International comparisons of 'big' health record data: application to cardiovascular diseases

Cardiovascular diseases (CVD) remain the leading cause of death in the UK and worldwide. With medical advancement in drug treatment and revascularization intervention, the acute hospital treatment phase shortens and the care for CVD has gradually shifted to primary care and community services. International comparisons can provide evidence of performance of health system in CVD care between countries, and thus facilitate learning from better performing systems.

One of the near-term grand challenges is to exploit international structured coded data arising from primary and secondary care to address the gap of putting scientific evidence on prevention and treatment into clinical practice to improve population health. While the quality and comprehensiveness of UK healthcare data is exceptional in world, there are remarkably few instances of UK health records being used in international research contexts. A central bottleneck is knowing what electronic health records (EHR) sources are available for research, and their strengths and weakness. There has been no concerted effort to make such data sources 'discoverable,' shareable or accessible to the international research community. Even the most simple questions (e.g. which countries have primary care data available for research, and which of these can provide meaningful data on smoking, blood pressure and cholesterol) do not have readily accessible answers.

With the support from the MRC Population Health Scientist Fellowship, in close collaboration with the Farr institute, I will expand international collaboration on linked population health records to investigate important research questions on cardiac care, and address some of the barriers in the field of health informatics. I will do this by addressing two interwoven aims:

(i) Clinical research: to address research questions in the prevention and progression of cardiovascular diseases across two or more countries, which can only be tackled using electronic health record data.

(ii) Informatics methods: To facilitate the discovery and documentation of contemporary EHR data sources relevant to cardiovascular disease research in countries with diverse healthcare systems by working in close collaboration with health informatics experts across the Farr Institute.

OBJECTIVES
(i) Clinical research objectives will be first to compare the extent to which adherence to major new guideline recommendations might be associated with reduction in a wide range of CVD in two or more countries. Second, to compare the incidence of CVD, such as heart failure in the short and long term after acute myocardial infarction between two or more countries.

(ii) Informatics method objectives will be first to build an inter-disciplinary, international network of investigators and resources seeking to make EHR data sources discoverable and sharable for cardiovascular research. Second, to facilitate the creation of an international searchable meta-data catalogue of EHR data from different health care settings through the Farr Institute and the Health Informatics Network.

Potential applications and benefits
By focused on CVD, research results are expected to provide insights on effective health system and care strategies to contribute to the health and wellness of the population. The research may contribute to knowledge by firstly providing the unique understanding the quality and outcome consequences of preventive and treatment care across countries. Secondly, and tightly interwoven with the substantive clinical questions, I will be able to contribute to the knowledge of methods by making complex data sources across countries more 'discoverable', and provide the contextual and metadata information required in order to scale replicable research using these data sources. The research will contribute to our knowledge on research collaboration, study coordination and information governance for international health informatics studies.

Aims and objectives
(i) Clinical research: to address research questions in the prevention and progression of cardiovascular diseases across two or more countries.
(ii) Informatics methods: To facilitate the discovery and documentation of contemporary electronic health records (HER) sources for cardiovascular disease (CVD) research in different.

Methodology
In clinical investigation, I will define strategies of handling data quality and comparability issues, harmonize patient inclusion/exclusion criteria, disease ascertainment, study variables, missing data and analytical protocol across different countries. Statistical analyses steps include: 1) descriptive statistics, 2) survival analysis and longitudinal analyses for observed and adjusted risk estimates, 3) prognosis analysis, 4) estimating missed opportunities burden and its association with excess adverse outcomes between countries. Between country analyses could be implemented via a) secure remote access, b) site visit, and c) a common protocol. Computationally intense analyses, including imputation and multi-level approaches will exploit the Farr high performance computing infrastructure.

In contributing to informatics method, I will be first to build an interdisciplinary, international network of investigators and resources seeking to make EHR data sources discoverable and sharable for cardiovascular research. Second, to contribute to the creation of an international searchable meta-data catalogue of EHR data from different health care settings through the Farr Institute and the Health Informatics Network.

Impact
By focused on CVD, research results may provide insights on effective health system and care strategies for better health and wellness of the population in the UK and worldwide. In methodology, the research will contribute to our knowledge on research collaboration, study coordination and information governance for international studies based on linked electronic health records.

According to the Chancellor of the Exchequer the UK has the 'world's best and most complete healthcare records'. An overarching impact of my fellowship is to illustrate specific examples of where this may be usefully true and where it is not. The research contributes to the UK's global leadership in EHR research and facilitates the shaping of EHR scientific agenda. By focused on CVD, the area with high burden, the research may contribute to the economic and societal gain of the UK through improving the health and wellness of the population. Individuals can fully engage in life and work with minimum absence or compromise of their capabilities due to diseases. The sharing, learning and transfer of innovations among countries may facilitate the funding leveraged from international partners to support scientists in the UK and worldwide.

By giving greater visibility to the types of data available, their quality and accessibility in different countries, and by working with colleagues in EHR4CR and other international initiatives, my fellowship will impact pharmaceutical industries in their trial pipelines at multiple steps, from initial lower cost feasibility estimates right through to real world evidence studies increasingly required for regulatory approval. The international collaboration established in the proposed research may strengthen the links with industrial partners of the Farr Institute, London (AstraZeneca for example) for novel national and international projects involving industry, funders and researchers across health sectors. I will ensure that my work is aligned with initiatives arising out of the Farr-ABPI (Association of the British Pharmaceutical Industry) forum.

Computing, internet and health record vendor industries: my approach - a 'bottom up' single fellowship involving close clinical collaboration can provide much needed insights for blue chip, small and medium enterprises wanting to develop products in different international markets.

Policy-makers nationally: I seek to leverage the close links with NHS England to inform NHS policy. Specifically the national cardiovascular intelligence network will be directly impacted by my results. With Farr Investigator Deanfield (NICOR) lead and Farr Cardiovascular lead I will seek to ensure that my findings are used to inform quality improvements in the prevention and treatment of CVD as well as initiatives to improve the quality of data collection (e.g. RCP Professional Standards).

International policy: my findings should inform international guidelines for CVD care and data collection standards (for example, the Cardiology Audit and Registration Data Standards).

The third sector: I will seek to work with the British Heart Foundation to complement and modernize their 'Health Statistics' using new opportunities from health record data.

The wider public, patients and their family may be the largest partners in the health system, and the largest long-term beneficiaries of the research. Study results may help to form training for patients or family to understand their health records, and engaging them in the design of better utilisation of care services for managing their health.

Healthcare professionals: I anticipate that my findings will encourage clinicians to develop and share best practice in a 'quality cycle' informed by a 'data cycle' so that improvements in the quality and outcomes of care go hand in hand with improvements in data quality and the way data are used to generate knowledge for decision support.

Timeline and assessment
Year 1: establishment of framework for investigator network, and meta-data for describing data within and across countries. Finalisation of which countries and which datasets will be analysed.
Year 2: completion of first international comparison analysis and manuscript submission.
Year 3: completion of further international comparison publications, communications to all beneficiaries.