Appropriateness of Emergency CT Utilization During the Initial Peak of the COVID-19 Pandemic

Brief Description of the Problem

The appearance of the novel coronavirus known as severe acute respiratory syndrome coronavirus 2 at the end of 2019, and the subsequent coronavirus disease 2019 (COVID-19) pandemic, which gripped most of the world by March 2020, required massive adaptations in the health care delivery system in the United States. During the initial peak of the COVID-19 crisis in the United States, from late March through the end of May 2020, there was nationwide concern about maintaining adequate hospital resource capacity, including conserving supplies of needed personal protective equipment and maintaining access to needed services, including CT scanning.

Although the gold standard for the diagnosis of COVID-19 has been established to be reverse transcription polymerase chain reaction (PCR) testing, in the early phase of the pandemic, limitations in test availability and concerns for test sensitivity led to a surge of interest in the use of radiography and especially for use of CT scanning for establishing the diagnosis, based on the characteristic CT findings of the atypical or viral pneumonia seen in COVID-19 [1] and to address the increased incidence of acute pulmonary embolism associated with COVID-19 pneumonia [2]. As evidence for and against the utility of CT scanning for the diagnosis of COVID-19 was collected and published, concern grew worldwide with regard to CT scanning resource utilization stewardship, because use of CT scanning in patients suspected of coronavirus disease had the potential to substantially limit access. This is because up to 90 min of scanner downtime might be required after each use in a patient suspected of having COVID-19 to allow for comprehensive cleaning of the equipment and recirculation of the room air to prevent virus transmission to subsequent patients. This downtime issue was of particular concern in the emergency department, because patients requiring immediate CT scanning commonly present with very little advance notice.

Accordingly, the Fleischner Society released guidance on the role of chest imaging, including CT scanning, during the COVID pandemic [3], and similar guidance was also released by an expert panel of the RSNA [4], advising against the routine use of CT for this diagnosis—guidance that was echoed by subsequent documents released by the ACR [5,6]. By April 2020, it had become clear that the evidence-based guidance was against the routine use of CT scanning to establish or confirm the diagnosis of COVID-19 and generally discouraged the use of CT in COVID-19 patients, except for high-acuity clinical indications that were not directly due to COVID-19 [7].

What Was Done

We retrospectively reviewed all consecutive patients suspected of having COVID-19 during April 2020 at our center to determine whether CT utilization in the emergency department was appropriate based on current guidelines. We wished to better understand the drivers of CT utilization in this population and also to assess the impact of CT scanning on the management of these patients. Our project constituted a “practice quality improvement” activity and was therefore institutional review board exempt.

A list was compiled of all consecutive adult patients suspected of having COVID-19 who underwent CT scanning from our emergency department during April 2020 (designated as “persons under investigation for COVID-19”). A subgroup of these were designated as being COVID+ (ie, persons presumed by the emergency physicians to be infected with COVID-19, although not all of these diagnoses were ultimately confirmed by positive PCR test results). We compared the documented rationale and clinical indications for the use of CT in these patients to available evidence-based guidelines, such as the ACR Appropriateness Criteria and the Wells Criteria for suspected pulmonary embolism, which had previously been adopted by multidepartmental consensus at our center [8], as metrics of utilization appropriateness. Additionally, all cases were scored as to whether the CT scan altered the presumptive diagnosis or significantly affected the management of these patients. Significance testing was performed using a χ² test.

Outcomes and Limitations

A total of 199 adult patients, 96 women and 103 men, were included in our series, of which 175 were considered to be persons under investigation and 24 were presumed positive for COVID. Most patients underwent CT scans of the chest (approximately 150 chest CT scans), many of which were performed as CT angiography studies for pulmonary embolism. The next most frequent type of scan was CT of the abdomen and pelvis, followed by CT of the head, some of which were performed in addition to CT scanning of the chest or abdomen. Results are summarized in Table 1 .

Table 1.

Results summary

Patient Classification	No. of Cases	CT Order Appropriate?	Tested COVID-19+	COVID-19 Testing Delay	CT Changed Diagnosis	CT Changed Management	D-Dimer Positive	CXR positive
PUI	175	141 (80%)	2.3%	33%	52 (29%)	53 (30%)	24%	7.4%
Presumed COVID	24	13 (54%)	38%	4%	8 (33%)	6 (25%)	45%	33%

CXR = chest x-ray; COVID-19 = coronavirus disease 2019; PUI = persons under investigation for COVID-19.

In our series, 33% of patients were noted to have had delayed PCR test results, which sets an upper boundary for CT scan requests that may have been motivated by delayed PCR testing, and very few CT scans (n = 4) were requested expressly for the sole purpose of confirming a presumed diagnosis of COVID-19 and were deemed as being inappropriate. Only 13 patients (7.4%) had a prior suspicious chest radiograph, and only 5% had documented low oxygen saturation, defined as Po ₂ < 90%, so neither of these factors seemed to significantly drive utilization of CT scanning in our series.

The majority of CT scans were deemed to be clinically appropriate on retrospective review, as shown in Figure 1 . Despite this, CT did not alter the presumptive diagnosis or clinical management in the majority of cases, with a change in management defined in our series as CT results having altered the treatment plan as documented in the clinician notes or when the CT scan detected unexpected disease, as well as the CT being positive for the diagnosis of concern, such as confirming a suspicion of pulmonary embolism (Fig. 2 ). Negative scans were scored as not having altered the diagnosis or management. The most commonly noted comorbidity (n = 14) for patients in our case series was a currently active cancer diagnosis in treatment.

Fig 1.

Fraction of appropriate versus nonappropriate CT requests studies in the two cohorts. Although a majority of CT requests were deemed appropriate in both cohorts, the fraction of appropriate requests in the persons under investigation (PUI) cohort was significantly greater than in the presumed coronavirus disease 2019 (COVID) group (P < .001).

Fig 2.

Fraction of cases in which the CT result altered the diagnosis or management. Changes in diagnosis or management (Dx/Mgt) occurred in a slightly larger fraction of cases in the presumed coronavirus disease 2019 (COVID) group; however, this difference was not statistically significant (P > .7). PUI = persons under investigation.

Our study is limited by being a single-center retrospective analysis. Despite this limitation, we believe that it illustrates the effectiveness of interdisciplinary scientific communication within the medical and radiological communities throughout the nation during the pandemic crisis, as well as a testimonial to the value of the scholarly professional societies of medicine and radiology who rapidly developed and disseminated the guidance that was broadly followed.