The 2019 coronavirus disease (COVID-19), caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), is rapidly spreading throughout the world. In the United States, the disease is expected to infect 20-60 percent of the population before the pandemic finishes its course. The recent experience in Italy has highlighted the critical need to ensure adequate capacity of inpatient and intensive care beds for what is likely to be a surge of seriously ill patients.
To assess the capacity of U.S. hospitals to care for the impending number of hospitalized patients, we merged hospital infrastructure data from the 2018 American Hospital Association (AHA) Annual Survey and the American Hospital Directory to estimate inpatient and intensive care unit bed capacity in non-federal acute-care hospitals overall in the US and in each of the 306 hospital referral regions. We calculated average inpatient and ICU occupancy rates from annualized inpatient bed days. We used population estimates from the U.S. Census American Communities Survey to project likely numbers of infected patients, hospitalization rates and ICU utilization rates based on existing data from recently published reports. We adjusted these rates across communities by proportion of people who were age 65 or over, as data suggests that hospitalization rates vary by age.
Based on work by Lispsitch and colleagues, we used a middle-level estimate of COVID-19 infection rate of 40 percent. We assumed lengths of stay based on published studies. We calculated the capacity gap between current bed occupancy and anticipated COVID-19 demand assuming six, 12 and 18 month transmission curves. In our primary model, we made the very aggressive assumption that 50 percent of currently occupied beds could be freed up to care for COVID-19 patients.
Nationally, based on 40 percent prevalence of COVID-19 over the course of the pandemic, we estimate that 98,876,254 individuals will be infected, 20,598,725 individuals will likely require hospitalization and 4,430,245 individuals will need ICU-level care. We found that inpatient and ICU bed capacity to handle expected patient volumes varied significantly by Hospital Referral Region (See exhibits 1 and 2).
If the infection curve is not flattened and the pandemic is concentrated in a 6-month period, that would leave a capacity gap of 1,373,248 inpatient beds (274 percent potentially available capacity) and 295,350 ICU beds (508 percent potentially available capacity). If the curve of transmission is flattened to 12 months, then the needed inpatient and ICU beds would be reduced to 137 percent and 254 percent of current capacity. However, if hospitals can indeed reduce current bed occupancy by 50 percent and flatten the transmission curve to 18 months, then the capacity needed would be reduced to 89 percent of inpatient and 166 percent of ICU beds. If the infection rate is only 20 percent (low end of current estimates), we would largely be able to meet the needs for inpatient care if we flatten the curve to 12 months.
Finally, we observed large variations in availability of both regular and ICU beds across communities (see exhibits 1 and 2). Some rural communities have adequate numbers of regular beds but often large shortfalls of ICU beds, whereas many more-populous communities have inadequate number of total beds but smaller shortfalls of ICU beds. Our findings clearly illustrate the public health imperative to both flatten the curve of transmission and recruit additional capacity of both regular and ICU beds.
The large shortfall in US inpatient hospital and ICU bed capacity raises important policy implications for current efforts to address COVID-19, as well as lessons for the inevitable next pandemic. First, to meet the needs ahead, hospitals need to develop contingency plans to expand hospital capacity. Postponing elective inpatient surgical admissions is clearly one relatively straightforward strategy, but all hospitals in the US should develop clear protocols to aid decision making on which types of procedures and clinical diagnoses could be safely postponed. Second, the expected shortage in mechanical ventilators and ICU beds suggest that hospitals need to further develop plans to transition inpatient operating rooms, ambulatory surgical sites, and post-anesthesia care units into flexible ICUs if warranted by the number of infected patients or rates of ICU care needed.
Lastly, we need a comprehensive strategy to expand the workforce to care for this large influx of patients, from keeping current providers healthy to recruiting recently retired physicians and nurses (possibly to take care of non-COVID patients to free up currently practicing clinicians to care for COVID patients). Massachusetts governor Charlie Baker’s recent announcement that the state would expedite medical license approvals for out-of-state practitioners to achieve an emergency increase in the physician workforce is one example of what such a strategy could look like. Other regulatory relief measures, such as granting temporary licenses to providers who might have retired in the previous five years and relaxing scope-of-practice laws, can further boost workforce capacity.
We are at an inflection point and clearly do not have the capacity to care for our population of COVID-19 patients if the infections occur quickly and there is a spike in acutely ill patients. However, spreading the disease out, and providing new strategies to expand the number of beds and the workforce, can help ensure that we get through this difficult period.