the disease. But, of course, restricting mobility typically comes with some economic burdens, and our study provides quantitative results to help authorities better understand the underlying tradeoffs in health gains versus such economic burdens.
Q: Your work in this case is from the United States. Do you think the findings are relevant for other countries too or are they very U.S.-centric?
In the spring, I worked with the government of Bahrain and analyzed interventions, which helped shed light on the set of policies that they could follow. But countries are different, and so we try to avoid recommending the same policy for all countries. In particular, policies that might work for one country might not be suitable for another country. Even within the United States, we find that one should avoid imposing the same set of policies for all states.
Having said this, it should be noted that the analytical methodology we have developed is the commonality. That is, our analytical models can be used by different states or even different countries to study what policies work best. Have they been imposing the correct set of policies since the pandemic started? Should they do something differently going forward? These are the types of questions we want to help authorities answer.
Banner photo by Joshua Lott; faculty portrait by Martha Stewart