Cardiac coronary tomography angiography (CCTA) use across geographical regions in the USA and the UK: a cross-sectional study

Introduction

Increasing use of cardiac imaging is a key driver of unsustainable rising healthcare costs.1 2 Imaging use varies internationally, as with many technologies, its use is highest in the USA, which performs 40% more MRIs than 10 other high-income countries—118 per 1000 population compared with 82 per 1000 population.2 For all types of CT scans, the USA mean was 245 CTs per 1000 population compared with 151 per 1000 population in other high-income countries.2 Appropriate use occurs when the chance of benefit outweighs chance of harm. Understanding imaging trends and what factors drive use of cardiac coronary tomography angiography (CCTA) can help inform the development of high value care initiatives that can promote appropriate imaging practices.

CCTA is a cardiac imaging test that captures a two-dimensional image of the heart, including the coronary arteries.3 There is significant radiation from this test, estimated to be 12 mSv per examination (approximately 600 chest X-rays), with concomitant cancer risk.4 5

There is a lack of agreement in the scientific community on the clinical value of the CCTA for evaluation of patients with suspected coronary artery disease (CAD).6 Many clinicians question whether imaging is helpful at all, particularly in patients who are at low risk of CAD,7 8 while other clinicians argue that CCTA’s association with reduced nonfatal myocardial infarctions supports its use.9 CT scanning is an example of supply-sensitive care, meaning a greater supply of CT scanners is associated with more use of CT scans.10

The history of Medicare coverage of CCTA is illustrative of the interplay of science and politics.11 In 2006, the Centers for Medicare and Medicaid (CMS) convened the Medicare Coverage Advisory Committee (now Medicare Evidence Development & Coverage Advisory Committee), an advisory panel of independent experts, to review the use of CCTA for diagnosing CAD.11 When the expert committee found no evidence of benefit for use of CCTA, which would generally lead to a determination of non-coverage, CMS instead declined to issue any national coverage determination (NCD), likely for political reasons.12 In the absence of an NCD, responsibility falls to local carriers (Medicare Administrative Contractor (MACs), online supplemental figure 1)13 for defining local coverage determinations (LCDs). LCDs are done through a region-specific process mirroring that of an NCD, which involves informal meetings and external requests to develop an LCD, consultations, a proposed determination and public comment period, optional input of a Contractor Advisory Committee, and notice of a final determination.14

Specialist societies have published appropriate use guidelines for CCTA. In 2010, the Journal of Cardiovascular Computed Tomography published appropriate use criteria (AUC) for CCTA: ‘ACCF/SCCT/ACR/AHA/ASE/ASNC/NASCI/SCAI/SCMR 2010 Appropriate Use Criteria for Cardiac Computed Tomography.’15 The (AUC) guidelines state that if a patient has an intermediate pretest probability, regardless of their ability to exercise or interpret their electrocardiogram (ECG), CCTA is appropriate. In the case where a patient has a low pretest probability, the appropriateness depends on if the ECG is interpretable and if the patient is able to exercise or not. In 2021, the American Heart Association published the ‘AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR Guideline for the Evaluation and Diagnosis of Chest Pain’ that recommends the use of CCTA for intermediate high risk patients with stable chest pain and no known CAD.16

The lack of a national coverage policy allows regional variation. For CCTA, the decision to not issue an NCD was rapidly followed by an increase in CCTA use among the Medicare population.11 17 When CMS declined to issue an NCD in 2006, CCTA use rose rapidly and continues to rise, as commercial payer coverage becomes more permissive.17 18

The UK relies on the National Institute for Health and Care Excellence (NICE) to issue guidelines for the National Health Service (NHS). NICE has recommended CCTA as first-line diagnostic testing for stable CAD since 2016, and there is a class I (strongest) recommendation from the European Society of Cardiology in 2019.19 20

We conducted a descriptive study evaluating regional differences in USA utilisation of CCTA for Medicare beneficiaries and comparing CCTA utilisation in the USA and England, the country in the UK with the largest population and highest rate of CCTA use.21 As CCTA use is often related to availability of CT scanners, we also looked at density of CT scanners across regions.10

We undertook this study to better understand current trends in CCTA usage and its relation to evidence and coverage policies in the USA and UK. The primary country of interest in this analysis is the USA. England and the UK are used as comparators.

Methods

Results

Discussion

We found significant differences in CCTA coverage policies across the USA and when compared with the UK’s NICE guidelines. There are also significant differences in CCTA orders per beneficiary at the regional carrier level. Conservative coverage policy in the USA and the UK did not always correspond to a lower rate of CCTA use, suggesting there are additional reasons to explain test ordering behaviour, including demographic factors and medical culture. For example, one of the MACs (First Coast) with the highest rates of CCTA orders per medical beneficiary covers Florida only, a state commonly found to be a high volume user of medical care.27

There was much regional variation of CCTA use in the USA and England. The USA slightly exceeds England in CCTA orders, although by a non-significant amount. Higher volumes of medical procedures are typical in the US healthcare, especially for supply-sensitive care such as CCTA; however, England has a more liberal coverage policy for CCTA than many regions in the USA. This may be related to the similarity of order rates in both countries even though availability of CT scanners is much lower in the UK than the USA. Among the 42 NHS regions, one region had >300 orders per 100 000 beneficiaries. Among the seven USA MACs, two regions had >300 orders per 100 000 beneficiaries. The smaller availability of CT scanners in the UK along with similar order rates to the USA indicates that each CT machine is being used more often in England than the USA. According to the OECD 2020 Health Equipment report, the USA has almost 40% of CT scanners in ambulatory care provider clinics while the UK has the majority of its CT scanners within hospitals. The centralisation of CT scanners in hospitals in the UK could be a driver in increased use per machine.

Our finding highlights the importance of having evidence-based outcomes data for diagnostic tests to help inform use and coverage so that testing and treatment can lead to improved health outcomes. Outcomes evidence is essential to best direct use of healthcare resources.

Limitations include lack of granularity in the dataset. It would be helpful to understand if there are a small number of individual physicians driving a majority of CCTA orders in certain regions or if there are widespread differences in CCTA ordering habits across all physicians in different regions. However, the CMS dataset lacks detailed clinical information. Additionally, variation in CCTA ordering may be due to underlying sociodemographic differences in Medicare beneficiaries in different regions. This could not be considered in our analysis due to lack of availability and thus, we cannot comment on appropriateness of use of CCTA. A long-standing limitation of CMS data is that it includes FFS claims only, as Medicare Advantage (plans offered by private companies and approved by Medicare) data are not publicly available. Finally, this is a population-level analysis and, therefore, existing associations need to be interpreted with caution.

The rapid increase in use of CCTA following CMS declining to issue a national coverage decision for CCTA in the USA in 2006, and the continued variability in CCTA utilisation across Medicare regions highlights the problems of non-evidence-based policy. A 2017 systematic review and meta-analysis found that CCTA, when compared with functional stress testing, has no significant impact on death or cardiac hospitalisation and leads to an increase in downstream invasive procedures.28 In addition, a large-scale observational study did not show benefit for patients undergoing initial testing strategy with stress or anatomical imaging compared with those undergoing an initial exercise stress test.29

The SCOT-HEART trial is often cited as justification to make CCTAs first line for stable angina.9 The SCOT-HEART 2018 paper concluded that CCTA use shows a significant reduction in their primary clinical endpoint, which was defined as: nonfatal myocardial infarction or death from coronary heart disease. Of note, disease-specific mortality was not statistically significant which suggests that nonfatal myocardial infarctions alone drove the significant difference in the primary outcome. Additionally, the SCOT-HEART trial randomised patients into either receiving standard care plus CCTA or standard care alone; 85% of patients received functional stress testing in both arms.28 The SCOT-HEART trial compares functional stress testing plus CCTA compared with functional stress testing alone.

Despite the lack of robust evidence and consensus on outcomes benefit and the radiation risk, CCTA use continues to grow.28 Imaging tests, such as CCTA, are often not held to similar clinical outcomes criteria of benefit as therapeutics. Additionally, often imaging tests such as CCTA lead to additional testing and downstream tests of unknown clinical value.30 More awareness and discussion of benefits, and selecting and prioritising outcome measures (e.g., non-fatal myocardial infarctions, cardiac deaths, cardiac hospitalisations, procedure-related complications, exposure to radiation) can help inform and direct national CCTA policy that is grounded in evidence. Systematic collection of CCTA outcomes data in a national database could help build a robust dataset used to drive future policy creation.

Conclusion

We found much variation in use of CCTA across the USA and England, and a fourfold higher prevalence of CT scanners in the USA. High variation in coverage policies across the USA and in utilisation of CCTA across the USA and England illustrates opportunities for better informed evidence-based policy. In addition, comparing USA to England CCTA utilisation while considering differences in permissiveness of coverage policies highlights the potential impacts of scanner availability within supply-sensitive care, such as CCTA use. Future studies should further evaluate areas with higher ordering rates and look at clinical outcomes to begin to understand the drivers of CCTA use and help to reduce low-value care.