Advertisement

Centralized triage of suspected coronary artery disease using coronary computed tomographic angiography to optimize the diagnostic yield of invasive angiography.

  • J.-D. Schwalm
    Correspondence
    Corresponding author: J-D Schwalm MD, MSc, FRCPC, Population Health Research Institute, Hamilton General Hospital Campus, DBCVSRI, 237 Barton Street East, Hamilton, ON, Canada L8L 2X2
    Affiliations
    Population Health Research Institute, McMaster University, Hamilton, ON, Canada

    Department of Medicine, Division of Cardiology, McMaster University, Hamilton, ON, Canada

    Hamilton Health Sciences, Hamilton, ON, Canada

    Department of Medicine, Niagara Health Services, St. Catharine’s, ON, Canada
    Search for articles by this author
  • Zachary Bouck
    Affiliations
    Epidemiology Division, Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada
    Search for articles by this author
  • Madhu K. Natarajan
    Affiliations
    Population Health Research Institute, McMaster University, Hamilton, ON, Canada

    Department of Medicine, Division of Cardiology, McMaster University, Hamilton, ON, Canada

    Hamilton Health Sciences, Hamilton, ON, Canada

    Department of Medicine, Niagara Health Services, St. Catharine’s, ON, Canada
    Search for articles by this author
  • Natalia Pinilla
    Affiliations
    Population Health Research Institute, McMaster University, Hamilton, ON, Canada

    Department of Medicine, Division of Cardiology, McMaster University, Hamilton, ON, Canada

    Hamilton Health Sciences, Hamilton, ON, Canada

    Department of Medicine, Niagara Health Services, St. Catharine’s, ON, Canada
    Search for articles by this author
  • Danielle Walker
    Affiliations
    Hamilton Health Sciences, Hamilton, ON, Canada

    Department of Radiology, McMaster University, Hamilton, ON, Canada
    Search for articles by this author
  • Nida Syed
    Affiliations
    Hamilton Health Sciences, Hamilton, ON, Canada

    Department of Radiology, McMaster University, Hamilton, ON, Canada
    Search for articles by this author
  • David Landry
    Affiliations
    Hamilton Health Sciences, Hamilton, ON, Canada

    Department of Radiology, McMaster University, Hamilton, ON, Canada
    Search for articles by this author
  • Ali Sabri
    Affiliations
    Department of Radiology, Niagara Health Services, St. Catharine’s, ON, Canada
    Search for articles by this author
  • Vikas Tandon
    Affiliations
    Department of Medicine, Division of Cardiology, McMaster University, Hamilton, ON, Canada

    Hamilton Health Sciences, Hamilton, ON, Canada
    Search for articles by this author
  • James Nkurunziza
    Affiliations
    Department of Medicine, Division of Cardiology, McMaster University, Hamilton, ON, Canada

    Hamilton Health Sciences, Hamilton, ON, Canada
    Search for articles by this author
  • Monica Taljaard
    Affiliations
    Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, ON, Canada

    Epidemiology and Community Medicine, University of Ottawa, Ottawa, ON, Canada
    Search for articles by this author
  • Tej Sheth
    Affiliations
    Population Health Research Institute, McMaster University, Hamilton, ON, Canada

    Department of Medicine, Division of Cardiology, McMaster University, Hamilton, ON, Canada

    Hamilton Health Sciences, Hamilton, ON, Canada

    Department of Medicine, Niagara Health Services, St. Catharine’s, ON, Canada
    Search for articles by this author
Open AccessPublished:November 18, 2022DOI:https://doi.org/10.1016/j.cjco.2022.10.009

      ABSTRACT

      Background

      Coronary computed tomographic angiography (CCTA) is preferable to invasive coronary angiography (ICA) for coronary artery disease (CAD) diagnosis in elective patients without known CAD.

      Methods

      We conducted a non-randomized interventional study involving two tertiary care centres in Ontario. From July 2018–February 2020, outpatients referred for elective ICA were identified through a centralized triage process and recommended to undergo CCTA first instead of ICA. Patients with borderline or obstructive CAD on CCTA were recommended for subsequent ICA. Intervention acceptability, fidelity, and effectiveness were assessed.

      Results

      226 patients were screened with 186 confirmed eligible, of whom 166 had patient and physician approval to proceed with CCTA (89% acceptability). Among consenting patients, 156 (94%) underwent CCTA first; 43 (28%) had borderline/obstructive CAD on CCTA and only 1 with normal/non-obstructive CAD on CCTA was referred for subsequent ICA against protocol (99% fidelity). Overall, 119/156 CCTA-first patients did not have ICA within the following 90 days (i.e., 76% potentially avoided ICA due to intervention). Among the 36 who underwent ICA post-CCTA per protocol, 24 had obstructive CAD (66.7% diagnostic yield). If all patients referred for and underwent ICA at either centre between July 2016–February 2020 (n=694 pre-implementation; n=333 post-implementation) had CCTA first, an additional 42 patients per 100 would have had an obstructive CAD finding on their ICA (95% CI=26–59).

      Conclusion

      A centralized triage process, in which elective outpatients referred for ICA are instead referred for CCTA first, appears to be acceptable and effective in diagnosing obstructive CAD and improving efficiencies in our health system.

      Keywords

      Acronyms:

      CAD (Coronary Artery Disease), ICA (Invasive Coronary Angiography), CCTA (Coronary Computed Tomographic Angiography), PCI (Percutaneous Coronary Intervention), CABG (Coronary Artery Bypass Graft)

      1. INTRODUCTION

      Coronary artery disease (CAD) is the leading cause of death worldwide.
      • Roth G.A.
      • Forouzanfar M.H.
      • Moran A.E.
      • et al.
      Demographic and epidemiologic drivers of global cardiovascular mortality.
      The gold standard for the diagnosis of CAD is invasive coronary angiography (ICA). However, the frequency of identifying obstructive CAD on ICA performed in elective patients without a history of CAD is low.
      • Patel M.R.
      • Peterson E.D.
      • Dai D.
      • et al.
      Low diagnostic yield of elective coronary angiography.
      Using the American College of Cardiology’s National Cardiovascular Data Registry, Patel and colleagues studied nearly 400,000 patients without prior known CAD undergoing elective ICA and found that 62% of patients had non-obstructive disease or normal findings.
      • Patel M.R.
      • Peterson E.D.
      • Dai D.
      • et al.
      Low diagnostic yield of elective coronary angiography.
      Despite lower population rates of ICA in Ontario, Canada, over 50% of elective ICA procedures find non-significant CAD (Corhealth Ontario Quarterly Quality Report Cards, unpublished). Unfortunately, this problem persists despite years of province-wide efforts including audit and feedback of cardiac center performance as the rate of non-obstructive disease on ICA remains unchanged (54% in 2018 to 51% in 2021- Corhealth Ontario Quarterly Quality Report Cards, unpublished). A prior retrospective analysis revealed that inappropriate referrals are not a significant contributor to the low diagnostic yield of ICA, as over 90% of patients meet criteria for this procedure according to current standards.
      • Ouellette M.L.
      • Beller G.A.
      • Löffler A.I.
      • Workman V.K.
      • Bourque J.M.
      Appropriate referrals of angiography despite high prevalence of normal coronary arteries or nonobstructive CAD.
      Nonetheless, the low diagnostic yield of elective ICA is concerning, as this test poses risks to patients (including death, stroke, myocardial infarction, and vascular access complications) and is costly to health care systems.
      • Tavakol M.
      • Ashraf S.
      • Brener S.J.
      Risks and complications of coronary angiography: a comprehensive review.
      Utilization of non-invasive diagnostic tests as a first-line approach are recommended to improve the diagnostic yield of elective ICA. Currently, approximately 60% of elective outpatients without known CAD undergo functional stress testing prior to referral to ICA (Ontario-59%, Hamilton Health Sciences-57%- Corhealth Ontario Quarterly Quality Report Cards, unpublished 2021). Coronary computed tomographic angiography (CCTA) is a non-invasive imaging test that provides an anatomic assessment of CAD with lower risk of serious complications and reduced health care costs compared with ICA.
      • Rudziński P.N.
      • Kruk M.
      • Kępka C.
      • et al.
      Assessing the value of coronary artery computed tomography as the first-line anatomical test for stable patients with indications for invasive angiography due to suspected coronary artery disease. Initial cost analysis in the CAT-CAD randomized trial.
      CCTA has the highest sensitivity for detection of >50% coronary stenosis among the non-invasive modalities currently available, with most estimates between 90% to 95%.
      • Haase R.
      • Schlattmann P.
      • Gueret P.
      • et al.
      Diagnosis of obstructive coronary artery disease using computed tomography angiography in patients with stable chest pain depending on clinical probability and in clinically important subgroups: meta-analysis of individual patient data.
      ,

      National Institute for Health and Care Excellence. Recent-onset chest pain of suspected cardiac origin: assessment and diagnosis. 2010;Clinical Guideline CG95. https://www.nice.org.uk/guidance/cg95.

      Furthermore, CCTA has 100% sensitivity and 95% specificity for detection of high-risk, prognostically significant coronary anatomy (including left main stenosis, three-vessel disease, or two-vessel disease with involvement of the proximal left anterior descending artery) among intermediate risk patients (i.e., 25-60% likelihood of a significant stenosis) undergoing ICA.
      • Sheth T.
      • Amlani S.
      • Ellins M.L.
      • et al.
      Computed tomographic coronary angiographic assessment of high-risk coronary anatomy in patients with suspected coronary artery disease and intermediate pretest probability.
      Increasing evidence suggests that first-line use of CCTA can improve the diagnostic yield of ICA. For example, Dewey et al. identified low-risk patients referred for ICA with atypical angina and randomized these patients to undergo either (1) ICA or (2) CCTA first with subsequent ICA dictated by the treating physicians.
      • Dewey M.
      • Rief M.
      • Martus P.
      • et al.
      Evaluation of computed tomography in patients with atypical angina or chest pain clinically referred for invasive coronary angiography: Randomised controlled trial.
      In the CCTA first group, ICA was performed in only 14% of patients (vs 100% in the ICA group) and a substantial increase in the diagnostic yield of ICA was observed (75% vs 15%).
      • Dewey M.
      • Rief M.
      • Martus P.
      • et al.
      Evaluation of computed tomography in patients with atypical angina or chest pain clinically referred for invasive coronary angiography: Randomised controlled trial.
      Furthermore, a recent randomized controlled trial demonstrated that among patients referred for ICA because of stable chest pain and intermediate pretest probability of CAD, the risk of major adverse cardiovascular events was similar in the CCTA first group and the ICA first group.
      • Maurovich-Horvat P.
      • Bosserdt M.
      • Kofoed K.F.
      • et al.
      CT or invasive coronary angiography in stable chest pain.
      Lower procedural-related complications were demonstrated in the CCTA first group.
      • Maurovich-Horvat P.
      • Bosserdt M.
      • Kofoed K.F.
      • et al.
      CT or invasive coronary angiography in stable chest pain.
      The United Kingdom’s National Institute for Health and Care Excellence and the American College of Cardiology/American Heart Association both strongly recommend a CCTA-first approach for patients with stable angina and suspected CAD.

      National Institute for Health and Care Excellence. Recent-onset chest pain of suspected cardiac origin: assessment and diagnosis. 2010;Clinical Guideline CG95. https://www.nice.org.uk/guidance/cg95.

      ,
      • Gulati M.
      • Levy P.D.
      • Mukherjee D.
      • et al.
      AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guideline for the evaluation and diagnosis of chest pain: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.
      Despite the accumulating evidence in favour of CCTA, few patients undergo CCTA prior to ICA in Canada. Difficulties with timely access to CCTA due to limited diagnostic infrastructures,

      Canadian Agency for Drugs and Technology in Health. “The Canadian Medical Imaging Inventory 2019–2020.” Canadian Journal of Health Technologies 1, no. 1 (January 2021): 1–215. Accessed March 15, 2022. https://cadth.ca/sites/default/files/ou-tr/op0546-cmii3-final-report.pdf.

      ,

      The Age of Imaging Equipment in Canada. Canadian Medical Imaging Inventory Service Report, Ottawa: CADTH, October 2021. Accessed February 11, 2022. https://www.cadth.ca/sites/default/files/attachments/2021-11/cmii_service_age_of_imaging_equipment.pdf.

      outdated Canadian guidelines recommending non-invasive functional testing,
      • Mancini G.B.
      • Gosselin G.
      • Chow B.
      • et al.
      Canadian Cardiovascular Society. Canadian Cardiovascular Society guidelines for the diagnosis and management of stable ischemic heart disease.
      a lack of centralized triage with coordination between ICA and CCTA testing, and practice inertia of referring physicians may all contribute to the current suboptimal use of CCTA and low diagnostic yield of ICA overall.
      • Cabana M.D.
      • Rand C.S.
      • Powe N.R.
      • et al.
      Why don’t physicians follow clinical practice guidelines.
      In the CarDIA study (CoronAry computed tomogRaphic angiography to optimize the Diagnostic yield of Invasive Angiography in lower risk patients), our aim was to improve the diagnostic yield of ICA and determine the number of ICAs potentially avoided by using CCTA first in elective patients with suspected CAD. Specifically, we implemented and assessed a centralized triage process for regional outpatients referred for ICA to two tertiary care centers in Ontario, Canada. Eligible patients were triaged to undergo CCTA first. We assessed (1) the proportion of eligible patients who did not undergo CCTA due to refusal from their referring physician or patients themselves (intervention acceptability), (2) the proportion of patients who underwent ICA within 90 days post-CCTA against recommendations of the triage process (intervention fidelity), and (3) the potential impact of first-line use of CCTA on diagnostic yield of ICA in elective patients (intervention effectiveness) based on an exploratory analysis using linked trial and administrative data.

      2. METHODS

      2.1 Design and setting

      We conducted a non-randomized interventional study between July 2016 and February 2020 (24 months pre-intervention; 20 months post-intervention) in the southwest health region of Ontario, Canada, which has a population of over 1.8 million people. Within this health region, there are 22 hospital sites including two cardiac centres—Hamilton Health Sciences’ General Hospital (referred to as Hamilton hereafter) and Niagara Health Systems’ St. Catharine’s Site (referred to as Niagara hereafter)—that are tertiary care centres where advanced cardiac procedures (including CCTA and ICA) are performed.

      2.2 Intervention and target population

      We implemented a centralized triage process within the regional ICA triage office located in the Cardiac Catheterization Lab (CCL) in Hamilton from July 2018 through February 2020. During the implementation period, triage staff aimed to screen the majority of non-urgent outpatient referrals for elective ICA at both the Hamilton and Niagara sites. Eligibility criteria also included variables that predicted a high quality CCTA (e.g. sinus rhythm). Eligible patients were non-urgent (elective) outpatients referred for ICA with a Canadian Cardiovascular Society Functional Classification Angina Score of 0 (asymptomatic), I (angina with strenuous, rapid, or prolonged exertion), or II (angina with moderate exertion) and the indication for ICA was to rule out CAD or cardiomyopathy. Patients were ineligible if they met any of the following criteria at time of referral: were <18 years old, male and >65 years old, or female and >75 years old; had a prior CCTA; were on anticoagulants (proxy marker for atrial fibrillation); had an elevated creatinine level (>150 mmol/L); diabetes mellitus on oral hypoglycemic or insulin; had high-risk exercise stress test or functional imaging (Appendix 1); known severe valvular disease being considered for valve surgery; or any known history of CAD including prior acute coronary syndrome, prior percutaneous coronary intervention (PCI) or coronary artery bypass graft (CABG) procedure. Potentially eligible patients identified by the triage process underwent a secondary review by one of the study’s two principal investigators (JDS or TS) to confirm their eligibility. Eligible patients were then considered for CCTA first instead of ICA, as initially referred.
      All referring physicians in the region were contacted by email prior to study commencement for consent to include their eligible patients in the novel diagnostic pathway. Eligible patients of consenting physicians were contacted by phone to obtain verbal consent for CCTA instead of ICA. If either the referring physician or patient refused CCTA, then ICA was arranged as per usual care. Patients were not contacted for study purposes beyond obtaining consent and standard procedures (e.g., booking of CCTA or ICA).
      All CCTAs were read by a level 3-trained cardiologist and radiologist. The intervention protocol dictated that all patients with a borderline or obstructive CAD finding on their CCTA were to be reviewed by the principal investigators (JDS or TS) and, where appropriate, booked for subsequent ICA. Patients with a normal finding or evidence of non-obstructive CAD on CCTA were not recommended to undergo ICA. Obstructive CAD was defined as left main artery stenosis of ≥50% or stenosis of ≥70% of a main epicardial coronary artery (i.e., >2 mm in diameter). CAD was defined as borderline, if the stenosis of a main epicardial coronary artery was 50-70%.
      For the first 5 months of implementation (July–November 2018), only patients referred to undergo ICA in Hamilton were screened and referred for CCTA (if eligible and consenting). Screening and referral of patients in Niagara for CCTA began in December 2018 after the site’s CCTA program was sufficiently established. This study was approved by the Hamilton Integrated Research Ethics Board (HIREB #4697).

      2.3 Participant screening and identification in administrative data

      We maintained a study-specific database (CarDIA) that catalogued screened patients, their assessed eligibility (yes or no), as well as their CCTA and ICA booking dates and results. Since only patients screened by the centralized triage process for CCTA eligibility during the implementation period were captured in the CarDIA database, we also collected data from the CorHealth Ontario Cardiac Registry, which contains demographic, clinical (e.g., comorbidities, history of cardiac procedures), administrative (e.g., referral date, referring physician, wait time data, patient’s wait location), and procedure-specific information (e.g., location, procedure findings and complications) for all adult patients undergoing ICA in Ontario.

      CorHealth Ontario. Data Holdings and Statements of Purpose List. Corhealth Ontaio Data Holdings. https://www.corhealthontario.ca/CorHealth-Ontario-Data-Holdings.pdf. Accessed: March 2022.

      Where necessary, records were linked across both databases using encoded patient identifiers.
      Using local CorHealth Ontario Cardiac Registry data from both participating cardiac centres, we sought to identify all outpatients with a non-urgent referral for ICA (i.e., similar to the target population for the intervention) between July 2016 through February 2020 (24 months pre- and 20 post-implementation). To be considered eligible for inclusion in the analysis, patients had to meet criteria based on their ICA referral record (Appendix 2), which were intended to approximate the clinical inclusion and exclusion criteria used to identify eligible patients through the implemented centralized triage process outlined in section 2.2. For patients with ≥2 unique ICA referrals meeting the eligibility criteria during the study period, we only selected their first referral record (earliest by date). This restriction was necessary as eligible patients who were referred for and underwent CCTA in the intervention period would not be included by the triage process for subsequent ICA referrals (i.e., these patients would not be referred for first-line CCTA again due to prior CCTA).

      2.4 Evaluation of acceptability and fidelity of the intervention

      Using the CarDIA database, we summarized the frequency (and proportion) of screened patients during the implementation period who were eligible for inclusion and recommended to undergo a CCTA instead of ICA as initially referred. To gauge acceptability of the intervention among physicians and patients, we calculated the proportion of eligible patients who did not undergo CCTA due to (1) refusal from their referring physician and (2) self-refusal.
      A process evaluation was conducted to further assess acceptability of the intervention by patients, referring physicians, and hospital administration (CCL triage staff). Specifically, we conducted telephone interviews with patients who underwent CCTA and/or ICA, online surveys with physicians, and semi-structured telephone interviews and a focus group with triage staff. Surveys, focus groups, and interviews consisted of Likert-scale statements and open-ended questions regarding attitudes and perceptions about the pathway, implementation challenges, and personal experience and satisfaction (Appendix 3).
      Intervention fidelity was assessed by calculating the proportion of patients with a normal or non-obstructive CAD finding on CCTA who were subsequently referred for ICA against protocol.

      2.5 Evaluation of intervention effectiveness

      Using the CarDIA database, we first tabulated the number of patients who underwent first-line CCTA during the 20-month implementation period but who did not subsequently undergo ICA within 90 days after their CCTA. This value was interpreted as the number of ICA procedures potentially avoided due to having CCTA first. Second, we calculated the diagnostic yield of ICA among all patients who underwent CCTA first during the implementation period.
      Finally, we assessed the potential change in diagnostic yield of ICA attributable to the intervention. For this assessment, we included all eligible outpatients in the CorHealth Ontario Cardiac Registry who were referred for, and subsequently underwent, ICA at either study site between July 2016 and February 2020 (24 months pre- and 20 months post-implementation). Those patients who could be subsequently linked to the CarDIA database and who had a recorded CCTA within 90 days before their ICA were considered “exposed” to CCTA first whereas all other patients were considered “unexposed”. Per patient, we noted whether their ICA found obstructive CAD (as defined in section 2.2). We then fit a logistic regression model with significant CAD finding on ICA regressed on terms for exposure (CCTA first; yes or no, several patient-level characteristics (age, sex (male or female), Canadian Cardiovascular Society Functional Classification of Angina score, primary indication for ICA to “rule out CAD” (yes or no), and procedure site (Hamilton or Niagara Site). Using post-regression marginalization, we obtained a risk difference (RD) comparing the marginal probability of a significant CAD finding on ICA between exposed (CCTA first) and unexposed patients.
      • Austin P.C.
      Absolute risk reductions, relative risks, relative risk reductions, and numbers needed to treat can be obtained from a logistic regression model.
      A corresponding 95% confidence interval (CI) was obtained via non-parametric percentile bootstrapping with 500 samples (with replacement).
      • Austin P.C.
      Absolute risk reductions, relative risks, relative risk reductions, and numbers needed to treat can be obtained from a logistic regression model.

      3. RESULTS

      3.1 Intervention acceptability and fidelity

      Based on the CarDIA database, a total of 226 patients were screened by the centralized triage process and identified as eligible by the triage staff during the implementation period (July 2018–February 2020) (Figure 1), with 40 (18%) of these patients considered ineligible for diagnostic CCTA upon secondary review by one of the two principal investigators (JDS or TS). Of the 186 confirmed eligible patients recommended to undergo first-line CCTA instead of ICA, 8 (4.3%) patients had their referring physician refuse CCTA on their behalf and 2 could not be contacted. Of the 176 patients we were able to contact, 10 (5.7%) refused CCTA. Therefore, overall acceptability was 89% (166/186). Reasons for CCTA refusal are summarized in Figure 1.
      Figure thumbnail gr1
      Figure 1Results of the centralized triage-based screening process for patients across both participating cardiac centres during implementation period (July 2018–February 2020). Notes: CCTA = coronary computed tomographic angiography; ICA = invasive coronary angiography; CAD = coronary artery disease.
      Of the 166 referred patients who consented to first line CCTA (109 at Hamilton and 57 at Niagara), 1 (<1.0%) did not attend their scheduled test, 4 (2.4%) had their CCTA cancelled due to a medical reason, and 5 (3.0%) had a non-diagnostic CCTA (Figure 1). Thus, 156 referred patients underwent a diagnostic first-line CCTA with the following results: 42 (27%) were normal; 71 (46%) found non-obstructive CAD; 7 (4.5%) found borderline CAD; and 36 (23%) found obstructive CAD. Overall, 42 (98%) of patients with a borderline or obstructive CAD result on their first-line CCTA were referred for ICA, with 36 (86%) of these patients subsequently undergoing this procedure. Given the distal obstructive CAD identified on CCTA, one patient with borderline CAD on CCTA was not referred for ICA and medical management was pursued. Only one patient of the 113 with a normal or non-obstructive CAD finding on their CCTA was referred for ICA within 90 days of their CCTA (99% fidelity). This patient had a calcium score of 0. Five weeks post-CCTA they were admitted with a non-ST segment elevation MI and underwent uncomplicated PCI. Retrospective review of their CCTA, revealed a missed severe focal non-calcified plaque in otherwise normal coronary arteries.
      The results of the process evaluation indicated further support for the intervention among physicians, triage staff, and patients. Thirty-two referring physicians (37% response rate) agreed that the centralized triage process accurately identified eligible patients (median=6: 1=”strongly disagree”; 7=”strongly agree”, range 3 to 7) and its implementation was both appropriate (median=6, range 3 to 7) and sustainable (median=6, range 3 to 7) in their setting. Physicians identified the key advantages of the centralized triage process to be (1) optimized testing and (2) reduction in unnecessary invasive procedures. Physicians identified the primary barrier to ordering a CCTA in eligible patients was the prolonged wait times. Five triage staff (100% response rate) participated in the process evaluation [focus group (n=3) and interviews (n=2)] and all expressed strong support for continued implementation of the triage process (median 7, 1=“strongly disagree” to 7=“strongly agree”) with 3 respondents indicating limited administrative support as a potential barrier. Eighty-seven patients were contacted and 30 (35% response rate) completed an interview. Patient respondents were satisfied with how their CCTA tests were scheduled and conducted. Furthermore, most patients reported being very confident that their test results were accurate (median 4, 1=“not confident” to 4=“very confident”).

      3.2 Intervention effectiveness

      3.2.1 Invasive angiograms potentially avoided due to CCTA first

      Among the 156 patients across both sites who underwent first-line CCTA during the implementation period, only 36 underwent ICA at either study site in the following 90 days (Figure 1). Therefore, 120 patients (77%) may have avoided an ICA due to undergoing CCTA first.

      3.2.2 Diagnostic yield of invasive coronary angiography with and without CCTA first

      Of the 36 patients who underwent ICA within 90 days post-CCTA during the implementation period, 24 (67%) were found to have significant CAD and 18 (75%) of these patients underwent revascularization (12 PCI and 6 CABG) (Figure 1).
      Figure 2 details our process of identifying eligible outpatients with an ICA referral to either cardiac centre in the health region during the study period using CorHealth Ontario Cardiac Registry data. Overall, we identified 1,027 patients who completed a diagnostic ICA as initially referred between July 2016 and February 2020. This sample included 31 of the 36 (86.1%) patients who underwent ICA within 90 days post-CCTA from the CarDIA database; the inability to link the other 5 patients who received CCTA first (i.e., within 90 days before their ICA) is likely owed to discrepancies in data entry between the CarDIA trial and CorHealth administrative databases. The other 996 participants identified in this sample did not undergo CCTA in the 90 days preceding their ICA (and would not be captured in the CarDIA database), with 70% (n=694) and 30% (n=302) of these patients referred for their ICA during the pre- and post-implementation periods, respectively. Primarily due to resource limitations, the triage staff were only able to screen 75% (226/302) of the eligiblepatients referred for ICA.
      Figure thumbnail gr2
      Figure 2Identification of eligible low-risk participants referred for an invasive coronary angiogram across both study sites between July 2016 and February 2020 using local CorHealth Ontario Cardiac Registry Data. Notes: ICA = invasive coronary angiography; CCS = Canadian Cardiovascular Society; CAD = coronary artery disease; VHD = valvular heart disease; MI = myocardial infarction; PCI = percutaneous coronary intervention; CABG = coronary artery bypass graft; CCTA = coronary computed tomographic angiography. *First-line CCTA means patient was screened by implemented triage process, referred for, and underwent CCTA as a first-line diagnostic test within 90 days before undergoing ICA. Note, only 31 of 36 patients with CCTA first in the CarDIA database () could be linked with a corresponding ICA record in the CorHealth data (n=1,027 patients).
      Table 1 describes the distribution of select baseline characteristics (assessed on ICA referral date) among eligible patients in the administrative database sample (n=1,027), overall and stratified by implementation period and exposure to CCTA first. A lower proportion of patients referred for ICA in the post-implementation period without first-line CCTA had angina with moderate exertion (38%) versus patients referred for ICA in the pre-implementation period (46%) or in the post-implementation period with a first-line diagnostic CCTA (45%). Furthermore, a larger proportion of referred patients in the post-implementation period underwent ICA at Hamilton (254/333 or 76%) versus the pre-implementation period (432/694 or 62%).
      Table 1Characteristics of 1,027 low-risk outpatientsa referred for and who underwent a diagnostic invasive coronary angiography at either study site, overall and stratified by implementation period and whether they had CCTA before the procedure – CorHealth Ontario Cardiac Registry data, July 2016 to February 2020.
      CharacteristicPre-implementation (Jul 2016–Jun 2018)Post-implementation (Jul 2018–Feb 2020b)
      Overall (n=1,027)No CCTAc (n=694)CCTA firstc (n=31)No CCTAc (n=302)
      Male patient, n (%)534 (52)366 (53)16 (52)152 (50)
      Patient age (in years), mean ± SD58.3 ± 9.157.7 ± 9.260.5 ± 7.259.4 ± 8.9
      Angina (CCS score), n (%)
      Asymptomatic (0)336 (33)224 (32)8 (26)104 (34)
      Angina with strenuous, rapid, or prolonged exertion (1)241 (23)149 (21)9 (29)83 (27)
      Angina with moderate exertion (2)450 (44)321 (46)14 (45)115 (38)
      Primary indication to “rule out CAD”, n (%)788 (77)538 (78)22 (71)228 (76)
      Planned site of procedure, n (%)
      Hamilton General Hospital686 (67)432 (62)23 (74)231 (76)
      St. Catherine’s Site341 (33)262 (38)8 (26)71 (24)
      Notes: CCTA = Coronary computed tomographic assessment; CCS = Canadian Cardiovascular Society; CAD = coronary artery disease.
      a All characteristics assessed as of referral date (all referrals occurred between July 2016–February 2020).
      b From July–November 2018, only patients at 1 of 2 participating sites were eligible for CCTA.
      c Participants who had a diagnostic CCTA (due to the intervention) within 90 days before their procedure were considered to have had “CCTA first”; otherwise, participants were considered to have “No CCTA” prior.
      Table 2 compares the unadjusted and adjusted diagnostic yields of ICA between patients with (n=31) and without a first-line CCTA in the 90 days preceding their ICA (n=996). Notably, the unadjusted diagnostic yield for elective ICA in CCTA first patients within this sample (22/31 or 71%) is nearly identical to the diagnostic yield among all 36 CCTA first patients in the CarDIA database (67%). After adjusting for all characteristics in Table 1 via multivariable logistic regression, we estimated using post-regression marginalization that if all patients in the sample had a CCTA before their ICA (versus if no patients underwent CCTA first), an additional 42 patients per 100 would have had an obstructive CAD finding on their ICA (RD=42%, 95% CI= 26% to 59%). As a sensitivity analysis, we restricted the analytic sample to the 333 participants whose ICA referral occurred within the implementation period (July 2018–February 2020); this resulted in an estimated effect of similar magnitude and precision (RD=44%, 95% CI=24% to 61%).
      Table 2Diagnostic yield for invasive coronary angiography (ICA)a based on first-line CCTA status* in 1,027 low-risk outpatients who underwent a diagnostic invasive coronary angiography at either study site – linked CarDIA and CorHealth Ontario Cardiac Registry data, July 2016 to February 2020.
      Diagnostic yield of ICAa
      ModelCCTA first*UnadjustedAdjustedbAdjusted RDb (95% CIc)Adjusted RRb (95% CIc)
      1 (n=1027)Yes22/31 (71%)700/1027 (68%)42% (26% to 59%) [reference]2.62 (1.97 to 3.34)
      No259/996 (26%)268/1027 (26%)[reference]
      2d (n=333)Yes22/31 (71%)225/333 (68%)44% (24% to 61%)2.84 (1.93 to 3.81)
      No71/302 (24%)79/333 (24%)[reference][reference]
      Notes: CCTA = Coronary computed tomographic assessment; RD = risk difference; RR = relative risk; CI = confidence interval.
      *Underwent a diagnostic CCTA (due to implemented triage process) within 90 days prior to ICA procedure.
      a Defined as the marginal probability of a significant CAD finding.
      b Adjusted for patient-level covariates (age, sex, and Canadian Cardiovascular Society angina score [0, 1, or 2], primary indication for ICA referral to “rule out CAD” [yes or no], and procedure site) using a logistic regression model, with estimated probabilities calculated using post-regression marginalization method as mean predicted probabilities of outcome (significant CAD finding on ICA) assuming (1) all participants in the sample received first-line CCTA compared to (2) if no participants in the sample received first-line CCTA.
      c Bootstrap 95% CI constructed using 500 samples with replacement. d Limited to participants whose ICA procedure had a referral date within the intervention period for either site (i.e., Jul 2018–Feb 2020 for Hamilton patients; Dec 2018–Feb 2020 for Niagara Site patients).

      4. DISCUSSION

      To our knowledge, we present the first evaluation of a novel centralized triage process that identified eligible outpatients referred for ICA and alternatively referred these patients for CCTA as a first-step diagnostic test. This novel triage process resulted in a >70% reduction in ICAs within the first 90 days after referral in this eligible group. The diagnostic yield of elective ICA in this population increased from 27% pre-implementation to 67% following implementation of the triage process. This novel process was deemed acceptable and sustainable by patients, physicians, and hospital administration. Furthermore, our centralized diagnostic pathway using CCTA-first has garnered attention by the Office of the Auditor General of Ontario, and has been cited in their Value for Money Audit, December 2021 (Page 35).
      This study is unique for several reasons. First, the diagnostic yield of ICA was improved in this patient population. While a negative ICA result may allow exclusion of CAD as a cause of symptoms and provide prognostic insight, this information is obtained at the cost of procedural risks (stroke, death, MI and vascular complications) and without any potential benefit from revascularization. As highlighted in a recent randomized trial, procedural complications were almost four-fold higher in the ICA versus CCTA group (1.9% versus 0.5%, (hazard ratio=0.26; 95% CI=0.13 to 0.55).
      • Maurovich-Horvat P.
      • Bosserdt M.
      • Kofoed K.F.
      • et al.
      CT or invasive coronary angiography in stable chest pain.
      Furthermore, ICAs consume resources that might be better allocated to alternate patients, particularly in resource constrained universal health care systems. It is preferable for selected patients to undergo a non-invasive assessment if similar diagnostic certainty can be achieved. Our study significantly increased the diagnostic yield of ICA. By applying this approach routinely, an additional 2 in 5 patients referred for ICA will have obstructive CAD, therefore increasing the effective utilization and value of ICA for the patient and the health care system.
      Second, this novel triage process could reduce risks to the patient and costs to the health system. Among the 156 patients with a diagnostic CCTA across both sites, only 36 were referred for and underwent ICA within the first 90 days (23%), meaning as many as 120 patients in our study avoided an invasive procedure by having CCTA first. While we did not have the ability to triage every eligible patient during the study period, our intervention did have an impact on the referral process. (We note that the discrepancy in total referrals during the post-implementation period is almost half the pre-implementation period and is likely owed to the 120 patients that were referred for ICA but only received a first-line CCTA instead according to the CarDIA database (Figure 1); furthermore, the pre-implementation period was longer than the post-implementation period).
      Approximately 4000 outpatient elective ICA are performed in one year in our study region, with an anticipated 12% of these procedures meeting the eligibility criteria for this study. If we applied the results of our study to this population, then as many as 370 unnecessary ICAs could potentially be avoided. Sustained implementation of this novel triage program over 3 years could prevent as many as 39 vascular complications and 1 myocardial infarction, stroke, or death.
      • Tavakol M.
      • Ashraf S.
      • Brener S.J.
      Risks and complications of coronary angiography: a comprehensive review.
      Furthermore, the estimated cost savings could be $1,665,000 (cost difference of approximately $1,500 between ICA and CCTA imaging).
      • Chang H.J.
      • Lin F.Y.
      • Gebow D.
      • et al.
      Selective Referral Using CCTA Versus Direct Referral for Individuals Referred to Invasive Coronary Angiography for Suspected CAD: A Randomized, Controlled, Open-Label Trial.
      Finally, the novel diagnostic algorithm in this study promoted a unique collaboration between speciality groups involved in the care of the patient i.e. Cardiology and Radiology. Combining the triage for diagnostic cardiac testing could streamline health care efficiencies and patient satisfaction. A centralized triage unit screened and booked eligible patients for CCTA as a first-line test versus ICA as part of this study. This screening was done by manual review of the ICA referral form to confirm eligibility based on previous research done at our centre.
      • Sheth T.
      • Amlani S.
      • Ellins M.L.
      • et al.
      Computed tomographic coronary angiographic assessment of high-risk coronary anatomy in patients with suspected coronary artery disease and intermediate pretest probability.
      The triage staff were able to select 82% of the final patient population (18% being removed following a more in-depth secondary review by the study investigators). In the real-world setting, this secondary review would not be necessary as this was in place as part of the ethics board suggested study protocol amendment. A centralized referral process has the potential to optimize not only referrals to ICA but also those to CCTA (i.e., triage could send patients directly to ICA instead of CCTA if they have a high pre-test probability of obstructive CAD and/or a high chance of a non-diagnostic CCTA). This bidirectional triage process would help ensure the right test for the right patient. The process evaluations confirmed that this novel triage process was supported by patients, referring physicians, and hospital administration.
      This study has some limitations that should be highlighted and would need to be addressed in order to support widespread and sustained implementation of this triage process. First, timely access to high-quality CCTA is limited across Canada. This limitation was raised in the process evaluation by the referring physicians in our study and confirmed in the literature.

      Canadian Agency for Drugs and Technology in Health. “The Canadian Medical Imaging Inventory 2019–2020.” Canadian Journal of Health Technologies 1, no. 1 (January 2021): 1–215. Accessed March 15, 2022. https://cadth.ca/sites/default/files/ou-tr/op0546-cmii3-final-report.pdf.

      ,

      The Age of Imaging Equipment in Canada. Canadian Medical Imaging Inventory Service Report, Ottawa: CADTH, October 2021. Accessed February 11, 2022. https://www.cadth.ca/sites/default/files/attachments/2021-11/cmii_service_age_of_imaging_equipment.pdf.

      The mean wait-time for an ICA is 3 weeks versus approximately 3 months for a CCTA at our two cardiac centres. This wait-time difference has a significant influence on diagnostic test selection and may be contributing to the low diagnostic yield of ICA. Second, as highlighted by the process evaluation, to sustain this triage process, added administrative support in terms of human resources would be required. To minimize the added workload and costs to the triage team, a point-of-care decision support program that helps referring physicians or triage staff select the most appropriate test based on current guidelines and the probability of obstructive CAD could be a potential solution. Our group has developed such predictive algorithms, and if implemented into existing triage pathways, diagnostic efficiencies could be significantly optimized.
      • Schwalm J.D.
      • Di S.
      • Sheth T.
      • et al.
      A machine learning–based clinical decision support algorithm for reducing unnecessary coronary angiograms.
      Third, this study was implemented in two tertiary care cardiac centres and led by a novel collaboration between two specialist groups. This experience is unique and my not be readily implemented in other centres across Ontario and Canada. Ongoing piloting of this triage strategy in multiple centres should be explored. Fourth, while our intervention resulted in a significantly higher diagnostic yield for elective ICA following CCTA (27% pre-intervention to 67% post-intervention, p<0.01), caution must be taken as this post-intervention group involves only a select population of patients that made it through CCTA to ICA (n=36). Fifth, while CCTAs are an extremely sensitive test, they are not perfect. We did have one patient in the CCTA-first group present with a NSTEMI requiring PCI 5 weeks following an unremarkable CCTA. While this is a unique case, with a calcium score of 0 that might have affected interpretation of the test, it highlights the potential limitations of non-invasive imaging for obstructive CAD and the need for broader experience with this strategy. A recent study highlighted that a CCTA first strategy compared to ICA in intermediate pre-test probability for CAD elective patients, demonstrates no difference in major adverse cardiovascular events (hazard ratio=0.70; 95% CI=0.46 to 1.07). Therefore, given the proven safety of this strategy, this single event, should be considered against the potential complications and costs of ICA in this population. Finally, while this centralized triage process is appropriate for our Canadian health care system, it may not be as relevant to health systems where the rates of normal coronaries on ICA are not high and where there is easy access to CCTA as a first line approach to elective outpatients with chest pain and suspected CAD.

      5. CONCLUSION

      A centralized triage process in which low-risk patients referred for ICA are instead referred for first-line CCTA is an acceptable and effective mechanism to diagnose obstructive CAD and to reduce unnecessary invasive tests, costs to the health system, and risks to patients.

      Acknowledgements

      We thank CorHealth Ontario for use of the CorHealth Ontario Cardiac Registry.

      REFERENCES

        • Roth G.A.
        • Forouzanfar M.H.
        • Moran A.E.
        • et al.
        Demographic and epidemiologic drivers of global cardiovascular mortality.
        N Engl J Med. 2015; 372: 1333-1341
        • Patel M.R.
        • Peterson E.D.
        • Dai D.
        • et al.
        Low diagnostic yield of elective coronary angiography.
        N Engl J Med. 2010; 362: 886-895
        • Ouellette M.L.
        • Beller G.A.
        • Löffler A.I.
        • Workman V.K.
        • Bourque J.M.
        Appropriate referrals of angiography despite high prevalence of normal coronary arteries or nonobstructive CAD.
        J Am Coll Cardiol. 2017; 69: 2673-2675
        • Tavakol M.
        • Ashraf S.
        • Brener S.J.
        Risks and complications of coronary angiography: a comprehensive review.
        Glob J Health Sci. 2012; 4: 65-93https://doi.org/10.5539/gjhs.v4n1p65
        • Rudziński P.N.
        • Kruk M.
        • Kępka C.
        • et al.
        Assessing the value of coronary artery computed tomography as the first-line anatomical test for stable patients with indications for invasive angiography due to suspected coronary artery disease. Initial cost analysis in the CAT-CAD randomized trial.
        J Cardiovasc Comput Tomogr. 2020; 14: 75-79https://doi.org/10.1016/j.jcct.2019.07.008
        • Haase R.
        • Schlattmann P.
        • Gueret P.
        • et al.
        Diagnosis of obstructive coronary artery disease using computed tomography angiography in patients with stable chest pain depending on clinical probability and in clinically important subgroups: meta-analysis of individual patient data.
        BMJ. 2019; 365: 1945https://doi.org/10.1136/bmj.l1945
      1. National Institute for Health and Care Excellence. Recent-onset chest pain of suspected cardiac origin: assessment and diagnosis. 2010;Clinical Guideline CG95. https://www.nice.org.uk/guidance/cg95.

        • Sheth T.
        • Amlani S.
        • Ellins M.L.
        • et al.
        Computed tomographic coronary angiographic assessment of high-risk coronary anatomy in patients with suspected coronary artery disease and intermediate pretest probability.
        Am Heart J. 2008; 155: 918-923
        • Dewey M.
        • Rief M.
        • Martus P.
        • et al.
        Evaluation of computed tomography in patients with atypical angina or chest pain clinically referred for invasive coronary angiography: Randomised controlled trial.
        BMJ. 2016; 355: i5441
        • Maurovich-Horvat P.
        • Bosserdt M.
        • Kofoed K.F.
        • et al.
        CT or invasive coronary angiography in stable chest pain.
        N Engl J Med. 2022; https://doi.org/10.1056/NEJMoa2200963
        • Gulati M.
        • Levy P.D.
        • Mukherjee D.
        • et al.
        AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guideline for the evaluation and diagnosis of chest pain: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.
        Circulation. 2021; 144: e368-e454
      2. Canadian Agency for Drugs and Technology in Health. “The Canadian Medical Imaging Inventory 2019–2020.” Canadian Journal of Health Technologies 1, no. 1 (January 2021): 1–215. Accessed March 15, 2022. https://cadth.ca/sites/default/files/ou-tr/op0546-cmii3-final-report.pdf.

      3. The Age of Imaging Equipment in Canada. Canadian Medical Imaging Inventory Service Report, Ottawa: CADTH, October 2021. Accessed February 11, 2022. https://www.cadth.ca/sites/default/files/attachments/2021-11/cmii_service_age_of_imaging_equipment.pdf.

        • Mancini G.B.
        • Gosselin G.
        • Chow B.
        • et al.
        Canadian Cardiovascular Society. Canadian Cardiovascular Society guidelines for the diagnosis and management of stable ischemic heart disease.
        Can J Cardiol. 2014 Aug; 30 (Epub 2014 May 28. PMID: 25064578): 837-849https://doi.org/10.1016/j.cjca.2014.05.013
        • Cabana M.D.
        • Rand C.S.
        • Powe N.R.
        • et al.
        Why don’t physicians follow clinical practice guidelines.
        JAMA. 1999; 282: 1458-1465
      4. CorHealth Ontario. Data Holdings and Statements of Purpose List. Corhealth Ontaio Data Holdings. https://www.corhealthontario.ca/CorHealth-Ontario-Data-Holdings.pdf. Accessed: March 2022.

        • Austin P.C.
        Absolute risk reductions, relative risks, relative risk reductions, and numbers needed to treat can be obtained from a logistic regression model.
        J Clin Epidemiol. 2010; 63: 2-6https://doi.org/10.1016/j.jclinepi.2008.11.004
        • Chang H.J.
        • Lin F.Y.
        • Gebow D.
        • et al.
        Selective Referral Using CCTA Versus Direct Referral for Individuals Referred to Invasive Coronary Angiography for Suspected CAD: A Randomized, Controlled, Open-Label Trial.
        JACC Cardiovasc Imaging. 2019; 12: 1303-1312https://doi.org/10.1016/j.jcmg.2018.09.018
        • Schwalm J.D.
        • Di S.
        • Sheth T.
        • et al.
        A machine learning–based clinical decision support algorithm for reducing unnecessary coronary angiograms.
        Cardiovascular Digital Health Journal. 2021; 3: 21-30https://doi.org/10.1016/j.cvdhj.2021.12.001