Outpatient visit trends for internal medicine ambulatory care sensitive conditions after the COVID-19 pandemic: a time-series analysis

The COVID-19 pandemic led to rapid changes in the delivery of healthcare services to meet patient needs throughout the pandemic while reducing the risk of exposure to SARS-CoV-2. Among other adaptations, healthcare systems rapidly transitioned from traditional in-person office-based visits to predominantly virtual care [1‐4].

Within Ontario, a Canadian province with a population of 14,500,000 and universal health insurance, adoption of virtual care was low before the pandemic [5, 6]. Barriers included challenges with reimbursement, policy issues, and limited access to technology [1, 7‐9]. Within Ontario specifically, virtual delivery of physician services were previously only covered by the Ontario Health Insurance Plan if delivered from a government-approved facility using a government-approved telemedicine platform [3, 5].^3,5 After the declaration of COVID-19 as a global pandemic, new temporary billing codes were issued on March 14th, 2020, which allowed physicians to be renumerated for outpatient visits conducted via telephone or videoconference at the same rate as an in-person visit which facilitated the widespread adoption of virtual care [3, 10]. While previous studies have examined changes in visit volumes and uptake of virtual care during the pandemic, these have generally been limited to primary care [4, 5, 10] or narrow scopes of care [11, 12].

To our knowledge, no study to date has assessed the COVID-19 pandemic’s impact on visit trends and utilization of virtual care across multiple internal medicine ambulatory care sensitive conditions (ACSCs). ACSCs are health conditions where high quality and timely care may prevent acute exacerbation, sequelae of disease, and emergency department visits or hospitalizations [13]. By assessing changes in outpatient visits for several ACSCs across a single institution, we can understand the relative impact of the pandemic and virtual care on different conditions. These findings may highlight areas where virtual care is better suited than others. Additionally, understanding care patterns for multiple ACSCs may help organizations plan future allocation of resources within an evolving healthcare environment in the ongoing pandemic with the goals of preventing clinical deterioration and avoidable inpatient visits.

Our objectives were to understand how the onset of the COVID-19 pandemic and introduction of virtual care billing codes impacted patterns of care and uptake of virtual care for outpatients with common internal medicine ACSCs at a large academic ambulatory hospital in Ontario, Canada. Furthermore, we sought to understand patient and visit factors associated with utilization of virtual care.

In this analysis of outpatient visits to internal medicine clinics for ACSCs at a large ambulatory hospital in Ontario, Canada, we found a dramatic shift towards virtual care after the onset of the COVID-19 pandemic, as well as several changes in visit trends for overall visits and by specific visit diagnoses. While new consultations decreased by 10% during the pandemic, this was offset by an increase in follow-up visits leading to an overall 15% increase in visits. Our results showed a transition from primarily in-person visits pre-pandemic to predominantly virtual care during the pandemic. Importantly, there were several factors associated with visit modality in the pandemic period, with patients who were older, men, and without a registered email account more likely to be seen in-person than virtually. There were also significant differences in uptake of virtual care by clinic and visit diagnosis, with less than half of heart failure visits being seen virtually, in contrast to diabetes care which was nearly entirely virtual.

The dramatic shift towards virtual care during the COVID-19 pandemic has been consistently shown in other studies across a number of jurisdictions, which is consistent with our findings [1‐5, 10‐12, 17‐19]. A study of visit trends to Veterans Affairs clinics in the United States during the first 10 weeks of the pandemic showed a decrease in in-person visits by 56% which was partially offset by an increase in telephone and video visits, but overall visits still decreased by 30% [1]. Similarly, other studies from the US and Canada have shown an overall decline in visit numbers after the pandemic [10, 20, 21]. These findings are in contrast to our findings of a significant increase in patient volumes during the pandemic after accounting for prior trends, which was driven by more frequent follow-up visits. The reasons for these differences in findings are unclear and may be due to our study having a longer follow-up period allowing for a correction in visits after an initial drop with the start of the pandemic or due to differences in populations and conditions studied.

A prior study demonstrated that new consultations were postponed while follow-up visits shifted to primarily telemedicine during the pandemic [22], which is consistent with our findings, but the literature on patterns in internal medicine is lacking. We do not have data on the reasons leading to decreased new consultations and increased follow-up visits; however, there are several potential factors. Decreased visits in primary care during the pandemic may have led to decreased referrals as demonstrated in Ontario [10] and other jurisdictions during the first wave of the pandemic [23‐25]. Other reported barriers, such as patient concerns about acquiring COVID-19 leading to delay seeking care [26] and decreased access to care [27] may have also impacted new visits. Regarding increased follow-up visits, telephone visits may have facilitated easier access to follow-up visits with a prior study showing that virtual follow-up visits were more efficient [28, 29]. Similarly, both patients and providers have reported feeling telemedicine visits worked best for follow-up visits [30]. Furthermore, with the introduction of billing codes for telephone visits, physicians may be booking more follow-up telephone visits for brief calls that would have previously occurred but not been remunerated. Increased follow-up visits may be due to physicians wanting closer follow-up because they are unable to examine their patients, patients seeking specialist care rather than going to emergency departments for exacerbations of their chronic illness, decreased patient access to primary care [10] leading to more frequent telephone visits with specialists, or physicians feeling more comfortable asking patients for frequent follow-ups because of the convenience of telemedicine [29]. Finally, delays in access to tests and procedures during the pandemic [27] may have led to patients needing to rebook earlier follow-up after completing delayed investigations.

To our knowledge, our study is the first to examine how visits for a range of common internal medicine clinic conditions changed during the pandemic. We found variable uptake of virtual care across different clinics and conditions. Patient visits for CHF, hypertension, and asthma each increased by 20%, whereas atrial fibrillation visits decreased by nearly 20%. We cannot discern the reasons for the variability by condition but there are several potential patient, provider, and system factors that may contribute. Our finding that patients who were older were less likely to use virtual care is consistent with past studies, [12, 21, 31, 32] as well as those showing that older patients are less likely to have internet and are slower to adopt technology [33‐35]. Also, sicker or older patients may require more frequent follow-up. For example, CHF patients who are at risk of exacerbation may need closer follow-up to remotely monitor for signs of exacerbation with reduced in-person visits for volume status exams. Similarly, hypertension may lend itself easily to virtual follow-up if patients monitor their blood pressure at home allowing for quick telephone visits to titrate medications [29]. Female patients were more likely to be seen virtually in our study, although findings from other studies on sex differences in virtual care uptake have been mixed [10, 21, 31]. In terms of provider factors, a prior study showed that physicians viewed virtual care as optimally suited for managing conditions that primarily involved counselling and were less reliant on physical exams, and in particular, that hypertension and diabetes were easily managed virtually [29]. Other provider factors that may contribute to variability include familiarity or lack thereof with virtual mediums or supplier-induced demand. Finally, system factors likely contributed to decreased atrial fibrillation visits because of decreased clinic hours and fewer presentations to the emergency department, which is the primary referral mechanism to this clinic.

There are several limitations of our study. First, this was a single centre study of select internal medicine clinics and conditions at an academic hospital under a universal healthcare system in Canada. Our findings may not be generalizable to other conditions or health care systems. Second, visit diagnoses were based on the most responsible diagnoses coded by the visit physician, and for some patients, multiple conditions could have been addressed in a single visit. It is possible that some patients followed for multiple conditions by the same provider (e.g. atrial fibrillation and heart failure) had varying visit diagnoses across different visits despite the same conditions repeatedly being addressed in the same visit. This may contribute to the trends in conditions seen in the pre-pandemic and pandemic periods. However, we expect this to affect a minority of patients and not substantially alter our findings. Third, we were unable to extract data on comorbidities and sociodemographic factors that may be associated with uptake of virtual care such as high disease burden, ethnicity, language, socioeconomic status, and under-housing. Other studies from the US have shown these factors to be associated with telemedicine use [21, 31, 36, 37]; however we were limited to data available within the EMR. Finally, while this study examined trends in outpatient internal medicine visits and uptake of virtual care during the pandemic, we did not examine quality of care, outcomes, or patient satisfaction.

The COVID-19 pandemic led to a transition in the delivery of healthcare in Ontario from almost entirely in-person visits to over 80% virtual care in a short time period. Virtual care remained the predominant model of care delivery a year after the onset of the pandemic, and there were changes in the types of office visits and visit diagnoses seen during the pandemic. Several patient and visit factors were associated with greater uptake of virtual care. This study contributes to the growing literature on the use of virtual care and the effect of the pandemic on health services delivery. Future research is needed to understand drivers at the patient, provider, and system level led to decreased new consultations and increased follow-up visits. Data from other centres is needed to understand if our observation that patients with certain conditions were more likely to be seen virtually is replicated and understand the drivers behind these differences. Similarly, future research is needed on the equity of virtual care and how certain groups may be disadvantaged by the rapid shift towards virtual care. Finally, given the persistence of virtual care one year from pandemic onset as well as the multiple extensions of the temporary billing codes by the Ontario government, we suspect that virtual care will continue to be utilized in the future as part of the physician’s toolbox to provide quality care. However, this speaks to the importance of future work to understand the impact of virtual care on quality of care, patient outcomes, patient satisfaction, and cost-effectiveness of outpatient care.