What are our limits? The best team in the best hospital can reach a point where patients’ needs exceed the institution’s capacity to provide comprehensive care. The location of this inflection point, and how best to measure it, has become an area of intense concern since the onset of the COVID-19 pandemic. Before the pandemic, acute emergencies with the potential to overwhelm ICU capacity were usually localized events, finite in both geography and location (at least in high-income countries). Even in a massive disaster such as Hurricane Katrina in 2005 or the Great East Japan Earthquake in 2011, there was the potential to transfer seriously ill patients out of the disaster zone to unaffected areas, although many of these evacuation plans failed in their implementation (1,2). The pandemic was different: following initial waves in China, Italy, South Korea, the United Kingdom, Washington State, and New York, no area was entirely spared. Theoretical discussions about the ethics of ventilator distribution became real and deadly serious (3,4).

Although our fears about running out of ventilators mostly did not happen, we did have sustained levels of critically ill patients with complex needs, high acuity, and often prolonged hospitalizations. The high volume of patients with COVID-19 left few hospitals untouched, especially during the winter of 2020–2021 and during the Delta and Omicron surges in 2021 and 2022. Nonetheless, the impact of these surges was not uniform, and some health systems sought to manage this uneven burden through structured transfers from more-burdened to less-burdened facilities. This system of transfers, known as “load balancing,” was implemented successfully in many jurisdictions with the added benefit of avoiding crisis standards of care, where resource limitations would inhibit the standard practice of effective critical care (5–7). Unfortunately, other regions were less successful in load balancing, where competition between hospitals and health systems, incomplete data, and inefficient communications acted as barriers to the safe and organized transfer of patients (8).

The use of load balancing is not simply a matter of ICU bed availability and staffing. Rather, the problem is that we have limits. Patient surges lead to higher mortality. Time spent caring for multiple acutely decompensating patients is time that cannot be spent considering the fine details of those patients’ care. It is time not spent ensuring that ventilators are optimized, fluid management is adequate, goals of care are addressed, and antibiotics are de-escalated. This is all intuitive. The challenge is to quantify it so that we can act on it.

Multiple studies have used novel metrics to measure the impact of strain on the ICU. Kadri et al (9) evaluated 558 hospitals in the United States from March to August 2020, using a surge index that used a weighted total of all patients in the hospital (with additional weight placed on patients in the ICU, requiring noninvasive ventilatory support, or on invasive mechanical ventilation), divided by the standard prepandemic bed capacity. Using this metric, an estimated 23% of total COVID-19 deaths in the study period were related to excessive hospital caseloads (9). A nationwide observational study in the United States using data from the Centers for Disease Control and Prevention and the Cybersecurity and Infrastructure Security COVID Task Force, conducted later during the pandemic (from July 2020 to July 2021), showed comparable findings: ICU mortality increases with bed occupancy. In this analysis, 12,000 excess deaths per week occur across the United States when total ICU occupancy exceeds 75%, and 80,000 excess deaths per week when occupancy reaches 100% (10). A strikingly similar result was identified in an analysis of 89 National Health Service Trusts in England between April and December 2020, with an odds ratio (OR) for the death of 1.23 for patients admitted to ICUs during times of greater than 85% ICU bed occupancy during COVID-19 surges.

These increased numbers of deaths are not exclusively due to COVID-19 but also due to deferred services, for example, cancer and cardiac care, illustrating the cascading effect of ICU strain on the entire healthcare system (11). It is worth noting that this association between mortality and strain is not unique to ICUs during COVID-19 surges; there is a large body of literature linking an increased risk of death to emergency department overcrowding (which is simply strain by another name) (12).

In this issue of Critical Care Medicine, the study by Pilcher et al (13) builds on this growing body of knowledge with their analysis of a comprehensive national database in Australia. Similar to the study by Kadri et al (9), the authors use a scoring system (the Activity Index) that assigns points to patients requiring 1:1 nursing care, invasive ventilation, renal replacement therapy (RRT), extracorporeal life support (ECLS), or COVID-19 isolation and then dividing by the total number of staffed beds. Once adjusted for patient-level risks, the authors demonstrate a nearly linear increase in mortality as the Activity Index increases, with an OR for in-hospital mortality increasing from a baseline of 1.0 at an index of 0, to 1.5 at indices of 0.5–1.0, to 2.0 at an index of 1.5, and so forth. Some of this increased mortality may be ascribed to patient acuity, despite the authors’ rigorous work to control for such factors; an ICU with an Activity Index of 0 has no patients on mechanical ventilation and no patients requiring 1:1 nursing, so mortality would be expected to be low. The increase in mortality above predicted levels makes this tool valuable, although its linear increase and the absence of a clear inflection point would make it challenging to operationalize from a systems perspective.

All of these scores use reasonable but somewhat arbitrary metrics, using methods that facilitate rapid calculation. One can debate why the surge index gives greater weight to invasive mechanical ventilation over noninvasive ventilation, for example, or why COVID-19 isolation alone is awarded an extra point in the Activity Index (when infection control precautions are not rare events in a typical ICU). One must also account for national differences in ICU nursing practice. Mechanically ventilated patients typically receive 1:1 nursing in the United Kingdom and Australia, for example, unlike the routine practice in the United States (although patients on continuous RRT or ECLS usually do). Despite these issues, the consistent results across multiple studies suggest a true association, one that cannot be explained away solely by patient-level factors such as acuity. This consistency in different epochs of the pandemic also supports strain as an independent risk factor. One could posit, for example, that higher mortality in studies conducted in early 2020 could be related to the experience of managing a new, highly acute disease. Some of the highest Activity Indices in the study by Pilcher et al (13), however, occurred in late 2021 and early 2022 (eTable 7 and eFig. 6 in [13]), at which time most critical care professionals were intimately familiar with the management of COVID-19.

Pilcher et al (13) have contributed a great deal in helping us answer the question posed at the beginning of this editorial: what are our limits? Taking the published data we have, it would appear that ICU mortality during the COVID-19 pandemic began to increase as ICU bed occupancy reached 75–85% of capacity, particularly when there were a high number of patients requiring invasive forms of organ support. This may be true as well during nonpandemic times. ICU strain can kill, and avoiding strain improves outcomes. As we learn to better define these inflection points for risk, we must simultaneously work to develop systems that permit load balancing to occur in an organized, proactive matter.

1. Sasaki H, Maruya H, Abe Y, et al.: Scoping review of hospital business continuity plans to validate the improvement after the 2011 Great East Japan Earthquake and Tsunami. Tohoku J Exp Med. 2020; 251:147–159 2. Aucoin RG: Hurricane Katrina—one hospital’s experience. Crit Care. 2006; 10:109 3. Maves RC, Downar J, Dichter JR, et al.; ACCP Task Force for Mass Critical Care: Triage of scarce critical care resources in COVID-19 an implementation guide for regional allocation: An Expert Panel Report of the Task Force for Mass Critical Care and the American College of Chest Physicians. Chest. 2020; 158:212–225 4. Daugherty Biddison EL, Faden R, Gwon HS, et al.: Too many patients...A framework to guide statewide allocation of scarce mechanical ventilation during disasters. Chest. 2019; 155:848–854 5. Lowe AE, Garzon H, Lookadoo RE, et al.: Avoiding crisis conditions in the healthcare infrastructure: 2 case studies in statewide collaboration. Health Secur. 2022; 20:S71–S84 6. Mitchell SH, Rigler J, Baum K: Regional transfer coordination and hospital load balancing during COVID-19 surges. JAMA Health Forum. 2022; 3:e215048 7. Berkeveld E, Mikdad S, Terra M, et al.; ROAZ Regional Network: Optimization of a patient distribution framework: Second wave COVID-19 preparedness and challenges in the Amsterdam Region. Health Secur. 2023; 21:4–10 8. Mandel-Ricci J, Belfi K, Shapiro J, et al.: Rethinking surge preparedness after COVID-19: Effective patient load balancing within health systems and beyond. Health Secur. 2022; 20:497–503 9. Kadri SS, Sun J, Lawandi A, et al.: Association between caseload surge and COVID-19 survival in 558 U.S. Hospitals, March to August 2020. Ann Intern Med. 2021; 174:1240–1251 10. French G, Hulse M, Nguyen D, et al.: Impact of hospital strain on excess deaths during the COVID-19 Pandemic—United States, July 2020-July 2021. MMWR Morb Mortal Wkly Rep. 2021; 70:1613–1616 11. Wilde H, Mellan T, Hawryluk I, et al.: The association between mechanical ventilator compatible bed occupancy and mortality risk in intensive care patients with COVID-19: A national retrospective cohort study. BMC Med. 2021; 19:213 12. Sartini M, Carbone A, Demartini A, et al.: Overcrowding in emergency department: causes, consequences, and solutions—a narrative review. Healthcare (Basel). 2022; 10:1625 13. Pilcher DV, Hensman T, Bihari S, et al.: Measuring the Impact of ICU Strain on Mortality, After-Hours Discharge, Discharge Delay, Interhospital Transfer, and Readmission in Australia With the Activity Index. Crit Care Med. 2023; 51:1623–1637