National Trends in Coronary Artery Disease Imaging

The University of Cambridge, University of Edinburgh and Royal Brompton and Harefield NHS Trust will use the data to better understand the impact of national guidelines on the investigation of stable chest pain (CG95), both on resource utilization of the different imaging modalities, and the resultant downstream morbidity and mortality. Such knowledge will allow the University of Cambridge, University of Edinburgh and Royal Brompton and Harefield NHS Trust to determine if the findings from such trials are being actualized in the routine clinical environment. Analysis of the imaging trends and their outcomes will inform future clinical practice, and further research in this area.

University of Cambridge, University of Edinburgh and Royal Brompton and Harefield NHS Trust are joint Data Controllers and University of Cambridge is the sole Data Processor. The data will be controlled and held solely by the University of Cambridge School of Clinical Medicine. Data processing will only be carried out by substantive employees of the University of Cambridge who have been appropriately trained in data protection and confidentiality.

Data on the use of diagnostic imaging tests pertaining to cardiac investigation is required from the Diagnostic Imaging Dataset (DIDS). Longitudinal data covering the period before and after the release of the guidelines will determine their impact on practice. DIDS has collected data from 2012 to present. The totality of this period will allow for comparison of the trends in imaging utilization in the 4 years before and the 2 years after the introduction of the guidelines. Coverage of the whole of England is required to determine regional trends, and areas of best practice. National data is to be unsuppressed aggregated data and split into Sustainability and Transformation Partnership (STP) geographical regions to allow for sufficient inter-regional variability to be examined.

Data from the Hospital Episode Statistics and Office of National statistics pertaining to admissions for fatal and non-fatal Myocardial Infarction (MI)s, cardiovascular death and all cause death, are required to determine the clinical impact of the new guidelines on the health of the population. To examine the trends in these outcomes in those who have undergone a cardiac diagnostic imaging test, these recorded endpoints in those who have undergone a cardiac test identified in DIDS is requested for the same period of durations as the DIDS data collection (2012-2018). Again, data is to be unsuppressed aggregated data at the level of STPs to allow for sufficient inter-regional variability to be examined.

Non-invasive cardiac imaging has assumed a central role in both the diagnosis and management of coronary artery disease. Since its introduction in 1963, single-photon emission computed tomography (SPECT) has become the mainstay for the investigation of coronary artery disease. However technological advancements in stress echocardiography (SE), positron emission tomography (PET), coronary computed tomographic angiography (CTA) and magnetic resonance imaging (MRI) have allowed for alternate strategies for the investigation of chest pain with comparable or superior diagnostic accuracy.

In 2016, the UK National Institute for Health and Care Excellence (NICE) updated their guidelines for the investigation of stable chest pain (CG95). The 2010 guidelines had recommended a risk-stratified approach of coronary artery calcium scoring with potential CTA for low-risk individuals; functional testing using SPECT, stress echocardiography, or MRI for intermediate-risk individuals; and invasive coronary angiography (ICA) for high-risk individuals. The updated 2016 guidelines recommended a common approach to the investigation for coronary artery disease using CTA in those with possible angina due to its high sensitivity and negative predictive value. The specificity and positive predictive value of CTA is however much more limited, with several studies reporting that coronary CTA is associated with increased downstream testing, although this observation has been inconsistent across studies. Furthermore, the 5 year follow-up of the SCOT-HEART trial demonstrated that routine CTA utilization resulted in a reduction in non-fatal MI, a finding replicated in a retrospective Danish registry.

There is thus a public interest in being able to better understand the impact of these national guidelines, both on resource utilization of the different imaging modalities, and the resultant downstream morbidity and mortality. Such knowledge will allow them to determine if the findings from such trials are being actualized in the routine clinical environment.

To achieve this goal, data on the use of diagnostic imaging tests pertaining to cardiac investigation is required from the Diagnostic Imaging Dataset (DIDS). Longitudinal data covering the period before and after the release of the guidelines will determine their impact on practice. DIDS has collected data from 2012 to present. The totality of this period will allow for comparison of the trends in imaging utilization in the 4 years before and the 2 years after the introduction of the guidelines. Coverage of the whole of England is required to determine regional trends, and areas of best practice. Data is to be unsuppressed aggregated data at the level of STPs to allow for sufficient inter-regional variability to be examined.

Hospital Episode Statistics and Civil Registration statistics pertaining to admissions for fatal and non-fatal MIs, cardiovascular death and all cause death, are required to determine the clinical impact of the new guidelines on the health of the population. To examine the trends in these outcomes in those who have undergone a cardiac diagnostic imaging test, these recorded endpoints in those who have undergone a cardiac test identified in DIDS is requested for the same period of durations as the DIDS data collection (2012-2018). Again, data is to be unsuppressed aggregated data at the level of STPs to allow for sufficient inter-regional variability to be examined.

No other data repository other than the combination of DIDS, HES and Civil Registrations will allow for such an examination of the guidelines on the totality of the population to which they apply.

The data will be controlled and held solely by the University of Cambridge School of Clinical Medicine and will only be accessed by those with a contract with the University of Cambridge. The University of Edinburgh and the Royal Brompton and Harefield NHS Trust have been involved in the formation of the study design and will be involved in the interpretation of the study results. Whilst they do not have direct access to the data, their input into the study and the decisions surrounding how the data is analysed by the University of Cambridge identifies them as joint Data Controllers.

The University of Cambridge are requesting national, unsuppressed, aggregated data split at the level of Sustainability and Transformation Partnership (STP)s, with linkage of DIDS with data from HES and Civil Registrations. This will allow for the goals described in the objective for processing section to be achieved. No data will flow from the University of Cambridge to NHS Digital. Data produced by NHS Digital will be transferred to and stored within the secure area of the University of Cambridge School of Clinical Medicine. Data processing will only carried out by substantive employees of the University of Cambridge who have been appropriately trained in data protection and confidentiality.

To achieve the study aims, the specific processing requests required for the current work are as follows:

From DIDS:

- Aggregate data at an STP level for each imaging code related to cardiac investigation. The data is requested in a monthly format with summated yearly totals for each code (monthly data is requested as the DIDS is incomplete for 2012 and linear regression will need to be performed to estimate total yearly numbers from the 8 months of data available from this year).

- Aggregate data of the number of people undergoing a second cardiac imaging procedure following an initial cardiac investigation, with the code of this subsequent investigation and the code of the initial imaging investigation.

- Through linkage with HES, the aggregate age and sex for each cardiac imaging code for each imaging code is requested for each year.

HES:

- For all patients identified from DIDS to have had a cardiac imaging investigation, data is requested on numbers of admissions for an acute coronary syndrome or myocardial infarction.

- Within the same cohort, outputs of the number of per cutaneous coronary intervention procedures and coronary artery bypass procedures is requested.

- Again this data is requested stratified by year and STP to allow for examination of temporal and regional trends.

Mortality:

- For all patients identified from DIDS to have had a cardiac imaging investigation, aggregate data is requested for all cause mortality, fatal myocardial infarction, fatal stroke, and codes related to cardiovascular death.

- Again this data is requested stratified by year and STP to allow for examination of temporal and regional trends.

Re-identification of individuals is not permitted.

The primary output of the study will be findings disseminated by publications in peer-reviewed journals and presented at medical conferences to academics, NHS national policy makers and on the web.

The study investigators plan to submit for publication of the results in The Journal of the American Medical Association (JAMA) or the British Medical Journal. It is also planned to disseminate this knowledge via scientific presentation of the abstract at the European Society of Cardiology. In all published outputs the University of Cambridge study investigators will ensure data is aggregated and have small numbers suppressed in line with the HES analysis guide.

The University of Cambridge anticipate the analysis to be completed by the completion of 2019. The goal will be for publication and presentation within 12 months of the completion of the project, allowing for the factors out with the study investigators control such as peer review, and article revisions and amendments all of which can introduce substantial delays.

The expected benefits to Health Care will be brought about by the widespread dissemination of the knowledge gained from the undertaken study as specified in section 5c.

Better understanding of imaging utilisation trends and which imaging modalities provide for the best patient outcomes will allow for better infrastructure planning to meet future demands . This will come about both through justification of changes in funding/fund allocation at a national level based on changes in changing demand and utilisation. At a more local level, such changes will be brought about through impact on future scanning technology acquisition, with a better more informed choice guiding the choices between the multiple medical technologies that are available. This will be managed at a health service provider level by providing the evidence necessary for local units to build business cases to expand local services based on current and projected growth.

Quantification of the impact of the initial imaging investigation on prevalence of downstream testing and cardiovascular events will allow for a more holistic assessment of the different imaging modalities. It will allow for assessment of whether the expected gains from the change in guidelines in the reduction of patient morbidity and mortality, as would be expected based on the results suggested from recent RCTs, are materialising in the real world. Replication of such findings outside the trial environment is important as it will substantiate that trial findings these can be translated into real world settings. If these benefits are not being actualised, it will stimulate further research to identify the barriers to this translation from the research realm to the clinic or lead to further research to understand why the expected benefits are not manifest. This will come in the form of future research grant applications, both by the current authors, and by others in the field.

Dissemination of the results will inform current healthcare practice and research goals , feeding into future advancements and improvements in the diagnostic pathway of coronary artery disease. This will come in the form of more robust guidelines to be derived by national and international societies of the best investigation strategy for the investigation of chest pain. Such guidelines are informed through robust literature reviews and assimilation of the extent of the current evidence. Publication and presentation will aid this, as will the authors involvement in clinical and imaging societies (such as the British Society of Cardiovascular Imaging) which also have representation and act as specialist interest groups in larger groups such as the Royal College of Radiologists where national guidance can be issued.

Finally, more accurate definition of changes in imaging modality utilisation will allow for a more informed targeting of training of the next generation of radiologists and cardiologists who will be responsible for meeting future imaging performance and reporting demands. This will be driven by sharing of the outputs with the British Society of Cardiovascular Imaging, British Society of Nuclear Cardiology, British Society of Cardiac MRI, British Society of Cardiology and Royal College of Radiologists who work in conjunction with one another as well as with the GMC in the guidance of current and future training needs of cardiac imaging.

Yielded Benefits is not a requirement for new applications.