Comments on the Rapid Access Chest Pain Assessment Clinic: An Australian Virtual Care Experience

To the Editor,

We thank the authors for their constructive comments regarding our article describing the Rapid Access Chest Pain Assessment Clinic (RACPAC) and the transition to a virtual model of care. We welcome the opportunity to clarify the definitions of outcomes, triage processes, diagnostic strategy, and cost assumptions.

• 1. Definition of repeat presentation and safety outcomes

  In our manuscript, repeat presentation encounters referred to patients who were discharged from RACPAC and subsequently reattended or were re referred to RACPAC for further chest pain evaluation. This endpoint did not capture emergency department reattendance, inpatient admissions outside the clinic pathway, or primary care visits, as our dataset was derived from the clinic database rather than the whole hospital and community-linked sources. We agree that this is an important limitation and should be considered when interpreting the reduction in repeat presentation encounters observed during the virtual period.

We also agree that reporting major adverse cardiovascular events, including myocardial infarction and cardiovascular death, would provide an important safety anchor. However, our study was observational and retrospective, and it did not include prospective follow-up or data linkage to capture adjudicated major cardiovascular outcomes. We therefore cannot report MACE reliably from the available dataset. We have acknowledged this limitation in the manuscript, and we agree that future evaluations of virtual chest pain pathways should prioritise outcome linkage to enable robust assessment of MACE and broader health service utilisation.

• 2. Increased waiting time and triage safeguards

  We observed an increase in median waiting time from referral to consultation during the virtual period (17 days–26.8 days), while annual clinic volume increased and attendance improved (86%–94%). Pre-COVID-19 Pandemic the median waiting time from referral to the first consultation was 17 days (5–21.5). We interpret the longer wait interval as multifactorial, reflecting increased demand, altered referral patterns during COVID-19, and system-level constraints during pandemic disruptions. Importantly, the clinic was designed for low-to-intermediate risk chest pain, and eligibility criteria included a negative troponin and the absence of ischaemic ECG changes, as illustrated in the workflow figure. Referrals were primarily generated from the emergency department and general practice after initial assessment, with baseline investigations requested at the time of referral.

Regarding triage safeguards, the operational model included a liaison nurse who supported administrative and clerical work, including triaging referrals and coordinating test bookings, with clinical oversight provided by the cardiology team at the hospital, including cardiology registrars who usually provide advice to the emergency department (where the majority of referrals originated) and ultimately oversight from the treating cardiologist. Patients with low cardiovascular risk or atypical chest pain could be discharged after follow-up without further testing, while patients whose results or symptom profile warranted escalation could proceed to additional testing and invasive angiography as required. This model is consistent with contemporary frameworks that recommend risk stratification using symptoms, ECG, and biomarkers to identify patients suitable for outpatient evaluation versus those requiring urgent escalation and inpatient management [1].

We agree that reporting the time from referral to test completion and from test completion to final diagnosis would further strengthen the interpretation of diagnostic latency. These time points were not captured as structured variables within our clinic dataset, which focused on referral and consultation timing, investigation types, and results. We view this as a useful design recommendation for future prospective data collection.

• 3. Test selection criteria, definition of a positive study, diagnostic yield, and sequential testing As described in the manuscript, the selection of the investigation modality was clinician-led and based on the cardiologist’s discretion, considering the patient’s cardiac risk profile. In routine practice, this decision is influenced by factors such as age (<40 years and >70 years more likely to have functional testing), pretest likelihood (higher risk more likely to have CTCA), ability to exercise, renal function, contrast allergy, heart rate and rhythm, and local availability and expertise [1–3]. Our model did not mandate a single protocolised pathway, which is why we reported the observed patterns of testing and downstream procedures in a real-world service. Important to highlight that we have noted that CTCA offers “This preventive value is supported by evidence from stable chest pain trials showing increased use of preventive therapies and improved outcomes when CT coronary angiography guides management, particularly in those with CT-defined coronary disease.

We can clarify the definitions of a positive study, as prespecified in the methods. For CT coronary angiography, findings were categorised by calcium score and stenosis severity. Atherosclerosis with a calcium score greater than 0 was categorised as positive, with stenosis further stratified. This definition intentionally captures non-obstructive plaque as a clinically actionable disease that warrants preventive therapy and risk factor modification, consistent with guideline-based management [1–4]. Stress echocardiography was positive when regional wall motion abnormalities were present, myocardial perfusion scanning was positive with perfusion mismatch, and invasive angiography was categorised by stenosis severity.

Sequential testing was explicitly reported in Table 2. Most patients underwent a single noninvasive test, with CT coronary angiography and treadmill stress echocardiography being the most frequently used first-line modalities. Among those requiring a second test, a substantial proportion proceeded to invasive angiography, and only a small minority required a third test. For downstream yield, we reported revascularisation rates following each initial modality, defined as subsequent PCI or CABG within the care pathway. Revascularisation occurred in 34.6% after CT coronary angiography, 26.7% after stress echocardiography, 20% after myocardial perfusion scanning, and 33.3% after invasive angiography.

• 4. Cost analysis assumptions and sensitivity analysis

  We appreciate the emphasis on generalisability. In our setting, the virtual clinic used existing cardiology department infrastructure, including consultation rooms, telehealth, and information technology systems, at no additional direct cost. Staffing was the principal expenditure, including the liaison nurse role that facilitated triage and bookings. Imaging and clerical support costs were not included because we assumed they were broadly comparable across inpatient and outpatient pathways for low- to intermediate-risk chest pain.

To address the request for sensitivity analysis, a simple one-way sensitivity assessment supports the robustness of the conclusion in our setting. Using the same framework as Table 3, varying clinic staffing costs by ±20% changes the estimated annual net savings from approximately AUD 265,000 to AUD 303,000, compared with the base case of AUD 283,663. Varying the admission cost benchmark by plus or minus 20% yields annual savings of approximately AUD 208,000 to AUD 359,000. While these are illustrative, they show that the direction of effect remains favourable across plausible ranges, with staffing and admission cost assumptions being the key drivers. We agree that jurisdictions with materially different telehealth costs, salary structures, or reimbursement models should perform local sensitivity analyses when implementing similar services.

In summary, we thank the authors for highlighting important areas for clarification. We agree that linkage to major cardiovascular outcomes and capture of diagnostic latency metrics would strengthen future evaluations. We hope the clarifications above assist readers in interpreting the scope, strengths, and limitations of our study, and in considering translation of virtual chest pain pathways to other settings.

Sincerely,

Marwan Shawki

On behalf of the authors