Impact of COVID-19 on hospital screening, diagnosis and treatment activities among prostate and colorectal cancer patients in Canada

We employed datasets from the Canadian Institute for Health Information (CIHI) (https://​www.​cihi.​ca/​en) on prostate and colorectal cancer screening, diagnoses, treatment, and mortality events quarterly from April 2010 to March 2021, with breakdown by age group (< 40, 40–59, 60–79, 80 +) and province (with the exception of British Columbia and Quebec due to opt-outs). Specifically, CIHI’s Discharge Abstract Database (https://​www.​cihi.​ca/​en/​discharge-abstract-database-metadata-dad) and National Ambulatory Care Reporting System (https://​www.​cihi.​ca/​en/​national-ambulatory-care-reporting-system-metadata-nacrs) metadata captures ICA-10-CA- and CCI-based administrative, day surgery, and clinical information on hospital discharges directly from acute care facilities or their respective health authorities. In the current study, diagnosis and treatment data included data breakdown by stage of cancer (non-metastatic vs metastatic). Moreover, treatment data included overall data on treatment classes (imaging, surgery, radiotherapy, pharmacotherapy) along with specific treatment interventions performed. Additionally, key COVID-19 intervention data by federal, provincial and territorial governments (i.e. closures of non-essential businesses, postponement of non-essential health services, state of emergency announcements, and vaccine implementation milestones) were obtained from CIHI’s publicly available COVID-19 Intervention Timeline in Canada dataset (https://​www.​cihi.​ca/​en/​covid-19-intervention-timeline-in-canada). The current study was approved by the CIHI’s Decision Support Services and ZS Associates Research & Development Excellence Committee.

A retrospective cohort study involving nine provinces in Canada was conducted. Specifically, aggregate, ICD code-based, semi-annual prostate and colorectal cancer patient journey datasets from 2010 to 2020 with breakdown by age group, disease stage, and province were obtained from CIHI whereby addition of all independent aggregate data (screening events, hospital admissions, treatment activities, patient expirations) or averaging of all independent aggregate data (median length of stay) was performed in half-year intervals. To understand the baseline screening, diagnosis, treatment activities, length of stay, and mortality among prostate and colorectal cancer patients prior to the COVID-19 pandemic, analysis was first carried out using data from April 2010 to March 2020. Subsequent inclusion of data from April 2020 to March 2021 in our analysis allowed the direct comparison of reported data pre- and during COVID-19. We also calculated 5-year averages from April 2010 to March 2015 and April 2015 to March 2020 as a baseline for pre-COVID-19 trend assessment and statistical analyses for each outcome. All descriptive statistical analyses were performed using R, version 4.2.0 (The R Foundation for Statistical Computing, Vienna) and data are reported and shown as mean ± SEM unless otherwise indicated. Specifically, two-tailed t tests were used to analyze normally distributed data for each mean comparison in our study. Furthermore, normality of the data distribution was tested using the Shapiro–Wilk’s test. A dataset with a p value > 0.05 from the Shapiro–Wilk’s normality test was considered to follow a normal distribution. In situations where the data is not normally distributed (a p value < 0.05 from the Shapiro–Wilk’s normality test), Mann–Whitney U test, a non-parametric test, was used for the mean comparison. Statistical comparison was considered significant at the p ≤ 0.05 significance level.

Available aggregate hospital data for prostate and colorectal cancer screening, diagnosis, treatment activities, hospital length of stay, and mortality were obtained using the most responsible ICD codes from CIHI. Detailed specification of each ICA-10-CA/CCI code used in the current is presented in Supplemental Table 9. The median length of stay was calculated where at least one of the ICD-10-CA most responsible diagnosis codes listed above was identified for prostate and colorectal cancer. The number of patient expirations was reported where a most responsible diagnosis code listed above for prostate and colorectal cancer was identified at the time of expiration. The analysis did not include data where age was unknown nor data from out of hospital (private/clinic practices) and abandoned procedures. Lasty, to comply with CIHI’s data privacy and confidentiality policies, in instances in which there were fewer than five reported cases, the reported values were suppressed to avoid residual disclosure of the identity of individuals or health facilities.

To establish the baseline trends in hospital cancer screening, admission, and treatment activities prior to the COVID-19 pandemic, we utilized data from April 2017 to March 2020 to assess the general trends and variations across the Canadian prostate and colorectal cancer landscapes. A 6-month average of 67,345 ± 2,047 colorectal hospital screening events and 6,779 ± 75 prostate and 12,226 ± 82 colorectal cancer hospital admissions were reported from April 2017—March 2020 prior to COVID-19 (Supplemental Table 1–3). To determine the up-to-date trends in cancer screening, we have excluded hospital prostate cancer screening data from our analysis due to the wide practice of performing the prostate specific antigen (PSA) blood test outside of the hospital setting. Nevertheless, the summarized data is presented in Supplemental Table 1. With the nationwide suspension of non-urgent medical procedures and services in March 2020 to increase COVID-19 patient intake capacity, a marked decrease in colorectal screening events (p < 0.0001 compared to baseline) as well as prostate (p < 0.0001 compared to baseline) and colorectal cancer hospital admissions (p < 0.0001 compared to baseline) was observed between April 2020 and September 2020 (Fig. 1a, e–f). At the regional level, a significant reduction in colorectal cancer screening was observed in AB/MB/SK (p = 0.006 compared to baseline) where an outbreak of COVID-19 was recorded between October 2020 and March 2021 (Supplemental Table 1). Notably, regional differences in colorectal cancer screening coverage on a per capita basis were observed with AB/MB/SK and the Atlantic provinces trailing behind ON by 0.05%-0.08% in population coverage (Fig. 1b).

Our analysis also indicated that a greater impact was reported for non-metastatic hospital admissions among prostate (9% decrease in metastatic vs 19% decrease in non-metastatic hospital admissions) and colorectal (13% decrease in metastatic vs 19% decrease in non-metastatic hospital admissions) cancer patients across all regions during the initial 6 months of COVID-19 (Supplemental Table 2, 3 & Fig. 1d, f). For prostate cancer, a significant decrease in metastatic hospital admissions was observed throughout in AB/MB/SK and ON (Supplemental Table 2). For colorectal cancers, a marked decrease in both metastatic and non-metastatic hospital admissions was observed across all regions between April 2020 and September 2020, and a quick recovery in hospital admissions was reported between October 2020 and March 2021 with non-metastatic hospital admission in ON and metastatic hospital admission in ATL exceeding its baselines (Supplemental Table 3). Interestingly, an unexpected increase in the number of non-metastatic colorectal cancer cases above baseline in the < 40 age group was observed during COVID-19 despite significantly reduced screening effort during this time.

We next performed a detailed landscape analysis to assess hospital treatment activities for prostate and colorectal cancers (Supplemental Table 4, 5). Between April 2017 and March 2020, a 6-month average of 1,487 ± 33 (24%) imaging-, 4,205 ± 41 (68%) surgery-, 397 ± 21 (6%) radiotherapy-, and 74 ± 5 (1%) pharmacotherapy-based interventions were reported for prostate cancer patients (Supplemental Table 4). Similarly, a 6-month average of 7,937 ± 36 (98%) surgery-, 26 ± 2 (0.003%) radiotherapy-, and 83 ± 15 (0.01%) pharmacotherapy-based interventions were reported for colorectal cancer patients (Supplemental Table 5). For prostate cancer, a significant decrease in imaging- and surgery-based interventions was observed in ON and ATL while AB/MB/SK managed to minimize the impact of COVID-19 on surgical and imaging treatment activities (Fig. 2a–c). Interestingly, a significant increase in radiotherapy and pharmacotherapy above its baselines was reported in ON, suggesting potential changes in patient treatment plans during COVID-19. In line with this observation, a narrowing gap between hospital admissions and treatment activities was observed for prostate cancer during COVID-19 (Fig. 2d). For colorectal cancers, a significant decrease in surgical interventions was observed across all regions during the initial months of COVID-19, nonetheless, a ~ 99% of surgical intervention volume was restored between October 2020 and March 2021 (Fig. 2e–g and Supplemental Table 5). Of note, a gradual increase in radiotherapies and pharmacotherapies above its historical baselines for colorectal cancers was reported in ON during COVID-19. A consistent gap between colorectal cancer hospital admissions and treatment activities was observed from April 2017 to March 2021, suggesting the possibility of treatment activities performed outside of the hospital setting (Fig. 2h).

Our analysis showed that there were natural variations in the median calculated hospital length of stay (LOS) for both prostate and colorectal cancers (Fig. 3a, b, e, f). Between April 2017 and March 2020, a 6-month average median calculated hospital LOS of 6.2 ± 0.4 days for metastatic and 1.9 ± 0.2 days for non-metastatic prostate cancer was reported (Supplemental Table 6). Similarly, a 6-month average of median calculated hospital LOS of 8.1 ± 0.1 days for metastatic and 6.2 ± 0.1 days for non-metastatic colorectal cancers was reported (Supplemental Table 7). While it is expected that metastatic cancer cases would require an extended hospital stay, our analysis showed that hospitals in ON consistently reported a lower hospital LOS than hospitals in AB/MB/SK and ATL for metastatic prostate cancer and non-metastatic colorectal cancers, albeit not significant (Fig. 3a, f).

To assess the variations in hospital LOS among prostate and colorectal cancer patients during COVID-19, we compared the median calculated LOS between April 2020 and March 2021 to its historical baseline data. For prostate cancer, a significant decrease of LOS was reported in ATL (Supplemental Table 6). An extended LOS was also reported for non-metastatic cancer patients likely from the 80 + age group in ATL between April 2020 and September 2020. For colorectal cancers, while a marked decrease in hospital LOS was observed for metastatic cancer patients in ATL, a significant increase in hospital LOS was observed for metastatic cancer patients in ON, suggesting a potentially overburdened cancer care system between October 2020 and March 2021 (Supplemental Table 7). Interestingly, our analysis also demonstrated that the hospital median calculated LOS increased proportionally as cancer patients aged from the 40–59 cohort to the 60–79 and 80 + cohorts (Fig. 3c, g). It is important to highlight that the changes in LOS observed in the current study are likely direct consequences of the precautionary COVID-19 measures and protocols instituted by the regional authorities. However, a follow-up investigation is warranted to examine whether LOS is impacted for cancer patients with delayed diagnoses.

Lastly, we assessed the changes in hospital patient mortality for prostate and colorectal cancers during COVID-19. Between April 2017 and March 2020, a 6-month average of 83 ± 4 prostate cancer patient expirations and 266 ± 6 colorectal cancer patient deaths were reported. For prostate cancer, a trend of decreasing patient hospital deaths was reported between April 2020 and March 2021. Similarly, a significant decrease in colorectal cancer patient hospital deaths across all regions was reported during COVID-19 with significantly fewer patient deaths reported from the 60–79 and 80 + age groups (Supplemental Table 8). Our mortality rate analysis showed that a trend of increasing prostate cancer mortality rate was observed in ATL over time whereas a relatively consistent, but slightly higher mortality was reported for colorectal cancers (Fig. 3d, h).

To further investigate the impact of service interruption due to COVID-19, we next performed the observed-to-expected (OE) analysis for prostate and colorectal cancer hospital screening, admissions, and treatment activities during the first two waves of COVID-19 in Canada. Figure 4 demonstrated the total procedural deficits generated between April 2020 and March 2021. For prostate cancer hospital admissions, AB/MB/SK and ATL reported a total deficit of < 10% of their respective annual baselines while ON reported a total deficit of 14% of its annual baseline (Fig. 4a). For prostate cancer treatment activities, all regions reported a total deficit of < 5% of their annual baselines (Fig. 4b). Similarly, AB/MB/SK, ON, and ATL reported a total deficit of 34%, 49%, and 27%, respectively, for colorectal cancer screening (Fig. 4c), accumulating to an estimated total of 58,000 + unperformed colonoscopies in Canada. For colorectal cancer hospital admissions, our OE results indicated that a 13% deficit in AB/MB/SK, 8% deficit in ON, and 12% deficit in ATL were reported (Fig. 4d). In line with the reported admission deficits, our results showed that AB/MB/SK reported an 8% deficit in colorectal cancer treatment activities and ON and ATL reported an 8% and 12% deficit in colorectal cancer treatment activities, respectively (Fig. 4e). All in all, an increasing OE ratio close to 1, though a good sign for our cancer care system, is likely insufficient to restore a healthy and efficient treatment environment for cancer patients.

More importantly, to help plan the incremental efforts required to take on all unperformed screening events and treatment procedures, an estimation analysis was performed to calculate the potential monthly capacity increase required to clear all unperformed cancer procedures (Fig. 5). Specifically, the potential monthly capacity increase as a percent of baseline was derived from dividing the total case deficit between April 2020 and March 2021 by the desired number of months to clear all unperformed procedures. Figure 5a demonstrated that ON and ATL reported a total of 545 and 106 unperformed treatment activities, resulting in a potential monthly capacity increase of 5% and 2% above baseline assuming a recovery time of 3-month is desired. Given the lower sense of urgency for clearing unperformed colonoscopies, an estimated monthly capacity increase of 11% in AB/MB/SK, 16% in ON, and 9% in ATL would be required to clear unperformed colonoscopies in 6 months (Fig. 5b). On the other hand, an estimated monthly hospital capacity increase of 5% in AB/MB/SK, 5% in ON, and 11% in ATL would be required to clear all unperformed colorectal cancer treatment activities in 3 months (Fig. 5c).