Management decisions of an Academic Radiology Department during COVID-19 pandemic: the important support of a business analytics software

Objectives To analyze the response in the management of both radiological emergencies and continuity of care in oncologic/fragile patients of a radiology department of Sant’Andrea Academic Hospital in Rome supported by a dedicated business analytics software during the COVID-19 pandemic. Methods Imaging volumes and workflows for 2019 and 2020 were analyzed. Information was collected from the hospital data warehouse and evaluated using a business analytics software, aggregated both per week and per quarter, stratified by patient service location (emergency department, inpatients, outpatients) and imaging modality. For emergency radiology subunit, radiologist workload, machine workload, and turnaround times (TATs) were also analyzed. Results Total imaging volume in 2020 decreased by 21.5% compared to that in 2019 (p < .001); CT in outpatients increased by 11.7% (p < .005). Median global TAT and median code-blue global TAT were not statistically significantly different between 2019 and 2020 and between the first and the second pandemic waves in 2020 (all p > .09). Radiologist workload decreased by 24.7% (p < .001) during the first pandemic wave in 2020 compared with the same weeks of 2019 and showed no statistically significant difference during the second pandemic wave, compared with the same weeks of 2019 (p = 0.19). Conclusions Despite the reduction of total imaging volume due to the COVID-19 pandemic in 2020 compared to 2019, management decisions supported by a dedicated business analytics software allowed to increase the number of CT in fragile/oncologic outpatients without significantly affecting emergency radiology TATs, and emergency radiologist workload. Key Points • During the COVID-19 pandemic, management decisions supported by business analytics software guaranteed efficiency of emergency and preservation of fragile/oncologic patient continuity of care. • Real-time data monitoring using business analytics software is essential for appropriate management decisions in a department of radiology. • Business analytics should be gradually introduced in all healthcare institutions to identify strong and weak points in workflow taking correct decisions. Supplementary Information The online version contains supplementary material available at 10.1007/s00330-022-08709-3.


Introduction
The international outbreak of novel SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2) [1,2] and the associated coronavirus disease 2019 (COVID-19) had crushing effects both on the global population and on healthcare systems due to its high transmissibility, elevated mortality among elderly and fragile patients, and lack of effective therapy [3]. In particular, the need for intensive care for many patients has put severe pressure on hospitals, whose emergency departments (EDs) and intensive care units (ICUs) were unprepared [4,5]. Moreover, since COVID-19's presentation is often an interstitial pneumonia, whose diagnosis and clinical evaluation rely heavily on imaging findings, radiology units have been at the forefront of the emergency, with a huge number of patients requiring chest imaging, either X-ray or highresolution CT [6][7][8].
The pandemic brought up two kinds of problems for radiologists: on one hand, the diagnostic aspects that required specific training given the peculiarities of X-ray and CT findings of COVID-19 pneumonia [9], and on the other hand, the unprecedented complex organizational problems [10]. Radiology departments were asked to completely reorganize the workload in a matter of a few days, by taking critical decisions regarding the management of acute emergencies as well as those of chronic oncologic and fragile patients.
Main urgent needs were (1) creating separate paths for COVID-19 and non-COVID-19 patients, to avoid crosscontamination of patients; (2) integrating a dedicated CT scanner in the COVID-19 path to improve diagnostic accuracy; (3) intensifying personnel (radiologists and radiographers) at the ED, particularly for night shifts; and (4) guaranteeing diagnostic exams in frail and cancer patients [10,11]. Putting these organizational changes into practice in a sudden and precise way was extremely complex and particularly true for healthcare organizations who lack real-time information on workflows and on modifications brought on by unexpected external events, such as, in this case, the COVID-19 outbreak. A solution might be the implementation of a business analytics (BA) tool to support management decisions [12]. BA refers to the procedural and technical infrastructure that collects, stores, and analyzes data produced by a company's activity by providing metrics and graphical dashboards for an objective real-time measurement of business efficiency [12].
In this study, we reported the response of a radiology department of an academic hospital during the two pandemic waves that occurred in Rome from March 9 to April 12 (weeks [11][12][13][14][15] and from November 2 to November 29, 2020 (weeks 45-48) [14]. Moreover, we examined the department governance choices guided by the support of a dedicated BA software, having the management of emergencies and the guarantee of continuity of care, especially for oncologic and fragile patients, as the two main objectives.

Study design
A retrospective review of the imaging volumes and workflows of the radiology unit of Sant'Andrea Academic Hospital in Rome for 2019 and 2020 was performed. The institutional ethical committee approved the study (ref. nr CE 5773_2020).
Our institution is a medium-size academic hospital; it counts 452 beds, 7 departments with 48 units, 523 doctors, and 1063 healthcare professionals. It is a center of excellence for thoracic surgery, medical oncology, and orthopedic surgery. Radiology unit includes three subunits: interventional radiology (IR); breast imaging and general radiology; and emergency radiology, located in a different floor of the hospital, embedded in the ED, which belongs to general radiology. Radiology personnel is made of 23 radiologists (4 dedicated to IR), 38 residents, 41 radiographers, and 15 nurses. Three digital radiology (DR) units, two interventional radiology (IR) suites, two multidetector computed tomography (CT) scanners, two 1.5-T magnetic resonance (MR) scanners, one MSK-dedicated low-field MR scanner, one ultrasound scanner (US), one mammography equipment, and one bone densitometry scanner (DXA) are available in the main department. The emergency radiology unit is fully equipped with DR, US, and MDCT scanners. The same scanners were available in 2019 and 2020 and no major changes in radiology personnel occurred.
With the COVID-19 outbreak, our hospital was selected by the Regional Healthcare System as one of the nine hubs in our region. As a result, dedicated COVID-19 beds were opened in intensive care unit (ICU) (nr, 32) and high-intensity (nr, 40) and low-intensity (nr, 76) care units.

Data collection and analysis
Data were collected from the hospital data warehouse (electronic records, RIS, and PACS data from 2015 onwards) and evaluated using a data analytics and business intelligence software operating since March 2019 (Radiology Command Center™, GE Healthcare), which provides real-time, actionable information accessible by staff and hospital leaders. Data from years 2019 and 2020 were aggregated both per quarter and per week during the selected time intervals, stratified by patient service location (emergency, inpatients, outpatients) and by imaging modality (DR, mammography, DXA, ultrasound, CT, MRI, and interventional radiology).
For the emergency radiology only, we collected metrics of the average number of examinations per hour, CT exam allocation between primary CT scanner (CT2) and secondary/ backup CT scanner (CT1), and median turnaround time (TAT). TAT was defined as the time from the request of the radiological examination from the emergency physicians to the password-secured electronic signature of the radiological report from a radiologist. Turnaround time was analyzed considering all radiological exams (global TAT), code-blue only patients (code-blue global TAT), CT examinations only (CT TAT), and CT performed in code-blue patients (codeblue CT TAT).
Data obtained by our data analytics and business intelligence software were exported to Microsoft Excel and descriptive statistics were performed to assess the number of patients and imaging exams in 2019 and 2020, divided by quarters. The total number and relative percentage of imaging exams (DR, mammography, DXA, US, CT, MRI, and IR) and patients, stratified as emergency, inpatients, and outpatients, were calculated, and evaluated.

Statistical analysis
All analyses and graphs were performed using commercially available software SPSS (IBM Corp. Released 2017. IBM SPSS Statistics for Macintosh, Version 25.0.: IBM Corp.). Quantitative variables have been expressed as mean ± standard deviations; categorical variables have been conveyed as frequencies and percentages. The Kolmogorov-Smirnov test was used to assess data distribution. In case of Gaussian distribution, data were tested with Student's t test, while the Wilcoxon test was applied for non-Gaussian distributed data. Percentages were compared by using the Pearson's chi-square test or in alternative Fisher's exact test. A p < 0.05 was considered to indicate a statistically significant difference.

Results
Total imaging volume in 2020 marked a 21.  Table 1).

Discussion
Our results demonstrate that, despite the pandemic and the decision of our regional healthcare system to include Sant'Andrea Academic Hospital among the nine regional COVID-19 hubs, and despite the significant reduction of about one-fifth of the total radiological examinations, our radiology unit was able to increase by around 12% the number of CT examinations for outpatients, particularly fragile and oncologic patients, in 2020 compared to 2019.
This result was the consequence of the different decisions about the opening and closure of radiological paths for COVID-19 and non-COVID-19 patients, taken during the first and the second pandemic peaks and driven by the availability of real-time data analytics about number and distribution of diagnostic examinations. However, caution is needed when interpreting our data, because the epidemiology of the pandemic in Rome and the role of our hospital within the regional healthcare network may have affected our results.  The pandemic in Rome came in two waves: the first one in March-April 2020 (weeks [11][12][13][14][15] and the second one in November 2020 (weeks 45-48).
At the beginning, alarming media reports regarding Rome together with a dramatic increase in COVID-19 patients admitted to the ED suggested to take the decision to separate the emergency diagnostic paths for COVID-19 and non-COVID-19 patients, to avoid cross-contamination. The radiology subunit embedded in the ED became dedicated to COVID-19 patients and had its own emergency radiology team, while a second emergency path for non-COVID-19 patients was set up in the main radiology department, on a different floor, including one of the two CT scanners usually available for in-/outpatients. The reason behind such logistic choice was mostly an emotional response to an unprecedented alert situation that called for extreme health safety measures, certainly not justified by either number of examinations or radiologist workload, at least in our hospital. In fact, number of examinations showed a progressive reduction starting from week 8 (February 17, 2020) when the government confirmed the first twenty cases and reported the first death, with a clear downwards trendline. On the 11th week (when on March 9, 2020, the national lockdown began), there was a 41% drop of outpatient examinations, compared to the 10th week, and the progressive decline of outpatient examinations, including CT for fragile and oncologic patients, continued until the 15th week.
Reasons for patients' behavior were probably the fear of contagion in the hospital environment and the cancelation of most visits and non-urgent surgeries imposed by the government. Probably, the decrease of CT for fragile and oncologic outpatients was also, at least in part, the consequence of the reduction of available slots due to the re-organization of emergency radiology paths. In fact, with the improvement of the pandemic, at the time of partial reopening of the national governmental lockdown (May 4th, beginning of week 19), the standard organization (emergency radiology for both COVID-19/non-COVID-19 emergency patients and main department devoted to in-/outpatients) was re-established and the number of outpatient examinations started recovering.
At the time of the second pandemic wave (weeks 45-48), data analysis of the previous three quarters of the year suggested the possibility to maintain the standard organization. In fact, the separation of the emergency paths in two different floors of the hospital was not justified by workload changes because an indepth analysis of the emergency radiology during the first pandemic wave showed an average reporting of 2.93 exams/h, around 25% lower than 2019 (3.89 exams/h). Emergency radiology workload not only decreased, but it became less complex with a more than three-fold increase in chest CT scans and a reduction of more than 70% of brain CT and more than 65% of abdominal-pelvic CT. These data are in line with what were reported among Italian stroke units [13] and surgical [14], orthopedic, and traumatology [15] EDs, where a dramatic decrease of hospitalization was observed.
A possible criticism at the decision of keeping a single path in the emergency radiology setup sharing CT for COVID-19 and non-COVID-19 patients might be related to difficult sanitation procedures and possible overlapping of urgent cases. In our experience, we established stringent internal CT protocols, allowing scans of COVID-19 patients consecutively and safely while sanitizing the CT scanner room at the end of the shift and not after each single CT examination [16]. Nevertheless, we had continuously monitored ER accesses to be alerting by any critical scenarios weekly, ensuring the right assistance for acute patients and avoiding any unjustified delays in diagnosis. Registered TATs confirmed our choice: overall median CT TAT in 2020 was 10 min (+ 11%) higher than in 2019 and median CT TAT in code-blue patients was 9 min (+ 13.8%) higher than in 2019. The overall increasing of TAT in 2020 could be explained by the procedures of sanitization after each shift of COVID-19 patients and by clinical condition of COVID-19 patients, who were often critical, needing a longer preparation time in the CT scanning room by the radiographers, nurses, and ICU doctors. Despite the higher number of ED patients admitted in November (nr 354) compared to those admitted in March-April (354 vs. 294; + 16.9%), CT TAT and CT TAT in code-blue patients during weeks 45-48 were not statistically different from those weeks 11-15. Emergency radiology workload slightly increased (3.48 exams/h) compared to the first peak (+ 16%), but it was still well below the standards of 2019 (3.89 exams/h), thus justifying our correct choice of keeping personnel allocation unchanged.
The risk of a simultaneous urgent case was balanced by the availability of a second backup CT in the main department, whose use did not affect the overall CT productivity.
Finally, the decision to keep the main department completely COVID-19 free, with both CT scanners available for in-and outpatients, allowed us to increase by + 12% the number of CT for outpatients in the 4th quarter.
In conclusions, in times of crisis, resilience is a fundamental quality to possess. It allows for change to happen rapidly and smoothly. Business analytics software is an extremely powerful tool that provides a precise insight of real-time changing scenarios. When applied to our radiology unit, it allowed us to reinforce managerial decisions and to support the changes in workforce planning during pandemic crisis, in which the main hospital goals were management of emergencies and the guarantee of continuity of care, especially for oncologic and fragile patients.