Introduction
Coronary artery disease (CAD) is a leading cause of morbidity and mortality worldwide and in Canada.12 For individuals with stable CAD, regular follow-up is crucial to monitor symptoms and optimise risk factors.3 Individuals are often followed by cardiologists through periodic in-person visits or, increasingly, by primary care providers.4 Although shifting care to primary care providers may reduce the burden on cardiology clinics, evidence indicates that specialist involvement is associated with higher-quality care and improved outcomes.5 Unfortunately, resource limitations often restrict the feasibility of providing in-person specialist follow-up for all eligible patients.6
The COVID-19 pandemic led to a dramatic expansion in the use of virtual care worldwide. In Ontario, virtual visits accounted for less than 5% of physician encounters before 2020, surging to more than 70% during the pandemic.7 Although this shift was initially driven by necessity, it highlighted the feasibility of virtual care as a supportive model for chronic disease management that could complement traditional in-person care.8 Multiple studies have since shown that virtual follow-up can achieve clinical outcomes comparable to in-person care, with consistently high levels of patient satisfaction.910
Virtual care has been successfully integrated into the management of several chronic conditions with promising results. In heart failure, telemedicine-based management has been associated with reductions in mortality and hospitalisations compared with usual care.11 Similarly, among individuals with type 2 diabetes, telemedicine interventions have resulted in meaningful improvements in glycaemic control relative to standard management.12 These findings demonstrate that virtual care can effectively support patients in achieving guideline-directed targets of therapy. However, despite these successes in other chronic conditions, limited evidence exists regarding the ability of virtual care to manage individuals with stable CAD.
To address this gap, we developed a virtual, team-based clinic for individuals with stable CAD at an ambulatory care hospital in Canada. By shifting routine in-person follow-ups to virtual, this model has the potential to prioritise these in-person specialist visits to unstable patients, reduce unnecessary hospital visits and shorten wait times for new referrals. In addition, the virtual model can enhance access to specialist care for patients with mobility limitations or those living in remote areas.
Aim
This project aims to implement and evaluate the feasibility and safety of an interdisciplinary nurse practitioner (NP)-led virtual clinic for delivering secondary prevention to individuals with stable CAD. We hypothesise that a dedicated virtual CAD clinic can support individuals in maintenance of GDMT and achievement of therapeutic targets safely, while also yielding high levels of patient satisfaction.
Methods
Study design and context
We conducted a feasibility study of a virtual, interdisciplinary CAD clinic from July 2021 to July 2024. The clinic was based at a university-affiliated, ambulatory hospital in Canada. The cardiology division, staffed by six cardiologists, serves a diverse patient population from the Greater Toronto Area.
A convenience sample was recruited based on eligible referrals and clinic capacity. The initiative was enabled by the existing virtual care infrastructure at the institution, including the MyHealthRecord Electronic Medical Record (EMR) system, which allowed for secure patient-provider communication and video visits. SQUIRE writing guide and SQUIRE checklist (see online supplemental material A) were used to adhere to SQUIRE reporting guidelines.1314
Patient and public involvement
Patients were not involved in the development of research questions, study design or outcome selection. Patients were involved in the study through participation in virtual visits and completion of patient satisfaction surveys, which informed evaluation of the programme. Study findings will be shared with clinical teams and used to inform future programme development and patient care delivery.
Eligibility and recruitment
Eligible patients were defined as adults ≥18 years with stable CAD, defined as ≥1 year post-myocardial infarction, post-percutaneous coronary intervention or post-coronary artery bypass grafting or diagnosed via coronary CT angiography, with left ventricular ejection fraction ≥50% and stable symptoms. Patients with documented left ventricular dysfunction or symptoms suggestive of unstable disease were excluded.
At the time of enrolment, participants had received at least 1 year of in-person follow-up with a cardiologist. The research team conducted EMR screenings and collaborated with cardiologists to identify and refer eligible, interested patients. After obtaining patient consent, a referral was placed in the EMR to transfer care from their original cardiologist to the virtual clinic.
Subsequent in-person follow-up with the referring cardiologist was cancelled, and those appointment slots were reassigned to other patients requiring in-person evaluation.
Virtual clinic intervention
The virtual clinic was led by an interdisciplinary team including a cardiologist, NP, cardiology fellow and pharmacist (as needed). Video visits were conducted weekly by NP or fellow (as available), with the option of converting to telephone if technical barriers. Time allotted for visits was 30 min for initial and 15 min for follow-up.
To ensure intervention fidelity, NPs and fellows followed a standardised visit template covering GDMT review, risk factor assessment and patient education. Staff received written and verbal guidance on use of the template. Timing of clinic was based on provider availability and clinic resources.
Follow-up frequency was individualised based on clinical needs, with at least one virtual visit annually. Self-reported GDMT maintenance was assessed during each video visit. Case review with the supervising cardiologist took place at the end of each weekly clinic.
The pharmacist would assist with asynchronous medication consultations with providers as needed.
Participants were required to have a baseline assessment, including a lipid panel with apolipoprotein B and lipoprotein (a), glycated haemoglobin, creatinine and high-sensitivity C-reactive protein unless these tests were performed within the last month prior to enrolment. Participants without a home blood pressure (BP) monitor were offered a validated BP device (Omron BP-745 - BP Monitor)15 at no cost once this was identified during the study, and they were instructed on proper home BP monitoring per Hypertension Canada Guidelines.16 For participants identified with knowledge gap around home BP monitoring, teach-back method was used during visits to ensure proper technique.17
Participants who developed new or worsening symptoms or had recurrent cardiovascular events during the study duration were returned to in-person visits, with additional testing as clinically indicated.
Data collection and measures
The study’s primary measures were threefold: 1) maintenance of guideline-directed medical care defined as the change from baseline to 1 year in the proportion of individuals self-reporting continued GDMT and achieving blood-pressure and lipid targets; 2) feasibility and safety: defined by the proportion of scheduled visits completed via video without escalation that necessitated in-person care; and 3) patient satisfaction, captured through an anonymous, purpose-built, two-question five-point Likert scale survey developed by the clinical team.
Clinical and demographic data were collected via retrospective chart review in the EMR at baseline, after each clinic visit and at 1 year follow-up. GDMT maintenance was based on self-report and was not verified against pharmacy dispensing or prescription records. A study-developed patient satisfaction survey, which had not undergone formal validation, was distributed electronically via an anonymous MyHealthRecord link following each video visit.
Statistical analysis
All analyses were conducted using SAS, V. 9.4 and were descriptive. Twelve-month analyses included participants who completed follow-up; outcome-specific denominators reflect the available data.
No imputation was performed. This small quality improvement project was not powered or designed for hypothesis testing; as such, p-values and inferential comparisons were not reported. Missingness was not assumed to be random.
Ethical considerations
This initiative was formally reviewed by institutional authorities and was deemed not to require Research Ethics Board approval. This study was approved through the hospital’s Assessment Process for Quality Improvement Projects. Participating cardiologists provided consent to have their patient rosters screened for recruitment.
All participants provided informed consent before enrolment. Consented patients selected their preferred communication method: MyHealthRecord or email via an e-consent form. Patients declining both electronic options were mailed a copy of the consent form, and all subsequent non-visit communications were delivered via mail or telephone.
Participants could withdraw from the study at any time without impacting their care. All data except surveys were stored in a password-protected Excel document in accordance with institutional privacy and research policies. Surveys were collected and managed using REDCap electronic data capture.
REDCap (Research Electronic Data Capture) is a secure, web-based software platform designed to support data capture for research studies.1819
Results
Participant flow and demographics
Ninety-one potentially eligible individuals were invited to participate with 66 individuals choosing to enrol between July 2021 and July 2024. At the time of data collection, 43 (65.2%) participants completed the 12 month follow-up and were included in the analysis (figure 1). Among the 43 completers, the mean age was 66.4 years (SD 7.77), and 33 (76.7%) were male (table 1).
At baseline, 34 participants (79%) had a prior revascularisation procedure and 16 (37.2%) had a history of prior myocardial infarction. Mean age at the time of diagnosis was 59 years (SD 9.55), and the mean time from event to first virtual visit was 6.97 years (SD 6.33) (See online supplemental material B for data).
Compared with completers, the 14 participants who were eligible for but did not complete 12 month follow-up were similar in mean age (68.3 vs 66.4 years) and the proportion who were male (78.6% vs 76.7%).
A total of 145 appointments were completed during the study period: 66 initial consultations and 79 follow-up appointments.
Guideline-directed medical therapy
At baseline, reported use of GDMT was high: 41/43 (95.3%) on acetylsalicylic acid, 40/43 (93%) on statins, 32/43 (74.4%) on angiotensin-converting enzyme inhibitors (ACE inhibitors) or angiotensin II receptor blockers (ARBs), 25/43 (58.1%) on β-blockers and 6/43 (13.9%) on sodium-glucose co-transporter 2 inhibitors. In addition, 6/43 (13.9%) were on ezetimibe and 4/43 (9.3%) were on PCSK9 inhibitors (proprotein convertase subtilisin/Kexin type 9).
At 1 year, rates of self-reported maintenance of GDMT remained high: 41/43 (95.3%) for acetylsalicylic acid, 40/43 (93.0%) for statins, 32/43 (76.7%) for ACE inhibitors/ARBs, 24/43 (55.8%) for β-blockers, and 6/43 (14.0%) for sodium-glucose co-transporter two inhibitors, ezetimibe 8/43 (18.6%), and PCSK9 inhibitor 5/43 (11.6%) (table 2).
Achievement of therapeutic targets
Therapeutic targets were defined according to the 2018 Hypertension Canada Guidelines16 and 2021 Canadian Cardiovascular Society Dyslipidaemia Guidelines.20 13 of the 43 participants who completed the 1 year follow-up had missing BP data. Nine did not have a home BP monitor and were subsequently offered one, one had a monitor but did not know how to use it and three did not regularly monitor their BP despite being advised to do so. Consequently, follow-up BP target attainment was calculated among the 30 participants with available measurements.
One patient had missing low-density lipoprotein (LDL) and triglyceride values due to not completing the blood work requested ahead of the visit.
At baseline, 21/43 (48.8%) met blood-pressure targets (<130/80 mmHg) compared with 23/30 (76.7%) at 1 year follow-up among those with available data. As 13 participants lacked follow-up BP measurements, these proportions were based on different denominators and do not represent a paired estimate of change. LDL targets (<1.8 mmol/L) were achieved by 30/43 (69.8%) initially and remained stable at 30/42 (71.4%) at follow-up.
The proportion of participants reaching triglyceride targets (<1.5 mmol/L) slightly decreased with 31/43 (72.1%) at baseline compared with 29/42 (69.1%) at 1 year (table 3).
Clinic feasibility and safety
Nine appointments (6.2%), including six initial consultations and three follow-up appointments, were completed by telephone. Seven were conducted by telephone because the participants were not enrolled in MyChart, one because the participant was locked out of their MyChart account and another because the participant’s computer was not functioning.
Of 145 completed visits, only 2 (1.4%) required escalation to in-person care, one due to a myocardial infarction after their initial virtual visit and another due to the development of unstable symptoms. However, the latter participant continued to follow virtually despite recommendations.
Patient satisfaction
A total of 58 post-visit patient satisfaction surveys were completed. 91.3% of participants were very satisfied or satisfied with their care, and 77.6% felt the virtual visit was equal to or better than in-person care.
Discussion
This study evaluated the feasibility of virtual care for secondary prevention in patients with stable CAD and yielded several important findings. First, participants received at least 1 year of in-person cardiology follow-up prior to enrolment, with high baseline GDMT use that was maintained following transition to virtual care. This suggests that the virtual clinic model may complement traditional in-person care in encouraging persistence with GDMT in this population.
Second, assessment of therapeutic target achievement at 1 year relative to baseline demonstrated maintenance of lipid control, while blood-pressure target attainment was 48.8% at baseline and 76.7% among participants with an available follow-up measurement. However, 13 of 43 completers lacked follow-up blood-pressure data, the baseline and follow-up denominators were unmatched, and the study had no comparator. As such, the extent to which this difference was attributable to clinic could not be determined.
Third, only 1.4% of visits required escalation to in-person care, supporting the feasibility of comprehensive follow-up through a virtual, interdisciplinary clinic. Finally, patient satisfaction was high, with over three-quarters reporting that their virtual experience was equal to or better than in-person care.
Our findings align with the growing evidence supporting the utility of virtual cardiac care while extending it in meaningful ways. Prior studies have primarily focused on cardiac rehabilitation, heart failure or post-acute coronary syndrome populations and have demonstrated that telehealth can maintain or moderately improve risk factor management compared with in-person care,2122 with high patient satisfaction.23 For example, Jin et al.21 reviewed 30 randomised controlled trials of tele-based secondary prevention interventions initiated post-acute coronary syndrome, reporting reductions in total cholesterol, LDL and smoking, along with improved exercise capacity in patients receiving virtual care.21 Similarly, in a review of 24 randomised controlled trials by Kavradim et al.,22 which focused mainly on patients early post-acute coronary syndrome or during cardiac rehabilitation who received virtual care compared with routine care, significant improvements in lipid management, medication adherence, blood pressure and smoking cessation were found.22 These studies are consistent with our findings, demonstrating the feasibility of maintaining high levels of GDMT prescription and therapeutic targets. Our study extends the evidence, showing that virtual care can be implemented not only in transitional or high-risk populations but also for ongoing management of stable patients.
Our clinic model also structurally differed from many other virtual cardiac care studies. Whereas much of the existing literature has examined telephone-only or app-based platforms,2425 we primarily delivered care through video visits. Sanderson et al.26 evaluated a telephone-based virtual cardiology follow-up model in Ontario during the COVID-19 pandemic and found it to be feasible and acceptable, with minimal need for in-person escalation.24 In contrast, Fuda et al.23 piloted a video-based model like ours and reported high patient satisfaction, with participants describing the approach as convenient and supportive of high-quality cardiac education at home.23 Together, these studies highlight that various virtual care formats can be both feasible and acceptable in cardiac care.
Our study adds novel evidence by focusing on the potential therapeutic impact video-based care can have in achieving risk factor targets. Additionally, our high levels of patient satisfaction, consistent with other telehealth studies,2427 provide further support for team-based virtual care involving cardiologists, nurse practitioners and pharmacists – an approach that has been less extensively evaluated.
Delivering virtual care offers both benefits and challenges. Operationally, virtual clinics can improve access to care, particularly for patients living in remote areas or those with mobility limitations. Several individuals voiced that the reduction in travel time and costs was a factor in continuing in the programme and should be more formally evaluated.
At the systems level, virtual care also frees up in-person clinic capacity for patients with unstable conditions, supporting more efficient resource allocation. The use of video conferencing preserves real-time, face-to-face interaction, which may enhance patient engagement. However, virtual care also presents several implementation challenges.
Obtaining vital signs and laboratory data is more difficult in a remote setting. By using the current EMR infrastructure, our clinic was able to send laboratory requisitions electronically. Additionally, patients without a home blood pressure monitor were provided one along with education around home BP monitoring.
However, having access to a BP cuff does not equate to consistent use and still requires patient engagement. Further efforts to encourage patient engagement and compliance over a virtual platform should be explored. Another challenge involves the hardware, software and technological literacy required to participate in virtual visits.28 To address this issue, telephone visits were available as an intermittent alternative.
Our study has several limitations. First, this was a small feasibility study involving a convenience sample from a single academic centre. Participants were recruited through convenience sampling from a single academic cardiology centre, were predominantly male (73%), were already engaged in specialist cardiovascular care and had high baseline medication use.
They may therefore have been more health-literate, motivated or adherent than the broader population with stable coronary artery disease, which may limit the applicability of our findings to more diverse settings. Nevertheless, the modest sample size was consistent with the study’s objective of assessing the feasibility of this model and informing future, larger-scale evaluations.29 Second, attrition occurred in 14 of 66 enrolled participants (21.2%); an additional nine had not yet completed their scheduled 12-month follow-up at the time of data collection (figure 1). Since the 12 month findings were based on participants with available follow-up data, they may be affected by attrition and selection bias.
Third, 13 of the 43 participants who completed follow-up did not provide 12 month BP measurements. Consequently, baseline and follow-up proportions were based on different denominators and should not be interpreted as a paired estimate of improvement. This may partly reflect that access to a home BP monitor was not systematically assessed at enrolment, as participants were offered a monitor only after it was identified during the study that they did not have one.
Future implementation could address this by assessing the need for a home BP monitor at enrolment and providing one before the first appointment. More broadly, reliance on patient-generated measurements may represent an implementation challenge for virtual care and limits interpretation of the observed difference in blood-pressure target attainment. Fourth, medication maintenance was self-reported and was not verified against pharmacy dispensing or prescription records.
The findings may therefore be affected by recall or social-desirability bias. Fifth, the purpose-built two-question patient-satisfaction survey was not validated. In addition, the anonymous nature of the survey does not allow for more granular interpretation (e.g. same individual submitting multiple surveys over time, individual may have not completed 1 year follow-up) and may skew patient satisfaction to more positive than is representative of the sample.
Finally, the absence of a comparator and the non-randomised design limit conclusions regarding the clinical effectiveness of virtual care relative to in-person follow-up. These limitations indicate the need for larger controlled studies with more complete follow-up to evaluate clinical effectiveness and generalisability. Despite these limitations, our study has several notable strengths.
It addresses an important gap in the literature by providing one of the first evaluations of a virtual secondary prevention clinic specifically designed for longitudinal care for individuals with stable CAD in a Canadian healthcare setting. The study also used a pragmatic design, enhancing external validity. Conducted within a general cardiology practice and integrated into routine clinical workflows, rather than within a tightly controlled trial environment, our findings likely reflect outcomes that are achievable in real-world practice.
This strengthens their relevance for broader implementation and scalability.
The findings of this study highlight several important directions for future research. Foremost, a randomised controlled trial comparing virtual and in-person follow-up for stable CAD patients is needed to validate these results and inform clinical practice. Such a study should be adequately powered to assess major adverse outcomes, including mortality, all-cause hospitalisation and quality-of-life measures.
Longer-duration studies are also warranted to evaluate the sustained impact of virtual care on long-term risk-factor management. In addition, future research should examine the economic implications of long-term virtual follow-up, including direct costs (eg, provider time and video platform infrastructure), indirect costs (eg, patient travel and clinic workflow efficiency) and downstream healthcare utilisation (eg, hospitalisations and emergency department visits). Finally, further work is needed to determine how virtual care technologies can be optimised and adapted to reduce inequities in access and outcomes.
Conclusion
In conclusion, this feasibility study demonstrates that a video-based, interdisciplinary virtual care model can be a feasible approach for the management of individuals with stable CAD. Despite the limitations above, this study highlights the potential this method can have on GDMT maintenance, while offering meaningful operational advantages for patients and the healthcare system. By focusing on a real-world, pragmatic setting, our findings add important evidence to support the integration of virtual care in secondary prevention.
Further research, including larger and more rigorous studies, is needed to determine effectiveness and assess long-term clinical outcomes that ensure equitable and cost-effective care.
Supplementary material