The Impact of Virtual Cancer Care on Chemotherapy Delivery and Clinical Outcomes in Colorectal Cancer Patients Receiving Systemic Therapy: A Pre- and Intra-Pandemic Analysis

(1) Background: The coronavirus 2019 pandemic has resulted in an abrupt transition to virtual oncology care worldwide. This study’s objective is to evaluate chemotherapy delivery and clinical outcomes in patients on systemic treatment for colorectal cancer before and during the pandemic. (2) Methods: Clinical data was collected on patients with colorectal cancer receiving intravenous chemotherapy at The Ottawa Hospital from June 2019 to March 2021. Patients were stratified by whether they were started on chemotherapy pre-pandemic (June 2019–January 2020) or intra-pandemic (February 2020–March 2021). Multiple regression analysis was used to compare outcomes between pandemic periods; (3) Results: There were 220 patients included in this study. The proportion of virtual consultations (1.2% to 64.4%) and follow-up visits (5.2% to 83.3%) increased during the pandemic. There was no difference in the incidence of treatment delays (OR = 1.01, p = 0.78), chemotherapy dose reductions (OR = 0.99, p = 0.69), emergency department visits (OR = 1.23, p = 0.37) or hospitalizations (OR = 0.73, p = 0.43) between pandemic periods. A subgroup analysis revealed no difference in outcomes independent of the presence of metastases; (4) Conclusion: These findings serve as an important quality-care indicator and demonstrate that virtual oncology care appears safe in a cohort of high-risk colorectal cancer patients.


Introduction
Colorectal cancer is the third most common cancer worldwide and second in Canada [1]. Multidrug chemotherapy regimens are widely used for treatment in both the adjuvant and metastatic setting. Fluoropyrimidine-based agents are the backbone of therapy and commonly used systemic treatment combinations include folinic acid/fluorouracil/oxaliplatin (FOLFOX), folinic acid/fluorouracil/irinotecan (FOLFIRI) or capecitabine/oxaliplatin (CAPOX) with or without monoclonal antibodies [2][3][4][5]. Although the routine uses of combination therapy is responsible for improved patient outcomes, it is also carries a significant risk of treatment related toxicity that can necessitate adjustment to the original chemotherapy plan to allow for recovery [6]. Treatment delays and dose reductions are common modifications. In severe cases, chemotherapy can be discontinued entirely, particularly in the adjuvant setting. Ideally, clinicians aim to provide the highest relative dose intensity possible to replicate the treatments received in clinical trials and optimize patient outcomes [7][8][9]. The delivery of full dose intensity chemotherapy is considered a quality-of-care indicator in oncology [10].
Traditionally, patients are seen regularly in clinic by their oncology team to screen for treatment related side effects and facilitate early management of chemotherapy toxicities. The emergence of coronavirus disease 2019 (COVID- 19) resulted in an unprecedent shift to virtual cancer care, including patients on active systemic therapy. This decision was based due to the high risk of complications related to COVID-19 infection [11]. Virtual care can be defined as any interaction between clinician and patient that is not performed in-person, primarily by video conferencing or telephone. In most cancer centers worldwide, the shift to virtual care was abrupt and took place over a matter of days [12].
In this study, we leverage the rapid adoption of virtual oncology care during the COVID-19 pandemic to evaluate its impact on chemotherapy delivery and clinical outcomes. We chose to study a cohort of colorectal cancer patients on active treatment with intravenous (IV) chemotherapy agents to assess the safety of virtual care in patients at high-risk of treatment related adverse events.

Study Design
We conducted a retrospective cohort study of patients diagnosed with colorectal cancer at The Ottawa Hospital Cancer Centre consecutively between 1 June 2019, and 31 March 2021 (n = 1,573) and treated with IV chemotherapy (n = 282). We excluded patients with stage 1-3 rectal cancer (n = 220), due to the various treatment options and multidisciplinary nature of treatment selection in this population.
We defined the pandemic period according to the date of chemotherapy initiation. Patients beginning chemotherapy from 1 June 2019 to 31 January 2020 were categorized as "pre-pandemic", while patients beginning from 1 February 2020 to 31 March 2021 were categorized as "intra-pandemic". Our center declared a public health state of emergency and shifted towards virtual care 22 March 2020. We selected 1 February 2020 as the pandemic period cut-off date to provide a buffer for patients started on chemotherapy just before the public health emergency declaration. Patients were followed until the discontinuation of first line chemotherapy or the end of the observation period (30 September 2021); outcomes that occurred after this point were censored.

Patient Characteristics
All patient information was gathered from the Ottawa Hospital's electronic medical records (EMR), including clinical notes, oncology treatment plans and pathology/radiology reports. Baseline characteristics including age, gender, functional status, body mass index and health comorbidities, as estimated by the modified Charlson Comorbidity Index, were collected. The Charlson Comorbidity Index was modified to omit points for the presence of malignancy [13]. Cancer characteristics such as stage, histology and molecular markers were also collected.
All cancer treatment data from The Ottawa Hospital Cancer Centre was available on the local EMR, regardless of whether chemotherapy was delivered at a satellite campus. The Ottawa Hospital Cancer Centre is the regional cancer center for its entire local health network and encompasses several satellites treatment sites where chemotherapy is delivered using the same EMR. Treatment data included chemotherapy regimen, initial dose, frequency of cycles, number of cycles planned, number of cycles received, and receipt of targeted therapies. Visit type was gathered from Ottawa Hospital administrative databases. Each visit was given a specific code depending on its type (consultation versus followup) and method of assessment (in-person versus virtual). Consultations were defined as the first outpatient assessment after diagnosis, while follow-ups were defined as any subsequent visit. Virtual visits consisted of telephone encounters and electronic video conferencing; these were grouped together for the purposes of analysis. Consultation data was not available after June 2021, based on when the health records department pulled the data for this study.

Defining Outcomes of Interest
Outcomes of interest were divided into chemotherapy delivery and clinical outcomes. Chemotherapy delivery outcomes included treatment delays, dose reductions and early discontinuations. Treatment delays were considered if a chemotherapy cycle was delayed ≥1 week from the planned treatment date, dose reduction if a chemotherapy cycle's dose was reduced by ≥10% and early discontinuation if intravenous chemotherapy was terminated prior to the planned stop date. Early discontinuations were only captured in patients undergoing adjuvant therapy, where there was a prespecified number of cycles. Clinical outcomes were composed of emergency department (ED) visits and hospitalizations. These were collected by review of oncology clinical notes, records of emergency assessment or admission notes at the Ottawa Hospital. ED visits to select community hospitals that share an EMR with the Ottawa Hospital were available as well. No survival outcomes were analyzed.

Statistical Analysis
Descriptive analyses were used to describe baseline and treatment characteristics of the study population. Statistical analysis was stratified by pandemic period: pre-pandemic vs. intra-pandemic, which are defined above. The relationship between categorical variables was assessed using the Fischer's Exact Test, while continuous variables were assessed using the Mann-Whitney U test. A binary logistic regression was used to evaluate the relationship between pandemic period and the outcomes of interest for binary variables, while generalized linear regression with a negative binomial (log linked) distribution was used for continuous variables.

Virtual Consultation and Follow-Up Visits
Patients were more likely to receive virtual care during the pandemic. Of the 85 consultations performed before the pandemic, 1 (1.2%) was virtual compared to 87 of 135 (64.4%) during the pandemic. Likewise, of the 463 follow-up visits performed before the pandemic, 24 (5.2%) were virtual compared to 1513 of 1891 (83.3%) during the pandemic. This represents a relative increase of 63.2 and 77.9% for virtual consultations and follow-up visits during the pandemic, respectively. It is important to note that no consultation data was available from July to September 2021. The trend of virtual care over time was evaluated, see Figure 1. The proportion of virtual follow-up visits appears stable throughout the pandemic period, while rates of virtual consultations increase early in the pandemic and fall over time.
The relationship of outcomes and pandemic period were evaluated, see Table 2. There were no significant differences in the incidence of chemotherapy delays (odds ratio (OR) = 1.11, p = 0.75), dose reductions (OR = 1.19, p = 0.58) or hospitalizations (OR = 0.72, p = 0.39) between patients receiving IV chemotherapy in the pre-pandemic versus intrapandemic period. Similarly, we did not find significant associations of the pandemic period and the mean number (rate ratio (RR) = 1.20, p = 0.39) or the cumulative length of treatment delays (RR = 1.25, p = 0.26). The number of ED visits over the course of treatment (RR = 0.97, p = 0.90) was similar between pandemic periods, as well. These analyses were controlled for age, chemotherapy type, modified Charlson comorbidity index, metastatic status and number of cycles received. 39 Point estimates were represented as odds ratios for binary outcomes (*) and rate ratios ( †) for continuous data. (Abbreviations: ED, emergency department; CI, confidence interval; pre, pre-pandemic period; intra, intra-pandemic period).
The relationship of outcomes and pandemic period were evaluated, see Table 2. There were no significant differences in the incidence of chemotherapy delays (odds ratio (OR) = 1.11, p = 0.75), dose reductions (OR = 1.19, p = 0.58) or hospitalizations (OR = 0.72, p = 0.39) between patients receiving IV chemotherapy in the pre-pandemic versus intrapandemic period. Similarly, we did not find significant associations of the pandemic period and the mean number (rate ratio (RR) = 1.20, p = 0.39) or the cumulative length of treatment delays (RR = 1.25, p = 0.26). The number of ED visits over the course of treatment (RR = 0.97, p = 0.90) was similar between pandemic periods, as well. These analyses were controlled for age, chemotherapy type, modified Charlson comorbidity index, metastatic status and number of cycles received. Point estimates were represented as odds ratios for binary outcomes (*) and rate ratios ( †) for continuous data.

Analysis of Outcomes Based on the Presence of Metastatic Disease
There were 112 patients with local disease and 108 with metastatic disease. The most common site of metastatic disease was the liver (68.5%), followed by peritoneum (30.6%), lung (17.6%) and bone (4.6%). There were no significant differences in the incidence of treatment delays, dose reductions or hospitalizations between pandemic periods in patients with nonmetastatic and metastatic disease. There were also no significant differences in the mean number of treatment delays, cumulative length of treatment delays or number of ED visits in either subgroup. The relationship of clinical outcomes stratified by the presence of metastatic disease is summarized in Figure 2.

Analysis of Outcomes Based on the Presence of Metastatic Disease
There were 112 patients with local disease and 108 with metastatic disease. The mo common site of metastatic disease was the liver (68.5%), followed by peritoneum (30.6% lung (17.6%) and bone (4.6%). There were no significant differences in the incidence treatment delays, dose reductions or hospitalizations between pandemic periods in p tients with nonmetastatic and metastatic disease. There were also no significant diffe ences in the mean number of treatment delays, cumulative length of treatment delays number of ED visits in either subgroup. The relationship of clinical outcomes stratified the presence of metastatic disease is summarized in Figure 2.

Discussion
Virtual encounters have been integral in delivering oncology care to patients during the COVID-19 pandemic. At most Canadian centers, virtual cancer care remains a principal method of assessment. The data presented in this article demonstrates an abrupt transition to virtual oncology care in April 2020 and sustained use over 1 year after the pandemic began. Follow-up visits remained predominantly virtual throughout the study period, while the rate of virtual consultations decreased later in the pandemic. It is commonly believed that initial care plan discussions should be performed in-person, which may explain why in-person consultations became more common later in the pandemic [14]. Furthermore, this study shows no significant difference in chemotherapy delivery or clinical outcomes in patients started on treatment during the pandemic. This serves as an important quality assurance measure demonstrating the safety of virtual oncology care during the COVID-19 pandemic, in patient undergoing active treatment for colorectal cancer.
Virtual cancer care was first introduced in Canada, albeit in a limited capacity, in 2003 [15]. Until the COVID-19 pandemic, virtual cancer care was being used sparsely even for residents living in remote geographies. In Ontario, problems with physician reimbursement and the mandated use of select, pre-approved video conferencing technology limited the role of virtual care in clinical practice [16]. These barriers were eliminated during the first wave of the COVID-19 pandemic, where Ontario Health Insurance Plan (OHIP) developed temporary billing codes for virtual care delivered via telephone or mainstream web conferencing software (i.e., Zoom, Skype, etc.) making virtual care practical for clinicians. Recently, an Ontario-wide consensus statement was released to guide virtual oncology care using a modified Delphi of 29 interdisciplinary panel members, demonstrating that momentum is beginning to be gained in virtual oncology care in Canada [17].
Despite being a pillar of clinical assessment throughout the pandemic, virtual cancer care is appealing beyond pandemic circumstances too. It offers the possibility of more accessible care to patients residing in remote geographies, including patients residing in Northern Canadian communities that often have barriers to cancer care access [18]. It is also convenient, reduces travel time and lowers the cost associated with accessing care for patients [19,20]. On a system level, virtual care has the potential of providing safe, efficient and cost-effective follow-up and survivorship care to a rapidly growing Canadian cancer population [21,22]. Canadian data suggests that both patients and clinicians are generally satisfied with virtual care [23]. Similar levels of satisfaction with virtual care have been observed internationally, as well [24,25].
Current evidence for virtual oncology care is strongest for low-risk follow-up and survivorship care [26]. This has led to debate regarding how virtual care compares to traditional in-person assessment for patients undergoing active treatment. An Australian study compared 89 patients undergoing chemotherapy receiving virtual follow-up to 117 patients followed in-person. They concluded no differences in the incidence of adverse effects of therapy or hospital admissions between the two groups [27]. A separate study, evaluating the safety and cost savings of telemedicine during the pandemic in a cohort of cancer patients receiving intravenous treatment, reported no differences in clinical outcomes over a 3-month observation period [20]. Our study adds to this data by reproducing safety outcomes in a population of colorectal cancer patients receiving, predominantly, combination chemotherapy. In addition, patients with metastatic disease are typically at higher risk for toxicity based on the higher cumulative dose of chemotherapy received and larger tumor burdens. The subgroup analysis presented above shows comparable chemotherapy delivery outcomes in patients with metastatic disease, suggesting that virtual oncology care appears safe even in a high-risk cohort of colorectal cancer patients.
There continues to be barriers impacting widespread use of virtual care into clinical practice. For instance, triaging patients who are appropriate for virtual assessment remains a challenge. Survey data suggests that health-care providers feel that there are inadequate guidelines available to distinguish when a patient requires an in-clinic assessment versus when they are safe for virtual care [28]. Serious illness discussion and palliative care are other areas where health-care providers feel less comfortable with a virtual format [26]. Currently, most virtual cancer care is being performed via telephone rather than platforms that allow video-conferencing in Canada [28,29]. The regular use of video-conferencing software may help bridge this gap by enhancing nonverbal communication thus improving the connection between clinician and patient during sensitive discussions. Other obstacles that need to addressed include privacy concerns, lack of technology expertise of patients and lack of physical examination [30]. Finally, implementation of virtual nursing care is an important consideration. Nurses working in oncology clinics play a major role in supporting clinicians with patient education, treatment counselling, patient assessment and caring for the psychological, emotional, and social aspects of cancer [31]. A standardized framework for delivering virtual oncology nursing care will need to be explored to prevent marginalization of their role.
This work may lack applicability to other virtual care models, as a single-centered retrospective study. That said, The Ottawa Hospital Cancer Centre is one of Canada's largest cancer centers. It has a catchment area of approximately 1.4 million people and operates multiple satellite chemotherapy units, as well as providing remote care to patients in the Baffin region of Nunavut. This study may also be limited by its sample size, which may have impacted the ability to detect small differences in outcomes between pandemic periods, particularly when stratifying by the presence of metastasis. There were restrictions in data availability outside of our EMR. The EMR used at the Ottawa Hospital Cancer Centre encompasses all oncology satellite treatment centers, therefore treatment quality outcomes should be captured in full. Many local community hospitals share an EMR with our center but patients who sought urgent medical attention outside of the network may have been missed. Furthermore, there was tendency towards differential use of chemotherapy regiments during the pandemic. For instance, CAPOX use approximately doubled in the pandemic period, which may reflect the results of the IDEA collaboration meta-analysis demonstrating non-inferiority of 3 months CAPOX to traditional 6 months of adjuvant therapy in the subgroup analysis or the emphasis on limiting addition chemotherapy chair time [32]. CAPOX is delivered in 3-week cycles compared to FOLFOX, which is delivered in 2-week cycles. This may also explain the tendency towards less cycles of chemotherapy being delivered in the pandemic period. Finally, virtual care is one of many factors impacted by the COVID-19 pandemic. Factors such as public health regulations, fear of infection, impeded access to non-oncology health services and COVID-19 infection were outside the scope of this study. In particular, fear of infection and hospital avoidance may explain why there a trend towards less hospitalization in the pandemic era.

Conclusions
Overall, this study demonstrates no significant differences in chemotherapy delivery, emergency department use or hospitalizations in patients with colorectal cancer treated with IV chemotherapy during the COVID-19 pandemic. These findings serve as an important quality-care indicator and demonstrate that virtual oncology care appears safe in a cohort of high-risk colorectal cancer patients receiving systemic therapy. Future work dedicated to other tumor sites and cancer centers would allow for broader application of these findings.  Informed Consent Statement: Informed consent not applicable. Patient data was collected passively from electronic medical records. No change to routine clinical practice was required for this study.

Data Availability Statement:
The data presented in this study are available on reasonable request from the corresponding author. The data are not publicly available due to restrictions in place to protect patient privacy.