We Need to Relocate ICU Patients Out of Covid-19 Hotspots

by Rodolfo Catena and Matthias Holweg - June 23, 2020

The stark national differences in the death rates of Covid-19 patients is much discussed and politicized. Germany and Italy, for example, have similar infection rates but dramatically different mortality rates; through June 20, in Germany, 4.7% of patients have died versus 14.5% of patients in Italy. Attempts to explain these disparate death rates generally point to differences in the effectiveness of testing regimes, the timing of social distancing measures, demographics, and the number of beds in intensive care units (known as ICU capacity).

Certainly each of these plays a role, not the least of which is ICU capacity; it makes intuitive sense that if a hospital runs out of ICU beds, more patients will die. But our research finds that this oversimplifies the situation. Simply increasing the number of available ICU beds in local hotspots - a common response to rising caseloads - may paradoxically result in increased mortality. The better response is to relocate patients early out of emerging hotspots in order to keep ICU loads well below the maximum capacity.

The health care system workload at the local county level - or even at the level of individual hospitals - can have a direct impact on Covid-19 mortality rates, which is often overlooked. The dramatically different mortality rates in Germany and Italy are a case in point: while both countries have had major outbreaks, Germany had not only more ICU beds per capita at the outset (48.7 versus 8.6 beds per 100,000 inhabitants), but it also managed to avoid any local overload of its health care system. In part, that’s because it had substantially more ICU capacity than Italy. But Germany’s national county-level ICU-bed registry tracking the availability of beds and ventilators also allowed it to prevent local overloading by allowing doctors to relocate patients swiftly in order to distribute the demand on ICU beds.

The impact of overload is vividly illustrated by the patterns of Covid-19 outbreaks in two regions in Italy: Lombardy, the worst affected region, which to date accounts for 48% of the country’s Covid-19 deaths, and Veneto, a neighboring region with similar demographics but far lower mortality rates. Both are wealthy regions with a similar health care system capacity (346.2 vs. 326.6 hospital beds per 100,000 residents respectively in 2018), and in both, the outbreaks started at virtually the same time.

Mortality rates of Covid-19 patients in both regions increased as the load on ICU capacity grew, as one would expect. But, strikingly, these rates began to increase long before ICUs were near full capacity, starting to uptick at about 40% and then, in Lombardy, launching into a dramatic increase when loads reached about 50% - just half capacity - and reaching nearly 5.5 deaths per 100,000 inhabitants. In Veneto, meanwhile, ICU loads remained below 50% and mortality rates remained nearly flat at less than 1.3 deaths per 100,000 inhabitants.

Our findings suggest that ICU overload is indeed a major factor driving Covid-19 mortality rates. COVID-19 patients can be complex and labor-intensive. Rising infection levels of medical personnel, ensuing shortages of qualified staff, and the resulting overwork and fatigue are all factors contributing to higher mortality.

This finding suggests that increasing the number of local ICU beds without commensurate increases in personnel and other resources - a common response in epidemic hotspots - may not be an effective way to decrease mortality, and may, in fact, increase it.

Instead of approaching Covid-19 as a local ICU management problem, health care leaders should think of it as a pan-regional (or pan-state) management problem from the start. It makes little sense to overload local ICUs while those in neighboring areas with much lower Covid-19 prevalence are underutilized. Given our data, distributing stable patients to even out ICU loads across affected regions (while avoiding sending patients to unaffected areas) seems an obvious remedy.

According to Fausto Catena, Head of Emergency and Trauma Surgery at Ospedale Maggiore in Parma, Italy, transferring stable Covid patients is both feasible and sensible. In Italy, a small number of patients were eventually transferred from the badly hit North to less affected regions in the South although these transfers were unfortunately too few and late to stem the loss of life in the Northern epicenters.

Jeffrey Lazar, Vice Chair of Emergency Medicine at New York’s St. Barnabas Hospital, told us that while ICU utilization within the city was not systemically collected, anecdotal information suggests that ICUs across the city all treated two to five times more patients than usual during the Covid-19 peak. Lazar likewise agreed that more evenly distributing ICU loads to reduce overutilization in individual hospitals would be sensible.

Clearly, obstacles must be overcome to accomplish this. In the U.S., the lack of integration of health care systems across regions and the required communication and coordination mechanisms will pose challenges. Further, unrelated or competing health systems may be reluctant to take on cases from elsewhere. And of course the logistics of safely transporting Covid-19 patients so as to prevent unduly spreading the infection must be addressed.

Yet as the experience in Germany shows, where there is the political will, it is possible to track the availability of ICU beds in a national system to facilitate a reduction in local overloads. The system was only created in mid-March in response to the Covid pandemic, and all hospitals are required by law to submit their ICU load every day. When local ICU constraints develop, doctors can request transfers from the state Emergency Medical Services command and control center. Even Bavaria, the most severely affected region in Germany (with a total case load of 39,395 cases, or 301 per 100,000 inhabitants) saw a maximum ICU utilization of only 61%. Germany in fact airlifted more than 120 patients out of the worst affected areas in Italy and France despite the equally significant number of cases within Germany itself.

Studies of early Covid-19 hotspots and simulation models all indicate we should expect recurring waves in the future that are very likely to manifest as local concentrations of infection, like those seen in Wuhan, Lombardy, and New York City. Health care leaders must look beyond their local outbreak and address rapidly rising ICU demand in hotspots before the ensuing workload leads to potentially preventable increased mortality. Digital technologies allow for close monitoring of cases as well as for the creation of online registers that can track ICU capacity across regions. Such data platforms can help enable the quick response needed when local capacity bottlenecks emerge. We continue to learn how to effectively treat and manage this pandemic, and the early transfer of Covid-19 patients out of hotspot areas may be the most effective measure to curbing excess Covid-19 mortality yet.

Rodolfo Catena is an Associate Lecturer in Healthcare Management at Arden University, UK. He can be reached at rodolfo.catena@gmail.com.

Matthias Holweg is the American Standard Companies Professor of Operations Management at Saïd Business School at the University of Oxford, UK. He can be reached at matthias.holweg@sbs.ox.ac.uk.