Strati ﬁ ed medicine using invasive coronary function testing in angina: A cost-effectiveness analysis of the British Heart Foundation CorMicA trial

coronary artery disease (CAD). However, up to half of patients with symptoms of ischaemia have no obstructive coronary arteries


Introduction
Coronary angiography is routinely performed in patients with suspected angina. Approximately four million elective coronary angiograms are performed each year in Europe and the United States [1,2]. However, up to half of patients with signs and/or symptoms of ischaemia have no obstructive coronary arteries (INOCA) presenting a diagnostic and therapeutic challenge for clinicians [2]. This large subgroup includes patients with suspected microvascular angina (MVA) and/or vasospastic angina (VSA) [3]. Coronary angiography is insensitive in visualising the micro-vessels [4]. Coronary angiography is also insensitive for coronary reactivity and a patient's susceptibility to vasospasm. The rationale for adjunctive coronary vascular function testing, in the form of an interventional diagnostic procedure (IDP) is as follows. Firstly, coronary angiography does not exclude a disorder of coronary vascular function, such as those relating to vasomotion or microcirculation [3]. Secondly, adjunctive coronary testing allows for the stratification of this hitherto undifferentiated subpopulation [3]. This is important because discrimination between microvascular angina, vasospastic angina, both, or neither, allows for specific and distinct treatment regimes. Finally, adjustive coronary function testing provides both patients and their physicians with prognostic information. Clinical guidelines relating to the management of patients with INOCA are limited by a lack of randomised controlled trials [5].
The uncertainty regarding the diagnosis of patients with INOCA presents a health economic challenge, both in terms of healthcare resource utilisation and of quality-of-life impact on patients. Patients with INOCA who remain undiagnosed are liable to re-present at primary or secondary care. This may result in additional GP consultations, outpatient cardiologist visits, diagnostic tests and repeat coronary angiography.
The British Heart Foundation (BHF) Coronary Microvascular Angina (CorMicA) trial was a randomised, controlled, developmental clinical trial of stratified medicine in patients with known or suspected angina undergoing clinically-indicated coronary angiography in the National Health Service.
The CorMicA study design and main trial results have been described in detail elsewhere [3,6,7]. In summary, 391 patients with angina undergoing invasive coronary angiography (standard care) were recruited between November 2016 and December 2017. 151 patients without angiographically obstructive CAD were randomised to either the intervention group (standard care, plus IDP-guided stratified medical therapy and cardiac rehabilitation) (n = 76) or control group (standard care, plus medical therapy and cardiac rehabilitation) (n = 75). The IDP consisted of guidewire-based assessment of coronary flow reserve, index of microcirculatory resistance, fractional flow reserve, followed by vasoreactivity testing with acetylcholine. Following the IDP, clinicians were able to update their diagnosis, if indicated, based on this additional information, and to tailor patient medical therapy and cardiac rehabilitation accordingly. The intervention resulted in a mean improvement of 13.6 units in the Seattle Angina Questionnaire summary score (primary outcome) at 12 months (95% CI: 7.3-19.9; p < 0.001).
The aim of this study was to undertake a cost-effectiveness analysis of the BHF CorMicA trial, using resource use and quality of life data from the trial, to estimate the potential cost-effectiveness of the use of the IDP among INOCA patients undergoing routine coronary angiography.

Overview
An economic evaluation was undertaken using clinical and resource use data obtained from the CorMicA feasibility trial and from the wider literature. We employed the standard methodology for an economic evaluation alongside a clinical trial [8]. A conceptual model was designed and presented to illustrate the patient pathway and to facilitate interpretation of the trial data.

Study perspective
The economic evaluation was undertaken from the perspective of the National Health Service (NHS) of the United Kingdom (U.K.) for cost year 2018/19, adhering to contemporary practice guidelines and the NICE reference case [9].

Time horizon
The economic evaluation estimated the cost and quality-of-life (QALY) gained, by trial arm, over the 1-year time horizon of the BHF CorMicA trial.

Discount rate
The discounting of NHS costs and health outcomes was not necessary for the economic evaluation as the CorMicA trial was limited to a 1-year time horizon.

Conceptual model of the CorMicA trial
We developed a conceptual model to illustrate the clinical pathway for each clinical strategy (Fig. 1).
The conceptual model begins with a patient with signs and/or symptoms of angina who has received coronary angiography and are subsequently diagnosed with no obstructive coronary artery disease (NOCAD). At this point, the standard care procedure would end.
In line with the CorMicA trial, patients may then receive either a "sham test" (control group) or the IDP (intervention group). In the control group branch of the model, patients receive the IDP but the results are not disclosed to the attending clinicians who are therefore blind to the IDP results. Patients then receive the standard of care for patients with signs and/or symptoms of angina but no CAD.
In the intervention group branch of the conceptual model, patients receive the IDP. Physicians were informed of the IDP results and were able to revise their initial diagnosis, which was based on the angiogram and all of the other available clinical information (e.g. prior non-invasive test results, medical history). Based on this new information, patients could either receive a diagnosis of positive or negative (for the purposes of the conceptual model, a positive diagnosis for any of the MVA or VSA endotypes is categorised as "positive"). Patients were then subsequently offered alternative medication and/or lifestyle modification advice depending on their revised diagnosis. If patients were subsequently diagnosed as "negative", anti-anginal therapies were discontinued.

Resource use and cost analysis
Resource use estimates were obtained from the CorMicA trial, published literature and expert opinion (the CorMicA clinical trial team) ( Table 1). Medication, participation in a cardiac rehabilitation programme, and IDP costs were included in the base case analysis.
Anginal medications were measured at baseline, 6 months and 12 months in the trial. Based on these data, we measured the proportion of the year a patient spent on each medication. We then estimated the proportion of the year spent on each mediation by the mean patient in each trial arm. Relevant unit costs were adjusted to annual costs for each medication and attached to the time spent on medication to obtain the mean cost of medication by trial arm. The cost of the IDP was included for all patients in the intervention group only. The unit cost of a cardiac rehabilitation cost was attached to the number of patients reporting attendance at a cardiac rehabilitation programme at 12 months. Medication cost, IDP test cost and cardiac rehabilitation programme cost were combined to obtain a mean total patient cost by trial arm.
All unit costs were collected or converted into UK pounds sterling (£) for the price year 2018/19. Unit costs were collected from routine sources such as the British National Formulary [10], Personal Social Services Resource Unit (PSSRU) [11] and NHS Reference Costs [12]. Some unit costs were obtained from NHS Reference Costs or from local sources (NHS Golden Jubilee National Hospital) ( Table 1).

Quality-of-life analysis
Quality-of-life (QoL) outcomes were based on the EuroQoL-5D (EQ-5D-5L) questionnaire and collected at baseline, 6 and 12 months in the CorMicA trial participants and expressed as health utility indices.
Quality adjusted-life years were calculated using the area-underthe-curve method (AUC) [13]. The AUC was estimated for each individual in the trial by weighting the time between EQ-5D health utility measurement in the trial by their health utility value over that time period.

Cost-effectiveness analysis
Mean cost and QALY gain per patient, by trial arm, was estimated by using a generalized linear model (GLM) and adjusting for potential confounding factors (age, sex, BMI and trial arm). The appropriate family for the GLM was selected based on the results of the modified Park's test. Our final cost model was based on the log link and gamma family and our final QALY model was based on the identity link and Gaussian family. There were few missing resource-use or EQ-5D data in the CorMicA trial (≤5%). Where missing data existed, multiple imputation by chained equations (MICE) was used to impute missing data [14]. All analyses were conducted using Stata 14 (StataCorp).
The outcomes of interest are mean costs, mean quality adjusted life year (QALY) gained, and the incremental cost-effectiveness ratio  (ICER), expressed as the cost per QALY. The estimated ICER is compared with a willingness-to-pay (WTP) threshold of £20,000 per QALY gained, which is a commonly accepted cost-effectiveness threshold used by the National Institute for Health and Care Excellence (NICE) in the United Kingdom.

Sensitivity analysis
The distribution of incremental mean costs and QALYs produced by a bootstrap analysis were presented on the cost-effectiveness plane. Costeffectiveness acceptability curves (CEACs) were used to present the uncertainty in the decision regarding the most cost-effective option, over a variety of monetary willingness-to-pay threshold.

Scenario analyses
Data on the number of previous coronary angiograms among patients enrolled in the CorMicA trial indicated that approximately one third (51 out of 151) of patients had "at least one previous coronary angiogram" [15]. This aligns with the experience of clinical experts involved in both the trial, and in the routine care of this patient population, which suggests that patients with signs and symptoms of angina, but no obstructive coronary artery disease are liable to represent in primary or secondary care.
Given the large potential cost associated with these rerepresentations, we modelled the impact of IDP testing on reducing the proportion of patients re-presenting to primary and secondary care with signs/symptoms of angina following coronary angiography. To implement this scenario in the model, we assumed that patients diagnosed as negative for microvascular or vasospastic angina following IDP testing did not re-present with suspected angina, and that a proportion of patients in the control group do re-present. For patients who represent, we proposed a typical "package of care" associated with a representation, including: one coronary angiogram, three GP appointments, two cardiology outpatient appointments and standard care diagnostic tests (U&E, lipids, Hba1c). In the CorMicA trial, data on repeat visits to primary care or hospital outpatient clinics were not prospectively captured, as such there is some uncertainty as to the true impact of adjunctive functional test in a real-world setting.
The unit costs associated with these resources are given in Table 1. Based on the utilisation of each resource required, we estimated a cost associated with an angiogram package of care of £1786. We estimated the impact on the mean cost per patient in the control group if one third, one fifth, and one tenth of patients re-present in primary care.

Ethical approval
The West of Scotland Research Ethics Committee approved the study (REC 1 reference 16/WS/0192).

Patient and public involvement
CorMicA researchers engaged with the patient population (service users) and the British Cardiovascular Care Partnership in the design of the CorMicA trial.

Results
The baseline characteristics of the patient population are given in Table 2. The median age in the study was 61 years old, with the majority of the study participants being female (73.5%). There was a high prevalence of cardiovascular risk factors. Participants had impaired quality of life at baseline, as reflected by EQ-5D-5L health status (mean: 0.60, s.d.: 0.29) and SAQ Summary Score (mean 50.8, s.d.: 18.1).

Resource use results
The mean time spent on medication was greater in seven out of the nine medications included in the intervention group, compared with the control group (Table 3). Thirty patients (40%) reported attending a cardiac rehabilitation programme in the intervention group, compared with 12 patients (16%) in the control group. All patients in the intervention group received the IDP test, compared with none in the control group.

Cost results
The total mean cost per patient in the control arm of the trial was £103 and was comprised solely of medication and cardiac rehabilitation programme costs ( Table 4). The total mean cost per patient in the intervention group was £568 and was comprised of the costs of the IDP, medication and cardiac rehabilitation programme. This equates to an incremental cost of £465 at 12 months associated with the intervention group.

Quality-of-life results
We estimated the mean QALY gain using the AUC approach. Patients in the control group had a mean QALY of 0.548 at 12 months, compared with 0.652 in the intervention group (Table 4). This equates to a mean QALY gain of 0.104 at 12 months in patients in the intervention group.

Cost-effectiveness results
The incremental cost per QALY gained at 12 months was £4500 (£2937, £33264) ( Table 4). Compared with a willingness-to-pay (WTP) threshold of £20,000 per QALY, this suggest that the use of stratified medicine with the IDP and linked medical therapy is a costeffective use of resources.

Sensitivity analysis
A visual illustration of the uncertainty surrounding the mean incremental cost and QALY gain in the intervention group over 1000 iterations of bootstrap analysis plotted on the cost-effectiveness plane is presented in Fig. 2. The x-axis represents the incremental QALY gain, and the y-axis represents the incremental cost, associated with the intervention group compared with the control group. The majority of the cost and QALY estimates are placed in the north-east quadrant of the cost-effectiveness plane, suggesting that the use of the intervention is associated with an incremental cost and incremental QALY gain.
The probability of the intervention being cost-effective, compared with the control group, is given for alternative willingness-to-pay for QALY gain thresholds in Fig. 3. At a WTP threshold of £20,000 per QALY gained, there is a 96% probability of the intervention being costeffective, based on the uncertainty described by our bootstrap analysis.

Scenario analysis
We explored the impact on our cost-effectiveness results of making an assumption regarding the number of re-presentations that might be avoided as result of the introduction of stratified medicine involving IDP testing with linked therapy. Including the cost of re-presentations has the impact of increasing the cost-effectiveness of the intervention, compared with the base-case scenario (Table 5). A reduction in the proportion of patients who re-present for angiograms has the impact of reducing the cost-effectiveness associated with the intervention. However, for all re-presentation scenarios considered, the intervention was still either highly cost-effective or cost-saving. Values are median (interquartile range), n (%), or mean +/− SD. a ASSIGN risk score.

Discussion
Our cost-effectiveness analysis of the CorMicA trial found that the introduction of stratified medicine, including an adjunctive interventional diagnostic procedure for the management of patients with signs and symptoms of angina when obstructive coronary artery disease is excluded by the angiogram, is likely to be associated with an increase in patient quality-of-life and an increase in health-related resource use over the trial period. With a mean incremental cost of £465 and incremental QALY gain of 0.104 per patient, we estimate an incremental cost-effectiveness ratio of £4500. Hence, our results find that stratified medicine in this patient population and setting is likely to be costeffective in a UK setting.
The CorMicA trial participants had substantial impairment in quality-of-life at baseline, as reflected by EQ-5D-5L health status [16] and SAQ Summary Score [17,18]. Our results have provided insights into the mechanisms of treatment effect. We have found that that the benefit is determined to an appreciable extent by changes in medication aligned to the endotype. Other factors, such as participation in cardiac rehabilitation and control of vascular risk factors, are also relevant. The potential reduction in repeat coronary angiogram procedures (and related cost saving) in patients with persisting symptoms of angina, as a result of introducing IDP testing, has the potential to increase the cost-effectiveness of IDP testing.
The burden of INOCA, particularly in women, is known to be significant [19]. Three quarters of the trial participants were female. Cardiology trials generally enrol more males than females, not least since coronary artery disease predominately affects men. Accordingly, CorMicA provides new data that is particularly relevant to heart disease in women.
To date, there are no comparable studies. CorMicA was the first, randomised controlled trial of stratified medicine in angina. INOCA includes a large sub-population, representing almost half of all-comers, presents an unmet clinical need, that are usually discounted from clinical trials. It is noteworthy that the health-related quality of life, as measured by EQ5D-5 L and SAQ scores, were lower in CorMicA (patients with NOCAD), compared with recent clinical trials of patients with obstructive coronary artery disease (ORBITA [17], ISCHEMIA [18]). This is the first study to investigate the health-related resource use implications of the introduction of adjunctive coronary function testing in patients with INOCA. There is no long-term evidence on the representation rate of patients with signs and symptoms of angina with a negative diagnosis of macrovascular angina following coronary angiography. In the CorMicA trial population, one in three patients had a history of prior coronary angiography indicating a persisting health burden and related demand on primary and secondary care. This can be explained by the limitations of the standard, angiography-guided care to reduce the health need of affected patients. The International Coronary Microvascular Angina (iCorMicA) trial which will randomise 1500 patients in Europe, is designed to address this evidence gap in multiple centres in different healthcare settings (ClinicalTrials.gov Identifier: NCT04674449). However, scenario analysis of our results suggests that, even when varying the proportion of patients expected to represent for angiograms, stratified medicine involving the IDP test is still highly likely to represent a cost-effective use of resources. Indeed, even if no patients re-present (base case results), this clinical strategy is still likely to be cost-effective. In addition, our analysis included only  costs attributable to the UK NHS healthcare systemwe did not attempt to capture non-healthcare system costs, such as productivity costs (both absenteeism and presenteeism). This cost to the individual and to society from such loses has been shown to be significant [20]. The diagnostic management of stable chest pain is also evolving. CT coronary angiography (CTCA) has very high sensitivity for coronary artery disease and moderately high specificity for discriminating obstructive lesions. CTCA is recommended as the first line test in patients presenting with stable chest pain in the National Health Service [21] and is increasingly adopted in Europe [22], including the United States [23]. The functional significance of coronary artery lesions may be assessed using adjunctive FFR-CT, however, since CTCA does not provide information on ischaemia (the metabolic consequence of angina), the implications of this strategy for patients with INOCA are uncertain. The CorCTCA trial is prospectively assessing this evidence gap and a health economic analysis will also be undertaken [24].
The quality-of-life improvement in the intervention group is believed to be a result of both pharmacological and non-pharmacological factors, including participation in cardiac rehabilitation. The IDP identifies INOCA in patients who otherwise would have received a false negative diagnosis and ineffective treatment. Our results suggest that the benefit is determined to an appreciable extent by changes in medication aligned to the endotype. Other factors, such as participation in cardiac rehabilitation and control of vascular risk factors, are also relevant. Further research aimed at understanding the mechanism by which this quality-of-life increase is generated seems warranted.

Conclusion
We undertook an economic evaluation of the CorMicA trial to estimate the cost-effectiveness of introducing stratified medicine using an adjunctive interventional diagnostic procedure with mechanisticallytargeted medical therapy for the management of patients with signs and symptoms of angina. Our results suggest that the introduction of this clinical strategy is likely to be cost-effective. The economic benefits of IDP testing increase further if we incorporate plausible assumptions regarding repeat angiograms in patients with persistent sign/symptoms of angina.

Data availability
The data underlying this article will be shared on reasonable request to the corresponding author.