Customize your JAMA Network experience by selecting one or more topics from the list below.
In late March 2020, the Centers for Medicare & Medicaid Services (CMS) released projections of US national health spending that predicted growth from the 2019 level of 17.8% of gross domestic product (GDP) to 19.7% over the next 10 years.1 Through no fault of their own, the CMS prognosticators are poised to take their place in history beside economist Irving Fischer, who announced that “stock prices have reached what looks like a permanently high plateau” shortly before the market crash of 1929 that marked the start of the Great Depression. The coronavirus disease 2019 (COVID-19) pandemic is likely to result in year-over-year changes in both health care spending and GDP that are without precedent. Because the ratio of these 2 numbers, the share of health care in the GDP, receives so much attention in public policy, it is worth thinking about how large these changes may be, and more importantly, what they mean.
The first step in projecting how COVID-19 may affect the health share of the GDP is to estimate how COVID-19 will affect health care spending. This will depend on the future extent, timing, and composition of the pandemic, so it is highly uncertain.
On one hand, in regions with a great deal of COVID-19 disease, hospitals are operating at or over capacity. The federal government, states, and hospitals have rushed to purchase ventilators and personal protective equipment and have taken bold steps to facilitate hiring of recent medical school graduates, retired medical workers, and physicians and nurses with out-of-state licenses. This surge in demand will drive up health care spending over the coming months. The range of estimates of the costs of this surge is very wide—somewhere between $34 billion and $500 billion in added private insurance spending, or between 3% and 40% of current spending,2,3 and between $7 billion and $30 billion each in additional spending for the Medicare and Medicaid programs, or between 1% and 5% of current spending in these programs.1 The differences in the percentage increases between private and public programs arise in part because baseline spending per enrolled person is much greater in public insurance, in part because per capita costs for COVID-19 are much lower in public insurance, and in part, in the case of Medicaid, because children and young adults have had lower rates of COVID-19 illness. All of these estimates depend critically on how many people are ultimately affected.
On the other hand, many clinicians have seen the demand for their services substantially decline or vanish altogether. Dentists, primary care physicians, outpatient service practitioners and centers, surgical specialists, and hospital departments that focus on elective procedures have all seen very sharp declines in demand. Between January and March, overall employment in health care actually decreased. Here, the timing of the pandemic is critical. If the pandemic is well controlled soon, the use of these services may simply be shifted into the fall. If not, many of the services (ie, visits and procedures) may never happen. These reductions in spending would offset at least a portion of the increased COVID-19–related expenditures.
Uncertainty about the composition of the pandemic-affected population further complicates the projections. Will those hospitalized be covered by Medicare, Medicaid, or private insurance—or will they be uninsured? The COVID-19 pandemic highlights the stunning differences in the prices paid by private and public payers in the US health care system. Commercial health insurers pay nearly 4 times as much for the kinds of care COVID-19 patients require than Medicare does, and nearly 5 times as much as Medicaid does.3 If the burden of this disease is disproportionately borne by poor and elderly persons, which appears to be the case, the effects on national health spending will be much lower than if most hospitalized cases are reimbursed through private insurance. In addition, if increased unemployment leads to large increases in the number of US residents who are uninsured, particularly in states that have not yet expanded Medicaid, non–COVID-19 health care spending may decline even after the pandemic is under control. The current employment-based health insurance system, combined with the lack of a coverage safety net, will exacerbate the effects of the pandemic on the most vulnerable people.
The midpoint of these various estimates suggests that the pandemic might plausibly lead to national health spending in 2020 that is 10% higher than in 2019. If GDP were unchanged, this increase in health care spending alone would increase the share of GDP devoted to health care spending by 10%; the share would increase from the current 17.8% to 19.6% in a single year.
But the effect of COVID-19 on the ratio of health care spending to GDP is likely to be even greater because of the consequences of the pandemic for the denominator: the nation’s output. The effect of COVID-19 on output likewise depends on the course of the pandemic and whether demand returns in the second half of the year. Current forecasts, which generally anticipate a resumption of economic activity by summer or fall, nonetheless project declines in GDP of between 2.4% and 8.7% for 2020 relative to 2019.4,5
Mathematically, an 8.7% reduction in GDP—unimaginable as recently as last month—would increase the ratio of health spending to GDP by 1.7 percentage points, even if health spending did not change. Combining a 10% increase in the numerator (health spending) and an 8.7% reduction in the denominator (GDP) yields a 20% increase in the ratio of the two. In other words, health care spending would increase by 3.7 percentage points, from 17.7% of GDP to 21.4% of GDP, in a single year. This 1-year change would be nearly twice as large as the 10-year forecast provided by the CMS actuaries just before the pandemic happened.1 Even if the decline in GDP is just 2.4%, reflecting an economy that moves quickly in the fall to make up ground lost in the spring, health care would comprise 20% of GDP next year, well above the amount projected for 2028.
Although these increases will inevitably attract substantial attention, it is critical to think through what these estimates do and do not mean.
The first consideration is the likely level of health spending in 2021 (the numerator). This amount will likely increase relative to last year, but, in present circumstances, it would be better if it were higher still: if more personal protective equipment were available, if ventilators were not in short supply, and, critically, if there were more effective ways to reduce morbidity and mortality among patients with COVID-19. Some of the care that will have been deferred or forgone is likely of low value, but much is surely not. It would be better if it had not been necessary to put off the useful elective procedures and routine care now deferred by the crisis. It would be better if people did not lose access to health insurance as they lost their jobs. The pandemic highlights that it is a mistake to think of health care spending as “bad”: a curve that must be bent. Some care is wasteful because it pays for care that is ineffective; but much care is not. The imperative to develop tools that will identify high-value care and design payment systems that reward that care is as strong as ever. The imperative to ensure that prices accurately reflect value is as strong as ever.
The second consideration is what will happen to GDP (the denominator). Slower growth in GDP has affected the ratio of health care to GDP in the past; between 2008 and 2009, health care as a share of GDP increased from 16.3% to 17.2%. But in that case, as in most earlier years, the evolution of GDP and that of health care spending were largely independent, except perhaps through the effects of unemployment on insurance coverage. The COVID-19 pandemic situation reflects an even deeper linkage. The cost of treating patients with COVID-19 is the smallest component of the economic burden of the pandemic. The much greater costs are the human consequence of disease for individuals and their families and the enormous cost of the precautions taken by individuals and societies to avoid this disease.6 Most of the economic costs of the pandemic are outside the health care system. In hindsight, it would have been well worth spending much more on health care, particularly in the form of disaster and pandemic preparedness, to control the pandemic and allow the economy to restart.
Corresponding Author: Sherry Glied, PhD, Robert F. Wagner Graduate School of Public Service, New York University, 295 Lafayette St, Second Floor, New York, NY 10012 (firstname.lastname@example.org).
Published Online: April 27, 2020. doi:10.1001/jama.2020.6644
Conflict of Interest Disclosures: Dr Glied reported receiving fees from and being a director of NeuroRx, a company with an interest in patent 8178489 for aviptadil, which is in a clinical trial for treatment of acute respiratory distress syndrome in COVID-19 patients. Dr Levy reported receiving personal fees from the FAIR Health Academic Advisory Board.
Glied S, Levy H. The Potential Effects of Coronavirus on National Health Expenditures. JAMA. 2020;323(20):2001–2002. doi:10.1001/jama.2020.6644
Artificial Intelligence Resource Center