Recovering lost opportunities in the management of critically ill patients boarding in the emergency department

Emergency department (ED) boarding of critically ill patients has been identified as a significant contributor to increased mortality for this patient population for nearly two decades.1, 2 The volume of critically ill patients continues to increase without a concomitant increase in intensive care unit (ICU) beds. We can simply accept that boarded ICU patients will continue to have worse outcomes, or we can seek alternate treatment modalities to address this issue. Various ED critical care models have been developed to optimize the care of these patients and reduce the burden of resource-intensive patients with critical illness on the ED.3-7 ED critical care teams may have a role in the management of these patients and have been previously associated with improvements in time to operative intervention and improvement in mortality. However, their expansion has been limited due to multiple factors including shortages of ED physicians with additional critical care training, space requirements, and initial startup cost. Consult teams can provide a flexible solution that can be quicker to implement compared to a dedicated ED ICU.6-9 Our adult ED has approximately 60,000 patient-visits per year and is licensed for 73 acute care beds. Intensive and intermediate care admissions make up 16% of annual visits. We piloted having ED critical care consult team (ED3CT), which is composed of six emergency physicians with additional training in critical care or resuscitation. ED3CT consultation was not mandatory and with the general expectation that the ED3CT would provide either direct takeover of the care or indirect care via consultation at the discretion of the primary ED attending. Thus not every patient with planned intensive or intermediate care admission received consultation. During the pilot, the EDC3T was planned Mondays, Wednesdays, and Fridays; a control group was identified by patients on Tuesday, Thursday, and Saturday. Inclusion criteria in both groups were patients in whom the original disposition from the ED was an intensive care or intermediate care bed. Intermediate care unit subtypes included medical/surgical progressive care, neuroprogressive care, cardiac-progressive care, and spine-progressive care. Admission criteria vary by specific unit subtype but are generalized as any nursing or laboratory intervention required with greater frequency than every 4 hours, vasopressor infusions at most 10 μg/min norepinephrine or equivalent, continuous sedative or anesthetic infusions, continuous insulin infusion, continuous electroencephalogram monitoring, acute needs for high-flow oxygen or noninvasive positive pressure ventilation or at physician discretion. Intensive care unit subtypes included medical ICU, surgical ICU, neuroscience ICU, cardiac ICU, and cardiovascular surgical ICU. Admission criteria to an intensive care level was similarly unit subtype specific, included progressive care criteria, and also included the generalized following: use of invasive mechanical ventilation, vasopressor infusions at greater than 10 μg/min of norepinephrine or equivalent, use of inotropes or inodilators, need for continuous renal replacement therapy, mechanical cardiac support, open surgical abdomen, and at physician discretion. We performed a retrospective chart review on patients that were identified as initially planned for admission to an intensive or intermediate care bed. There was a significant reduction in ED3CT shifts during the months of December 2020 and January 2021 due to staff redeployments to handle a regional COVID-19 surge. We identified 502 patients during the study period. Of these, 96% (484/502) had complete data. All patients were initially managed by standard ED staffing, 26.1% (126/484) of patients underwent a transition of care to the ED3CT after being identified as critically ill. We explored total hospital costs, ED professional charges, procedures, and ED length of stay (LOS; Table 1). We also examined downgrades (defined as patient originally intended for intensive care with disposition changed to intermediate care, inpatient wards, or discharge; or patients originally intended for intermediate care, with a disposition change to hospital wards or discharge). Eighteen of 126 (14.2%) of patients managed by the ED3CT were downgraded compared to only 13/358 (3.6%) of those in the control days. Of patients managed by the ED3CT, 50 were originally intended for intensive care with one (2%) downgraded to intermediate care and none (0%) to hospital wards, while 76 patients were originally intended for intermediate care with 17 (22%) downgraded to floor. In the control group, 124 were originally intended for intensive care with three (2.4%) downgraded to stepdown and none (0%) to floor, while 234 were intended for intermediate care with 10 (4.2%) downgraded to floor. After the introduction of the ED3CT and retrospective review of our critically ill patients, we found an increase in documented critical care time and procedural billing (Table 1). We also found an increase in patient downgrades. While the retrospective nature of this review limits the authority of these data, they are thought-provoking for several reasons. First, we found that the patients managed by the ED3CT were approximately four times more likely to be transitioned to a lower intensity of care after being initially intended for an ICU or intermediate care unit. This is a practice nearly exclusive to the ED3CT in our department. This may owe to the fact that some critically ill patients can be stabilized with early, focused, aggressive ED-based critical care. This practice is meaningful as leading to better utilization and allocation of hospital resources. A study on short-stay ICU admissions found that 37% of critically ill patients spent less than 24 h in the ICU.8 It is conceivable that if a portion of these patients can be managed in the ED setting that it may lead to better allocation scarce hospital resources (e.g., ICU and intermediate care beds), reduced costs, and improved care by reducing boarding for patients with the greatest need.6 Assuming approximately 11,000 intensive and intermediate care admissions at our institution, a 14.2% downgrade rate would reduce intensive and intermediate care admissions by 1562 per year. This has important system implications given these beds are often hospital throughput bottlenecks. The ED3CT was deployed with attending emergency physicians with fellowship training in critical care only; no additional nursing staff or training was required. Nurses did not require additional training as all care was performed within the emergency nursing scope of practice. ED3CTs may be readily implementable with relatively low startup costs when considering no additional nursing or infrastructure costs. Second, there was an approximately 40% difference in professional charges between the patients managed by the ED3CT and those by ED physicians without additional critical care training. This difference is multifactorial but potentially related to differences in the total number of procedures, critical care billing, and level of care provided. We do not have outcomes data for these patients; however, billing data show significantly more time spent at the bedside of these patients based on CPT codes. Interestingly, when the financial results are extrapolated over a year, the lost billing could be as high as $5 million annually (based on annual volume of approximately 60,000 patients with an approximately 16% ICU admission rate). No matter the reason for the differences in the billing, the absolute difference shows the financial opportunities lost when critically ill patients board in our ED. Finally, we saw a 100% increase in procedures performed per patient for those patients managed by ED3CT. This has positive implications for patient safety and quality care, billing, and importantly, graduate medical education and training in both emergency medicine and critical care. As an academic ED, the ability to provide bedside education is important. In graduate medical education, most critical care education occurs in the ICU, in simulation labs, or during didactics, the latter of which do not expose trainees to clinical nuance and complex decisions. At our facility medical students and residents have opportunities to work with the ED3CT as electives, enabling one-on-one education at the bedside of critically ill patients. Furthermore, given the existence of the ED3CT, we are able host a resuscitation and emergency critical care fellowship. While assigning value to this educational opportunity is contextual, there were clearly more opportunities to complete complex procedures (e.g., extubation, central venous access, intubation, arterial line insertion, bedside ultrasonography, bronchoscopy, chest tube insertion, etc.) with the ED3CT. There are several major limitations to this report. First, these data are generated by a retrospective chart review of a single center. Second, our division has six emergency physicians with additional formal critical care fellowship training, which may not be readily found at other institutions. Third, the EDC3T consultation was at discretion of the primary ED attending. This potentially introduces selection bias due to a nonconsecutive sampling. Fourth, at our institution it is the responsibility of the ED physician to assess for downgrades and not the intensivist; this makes downgrades an emergency physician–driven process as opposed to an intensivist-driven process. This may limit generalizability at institutions where intensivist-driven downgrades are common. Finally, patients managed by the EDC3T had slightly longer ED LOS and total hospital charges. However, existing literature suggests ED critical care interventions are overall associated with reduced hospital LOS despite the increase in ED LOS.9, 10 In conclusion, the addition of an ED3CT on various days to our ED resulted in multiple opportunities to improve patient care, billing, education, and improved resource availability and allocation for patients with critical illness. As the burden of critical illness continues to accumulate without a proportion increase in ICU beds, our data suggest that there may be benefit to both the patient and the hospital to developing and implementing ED3CTs. Matthew Johnson assisted with the analysis and interpretation of the data and drafted the letter. Eric Segev organized and analyzed the data. Alexander Bracey, Sean P. Geary, and Luke Duncan collected data, assisted with the analysis, and critically reviewed the letter. Christopher Hanowitz assisted with collection of the data. Denis Pauzé critically revised the letter. Gregory P. Wu assisted with data acquisition and data interpretation, conceived the letter, and reviewed and revised the letter. Gregory P. Wu takes responsibility for the letter in its entirety. The authors declare no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request.