Changes in the top 25 reasons for primary care visits during the COVID-19 pandemic in a high-COVID region of Canada

Purpose We aimed to determine the degree to which reasons for primary care visits changed during the COVID-19 pandemic. Methods We used data from the University of Toronto Practice Based Research Network (UTOPIAN) to compare the most common reasons for primary care visits before and after the onset of the COVID-19 pandemic, focusing on the number of visits and the number of patients seen for each of the 25 most common diagnostic codes. The proportion of visits involving virtual care was assessed as a secondary outcome. Results UTOPIAN family physicians (N = 379) conducted 702,093 visits, involving 264,942 patients between March 14 and December 31, 2019 (pre-pandemic period), and 667,612 visits, involving 218,335 patients between March 14 and December 31, 2020 (pandemic period). Anxiety was the most common reason for visit, accounting for 9.2% of the total visit volume during the pandemic compared to 6.5% the year before. Diabetes and hypertension remained among the top 5 reasons for visit during the pandemic, but there were 23.7% and 26.2% fewer visits and 19.5% and 28.8% fewer individual patients accessing care for diabetes and hypertension, respectively. Preventive care visits were substantially reduced, with 89.0% fewer periodic health exams and 16.2% fewer well-baby visits. During the pandemic, virtual care became the dominant care format (77.5% virtual visits). Visits for anxiety and depression were the most common reasons for a virtual visit (90.6% virtual visits). Conclusion The decrease in primary care visit volumes during the COVID-19 pandemic varied based on the reason for the visit, with increases in visits for anxiety and decreases for preventive care and visits for chronic diseases. Implications of increased demands for mental health services and gaps in preventive care and chronic disease management may require focused efforts in primary care.

total physician visits, office visits and virtual visits before COVID-19 with trends after pandemic-related public health measures changed the delivery of care, according to various patient and physician characteristics. We used interrupted time series analysis to compare trends in the early and later halves of the COVID-19 period.

Study design and setting
We conducted a retrospective, population-based analysis using linked health administrative data to assess changes in total visit volume and visit type in primary care for all residents in Ontario, from the pre-pandemic period to the period during which COVID-19-related restrictions on in-person consultations came into effect in mid-March 2020. We compared trends for the months of January through July for 2019 and 2020 to account for seasonality. Ontario is Canada's most populous province, with a 2020 population of 14 745 040 people. 7 All permanent residents have full coverage for necessary physician and hospital services, including primary care visits, without copayments or deductibles.

Data sources and collection
We used the following administrative databases: the Ontario Health Insurance Plan (OHIP) database for physician claims; the Registered Persons Database, which is Ontario's health care regis try for eligible patients; Client Agency Provider Enrolment tables for patients in primary care enrolment models; the Corpor ate Provider Database for physicians in patient enrolment models; and the ICES Physician Database for physician characteristics. These data sets were linked using unique encoded identifiers and analyzed at ICES. ICES is an independent, nonprofit research institute whose legal status under Ontario's health information privacy law allows it to collect and analyze deidentified health care and demographic data, without consent, for health system evaluation and improvement.

Study population
For this study, primary care physicians were defined as family doctors and general practitioners, but not pediatricians or general internists. We included only comprehensive primary care physicians, defined as those meeting minimum visit levels, billing mainly primary care codes and providing a diversity of core primary care services. 8 Office and home visits were defined using relevant billing codes. Virtual care was defined as any primary care physician billing for telephone or video visits using either the temporary new virtual care codes or existing telemedicine codes. The temporary codes did not distinguish telephone from video visits and did not include asynchronous care such as email or text. Throughout both the 2019 and 2020 study periods, the Ontario Telemedicine Network supported video visits in secure, physical host sites. As of Nov. 15, 2019, video visits with patients in their home or other locations were also supported. From Apr. 1, 2020, onward, specific codes for virtual care using video through the Ontario Telemedicine Network were introduced, at the usual rate for office care (Appendix 1). Office, home and virtual visits were limited to 1 per patient per physician per day.

Patient and physician characteristics
We collected data on the following patient characteristics: age, sex, neighbourhood income, first-time registration for health care coverage, rurality, primary care enrolment model and expected health care use. We collected information on age and sex from the health care registry. We derived neighbourhood income using postal codes and the 2016 Canadian Census, divided into quintiles of equal size, with quintile 1 having the lowest income and quintile 5 the highest. We used first-time registration for health care coverage within the previous 10 years as a proxy for recent immigration, as most recent registrants are immigrants. We determined rurality using the Rurality Index of Ontario (RIO), 9 including large urban (RIO score of 0), urban (RIO score 1-9), small urban (RIO score 10-39) and rural (RIO score ≥ 40). Primary care enrolment models included enhanced fee-for-service, blended capitation, blended capitation with an interprofessional Family Health Team, not in a patient enrolment model (fee-for-service only) and not having sufficient visits to be attributable to a primary care physician. 10 Expected health care use was assessed using the Johns Hopkins Adjusted Clinical Groups Resource Utilization Bands, with 0 being no health care use and 5 being the highest expected use. 11 We also collected data on the following characteristics of primary care physicians: sex, age, country of graduation and panel size, which we derived from provincial and ICES physician databases. We analyzed visit rates and type of visit by these various patient and physician characteristics.

Statistical analysis
We calculated weekly counts of total, office, home and virtual visits from January to July of 2019 and 2020, focusing on visit counts in the COVID-19 period (Mar. 11 to July 28, 2020) and the pre-COVID-19 comparison period (Mar. 12 to July 29, 2019). We calculated mean daily rates of visits per thousand people in the COVID-19 period compared with the pre-COVID-19 period. The denominator for rates included all Ontario residents registered with the provincial health insurance plan in each period. We used 2-sample z tests to compare visits in Mar. 11 to July 28, 2020 with those in 2019 for the same time period. Because major changes in visits were still occurring between Mar. 11 and Mar. 31, 2020, we conducted interrupted time series analyses from Apr. 1 to July 28, 2020, divided into roughly equal earlier (Apr. 1 to May 26, 2020) and later (May 27 to July 28, 2020) COVID-19 time periods. We calculated average daily physician visits in each week, accounting for long weekends. We did not detect any significant autocorrelation using the Durbin-Watson test in any outcome, so we performed ordinary least squares interrupted time series regression. We stratified all analyses by sex; sexspecific data are not reported because no major differences were found, but are available on request.

Ethics approval
The use of data in this project was authorized under section 45 of Ontario's Personal Health Information Protection Act, which does not require review by a Research Ethics Board. Table 1 summarizes the characteristics of all Ontario residents eligible for health care during the 2020 COVID-19 period (n = 14 574 884) and the 2019 pre-COVID-19 comparison period (n = 14 388 566), and the number of primary care visits by Ontarians according to visit location. The characteristics of the primary  to July 28, 2020), total visits were 28.0% lower than before COVID-19; office visits were 79.1% lower than in the pre-COVID-19 period, and virtual visits constituted 71.1% of all visits (Table 3).

Results
Comparing the COVID-19 period with the corresponding period in 2019, total visits declined by 28.0%, but with relatively smaller decreases among women (26.3%), older adults (23.9% among those age 65-74 yr and 19.1% among those age ≥ 75 yr), long-term OHIP registrants (25.1%), patients who could not be attributed to a primary care physician (10.2%) and patients with high expected health care use (18.2% among those in the second highest group and 8.3% among those with the highest expected use). Rural residents had a 6.4% increase in total visits. The greatest declines were among children (50.5%) and those with low expected health care use (55.7%) ( Table 3).
When evaluating the change in total visits by physician characteristics, which decreased by 27.1% overall, there were relatively lower decreases among patients seeing female physicians (25.3%), older physicians (20.8% among those age 65-74 yr and 20.1% among those age ≥ 75 yr), Canadian graduates (24.7%) and those with large panel sizes (21.0% for those with ≥ 2500 patients) ( Table 4).
Virtual care constituted 71.1% of all visits in the COVID-19 period, with higher proportions of virtual care visits among women (72.4%), adults aged 65-74 years (73.4%), those in the highest income quintile (74.3%), long-term OHIP registrants (72.8%), patients cared for in Family Health Teams (75.7%), and those with the highest expected health care needs (73.1%). The lowest use of virtual care was among children (57.6% of all visits), those not in a patient enrolment model (63.4%), those who could not be attributed to a primary care physician (62.5%), rural residents (60.6%) and those with the lowest expected health care use (61.2%) ( Table 3). Higher proportions of virtual care were provided by female physicians (75.4%) and physicians aged 30-44 years (75.0%) ( Table 4).
Weekly counts of visits are depicted in Figure 1, showing a precipitous decline in office visits and a large increase in virtual visits in mid-March 2020. The lowest number of office visits occurred in mid-April; by late July, the numbers had more than tripled, but still remained far below the pre-COVID-19 period. Total visits gradually increased from mid-April onward but remained lower throughout the COVID-19 period than in 2019, reaching 83.4% of the 2019 level by the end of the time period.   Interrupted time series analyses comparing Apr. 1-May 26, 2020, with May 27-July 28, 2020, showed that average daily visits increased significantly by 1509 visits per week (95% CI 957-2061) in the earlier time period, with a significantly lower increase of 517 visits per week (95% CI 55-979) in the later time period. There was an upward trend in average daily office visits that was significantly greater in the later time period, with an increase of 993 visits per week (95% CI 778-1208) in the earlier time period and 1410 per week (95% CI 1230-1590) in the later time period. Virtual visits increased significantly in the earlier COVID-19 period (512 visits per week, 95% CI 132-892), but declined significantly in the later COVID-19 period (897 visits per week, 95% CI 579-1214) (see Appendix 2 for details, available at www.cmaj.ca/lookup/doi/10.1503/ cmaj.202303/tab -related-content).

Interpretation
We describe sudden, striking shifts in primary care patterns in Ontario in the early months of the COVID-19 pandemic, when physical distancing directives necessitated shifts in the way health care was delivered. We found an almost 80% decrease in office visits and a 56-fold increase in virtual visits, changing most dramatically in mid-March 2020. Although it will take time to fully understand the impact of these changes and their effects on different groups, our early findings provide some reassurance that the groups with the highest care needs, including those older than 65 years and those with higher levels of morbidity, maintained relatively higher levels of care overall. Virtual care increased markedly for all groups, with relatively small differences across patient and physician characteristics.
Trends over time showed some recovery of office visits after the initial precipitous decline, but not back to the previous year's baseline by the end of July 2020. The COVID-19 period, from mid-March to the end of July 2020, was not a homogeneous period, with gradual and regional lifting of restrictions from early May to mid-July. 12 Trends showed increasing office visits over this time period, with greater increases in June and July, along with decreases in virtual care. Although valuable services were undoubtedly lost, it is likely that unnecessary visits and low-value care were also reduced. The lower levels of virtual care seen among children and in rural areas may warrant further attention.
Other jurisdictions have also reported decreases in the use of primary and ambulatory care, and rapid increases in virtual care during the early phases of the COVID-19 pandemic. [13][14][15][16] A large decrease in physician services was documented in 3 Canadian provinces in March and April, with subsequent increases in May and June 2020, 17 and a partial rebound in services after a decline has   also been seen in the US. 18 One study noted that blood pressure and cholesterol assessments, and new medication visits, declined substantially. 19 Most studies described overall trends and did not assess equity in changes in access, although pre-COVID-19 studies have documented inequitable access to virtual care. 20 Telemedicine and virtual care have been established in Canadian health care for decades, especially for Northern and remote regions, but until 2020, virtual visits comprised a very small proportion of all care. Virtual care has the advantages of reducing the impact of health care delivery on the environment and eliminating the potential for exposure to pathogens such as severe acute respiratory syndome coronavirus 2. Its disadvantages include the inability to perform most physical examinations or procedures, difficulty establishing new therapeutic relationships, dealing with some complex mental health issues, missing body language and nonverbal cues and lacking the full degree of comfort and support that can be provided in person. Educational attainment, digital literacy, age, rurality, language and culture all contribute to the "digital divide." 20,21 Concerns have been raised about virtual visits, including privacy, continuity of care and equity of implementation. 21,22 An Ontario pilot study of virtual care showed that, when provided with a choice of audio, video or text messaging on an integrated platform, more than 90% of visits occurred using asynchronous, secure text messaging followed by audio, yet there are no billing codes for text messaging. 28 Despite extensive use of virtual care in the COVID-19 pandemic, [23][24][25][26][27] the appropriate role of virtual care remains to be determined. It is not possible to separate the role of physician funding for virtual care from the impact of the pandemic itself in the major uptake of virtual care. However, the pandemic lockdown was undoubtedly an important driver of the large decline in office visits. The longer-term prospects for funding of virtual care, including which modalities are funded, likely rest on its impact on access to care, quality of care and costs, all of which hold promise but require further investigation and policy development. How care is provided also has large workforce implications that require further exploration. Canadians appear to be highly satisfied with virtual care and up to one-third would like virtual care to be the first point of contact after the pandemic. 29 There is support for virtual care to be covered by employer health plans, posing challenges to continuity of care and equity, if virtual care is not publicly funded in the future. 30

Limitations
Strengths of this study include its population-wide coverage and use of recent data, but it also has several limitations. We did not assess reasons for visits and therefore could not assess the need or value of forgone visits. It was not possible to separate telephone visits from video visits using temporary billing codes, and we could not assess the use of email or secure messaging because no billing codes were available for those services. The extent of care provided by phone before the pandemic is not well understood, so the increase in virtual care may have been overestimated. The Johns Hopkins Adjusted Clinical Groups system relies on health care use, so the disruption in care during the COVID-19 pandemic could have affected these measures. Billings for July 2020 were those submitted in August 2020, so some incompleteness is expected, especially for the final week of July. COVID-19 has disproportionately affected racialized communities, and the lack of routine data collection about ethnicity hinders our further understanding of these disparities, including forgone care and the use of virtual care. 31 The available data do not include primary care provided by nonphysicians. A greater understanding of care provision by primary care providers is needed, including the reasons for changes in care and an assessment of physicians who may have completely stopped in-person visits or ceased practising entirely. How and why both physicians and patients made choices about modality of care requires further attention, as do reasons for those choices, such as a lack of adequate personal protective equip-ment. Other Canadian provinces and territories implemented different billing codes to support virtual care, so our findings are not generalizable to other Canadian jurisdictions.

Conclusion
We report preliminary data on the extent to which office and virtual primary care changed during the initial months of COVID-19, and how this varied by type of patients and physicians. We found sudden and dramatic decreases in office visits and large increases in virtual care, with an overall substantial decrease in care provided. These changes affected patient and physician subgroups differently. The determinants and consequences of these major shifts in care, and for which patients and providers, require further study. J a n . 7 J a n . 1 4 J a n . 2 1 J a n .

No. of primary care visits
Week O ice 2020 Home 2020 Virtual 2020 Total 2020 Total 2019 Figure 1: Weekly primary care total visits and visits by type (office, home or virtual), January to July 2020, and total visits, January to July 2019, Ontario, Canada.