A Health and Biomedical Informatics Research Strategy for Scotland

1 Introduction

Information from electronic health records (EHR) has been extensively used in health and biomedical research in Scotland. It has enabled the development of novel approaches to life course epidemiology, cost-effective evaluations of large-scale health service and public health interventions, efficient long term follow-up of clinical trials, and important work in pharmaco-epidemiology and a wide range of clinical research areas. In brief, it has helped to transform the safety, effectiveness and efficiency of the health services that we provide. Important examples include:

• Revitalisation of historic cohorts: linkage of information collected in childhood on a series of Scottish birth cohorts with the Scottish Morbidity Record (SMR: see Box 1) and other sources has enabled researchers to establish a number of valuable life course studies, with very high levels of ascertainment of mortality, cancer incidence and hospital admission^[4],^[5]. A particularly valuable feature of these cohorts is the data they contain on cognitive ability in childhood, and work based on them has been instrumental in establishing the discipline of cognitive epidemiology^[6],^[7].
• The impact of the 2006 ban on smoking in public places: information on hospital admissions, again from the SMR system, has been used to assess the impact of the ban on acute coronary syndrome^[8] and childhood asthma admissions^[9], and on pregnancy complications^[10]. The marked reductions observed in these outcomes strengthen the case for public smoking bans, by showing that its positive impact extends to a wider range of outcomes than anticipated. The research has also shown that some of adverse consequences postulated as a result of more smoking within the home have not occurred.
• Evaluation of routine screening for bowel cancer: Bowel cancer screening has been found to be effective in reducing mortality in randomised controlled trials, but such trials are an imperfect guide to the benefits that would be found if the treatment were carried out routinely. To see whether the benefits could be replicated when bowel screening was implemented on a larger scale, researchers used the Community Health Index (CHI) number to link information on participation in screening in three Scottish NHS Boards to Cancer Registry, Scottish Morbidity Records and mortality data. They found a 10% overall reduction in bowel cancer deaths, and a 27% reduction in those who took up the offer of screening. The research also identified ways in which the programme could be improved to increase uptake and maximise the benefits from screening^[11],^[12].
• Long-term follow-up of the WOSCOPS trial: The West of Scotland Coronary Prevention Trial was a path-breaking trial of the use of statins for primary prevention of coronary events, which found reductions in acute myocardial infarction and coronary deaths after five years of treatment. Participants in the trial were followed up for a further five years via their primary care records and (by electronic data linkage) their SMR records. At ten years' follow-up, patients in the active arm of the trial had a lower risk of non-fatal myocardial infarction, fewer cardiac deaths and lower overall mortality^[13].
• Understanding drug safety and effectiveness: Linkage of Tayside prescribing data with information on traffic accidents obtained from police records has shown a link between use of benzodiazepines, but not tricyclic or selective serotonin re-uptake inhibitor (SSRI) antidepressants and risk of road accident^[14]. Again in Tayside, linkage of prescribing and cancer registry data in an observational cohort of Type 2 diabetes patients showed that patients prescribed metformin, as opposed to an alternative course of therapy, had a lower incidence and a longer median time to a diagnosis of cancer. The results provide a strong rationale for conducting randomised trials of metformin in subjects with a high risk of developing cancer^[15].

Studies like these have positioned Scotland as a leader in research using routinely collected health data. The large-scale investment now being made in health informatics research in Scotland and elsewhere in the UK, coupled with novel ways of capturing biomedical and clinical data, creates huge new opportunities. To the research areas listed above we can add:

• The impact of new cost-effective genomic technologies, not only as a research tool to elucidate mechanisms and susceptibility to disease in populations but also for the tailored treatment of people living with cancer and other chronic diseases (precision medicine)
• The diagnosis of rare diseases and subsequently for the clinical evaluation of individual risk of disease, and:
• The targeting of prevention, diagnosis and treatment

But to realise the full potential of these opportunities, there are challenges in creating the optimum configuration of infrastructure, governance principles on access to data, the quality and consistency of that data and engaging the public, and customers. Much work is already in train to address these issues.

However we believe it is necessary to co-ordinate the activities to deliver these opportunities by promoting a vision of where Scotland aims to be, and what needs to be done to achieve those aims.

The Vision

Our vision is for Scotland to set an international standard for the safe and secure use of EHRs and other routinely collected population-based data for research purposes. We will achieve this by:

• Creating the national focus for high-quality research using EHRs within a federal network of accredited safe havens, trusted environments where research can take place, across Scotland
• Contributing to the development of a streamlined, efficient, proportionate governance system for health informatics research that has public engagement and public involvement at its heart
• Playing a leading role in the new UK Farr Institute of Health Informatics Research, and collaborating in international initiatives such as the Global Alliance for Genomic & Health that is looking at sharing genomic and clinical data internationally
• Developing innovative training programmes to build research capacity and capability within the NHS and beyond
• Advancing methodological development in data linkage, manipulation, and analysis
• Creating opportunities to link datasets that are not yet linked routinely for research, such as Scottish general practice (GP) records, the Scottish nationwide prescribing dataset, disease registers, and non-health datasets
• Collaborating through the UK-wide network, to promote the establishment of 'deeply phenotyped' cohorts using EHRs linked to biologic datasets to advance the field of stratified medicine
• Developing partnerships with research organisations, policy makers and industry, in order to accelerate the translation and impact of health informatics research
• Supporting the development of the Scottish Government's policy Joined Up Data for Better Decisions24 through delivery of a National Data Linkage Framework
• Promoting greater convergence of research and eHealth strategies to ensure the ability to efficiently use and share routinely collected NHS data for research and other purposes is built into the specification of new clinical information systems

In Chapter 2 we have described the features of the health informatics research landscape in Scotland; Chapter 3 sets out how that landscape needs to develop to deliver our Vision.