Operational Impacts of the Coronavirus Disease 2019 (COVID-19) Pandemic and Early Recovery Trends: University of Washington Anatomic Pathology Experience

Accessioned pathology cases within the UW hospital system were enumerated using the reports functions of the PowerPath (version 10.0.1.39, Sunquest Information Systems, Tucson, Arizona) laboratory information system. Human subjects research was approved by the University of Washington Institutional Review Board, IRB protocol No. 2837. Total accessioned cases enumerated included all in-house, slide review, and consultation cases from the following diagnostic pathology services: autopsy, breast, bone and soft tissue, cardiac, cytology, cytogenetics, dermatopathology, gastrointestinal/hepatic/pancreatic, genitourinary, gynecology, head and neck/pulmonary, hematopathology, neuropathology, and medical renal. Cases were tallied in total and by subspecialty service during 4 major time periods in 2020: before a ban was placed on nonurgent procedures (2020 baseline, January 1–March 15), a strict ban on nonurgent procedures (March 16–April 17), a less-strict ban on nonurgent procedures (April 18–May 17), and no ban on nonurgent procedures (the initial recovery phase, May 18–June 27). As a historical control, the total cases from January to June 2019 were also tallied. Using the service code search function in PowerPath, billable cases by different current procedural terminology (CPT) codes were numerated during the different time periods. Intraoperative consultation was assessed in the 4 weeks before the strict ban and during the strict ban by searching for cases performed that contained the words “intraoperative consultation” in the report (a required text field in all cases with intraoperative consultation). The trends of COVID-19 in King County, Washington from January to June 2020 were drawn from the King County Public Health Web site, including the number of positive cases per day and the number of hospital admissions in the county per day.⁷ 

Using Microsoft Excel 2020 and GraphPad Prism version 8.4.2 (GraphPad Software, San Diego, California), descriptive and inferential statistics were performed. In most instances, the mean number of cases per workday per week or month was calculated by dividing the total cases for the week by the number of nonweekend and nonholiday days in the week or month. Using a 2-tailed unpaired t test, the mean total cases per workday per month from March to May were compared between 2019 and 2020. This test was also used to compare intraoperative volume both before and after the strict ban on nonurgent procedures. A 1-way analysis of variance and Tukey multiple comparison tests were used to assess differences in the mean cases per workday per week during the 4 different time periods in 2020. The Pearson r was used to evaluate correlation between the number of accessioned cases and COVID-19 trends in King County.

To estimate gross loss in billable revenue, the physician fee schedule for each CPT code analyzed was drawn from the Centers for Medicare and Medicaid Services Web site. For each time period, the average number of cases per workday billed with each CPT code was multiplied by the physician fee schedule for that CPT code within the Seattle/King County locality.¹⁰  This only grossly estimates billable income and does not capture either discrete data regarding CPT codes by case or actual billed or collected income.

Compared with 2019, the mean total cases accessioned per workday in 2020 did not differ when averaging January and February (2-tailed unpaired t test, t₂ = 0.59, P = .61), but was significantly lower than the March to June average (2-tailed unpaired t test, t₄ = 3.71, P = .02). During the COVID-19 surge in April 2020, the mean cases per workday was 41% of the volume in April 2019. In June 2020, although case volume had somewhat recovered, 2020 still had on average 32 fewer cases per workday compared with June 2019, which amounted to a 11% relative decrease in case volume for the month of June (Figure 1).

As shown in Figure 2, the weekly mean accessioned cases per workday was maintained during the initial phase of the COVID-19 pandemic, but dropped dramatically when a strict ban on nonurgent procedures was implemented. The overall nadir was 76 cases per day 3 weeks into the strict ban, a 73% reduction from the 2020 mean preban baseline of 285 cases per day. Case volumes showed partial but not significant (P = .48) recovery with the introduction of the less-strict ban. When all restrictions were lifted, case volumes eventually plateaued at a mean of 253 cases per day during the last 3 weeks analyzed (June 8–28), representing an 11% reduction from the 2020 prepandemic baseline. Overall, there was a significant reduction in the weekly mean cases per workday for all 4 time periods compared with the prepandemic time period (Figure 3; 1-way analysis of variance, F_3,21 = 65.4, P < .001), and most of the time periods were significantly different from each other (Table 2).

Overall accessioned cases negatively correlated with COVID-19 hospitalizations. As shown in Figure 2, there was a strong negative correlation between accessioned daily cases and COVID-19 admissions at hospitals in King county (r₂₃ = −0.76, P < .001) and an even stronger negative correlation with new daily positive COVID-19 cases in King County (r₂₃ = −0.88, P <.001).

As shown in Table 3, all subspecialties had at least a 43% decrease in case volume during the strict ban, with the largest drops in the dermatopathology (−77%), gastrointestinal pathology (−71%), and medical renal (−68%) services and the smallest reductions in the breast (−43%), hematopathology (−44%), and head and neck/pulmonary (−49%) services.

Most services had a similar path to recovery: a slight increase in volume during the less-strict ban and a greater increase during the no-ban period, but not yet at baseline. The dermatopathology and gastrointestinal/hepatic/pancreatic pathology services highlight this trend well (Figure 4). The services that came closest to baseline volumes during the no-ban time period included cytology (−6%), gynecology (−12%), genitourinary (−12%), breast (−13%), and head and neck/lung (−14%). The cardiac service actually had an 8.5% increase in case volume compared with baseline during the no-ban time period, the only service with an increase in volume.

As seen in Table 4, all assessed billable CPT codes had at least a 27% reduction in case volume during the strict ban, with the largest reductions in 88304 (−73%) and 88305 (−63%) and the smallest reductions in 88321 (−27%) and 88342 (−40%).

Most CPT codes then had a slight increase in volume during the less-strict ban and a greater increase during the no-ban phase but not yet at baseline. However, some, like 88305, had relatively the same volume during the strict and less-strict bans but then dramatically increased during the no-ban phase (Figure 5). Code 88321 actually had fewer cases during the less-strict ban (−49%) than the strict ban (−27%). The CPT codes that came closest to baseline during the no-ban time period were 88304 (−6.3%), 88342 (−7.9%), 88309 (−11%), 88313 (−11%), and 88307 (−12%).

As highlighted in Figure 6, there was a mean of 8.5 intraoperative cases per workday in the 4 weeks before the strict ban and a mean of 5.5 cases per workday during the strict ban, which was significantly different (2-tailed unpaired t test, t₆ = 2.61, P = .04). However, there was no significant difference in the parts (separate specimens) needing intraoperative consultation; there was a mean of 16 parts per workday 4 weeks before the ban and 14 parts per workday during the ban (2-tailed unpaired t test, t₆ = 0.702, P = .51).

For the CPT codes analyzed, compared with the average $45 985 billable per day in 2020 before March 16, there was on average a $22 273 (−48%) loss per day during the strict ban, a $20 702 (−45%) loss per day during the less-strict ban, and a $7991 (−17%) loss per day during the no-ban phase (Table 5).

The effects of the COVID-19 pandemic on reduction in volume in specific sectors of health care have been demonstrated in other studies, with examples including but not limited to breast surgeries, breast screening, trauma cases, emergency visits, imaging, and plastic surgery emergencies.^11–15  Similarly, our study shows a reduction in specimen volume that occurred across all aspects of the clinical service. This reduction primarily followed the trend of institutional and governmental polices on nonurgent-procedure bans, which were being informed by the number of COVID-19 cases. Not surprisingly, the greatest reduction in volume was during the strict ban. All subspecialties had at least a 43% reduction in volume; the highest impact was on the dermatopathology (−77%) and gastrointestinal pathology services (−71%). This is likely because these services tend to have high volumes of nonurgent biopsies. The CPT code 88305, typically applied to biopsies, had a substantial decrease in volume (−63%). These findings are congruent with other reports of a 90% decrease in endoscopy procedures and the restriction of access to dermatology practices during the pandemic.^16,17 

Although not as dramatically, larger specimens also had a decrease in volume: CPT codes 88307 and 88309 had a 40% and 41% reduction, respectively. There were also specific subspecialties that had less of a decline, such as breast and head and neck/pulmonary, 2 services with a substantial amount of surgical oncology resections at our institution. These findings are consistent with recommendations to avoid delays in many surgical oncologic cases,^18,19  and also with a study that found oncologic breast cases had less volume loss than nononcologic breast cases.¹¹  The continuance of complex oncologic surgeries is also an explanation for the relatively stable demand for the intraoperative consultation service before and after nonurgent procedure restrictions. Although the total number of surgical cases needing intraoperative consultation declined, the total number of parts to cases needing intraoperative consultation remained relatively the same, which we infer to reflect continuance of complex oncologic resections requiring several margin assessments (such as composite resections in the head and neck).²⁰  If facing a nonurgent procedural ban, other institutions with a high volume of complex oncologic cases may also experience less impact on larger specimens and relative preservation of the demand for intraoperative consultations.

Our AP laboratories also saw a decline in special stains (CPT 88312, 88313), immunohistochemistry (CPT 88342), immunofluorescence (CPT 88346), and electron microscopy (CPT 88348). Immunohistochemistry had less of a relative decline than the other services. The decline in immunofluorescence coincides with the decline in medical renal and dermatopathology specimens. Similarly, the decline in electron microscopy is directly attributable to the decline in medical renal cases.

As the ban on nonurgent procedures became less strict, there was mostly an increase in all services and CPT billable cases. However, slide reviews (CPT 88321) experienced a greater decline when restrictions were lessened (−27% strict versus −49% less strict). Because slide reviews are done for referral cases from other institutions, this lag in impact reflects the inherent delay in referring patients. Additionally, CPT code 88305 was virtually the same volume during both the strict (−63%) and less-strict bans (−63%), indicating that the bulk of the biopsy procedures are considered nonurgent procedures.

When all bans were lifted, the overall specimen volume rapidly increased and plateaued at 89% of baseline during the last 3 weeks analyzed. The reasons for this are likely multifactorial: patients may be intentionally avoiding medical care during the pandemic,^21,22  care centers must decrease the number of patients seen in order to maintain appropriate COVID-19 precautions, and more patients are being seen by telemedicine.²³  Another reason could be that health care resources are being prioritized to provide care for COVID-19 patients, although the number of COVID-19 inpatients was relatively low in King County (fewer than 10 cases) during the last 3 weeks analyzed. We also hypothesize that clinicians, in an effort to mitigate SARS-CoV-2 exposure, might have adapted their practice habits to perform fewer procedures that have a low pretest probability of actionable diagnostic findings on tissue sampling. Future studies are needed to look into the factors impeding a complete recovery to pre–COVID-19 volume and/or if this trend persists.

The major limitation of this study is its generalizability. The decline in specimen volume appears to be largely driven by institutional and legal policies on nonurgent procedures, as the volume precipitously fluctuated when these policies were changed. Therefore, our study is best generalizable to other pathology services facing similar types of restrictions placed on nonurgent procedures. We cannot predict how specimen volume would have organically changed as COVID-19 cases fluctuated. However, nonurgent procedural bans are likely when COVID-19 becomes a burden on the health care system, as this is a recommended policy to implement.^24–26  Overall, our findings are best extrapolated to other academic AP departments or private laboratories that serve large tertiary medical centers that are facing similar bans on nonurgent procedures.

As the impact of COVID-19 and subsequent polices on the volume and workflow were unprecedented and unknown, few changes were made to AP personnel except for within our training programs. During the strict and less-strict bans, residents were primarily kept at home to engage in virtual lectures, virtual sign-outs, and research. Only one resident a day was brought in to assist with intraoperative consultation. The AP medical student education team rapidly implemented a virtual rotation for the medical school, which had its clinical clerkships cancelled and had 70 students enroll in this virtual course.⁵  Lastly, in roughly 3 weeks UW neuropathology conducted and published a rapid-validation study aimed at deploying whole slide imaging technology for issuing diagnoses at a departmental level as part of our regional COVID-19 response.³  This minimal-resource-use study was conducted during a time period of relaxed Centers for Medicare and Medicaid Services and US Food and Drug Administration enforcement of policies regulating remote review of pathology slides.³ 

The COVID-19 pandemic has provided an opportunity for us to retrospectively analyze the impact of a major pandemic on health care delivery in AP. Our analysis reveals how necessary governmental and institutional regulations can impact operations, and provides insight into proactive strategies to address that may be of value in resource planning and management. These strategies should be data driven, basing decisions on evidence defining which areas are most and least likely to be impacted. This will become increasingly important as ongoing budget constraints related to diminished revenue from the first wave of COVID-19 cases hamper workforce retention, capital investment, and quality management.²⁷  We anticipate that our data analysis will help us monitor recovery and predict future impacts on volume and workload. Future studies will be needed to determine the long-lasting impact of this pandemic and whether such analyses will aid in taking proactive measures to mitigate the impact of such crises on pathology services.