Article Text

PDF

Engaging with the clinical data transparency initiative: a view from the National Institute for Cardiovascular Outcomes Research (NICOR)
  1. Chris P Gale1,2,
  2. Clive Weston3,
  3. Spiros Denaxas4,
  4. David Cunningham5,
  5. Mark A de Belder6,
  6. Huon H Gray7,
  7. Roger Boyle5,
  8. John E Deanfield5,
  9. on behalf of the NICOR Executive
  1. 1Centre for Epidemiology and Biostatistics, University of Leeds, Leeds, UK
  2. 2Department of Cardiology, York Teaching Hospital NHS Foundation Trust, York, UK
  3. 3College of Medicine, Swansea University, Swansea, UK
  4. 4Clinical Epidemiology Group, University College London, London, UK
  5. 5National Institute for Cardiovascular Outcomes Research, University College London, London, UK
  6. 6Cardiothoracic Division, The James Cook University Hospital, Middlesbrough, UK
  7. 7Departmen of Cardiology, University Hospital of Southampton, Southampton, UK
  1. Correspondence to Professor John E Deanfield, National Institute Cardiovascular Outcomes Research, National Centre for Cardiovascular Prevention and Outcomes, 170 Tottenham Court Road, London W1T 7HA, UK; j.deanfield{at}ich.ucl.ac.uk

Statistics from Altmetric.com

The availability of centrally analysed, comprehensive outcome data from cardiovascular audits has underpinned the success of the National Service Framework (NSF) for coronary heart disease1 in achieving a substantial reduction in morbidity and mortality from cardiovascular (CV) disease in the UK.2

The UK Government is now intending to implement ‘transparency initiatives’ within the National Health Service (NHS)—reporting outcomes for General Practice, hospital complaints and satisfaction surveys and healthcare team performance.3 Routine (‘raw’) clinical data will be made publicly and commercially available, so that independent parties will have open access to anonymised data relating to all recorded components contributing to the quality of CV care. The coalition Government has promised ‘an openness about results that we have never experienced before’, in an attempt to ‘reform public services, foster innovation and empower citizens’.3 While there is likely to be little opposition to policies aimed at improving quality of care for patients with CV disease, strategies for managing national clinical data present a number of challenges if they are to impact on the drive towards further reducing the burden from CV disease in the UK. The National Institute for Cardiovascular Outcomes Research (NICOR) runs a programme that engages with aspects of the Government's agenda. It addresses the particular challenges of managing large complex clinical datasets, while minimising any unintended consequences of releasing raw data into the public domain.

The National Institute for Cardiovascular Outcomes Research (NICOR)

NICOR is an umbrella organisation for a number of national registries created by professional CV societies. It is based at University College London (UCL), importantly offering linkage of national cardiovascular databases and the potential for increased international comparative data analysis. Established by Professor Sir Bruce Keogh, in 2006, it is currently led by Professors Sir Roger Boyle and John Deanfield, and receives grant funding from the Department of Health. The role of NICOR is to manage these national clinical audits, regularly report clinically validated outcomes to the public, participating organisations and regional and national health authorities, and to facilitate research using the data.

Through collaborative and transparent working partnerships with clinicians, database managers, analysts, health services researchers, patients and the public, NICOR seeks to provide healthcare service users and providers in England and Wales with high quality, valid, generalisable and timely information on the quality of CV care. Cross-sectional data are currently collected on coronary percutaneous interventions (British Cardiovascular Intervention Society), adult cardiac surgery (Society for Cardiothoracic Surgery), and acute coronary syndromes (Myocardial Ischaemia National Audit Project, MINAP),4–6 as well as cardiac rhythm management (Heart Rhythm UK), congenital heart disease (British Congenital Cardiac Association), and heart failure (British Society for Heart Failure). These data are used not only for audit, performance evaluation and research purposes, but are also referenced by policy makers and regulators to inform on the delivery of clinical services. NICOR is, therefore, well positioned to address the many complex issues that lie behind the ‘transparency initiative’ as they apply to CV disease.

NICOR collective achievements

Under the auspices of NICOR, the established and emerging CV audit databases each collect prospective clinical data on thousands of NHS patients, mainly from hospitals in England and Wales. In the case of MINAP, this extends to over 800 000 acute coronary syndrome events (figure 1). After implementation by the NSF for coronary heart disease in 2000,1 these data have been increasingly used to underpin service evaluation and implementation of change,7–11 performance reporting12 ,13 and research.14–18 Their analyses have revealed, for example, regional variation in mortality rates,19 the association between hospital performance and volume of cases of primary percutaneous coronary intervention,20 and substantial reductions in hospital mortality after acute myocardial infarction over the last 8 years.2

Figure 1

Plot of acute coronary syndrome events and hospitals recorded in the Myocardial Ischaemia National Audit Project (MINAP), 2004–2007. Black: acute coronary syndrome events. Red: hospitals.

International comparisons of CV treatment and outcomes using NICOR's data and corresponding national registries in Sweden21 have recently been initiated. In addition, the full potential of the national audit data will soon be realised through linkage to create one of the largest CV databases in the world. This promises a unique insight into real pathways of care, crossing specialty boundaries (acute cardiology, cardiothoracic surgery, elective coronary intervention, pacemakers, cardiac rehabilitation, heart failure, transcatheter aortic valve replacement) and including socioeconomic status, genetics, primary care and tracked cause-specific mortality.

NICOR developments and challenges

Data collection relies on the enthusiastic support of professional societies, clinicians and audit personnel. Reliable management and interpretation of national data to ensure robust outputs, requires among other things a sophisticated platform and suitable expertise for handling the analysis of complex hierarchical data. For this reason, NICOR was recently relocated within the academic environment of UCL's new Centre for CV Disease Prevention and Outcomes. This will enable succinct and timely delivery of standardised outcome information to the multiple stakeholders, including provider organisations, regulators and funders. The co-location and standardised analytical approach, provide exciting opportunities to link data across audits, and to track and assess the full patient journey through primary, secondary and tertiary care. This activity will be enhanced by plans to link data from the new Clinical Practice Research Datalink, and to obtain cause-specific mortality from the Medical Research Information System, which is part of the NHS Information Centre.

A further advantage of NICOR is the richness of the clinical data collected, which is achieved largely by the input of clinically trained staff, under supervision of the specialist national societies. When coupled with the data analysts' familiarity with the various datasets, this allows flexible and sophisticated risk adjustment which is much more robust than can be achieved using data such as those from Hospital Episode Statistics (HES), which are arguably more appropriate for administrative than clinical purposes. The international GRACE risk score for the evaluation of outcomes after acute coronary syndrome, locally validated in a large population (>50 000) by use of MINAP data,22 and endorsed by NICE,23 is a good example of national clinical data being used to improve patient care. It is anticipated that this approach will be extended to other clinical areas, such as the NHS National Health Checks, to provide a solid foundation for the development of new prevention initiatives.

The NICOR approach to data analysis has the potential to contribute greatly towards clinical trials in the UK. For example, it is possible to perform pseudo-randomisation of patients to test an historical evidence-based intervention in practice, to help inform clinical trial design and to quantify the economic and clinical consequences of non-adherence to care. Moreover, prospective clinical trials can be embedded in national registries, an approach which has already been successfully implemented in Sweden.24 This has a number of advantages including rapid recruitment at scale, ‘real world’ inclusive data and much lower costs when compared with the traditional design and conduct of randomised trials.

Translating a political call for wider, unfettered use of readily available clinical data into reality raises a number of issues. First, data collected by healthcare professionals from NHS patients as part of the national clinical audit, is highly sensitive. A careful process is required to ensure patient confidentiality, while maximising the value of the information contained in the national registries. In addition, ‘raw’ (uncleaned) data may not be ‘fit for purpose’, and erroneous conclusions can be drawn if they are handled incorrectly. Incomplete case ascertainment, incomplete data, measurement error, implausible/incorrect data, digit preferencing, phenotype and treatment definition variation, fuzzy matching, data duplication and the nested nature of the data make for complex databases.19 ,25 NICOR recognises these pitfalls, and has invested in specialists to help minimise biases, so that cleaned data may be offered to analysts for study. The need for both, transparency of primary data and the importance of data management and optimisation before release, have been highlighted by the recent debate over the publication of the NHS risk register.26 At the analytical stage, furthermore, there are many hurdles to overcome, and the modelling of the data requires ongoing development and sustained investment for best output.

Much time has been spent debating the thorny issue of data ownership. While the Secretary of State for Health carries final responsibility for the use of data within the NHS, it may be argued that the raw audit data is ultimately owned by patients. However, there is also an element of intellectual property associated with specialist data-cleaning and linkage to other datasets, and so, NICOR has a claim on some intellectual property ownership. Patient representatives accept that the responsibility for the analysis and reporting of information lies appropriately with those who have the necessary expertise, and NICOR's practise adheres to the Government's proposals that CV data should be made available to external parties.

Indeed, NICOR has welcomed, and supports, applications for the use of data from researchers, industry and healthcare professionals. Typically, researchers experienced in complex observational data and with funding from grant-giving bodies, have been provided with a ‘fit for purpose’ dataset. Moreover, NICOR regularly assists and works with other parties who may not have such support (figure 2). Care has been taken to ensure that this policy for data flow, analysis and availability has been endorsed by patients and the public, whose representatives are part of NICOR's strategic development. Additionally, the CV audits regularly engage with, or typically have, patients and the public as members of their Steering Group. Frequently, research applications for the use of national audit data have been vetted by a Patient and Public Involvement group as part of a grant application. For all these reasons, NICOR aligns with the core open Government principles of transparency, accountability and technology and innovation.27

Figure 2

The National Institute for Cardiovascular Outcomes Research (NICOR) and transparent cardiovascular data flow.

Conclusion

The development of NICOR and the CV registries, supported by specialist societies, is due to the considerable efforts of those who collect the data, and of the clinical leads that have devoted so much of their own time to its analysis and dissemination. Much of this earlier work received little funding and, as a consequence, the full potential of these audits is yet to be achieved. With some increase in funding, the promotion by Government of local data collection, and with its relocation within an academic environment, NICOR can make a substantial contribution towards the use of such data; to inform research and the quality of patient care. It has already enabled tracking of key measures of adherence to the ‘standards’ that were set in 2000 in the NSF for Coronary Heart Disease. The present coalition Government has placed great emphasis on commissioning and delivering healthcare, based principally on the quality of clinical care delivered, relying less on other metrics of delivery, such as waiting time. These ‘Quality Outcomes Frameworks’ in the NHS, Social Care and Public Health28 will only contribute to better service delivery if outcomes are measured and reported accurately and comprehensively. In the field of CV disease, NICOR can make an important contribution to these objectives, and maintain the momentum in reducing CV risk and improving patient care.

Acknowledgments

The authors gratefully acknowledge all the hospitals for their contribution of data to the national cardiovascular audits.

References

View Abstract

Footnotes

  • Funding CPG is funded through the National Institute for Health Research (NIHR) NIHR/CS/009/004. The extract from the MINAP database was provided through the MINAP Academic Group. We acknowledge funding from the British Heart Foundation PG/07/057/23215.

  • Competing interests None.

  • Provenance and peer review Not commissioned; internally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.