Association of the COVID-19 Pandemic With Patterns of Statewide Cancer Services

Abstract

The first cases of coronavirus disease 2019 (COVID-19) in the United States were identified on January 20, 2020 (1). By the end of March, there were more than 100 000 cases and 2000 deaths nationally (1). Stay-at-home orders and suspension of nonessential medical services were enacted in many states. Several reports based predominately on health insurance claims data described dramatic declines in cancer screening (2-4), diagnosis (2,5), and treatment (4). Early reports during the spring are limited, however, for comprehensively tracking trends because of the timing required for insurance claims adjudication, select providers and patient populations (eg, patients aged ≥65 years), and losses of employer-sponsored health insurance coverage resulting from unemployment.

Cancer incidence data in population-based registries lag 2 to 3 years from diagnosis to release date, limiting utility for timely assessment of the impact of the COVID-19 pandemic on cancer detection and care. Surveillance, Epidemiology, and End Results (SEER) registries, however, have captured real-time cancer pathology reports for many years; most solid tumors are pathologically confirmed, and all surgically resected specimens have corresponding pathology. This study used real-time electronic pathology data from 2 population-based SEER registries to assess patterns in cancer detection and surgical treatment during the pandemic.

Electronic pathology reports from the Georgia and Louisiana SEER registries were selected from the pathology laboratories that consistently transmitted reports from January 2018 to December 2020 (91.2% and 86.4% of all reports, respectively). Each report provides specimen collection date and patient birth date. Anatomic site of the cancer was classified with a validated algorithm that uses natural language processing (6). Trends for breast, colorectal, lung, and prostate cancers were assessed separately because these are the most common cancers and algorithm performance for these sites was high based on historical 2018 registry data, the most recent year with complete incidence data. The average number of pathology reports per cancer case in 2018 was highest for breast (n = 3.4), followed by lung (n = 2.6), colorectal (n = 2.3), and prostate (n = 1.2) cancers, reflecting differences in staging procedures and surgical management.

Pathology reports were grouped biweekly by specimen collection date for each age group and by cancer site. The number, distribution, and patterns of biweekly reports were compared in 2020 with the same period in 2019. Biweekly COVID-19 mortality rates per 100 000 people for 2020 in Georgia and Louisiana were obtained from the US Centers for Disease Control and Prevention(https://data.cdc.gov/NCHS/Weekly-Provisional-Counts-of-Deaths-by-State-and-S/muzy-jte6). We generated 95% confidence intervals for numbers of reports and the percentage difference in 2020 and 2019 by using a simulation method based on a Poisson distribution with 1000 replications (7). This study used deidentified, routinely captured surveillance data covered under existing registry authorizations and institutional review board approvals.

There were 29 905 fewer pathology reports in 2020 than in 2019, representing a 10.2% decline (Table 1). Absolute declines in the number of reports were greatest among adults aged 50 years or older (n = 23 065); percentage declines were greatest among children and adults younger than 18 years of age (38.3%). By cancer type, absolute declines were greatest for breast (n = 5947) and lung (n = 5844) cancers; percentage declines were greatest for lung (17.4%) and colorectal (12.0%) cancers.

Between January and early March 2020, biweekly numbers of pathology reports were similar, though occasionally higher, than those during the same period of 2019 (Figure 1; Supplementary Figure 1, available online). The nadir was early April 2020, with 42.8% fewer reports than in April 2019. Declines coincided with peaks in biweekly COVID-19 mortality rates. Numbers in 2020 never consistently exceeded those in 2019 after the first decline in March.

Figure 1.

Cancer Pathology Reports: 2019 and 2020. This figure shows the number of cancer pathology reports in Georgia and Louisiana in biweekly intervals to allow comparison of the same period in 2019 as in 2020. Numbers of pathology reports (blue bars) are shown with the primary y-axis, the coronavirus disease 2019 (COVID-19) mortality rate per 100 000 population in Georgia and Louisiana (light orange area) is shown with the secondary y-axis, and the percentage change in the number of pathology reports between 2019 and 2020 (red solid line) is shown with the tertiary y-axis. Shutdown policies announced in March 2021 are indicated by the vertical bar.

Open in new tabDownload slide

Biweekly patterns of pathology reports in 2020 and 2019 were similar by cancer site, with some variation in timing, magnitude, and duration of declines (Supplementary Figure 2, available online). Biweekly patterns in 2020 and 2019 were similar by age group (Supplementary Figure 3, available online). Numbers and patterns of reports were similar in 2018 and 2019 overall (Supplementary Figure 4, A and C, available online) and in Georgia (Supplementary Figure 4, B, available online) and Louisiana (Supplementary Figure 4, C, available online), suggesting that reports in 2019 were consistent with historic patterns and 2019 was not itself an unusual year.

This study provides evidence for adverse effects of the COVID-19 pandemic on delays in cancer detection and surgical treatment using real-time pathology data through December 2020 from population-based cancer registries. Overall, there were nearly 30 000 fewer reports in 2020 than in 2019, representing a 10.2% decline over the 12 months of 2020, the longest observation period to date for studies evaluating the effects of COVID-19 on cancer-related trends. Patterns of declines were generally similar to those reported elsewhere (2,5), with the greatest differences in April 2020 relative to April 2019, the first peak in COVID-19 mortality rates. Declines in pathology reports in August, November, and December coincided with later peaks in COVID-19 mortality rates. Notably, numbers of reports through December 2020 did not consistently exceed 2019 levels after declines in March, which might be expected if a backlog of cancer diagnoses were resolved. We observed declines in cancers with effective screening tests, including breast, colorectal, and lung cancers. Furthermore, we observed substantial declines across cancer sites and age groups without effective screening tests, including children and young adults, suggesting delays in routine well-child and primary care, evaluation of signs and symptoms, and treatment initiation for most cancers. Ongoing evaluation of trends will be important.

Increases in the uninsured population throughout 2020 (8) may adversely affect timely return to routine care, regular cancer screening, evaluation of signs and symptoms, and cancer diagnosis. Fears of COVID-19 infection may have reduced initiation and completion of cancer treatment. Ongoing monitoring of population-based data will be critical for informing public health efforts to minimize any lasting adverse effects of the COVID-19 pandemic on delays in cancer diagnosis, shifts to later-stage diagnosis, receipt of timely treatment, and survival. Evaluation of racial/ethnic, socioeconomic, and geographic disparities in care and outcomes will be important as these data become available.

This study has limitations. Data are from only 2 states, although both states have urban and rural regions and diverse populations. Additionally, data are from pathology reports, not all cancer diagnoses, and the number of reports per cancer case varies by site in historical data, although this ratio should be similar across years, allowing for trend comparisons. Nonetheless, the substantial declines in cancer pathology reports in 2020 from the population-based data used in this study suggest substantial delays in cancer diagnosis and treatment as an indirect result of the COVID-19 pandemic.

Funding

No specific funding was provided for this research.

Notes

Role of the funder: Not applicable.

Disclosures: Authors declare no conflicts of interest.

Author contributions: All authors contributed to the conceptualization of the study and the writing and review of the manuscript. Data curation and/or analyses were conducted by Serban Negoita, Jennifer Stevens, Linda Coyle, and Jingxuan Zhao.

Prior presentations: Preliminary findings will be presented at the 2021 American Society for Clinical Oncology Annual Meeting, June 4, 2021.

Data Availability

Electronic pathology report data from the Georgia and Louisiana SEER registries are not part of the publicly available SEER data release.

References