Project Title Project Lead Lay Summary Date Issued to IGRP Outcome Date of IGRP approval Health Care Utilisation, Care Pathways and Ann John, Associate Professor Public 19/10/2015 Approved 23/12/2015 Educational Status in Children and Young People with Mental Health, Swansea University Mental health issues are of growing concern and a source of controversy in children and young people (CYP) in the United Kingdom (U.K). They have been found to contribute to adverse life outcomes (such as poor Adolescent Mental Health (AMH) problems with a educational achievement) and serious disruptions to CYP's lives and those of their families . It is estimated from Office of National Statistics population survey data that, at any one time, four in 100 of CYP, have a focus on Self harm, Eating Disorders and Anxiety and clinically diagnosable or relevant emotional disorder (anxiety or depression). Some disorders such as anxiety may start when children are as young as 11 years. Earlier age of onset appears to be associated with Depression) greater severity and poorer long term outcomes . Concerns are heightened by the persistence of childhood or adolescent mental health issues into adulthood where up to a fifth of the adult population may be affected by a common mental disorder at any one time. Many of these conditions are managed in primary care and there are issues about the transition of care between primary and hospital care and appropriate settings to deliver care to children. There are also issues relating to those that do make contact but are lost to services from childhood to adulthood and between primary care and hospital care. We will explore differences in outcomes for these children.This project develops work already done in Wales to across the UK looking at levels of these disorders and with a particular focus on self harm, eating disorders and anxiety and depression, patterns of healthcare contacts and educational attainment. We will explore the outcomes for those children who attend services and how they are manged with a view to informing interventions. This project has been funded by the Healthcare Quality Improvement partnership (HQIP) so has a unique opportunity to inform and impact on services. Follow—up study of patients with latent autoimmune Dr Mark D Atkinson, Research Officer, 13/10/2015 Approved 23/12/2015 diabetes in adults (LADA) and type 2 diabetes. School of Medicine, Swansea The majority of cases of diabetes after the age of 35 are type 2 diabetes, but there is a significant proportion of these (approx. 4%) who have a type of diabetes more like type 1 called Latent Autoimmune Diabetes in University Adults (LADA), who require insulin treatment more quickly. These can be distinguished with an antibody test (GADA). We have a group of people from South Wales already in SAIL who have had the antibody test don so we know who has type 2 diabetes and who has LADA. This was a pilot study carried out between 2004 and 2007 and GADA tests continued for a few years. GADA testing is not routine in Wales or any other part of the UK. We will be able to see, from GP and hospital data, how these 2 groups have been treated and what their glucose levels are. We can also see the various complications they may have developed. From this we can determine whether LADA patients received best treatment and compare their outcomes with type 2 patients. Although test results are often present in the GP data, some may only be found in the Morriston laboratory dataset (PAMOV) Another aim of the research is to attempt to distinguish LADA patients from type 2 patients in the absence of GADA tests. By looking at a wide range of factors in our cohort of known LADA and type 2 patients, we will try to find methods of distinguishing between the two. Factors would be, for example; BMI, presence of other autoimmune conditions including thyroid conditions, time from diagnosis to first prescription of insulin. Once a way of distinguishing between the two types of diabetes has been established, this method can be used to identify cases of LADA in the absence of GADA tests. We hope to test the idea that these patients may have worse outcomes than type 2 patients as their true condition is not known when treatment starts. If this is the case, we would hope to use the results to recommend the wider use of GADA testing.
Tagged Electronic Cohort Cymru (TECC) Professor Sinead Brophy, Swansea In line with the aims of the National Centre for Population Health and Wellbeing Research, the Tagged Electronic Cohort Cymru (TECC) can help us to better understand illness and the effects of medications used to 18/09/2015 Approved 23/12/2015 University treat disease. The cohort will be developed from the linkage of routine data collected in Wales. Linking data from general practitioner records, hospital records and disease specific datasets (for example, for cancer and arthritis) will create a vast dataset of adults, comprising healthy individuals and those who may present with different conditions. This dataset allows researchers to report the number of people who have certain illnesses, who the illnesses are most likely to affect and when, as well as identifying illnesses that often exist together. The medications given to treat conditions will be examined, including associated adverse effects and interactions of medications. The dataset can answer basic questions about conditions in the Welsh population. TECC allows rapid analysis and can quickly support existing research ideas and generate new ones regarding illnesses, medical procedures and medications administered.
Treatment of depression in patients with epilepsy Owen Pickrell, Clinical Lecturer It is known that people with epilepsy are more likely to suffer from psychiatric illnesses, especially depression, than the general population. There is some evidence that depression in people with epilepsy may be 24/08/2015 Approved 28/09/2015 (Neurology), College of inadequately treated despite treatment recommendations, although there have not been any large‐scale studies to our knowledge. MedicineSwansea University Building on previous depression (led by Ann John) and epilepsy (led by Mike Kerr) research using the SAIL databank, we aim to study the treatment of depression in people with epilepsy. We will compare anti‐ depressant prescribing rates and secondary care referral rates in people with and without epilepsy. Identification of an inequality in such treatment in people with epilepsy could partly explain the high rate of
significant depression‐related problems, such as suicide attempts, in this population. The identification of any inequalities in treatment could also help to prioritise and guide service provision for people with epilepsy.
Understanding the determinants of educational Chris Taylor, WISERD Professor, Understanding the determinants of educational achievement has been the focus of extensive research in recent years. However, much of this analysis is significantly constrained by a lack of information relating to 24/08/2015 Approved 28/09/2015 achievement and the association with health and well‐ WISERD, Cardiff University the background and educational experiences of children. Indeed, in Wales there is currently no ‘on‐entry’ assessment of a child’s educational development as they start school. Understanding the factors that being by age 14 in Wales. potentially influence children’s educational achievement is therefore important as this could contribute to the Welsh Government’s recommendations and policies to tackle inequalities in educational achievement
as well as raise educational standards.
In 2012/13 the Welsh Government introduced annual national reading and numeracy tests for all children in years 2 to 9 (aged 7‐14). This project will be the first to use this education data, and will combine it with
information from the Millennium Cohort Study (MCS). The education and MCS data will be linked to administrative health records to explore what determines educational achievement and the association with children’s health. In addition, this project provides a unique opportunity to explore the benefits and usefulness of analysing linked administrative health records and survey data (such as the MCS), therefore contributing toward a better understanding on the factors influencing educational achievement with particular interest on the relationship with health.
Previous research in Wales using the education and MCS data has been able to demonstrate, for example, the limitations of the free school meal indicator to properly identify factors that are associated with low levels of educational achievement. Unfortunately, this previous analysis has only studied educational achievement up to the age of 7 years (and the end of Key Stage 1). This project aims to extend this analysis to
study children’s educational progress and levels of achievement by the age of 14 years (at the end of Key Stage 3). . In addition, linkage with health‐related data (based on administrative health records) will allow to
further explore the determinants of educational attainment and the association with children’s health. Variation in secondary care costs for patients Kendal Smith, Head of Financial Cwm Taf Local Health Boards has worked with the Public Health Observatory and NHS Wales Information Service beginning to ascertain which factors affect the variation in costs of patients undergoing similar 24/08/2015 Approved 28/09/2015 undergoing similar procedures Flows, Cwm Taf University Health procedures. Previous work has already indicated that it is possible to measure differences between organisations in the ‘additional’ cost of treating patients who have diabetes when compared to patients without Board diabetes. The cost of treatment and the length the patient stays in hospital are also affected by the age of the patients, the number of co‐morbidities they have but also whether they are treated on specialist wards. This shows that the PEDW dataset which is routinely collected by the organisation contains differences in performance depending on the patient pathway.
The purpose of working with SAIL is to utilise the links between the primary care and secondary care data. The plan is to ascertain whether the variations mentioned above can be correlated with (1) particular programmes of treatment in primary care or (2) characteristics of patients recorded in the primary care dataset. This project would specifically look at diabetes to measure 1 above and whether for example the prescribing of a particular drug in primary care, the frequency of an HBA1c blood test or the number of GP appointments prior to admission affects the length of stay in secondary care when these patients are admitted. The second strand of the research will be to measure whether the Body Mass Index or deprivation affects the length of stay in secondary care.
This above approach, by utilising the unique characteristics of the SAIL database would begin to inform the operational managers which programmes in Primary Care most influence secondary care costs. By concentrating on these 2 factors this project could prove the concept that SAIL can be utilised for assessing which programmes most influence costs for other conditions.
Linking Bacterial Genome and Patient Records Pilot Professor Samuel K Sheppard, Bacterial infections impose a major global public health and economic burden on humans and animals. Bacteria can infect in many ways, through food consumption or interpersonal or environmental transmission. 24/08/2015 Approved 28/09/2015 Swansea University Medical School, Hospital‐acquired infections (HAIs) are among the leading causes of severe illness and death in Wales, and the rest of the world. These infections are acquired during hospital visits, and are commonly associated MRC CLIMB Centre, with assisted ventilation, intravenous catheters, indwelling devices, and other invasive surgeries. Severe infections are often fatal but many of these deaths could be prevented with a better understanding of the organisms responsible and the epidemiology of infection. Despite Government commitments to prevention and control programs to reduce HAIs, they remain common, occurring in approximately 6.4% of acute patients of hospitals in Wales. The incurred cost to the NHS is estimated to be £50m/year, but importantly there is evidence that disease is not evenly distributed among patients with some patient groups at increased risk of infection – for example those visiting busy hospitals in low income areas. Major risk factors for contracting HAIs have been identified but the reasons for inequalities in infection are not fully understood. Most HAI‐causing bacteria are common in/on healthy individuals with a subset of strains causing infection under certain circumstances. Therefore, one reason for inequalities is the association of certain demographic groups with particular lineages – such as virulent strains. This type of investigation can also be applied to bacterial outbreaks, mainly food‐borne or environmental, whereby bacteria such as pathogenic or antibiotic‐resistant Escherichia coli, can infect and spread within a community in a very short period of time. Additionally, we have a particular interest in the recent 2013‐2014 scarlet fever outbreak which occurred in the UK, including Wales. Scarlet fever is a childhood disease caused by Streptococcus pyogenes, associated with throat infections and rashes that can develop into more serious complications.
Our team brings together expertise in microbiology and genomics with expertise in informatics, epidemiology and machine learning/pattern recognition. We currently hold a collection of over 5,000 isolates of clinical isolates from various species (mainly Campylobacter, Staphylococcus and Escherichia coli) collected from various clinical and environmental sources that include Public Health Wales (PHW) microbiology at Singleton Hospital (Swansea) and Heath Hospital (Cardiff). Our laboratory routinely sequences the whole genomes of these isolates and adopt comparative approaches to unravel specific genetic associations with virulence or host association. This study aims to identify specific markers (e.g. genes associated with antibiotic resistance) in specific groups of bacteria that are predicted to play a role in virulence and infection and combine this with anonymised patient clinical information to improve knowledge regarding outbreaks and hospital acquired infections.
This study, will form a proof‐of‐concept as to how bacterial genome data can generally be linked with anonymised health data, something which can be extremely useful to the public health community. We will address:
1. How bacterial metadata should be structured in order to link to health records. 2. How specific genetic elements (identified independently from the linkage), comprising mutations in common genes or genes that have been acquired horizontally by bacteria (e.g bacteria to bacteria transfer of antibiotic resistance or colonisation/virulence factors, rather than inheritance), should be associated with specific characteristics of the people infected.
Children with Cerebral Palsy (CP) Alison Kemp ,Professor of Child 24/08/2015 Approved 23/12/2015 Health, Cardiff University Cerebral Palsy (CP) is a disorder of movement and posture resulting from a non‐progressive injury to an immature brain, affecting two to three per 1000 live births. Associated conditions can include epilepsy, communication and learning difficulties, hearing loss, visual impairment, behaviour issues and feeding difficulties. Children with CP frequently develop respiratory, digestive, musculoskeletal and nutritional
disorders that can require hospitalisation. As CP is not a common condition, large samples of children with CP are rarely studied The availability of routinely collected data covering the entire child population of Wales, provides an excellent opportunity to describe the nature and extent of health needs and healthcare utilisation of children with CP. Cerebral palsy is wide ranging in severity and so educational experiences of these children vary wildly. Some children with mild forms of CP may flow through mainstream curriculum with minimal intervention and support, whilst those with complex needs require tailored support within a
mainstream setting or a special school placement. We aim to provide an epidemiological overview of children with CP and to identify any unmet need. Evaluating the impacts of Welsh Government funded Sian Morrisson‐Rees, Researcher, The World Health Organisation recommenda minimumindoortemperatureof18 degrees and recognise thatliving in cold and/or damp house may be harmfulto health. Inadequate indoortemperatures increase the 31/07/2015 Approved 28/09/2015 schemes designed to improve the energy efficiency of ADRC, Swansea University risks of mortality and excess winter deaths. Also there is evidence to support that living in cold homes has an effect upon cardiovascular, respiratory and mental health. the homes of low income households on health Rising fuel costs in the past decade have led to an increase the numberof people living in fuel povertyand inadequately heated houses. Fuel poverty is particularly acute in Wales; in 2011 UK estimates indicate outcomes through the use of existing data. that 29% of households in Wales were in fuel poverty compared to the UK figure of 17%. The Welsh Government’s fuel povertystrategy sets outto reduce the numberof people in Wales living in fuel povertyorat risk of fuel poverty.A key action in the strategy is to develop programmesto improvethe energy efficiency of homes and to targetthese improvementsat householders most in need and living in the most inefficient homes. A series of Welsh Government initiatives were launched as partof the Home Energy Efficiency Scheme (HEES). HEES providedhome efficiency improvementsto those most likely affected by fuel poverty,including low income and vulnerable households from 2000 to 2011, when it was replaced by the NEST programme. To improve the energy efficiency of homes, the measures provided by HEES and NEST included insulation and heating upgrades, such as a more efficient boiler. The overall aim of the programmeis to evaluate the Welsh Governmentfundedschemes designed to improvethe energy efficiency of the homes of low income households on whether it had an impact upon the health of those most likely to be living in fuel poverty.
Paramedic Supplied ‘Take Home’ Naloxone: a Mr Chris Moore, Clinical Support Lead 31/07/2015 Approved 28/09/2015 randomised feasibility study Welsh Ambulance Services NHS Trust Naloxone is an extremely effective antidote which is used to reverse the symptoms of an acute opioid overdose. It is administered routinely by clinicians in emergency departments and in the prehospital setting by ambulance paramedics and emergency medical technicians Since the mid 1990s, interest has grown in providing THN to opioid users, families and carers as an intervention to reduce drug related deaths. In 2009, the Welsh Assembly Government introduced a national THN initiative, following a positive evaluation of a ‘demonstration’ project, where a small number of drug service provider sites in Wales supplied THN training and kits to known opioid users. There is a general consensus amongst professionals who work within the substance misuse community, that those at highest risk of near‐miss or fatal opioid overdose are probably not engaged with traditional treatment services. Indeed, data from the Welsh Regional Confidential Review Panels into drug related deaths (since 2006), have shown that the highest number of deaths occur amongst those that have never accessed or have dropped out of such treatment services. Accessing these individuals requires an innovative approach that takes the intervention direct to the point of acute care. As a spin‐off to the aforementioned work in Wales, a cluster randomised feasibility study was developed to test whether it is feasible for paramedics to provide a brief training intervention and THN kit to patients they have treated and have subsequently recovered from an opioid related overdose; and assess the safety, clinical and cost effectiveness of THN distribution by paramedics. Specific objectives are to:
1. To assess acceptability of and compliance with a new intervention: paramedics supplying THN kits to 999 patients they attend for an opioid overdose 2. To assess the feasibility of paramedic and patient recruitment and follow up in order to determine whether a definitive trial is indicated 3. If so, to provide data to inform trial planning
Our primary outcome is to report on the number of THN kits distributed by trained paramedics during the one‐year study period. Secondary outcomes include establishing the actual/potential number of opioid related contacts during the study period; the processes of care including the number and proportions of patients offered/not offered the intervention; accepted/did not accept the intervention; whether the patient attended the emergency department (ED) or not; subsequent emergency healthcare contacts (999 call or ED attendance or hospital admission) or death for any reason within 12‐months of the index call. We planned to follow‐up of the actual/potential opioid overdose patients anonymously in order to achieve the primary outcome for our target sample size. This cluster randomised study did not seek consent to use routine data for anonymised follow‐up of those patients who declined or were not offered the intervention, because of the significant individual situational and environmental challenges associated with these types of incident. These challenges were discussed and considered by the South East Wales Research Ethics Committee (Panel C), and approval to proceed with this approach was granted.
Antithrombotic Selection in Percutaneous Coronary Daniel Harris, Lead Pharmacist Coronary Artery Disease (CAD) is the leading cause of death in the UK. The care of patients with CAD frequently involves the use of antithrombotic medication (such as aspirin). Atrial fibrillation (AF) is the most 30/07/2015 Approved 12/10/2015 Intervention‐Evaluation of outcomes (A‐SPICE) Cardiac & Stroke. Honorary Research common type of abnormal heart rhythm. The management of AF usually requires oral anticoagulants such as warfarin to reduce the risk of stroke. The risk factors for CAD closely mirror those for Atrial Fibrillation (AF) Fellow, ABMU LHB and these diagnoses often coexist. In patients with CAD complicated by AF it is sometimes necessary to use combinations of aspirin and warfarin to reduce the risk of stroke from AF and blockage of the coronary arteries from CAD. However, this approach is associated with an increased risk of severe bleeding.
The approach to managing these patients is poorly documented hence we aim to study the management of patients with CAD with particular attention to those with coexisting AF, record their medical management, presence of co‐morbidities and identify key features that may lead to poor outcomes in this patient group.
Examining the link between family health events and Philip Murphy, Professor of 29/07/2015 Approved 28/09/2015 pupil performance indicators in Wales. Economics, School of Management, School outcomes are known to play a significant part in determining not only the range of opportunities available to individuals in their own lives but also the future lives of their children. As a result there has been Swansea University much interest, and research undertaken, to identify the key factors determining educational attainment. This proposal will add to this body of knowledge by looking at a relatively under researched area in terms of t UK evidence base: namely untangling the nature of the complex relationship that exists between major episodes of parental and pupil ill‐health (including mental health issues), the timing of these events, and their
effect on subsequent levels of pupil attainment. The analysis will focus initially on outcomes within Wales and use data from the Annual School Census and Pupil Attainment records, which can be linked to administrative Welsh Health Records currently held in Swansea under the Secure Anonymised Information System (SAIL). Whether such events have temporary or persistent effects has important implications for the
design of appropriate policy interventions that seek to permit all children to realise their true potential and so contribute most effectively to the creation of a vibrant economy. This broadening and deepening of the skill base is a goal shared by the Welsh Government as outlined in the policy document Economic Renewal: a New Direction (2010).
Participation in post compulsory education training in Rhys Davies, Research 29/07/2015 Approved 28/09/2015 Wales Fellow/Associate Director, This research project aims to look at how pupils in Wales progress from secondary school, into sixth forms and further education colleges (post‐16 education). Specifically, it seeks to identify different learning
Administrative Data Research Centre pathways into post compulsory education and to look at factors which affect vocational learning over academic qualifications. Wales (ADRC‐W) By linking primary, secondary and further education data together, it is possible to trace individuals through the education system to entry to post‐compulsory education. Within Wales, such a database will allow us
to investigate the relative impacts of the social characteristics of individuals, their previous educational attainment and their progression through the education system, in order to create a classification of types of learning pathways. In doing so, it is hoped that distinct pathways into and through post‐compulsory education will result. Secondly, the impact of various characteristics on entry into vocational or ‘academic’ learning in post‐compulsory education will be analysed – for example, the impact of absenteeism and attendance on different types of education. Health Wise Wales Dr Shantini Paranjothy, Clinical Senior HealthWise Wales is a national initiative to engage large numbers of people in Wales with health and social care research. The aim is to improve the health and wellbeing of the Welsh population, by putting 29/07/2015 Approved 29/09/2015 Lecturer, Cardiff University people at the heart of knowledge creation and translating new knowledge into policy and practice.
HealthWise Wales will develop a dedicated research programme focusing on population health and wellbeing. This large cohort study based on the Welsh population will enable the investigation of the widest possible range of social and environmental determinants of health and social outcomes during the life‐course. These data can then be used to inform the development of interventions to reduce health inequalities and improve health and wellbeing of future generations. Healthwise Wales will engage the population of Wales to become actively involved in research to improve health and wellbeing, and provide a platform for research, policy and service development and evaluation. Healthwise Wales offers an open invitation to anyone aged 16 years or above who is living in Wales or receiving healthcare in Wales to take part. Participation involves completing web‐based questionnaires every 6 months. Consent is obtained to contact participants every 6 months and to access their routinely collected healthcare data and NHS records for population health research studies.
Investigating the prodrome of type 1 diabetes in Julia Townson, Research Fellow, Investigating the prodrome of type 1 diabetes in childhood as it presents to Primary Care to predict earlier diagnosis and reduce ketoacidosis at presentation, using pseudoanonymised linked Primary and Secondary 23/07/2015 Approved 28/09/2015 childhood as it presents to Primary Care to predict SEWTU, Cardiff University Care data. earlier diagnosis and reduce ketoacidosis at presentation, using pseudoanonymised linked Primary and Secondary Care data.
Epidemiology of dental procedures carried out under Hywel Jones, Research Associate Admission of children to hospital for essentially preventable conditions such as dental caries constitute a public health issue of concern. The project is designed to quantify the extent of the problem in Wales. It 23/07/2015 Approved 28/09/2015 general anaesthetic in children (Statistician), Cardiff University mirrors to some extent an earlier study undertaken in England but seeks to build on it to look in more detail at the children’s socio‐demographic and medical background. The project aims: i. to describe the characteristics of children hospitalised with a dental condition; ii. to see how they may have changed over time and how they relate with deprivation; iii. to examine the use of general anaesthetics in their treatment; and iv. to describe the previous encounters of those children with GPs, compared with the encounters of children not subsequently hospitalised with a dental condition; v. to model the time to hospitalisation, adjusting for perinatal characteristics and area level deprivation.
Assessing the health impacts of adults’ participation Dora Pouliou, Research Officer, 23/07/2015 Approved 28/09/2015 in sports in Wales: investigating the mediating role of Administrative Data Research Centre accessibility to sports facilities. Wales, Swansea University The “Programme for Government” is the Welsh Government’s plan of action. One of its commitments is to widen access and participation in sports programmes and facilities in order to increase physical activity which in turn improves our health and well‐being. In particular, research has found that individuals participating in sports are more likely to engage in health enhancing behaviours (e.g., non‐smoking and healthy diet), and less likely to engage in antisocial and violent behaviours , . Previous studies have also highlighted the positive influence that access to sport facilities may have on sport participation.
While a positive association between participation in sport and self‐assessed health and life‐satisfaction has been reported for England and Scotland , research on the health impacts of sport participation has been limited in Wales. Sport Wales, the national organisation responsible for developing and promoting sport in Wales, have collected data on sport participation and published relevant statistics , however these have not been linked with any health outcomes and/or health behaviours.
Data are now available to statistically explore the relationship between taking part in sport, accessing sporting facilities and key measures such as quality of life (self‐reported health and well‐being), health
outcomes (from routine health records e.g. sport‐related injuries, co‐morbidity index, number of hospital admissions) and health‐related behaviours (both self‐reported and routinely‐collected smoking, alcohol consumption, diet). This project provides an opportunity to explore the benefits and usefulness of analysing linked administrative and survey data and will therefore help us better understand the relationship and
pathway between the accessibility of sport facilities, sport participation and health for adults in Wales as well as whether accessibility to sports facilities contributes to inequities in health outcomes.
Emergency Medical Retrieval and Transport Service Professor Ronan Lyons, Professor of 22/07/2015 Approved 28/09/2015 (EMRTS) evaluation Public Health, Swansea University Evaluation of the Emergency Medical Retrieval and Transport Service (EMRTS) Cymru. This is a new service for seriously injured and ill people across Wales whereby rapid help will be provided by specialist doctors transported by helicopters and ambulances. The intention is to improve the survival of seriously injured or very ill people, including those suffering cardiac arrest (heart stopped). The service started in July 2015 and
will reach an estimated 300 people every year. In some parts of the world, but not everywhere, similar services have led to improved survival. This proposed evaluation would bring together complex data on around 600 patients treated by the service over two years and compare outcomes with a group of similar people before the service was introduced or when it is not operational, such as at nightime or during bad weather when the helicopters cannot fly. The data would come from datasets already in SAIL such as emergency department, inpatient and GP datasets, new data collected by the EMRTS service and the Welsh Ambulance
Service Trust, and data already collected by hospitals as part of national audits into the quality of care by intensive care and trauma services. The evaluation would compare survival, quality of life and functional abilities (ability to care for one self) in groups receiving or not receiving the service. This evaluation is intended to inform the further development of the service. The evaluation team, currently comprised of experts trauma, critical care and evaluation, will recruit two members of the public to join them in further designing the evaluation, its conduct and dissemination.
Cardiovascular risk, alcohol‐related harm and mental David Fone, Professor of Health 21/07/2015 Approved 28/09/2015 health: total population pilot electronic record‐linkage Sciences Research, Cardiff University We are investigating the linkages between three of the greatest threats to the health of the adult Welsh population, namely coronary heart disease, excess alcohol and poor mental health. We have so far cohort study investigated these three health outcomes in separate projects in the SAIL databank. We have now planned a new project to link these anonymised studies together and investigate further the links between heart
disease, the adverse effects of excessive alcohol consumption and poor mental health. We will assess whether people with a high risk for heart disease have a higher risk of an hospital admission caused by alcohol and the role of poor mental health in this process. This will give use preliminary data on which to prepare a full application to a major UK research funding body. Wales Electronic Cohort for Children (WECC): Phase 3 Professor Ronan Lyons, Professor of The Wales Electronic Cohort for Children (WECC) has been in operation since 2007. A cohort study is a longitudinal study (a study that follows individuals over time) of a group of people with a common characteristic 17/07/2015 Approved 24/07/2015 Public Health, Centre for Inovative WECC involves creating a cohort by linking anonymised data from multiple datasets on all children living in Wales born since 1990. The datasets include births, deaths, hospital admissions, outpatient attendances, Swansea University emergency department attendances, congenital anomalies (malformations at birth), cancer, national community dataset (e.g, immunisations and height and weight), school attainment (Key Stage 1‐5 results from school entry to A levels), and with links to data on children participating in a number of child cohorts, such as the Welsh Study of Mothers and Babies (WOMBS), Growing Up in Wales, the Brecon Cohort (children with Type 1 Diabetes) and the Millenium Cohort Study (MCS). Analyses carried out in the first two phases of WECC demonstrated the impact of being born small or prematurely on a range of health indicators and the impact of some medical conditions, such as brain injuries or epilepsy, on assessments of educational attainment at age seven. The plan for the third phase of WECC is to include data on children born up until 2020 and to continue the research on the impact of different health conditions occuring during development on subsequent measurese of health and educational attainment. Educational attainment is a key factor that influences health and wellbeing through out one's life. The purpose of these analyses is to identify factors that hinder or help protect children from ill health, injury or not doing well at school that would inform the development of policies and interventions. WECC is also designed to support the evaluation of existing or new activities, services, policies and practices that are designed to improve the health and wellbeing of children. These could be aimed at pregnant women, such as the BumpStart project in Cwm Taf Health Board that aims to reduce excess weight gain in pregnancy, children themselves e.g. physical activity interventions, households e.g. improvement in energy efficiency, or based on areas, e.g. Flying Start. Phase 3 of WECC will cover many diseases and disorders, including asthma, cancer, congenital anomalies, dental health, diabetes, infection, hearing and vision problems, and injury as well as broader concepts of health, wellbeing and educational attainment. This phase of WECC builds upon the earlier developmental phases and is being undertaken by the National Centre for Population Health and Wellbeing Research, a group that includes public health and child health researchers from Bangor, Cardiff and Swansea universities, Public Health Wales NHS Trust and Children in Wales, the leading 3rd sector organisation in this field.
Examining contributory factors in road traffic collision Charles Musselwhite, Associate Older people, especially those aged over 75 years, are more likely than middle aged people (aged 25‐65 for example) to be involved in road traffic collisions as a driver and as a pedestrian. Older people have 17/07/2015 Approved 18/08/2015 data involving older people in Wales Professor in Gerontology, Centre of different types of collisions than younger people including issues involving: Innovative Ageing Swansea • Junctions; University • merging traffic; • right‐hand turns (when driving on the right‐hand side of the road) and; • busy traffic (see Clarke et al., 2009 for review)
This project will add to this knowledge‐base by providing additional data from STATS19 (the form filled in by police at the scene of a road traffic collision when someone has been injured; see http://www.adls.ac.uk/department‐for‐transport/stats19‐road‐accident‐dataset/?detail).
Using Wales as a case study, this project will examine the STATS19 data for people aged over 65 involved in collisions, looking at contributing factors such as location, weather, time of day etc. Comparisons will be made with people aged under 65 to identify any salient factors for the older group. Information relating to the collision could be used to flag up at risk individuals who may need additional training or help with driving cessation.
The project will also make comparisons between STATS19 and emergency department records to investigate correlations in reporting of road traffic collisions.
Living Well Living Longer Evaluation William King, Registrar in Public The Living Well Living Longer programme in Gwent aims to find individuals in more deprived areas who are at high risk of developing cardiovascular diseases, such as heart attack and stroke, so that they can be 10/07/2015 Approved 03/08/2015 Health Medicine and WCAT Clinical offered the best available treatments for preventing these diseases. Fellow in Public Health, Public Health Wales/Cardiff University As part of this program, it is important that an assessment is made of whether as a result of this process individuals get those treatments appropriately and whether the risk of cardiovascular disease is reduced for
people who have been through the programme. In order to make this assessment, we would like to use the SAIL databank to compare the levels of treatment, disease and death in those who have been through the programme compared to those in a comparison area in which the programme did not take place.
Welsh Government, NHS Wales and Novartis Bob Hudson, Director, NHS Wales The Welsh government, NHS Wales and Novartis Pharmaceuticals have engaged in a broad partnership with the aim of developing programmes that will improve patient health and the performance of the Welsh 02/07/2015 Withdrawn/Not Approved N/A Partnership Project to Improve Patient Outcomes in Health Collaborative healthcare systems in two disease areas: heart failure and lung cancer (with the potential for eye care and advanced breast cancer to follow). Wales In order to understand the areas of greatest need within each of the disease areas, both expert opinion as well as hard data from within the NHS is crucial. The project team will analyse hospital, GP and diagnostic
data to better understand the factors affecting the entire patient pathway in order to identify areas for improvement and design interventions that will ultimately result in improvements in patient benefits. Examples of issues that the interventions could address include time to presentation, referral and diagnosis, length of hospital stay and in‐patient mortality. The current data request to SAIL is to facilitate the diagnostic and desig phase (Phase 2) of the project as described above.
The next part of the project (Phase 3) which is not covered by this application will involve implementing the selected proposals and prospectively evaluating impact on the selected outcomes through data analysis.
This will be covered in a separate and subsequent application and is not discussed in any further detail.
Feasibility study exploring the use of data linkage in Laura McGinn, Research Officer, The proposed work here is a feasibility project aimed at exploring the possibility of linking Local Authority Supporting People (SP) administrative data to health (GP, A&E, outpatient and hospital admission) data in the 16/06/2015 Approved 24/07/2015 the evaluation of the Supporting People Programme Swansea University SAIL databank. We will use the standard Secure Anonymised Information Linkage (SAIL) system (Swansea University) which will allow linking of the datasets in such a way that individuals remain anonymous.
The feasibility stage will look at the range and the quality of the available SP data and to see how many SP service users match with health records already held in the SAIL databank. If possible we will also investigate patterns of A&E/hospital/GP attendances for SP service users before and after service use.
The feasibility project will also report on the kinds of analysis likely to be possible in future, including whether a control group can be created and whether it is likely to be feasible to estimate the amount the SP Programme saves the NHS. Well‐being and Health Enhancement for Arthritis and Professor Sinead Brophy, Swansea Musculoskeletal diseases are the most common cause of disability in the UK. Working within the National Centre for Population Health Research (NCPHR), the Welsh Arthritis Research Network (WARN) aims to use 08/06/2015 Approved 22/07/2015 Rheumatism (WHEAR) University routinely collected data, such as general practitioner and hospital records to fully assess arthritic conditions, such as Rheumatoid Arthritis. The impact on patients, healthcare providers and society will be assessed. The project will also identify conditions that may affect patients alongside arthritis and factors that may increase the risk of developing arthritis (such as hereditary factors) as well as the efficacy and potential side‐ effects associated with treatments of the conditions. Greater understanding of these issues will improve the measurement of the burden of arthritis and will allow us to identify ways in which we can improve the outcomes for individuals living with arthritis.
Treatment choices for managing glucose control in James Chess, Honorary Clinical Rationale: 01/06/2015 Approved 24/07/2015 impaired renal function Senior Lecturer and Consultant Patients with chronic kidney disease and diabetes mellitus have a higher risk of heart disease. Treatment therapies for diabetes targeting glucose control are limited. The treatment choices are primarily insulin and Physician, Swansea sulphonylurea which may cause hypoglycaemia, especially in patients with kidney impairment. Metformin is an oral drug typically used for glycaemic control to manage type 2 diabetes. Current recommendations University/ABMU Health Board suggest that metformin and exenatide should be avoided in patients with elevated creatinine levels or a poor estimated glomerular filtration rate, which indicate impaired kidney function. A recent study conducted for Hywel Dda and Abertawe Bro Morgannwg Diabetes‐Renal Group examined the treatment choices for glycaemic control in a sample of secondary care hospital patients with moderate or severe kidney disease. Estimated glomerular filtration rate was found to be significantly higher in the metformin group, indicating better kidney function. No significant differences were observed for cholesterol, LDL‐ cholesterol, HDL‐cholesterol, triglycerides, systolic blood pressure, albumin‐creatinine ratio, body mass index and the use of angiotensin converting enzyme inhibitors, statins and aspirin between the two groups (metformin / other therapies). Metformin is a cost‐effective and safe oral treatment for type 2 diabetes. Further reporting and publication is required on its use in patients with moderate or severe renal impairment to determine its safety. It is suspected that the use of metformin in patients with renal impairment is not uncommon.
Aims and objectives: • To examine the relationship between diabetes therapies and chronic kidney disease. • To examine blood sugar level control and cardiovascular risk factors in relation to chronic kidney disease. • To examine the use of Aspirin, statins and angiotensin converting enzyme inhibitors in those with chronic kidney disease. • To examine the prevalence of chronic heart / cardiovascular disease in those with chronic kidney disease. The project hopes to influence further practice and guidance for patients with type 2 diabetes and renal impairment.
Method: The project hopes to search the SAIL databank for patients with type 2 diabetes, using GP records, outpatient records and Abertawe Bro Morgannwg Health Board pathology datasets. Information included in the search would include:‐ creatinine, estimated glomerular filtration rate, age, gender, weight, height, cholesterol types, blood pressure types, diabetes therapies, coronary heart disease status, cardiovascular disease status, duration of diabetes. The search will be limited to adults having type 2 diabetes mellitus, and will exclude those with type 1 diabetes mellitus.
Results: Data will be exported into excel and transferred to SPSS for statistical analysis. The results will be disseminated locally (Local diabetes research forum, Diabetes‐Renal meeting, Welsh Endocrine Diabetes Society) and nationally (Diabetes UK Annual Professional Conference). A manuscript will The use of antipsychotics for people with learning Professor Michael Kerr, Cardiff This study examines the characteristics of people with learning difficulties (mild and severe) who are prescribed antipsychotics. This will be performed among children (under 18) and adults. The impact will be 27/05/2015 Approved 23/08/2015 difficulties University determined by comparing the people with antipsychotics and learning difficulties (severe and mild) to people on antipsychotics and to people with learning difficulties (severe and mild). This comparison will allow the impact of antipsychotics on the lives of people with learning difficulties to be determined.
In addition, the study will use new statistical analysis methods (data mining) to examine clusters of diagnosis and symptoms that are associated with the use of antipsychotics. For example, antipsychotics may be
associated with higher numbers with diabetes but the non‐use of antipsychotic may be associated with higher likelihood of injury.
Childhood Obesity and Neighbourhood Environments: Sarah Rodgers, Associate Professor in This pilot project will investigate the opportunity to combine the strengths of the Medical & Environmental Data Mash‐up Infrastructure project (MEDMI) and SAIL (The Secure Anonymised Information Linkage Datab 22/05/2015 Approved 22/07/2015 Integrating MEDMI and SAIL Epidemiology, Swansea University at Swansea University) to link environmental, health and socio‐economic data. Specifically, the project will study the association between neighbourhood characteristics by measuring distances in a Geographic Information System (GIS) to greenspace such as public parks, woodlands and the coast and beaches. These distances will then be linked in SAIL to childhood Body Mass Index (BMI) to see if there is a link between proximity to opportunities to exercise and BMI. Using SAIL Data for Cedar Evaluations for NICE Dr Grace Carolan‐ Rees, Director, 18/05/2015 Approved 27/07/2015 Cedar, Cardiff & Vale University NICE provides guidelines to the NHS about which medicines, procedures, medical devices, etc. are the best to use. To contribute to an evidence base upon which recommendations are made, NICE commissions Health Board work from External Assessment Centres to evaluate existing evidence or to conduct new research. Cedar conducts these evaluations and would like to use SAIL as a source of data to provide evidence. SAIL contains a large amount of information about many procedures and devices that are used, as well as a wealth of detailed information about individual's health and care, making it an ideal resource for this purpose.
Even when SAIL does not contain information about the device or technology in question, it may be possible to import data about it and then use SAIL's health and care data to measure the effectiveness. Alcohol misuse in the militaryAlcohol misuse in the Dr Laura Goodwin, Lecturer in There are a number of factors about the UK military which could put them at a greater risk of physical health problems and mortality than the general population. There are very high levels of alcohol use and drinking 01/05/2015 Approved 21/07/2015 military: Should veterans' healthcare services focus Epidemiology, King's Centre for is an accepted part of military culture. The King’s Centre for Military Health Research (KCMHR) cohort study was set up to look at a representative sample of UK military personnel, and collected data on mental healt on the longer‐term physical, rather than mental, Military Health Research, King's alcohol use and deployment experiences. The study found that 13% of the military are abusing alcohol at a level which is hazardous to their health and overall, levels of alcohol use are much higher than the general consequences of deployment? College London population at all ages. Research findings from the general population have shown that both alcohol use and mental health problems are associated with poorer physical health outcomes. Therefore, it is predicted th current and future healthcare use for physical conditions will be high in the UK military population. Identifying the future healthcare needs of military personnel and veterans is a current priority for the UK government. However, the current datasets that are available to look at use of healthcare services (and the reasons why) do not identify whether or not an individual is in the military. The proposed study will be able to identify military personnel in existing NHS datasets for England, Wales and Scotland that include secondary healthcare records, by linking to the KCMHR cohort data. In NHS hospitals, data is recorded each time someone is admitted to hospital as an inpatient, or has an outpatient visit or visits an accident and emergency department. The reason for the hospital visit is coded using an existing international classification of disease and health related problems (ICD‐10). This data is collated from all hospitals and is called Hospital Episode Statistics (HES) in England. In this study, English, Scottish and Welsh data will be linked with the KCMHR cohort study, providing information on approximately 10,000 military personnel. This will offer a unique opportunity to study the health of UK military personnel and will provide data on the health problems for which they most commonly seek help. We will be able to identify whether physical health conditions or accidents and injuries are most common. A further aim for the study will be to look at whether we can predict if an individual is likely to be admitted to, or visit hospital, for a physical health condition or an accident or injury, by using the information from the KCMHR cohort about how much alcohol they drink and whether they have a mental health problem. We will also be able to look at the costs to the NHS resulting from high levels of alcohol use and mental health problems within the military. It is a high social, health and political priority to identify the future health care needs of the UK military (serving and ex serving), so that the NHS commissioning board have the data they need in order to plan specialised commissioning of health care services.
Recurrent admissions in children with diabetes Diana Wasag, FY1 Academic Doctor, 01/04/2015 Approved 24/06/2015 The aim of the project is to identify children with diabetes who require recurrent hospitalisation. Our goal is to identify possible risk factors for recurrent hospital admissions, such as the impact of living with a single Wrexham Academic Unit /Cardiff parent, having a parent with diabetes and coming from low income level family. Additionally, we aim to look at the impact of changing the healthcare provider from paediatric to adult and see if this change affects th University number of admissions to hospital. Moreover, we are going to look into the length of stay in the hospital required in order to predict if there are any risk factors for long hospital stay.
Electronic Longitudinal Alcohol Study in Communities Simon Moore, Professor of Applied The overall goals of this project are to: 17/03/2015 Approved 30/03/2015 (ELAStiC) Clinical Research & Public Health Cardiff University 1) understand how people come to misuse alcohol and suffer the harms from alcohol misuse by using information from different sources;
2) develop ways to analyse the information by making the most of the skills in the research group and recent research in this area;
3) make the research relevant to policy makers and practitioners by looking at the specific areas identified by the third sector agencies and other advisory bodies through the project’s previous consultation (for example, the secondary impacts of parents' alcohol use on children); and
4) deliver meaningful, accessible and timely advice to practitioners and policy makers via a Study Steering Committee with representation from the third sector, NGOs and other advisory boards and through an on line learning community that allows policy makers, practitioners and the third sector to access the project’s outputs.
Controlled Antenatal Thyroid Study – Obstetric Dr Peter Taylor, Welsh Clinical Thyroid hormones are essential for health and are particularly important in pregnancy. It is well established that profound abnormalities in thyroid hormone levels (low or high) result in worse outcomes for both mo 12/03/2015 Approved 28/04/2015 Outcomes Academic Trainee in Diabetes and and child in pregnancy. These can include increased risk of miscarriage, increased risk of requiring a Caesarean section, higher likelihood of needing special care as a baby and potentially lower IQ in childhood. Endocrinology, Cardiff University Fortunately thyroid hormones are easy to measure and treatments for thyroid disease, particularly in individuals with low thyroid hormone levels are safe and inexpensive.
What is less clear is when to treat individuals, as there is growing evidence that we should be treating more women with more borderline thyroid function. To use a more familiar example thyroid hormones are
measured on a continuous scale like blood pressure and cholesterol, we change at what level to treat blood pressure and cholesterol depending on special circumstances such as diabetes or previous
cardiovascular disease. Thyroid hormones can be thought of in a similar fashion, with pregnancy as a special circumstance, however unlike blood pressure or cholesterol we do not have the data to say at what precise level thyroid hormones should be treated in pregnancy. Studies in pregnancy are notoriously difficult to conduct, as many women are concerned about taking any medication in pregnancy, furthermore although thyroid hormones levels are not routinely measured in pregnancy, many clinicians would treat patients if any borderline function was identified, making a clinical trial difficult to conduct. There is a pressing
need for a study in this area as there is considerable debate as to whether there should be universal thyroid screening in pregnancy. This has major implications for both UK and global management of pregnant
women.
Fortunately several years ago there was a large thyroid study undertaken in South Wales – the controlled antenatal thyroid study. This study measured the thyroid function of over 14,000 women pregnant women in the first trimester and treated half of women found to have lower thyroid hormone levels. This study however only assessed the IQ of the offspring in women who had lower thyroid hormone levels (around 700) therefore the majority of women were not followed up. Crucially it did not assess pregnancy outcomes (such as stillbirth, premature delivery, need for caesarean section).
Evaluation of risk factors associated with the Lavinia Margarit, Consultant This will be a study using information provided by The Secure Anonymised Information Linkage (SAIL) Databank developed at Swansea University. The study will estimate the main differences of characteristics as 12/03/2015 Approved 15/05/2015 development of endometrial cancer in patients taking Obstetrics ‐ Gynaecology, Obstetrics age, BMI, ethnicity, social status, previous medical history, between patients that develop endometrial cancer (EC) and those that do not develop EC. In previous research we have found a difference in expression of Tamoxifen. and Gynaecology Department, substances called biomarkers in the blood and endometrium from patients with EC compared to controls. The treatment with Tamoxifen given for breast cancer is considered to be linked to the development of EC. Princess of Wales Hospital, ABMU From existing research we are unsure of the prognostic of EC in patients that have been taking Tamoxifen. Increased surveillance in Tamoxifen users might lead to early diagnosis of EC in such women. This study will try to identify on a large anonymous cohort the potential risk factors associated with the development of EC in Tamoxifen users. Visualising the epidemiology of Parkinson's disease Dr John V Hindle, Associate Clinical 09/03/2015 Approved 28/04/2015 in Wales (VisEpPD). Director, NWORTH, Honorary Senior Lecturer and Consultant Physician, The intention of this project is to create a visual representation of Parkinson's disease occurrence in Wales. Using a mapping tool such as leafletJS http://leafletjs.com/ (accessed 27th November 2014) we will North Wales Organisation for 'layer' data pertaining to the residence of individuals diagnosed with Parkinson's disease in Wales and demographic information about those with the diasease, such as age, age at diagnosis and where possible Randomised Trials in Health and occupational background. This information will enable us to visually demonstrate where in terms of geographical areas, there appear to be more people living with Parkinson's and whether more people of certain Social Care social groups (as derived from the Welsh Demographic Service). Visualising Parkinson's incidence on a map of Wales will also us to identify whether specialist services for people with Parkinson's are best located for purposes of accessibility and whether any denser areas of disease occurrence do or do not exist. Validating Epilepsy and Neurological Case Definitions Dr Owen Pickrell, Clinical Lecturer Our epilepsy research in SAIL to date relies on using diagnosis codes from general practitioners (GPs) to identify people with epilepsy. This method has not been extensively validated and there are concerns that 17/02/2015 Approved 24/06/2015 within SAIL (Neurology), College of Medicine, this method of identifying people with epilepsy is not sufficiently accurate. Swansea University We will make sure that our method for identifying people with epilepsy in SAIL is accurate by carefully comparing GP diagnosis codes with definite diagnosis from hospital specialists. We will do this by importing and anonymously linking within SAIL lists of around 100 people definitely with and without diagnosis of epilepsy. The lists will be obtained from hospital registries with the appropriate consultants permission and will remain within the NHS until appropriately anonymised using the usual SAIL methods. If successful, we would like to extend this method for other neurological disorders such as Parkinson’s disease.
Osteoporosis and fracture risk: are we achieving Professor Ronan Lyons, Professor of Fractures in older people are generally caused by loss of bone strength ‐ a process called osteoporosis. These fractures are very common and lead to much pain and suffering, disability, loss of independence, and 06/02/2015 Approved 24/03/2015 secondary prevention? Public Health, Swansea University death (especially hip fractures). Caring for such people also utilises a lot of NHS resources. On an average day patients with hip fractures occupy about 400 hospital beds in Wales, with surges in cases contributing to cancelled elective procedures. The proposed study aims to create an anonymised dataset of people in Wales with a presumed osteoporotic fracture (based on age). Then, to use this to quantify the incidence and distribution of cases in Wales, measure trends in the uptake of interventions designed to prevent second fractures and assess the potential to do more to reduce the scale of this problem and its impact on individuals, the NHS and wider society. The relationship between Adverse Childhood Mark Bellis, Director of Policy, This study will examine the relationship between exposure to adverse childhood experiences (ACEs; including parental mental health issues; parental alcohol use; one‐parent household; childhood victimization; 27/01/2015 Approved 24/03/2015 Experiences and health and educational outcomes Research & Development, Public being born to a teenage mother) and healthcare utilization and educational outcomes. We propose to use the Wales Electronic Cohort for Children (WECC) for this analysis. Previous research has demonstrated a link Health Wales between exposure to adverse experiences in childhood and negative health, social and educational outcomes in adulthood. Although this link is well established, what is less well known is the impact of exposure to adverse experiences on childhood outcomes. In addition, the impact of ACEs has traditionally been examined using retrospective self‐reported surveys. The present study aims to outline the epidemiology of ACE using routinely collected data. We will also examine the influence of age of ACE exposure on outcomes, as this has not been previously reported. For a detailed description of background and methodology, please see the attached project outline. Creating and utilising a Wales Asthma Observatory to Dr Gwyneth Davies, Respiratory Studies on asthma frequency and associated impact and costs have tended to provide high‐level national snap‐shot pictures. Whilst these have been useful, we need to understand how these considerations are 14/01/2015 Approved 10/04/2015 support health policy, health service planning and Physician & Clinical Associate changing, both at a national and a small‐area level (e.g. Health Board), to inform national/local health policy and planning. clinical research Professor, Swansea University Traditionally, health data has been managed by computer experts. However we now need doctors trained in informatics to supply useful asthma data. Our clinically‐trained doctoral student will have access to a large amount of recent anonymised asthma data (e.g. GP visits, admissions, medications). A large dataset of asthma patients will form an observatory which can be regularly updated. This will initially be developed in Wales and could then be applied UK‐wide. It will allow important questions on asthma to be answered e.g. how common/severe is asthma in different areas, and the impact (e.g. hospital admissions, deaths) and costs of asthma.
The student will identify gaps in information and come up with systems to fill those gaps. These are likely to include test results, which will facilitate the identification of asthma patients to take part in important new treatment trials.
Understanding the impact of asthma on individuals and associated costs is important in planning effective treatments and highlighting the resource needs of patients. A state of the nation asthma observatory will
provide crucial up‐to‐date monitoring of key asthma outcomes and costs. This will allow improvements in asthma care and services and provide a platform for testing new strategies and treatments in asthma.
School Health Research Network Scoping and Simon Murphy, Co‐Director, DECIPHer,This study examines whether early stage education progress correlate with teenage health problems. In particular we will examine the following cases: those achieving neither key stage 1 or key stage 2, those 22/12/2014 Approved 16/02/2015 Feasibility Study Cardiff University achieving Key stage 1 but not key stage 2, those achieving key Stage 2 but not key stage 1, those achieving both key stage 1 and key stage 2. The measure of poor health in the teenage years will be self reported smoking, self reported drinking, self reported general health. All these health measures will be collected from the Health Behaviour in School Aged Children Survey. The findings from the work will complement the
findings from a larger study examining hospital admissions for teenagers compared with their primary school education.
Health IRIS; how can we improve the health and well Chris Poole, Clinical Science Liaison, 22/12/2014 Approved 15/02/2015 being of our population through the better use of Digital Health Labs Limited The Abertawe Bro Morgannwg University (ABMU) Health Board has commissioned Digital Health Labs, a Welsh software company, to develop flexible web‐based tools allowing rapid processing of health data held population health information? in the SAIL Databank by non‐programmers. In addition to giving a general overview of health in Swansea, Neath, Port Talbot and Bridgend, these tools will enable ABMU health researchers to select any group of
patients and gain a detailed understanding of the different ways in which their health is managed. This will help health managers to understand not only the best ways to treat and manage illness but also to understand factors that may predispose certain people to disease, providing targeted opportunities for keeping those people well.
Feasbility asssessment data for CIPHER Dr Shantini Paranjothy, Clinical Senior Farr Institute CIPHER is studying various aspects of infection in order to improve the health of people (e.g. looking at whether efforts to reduce the use of antibiotics has had any effect on the how resistant various 17/12/2014 Approved 15/02/2015 Lecturer, Institute of Primary Care & diseases are to these drugs). Public Health, Cardiff University Researchers from Farr Institute CIPHER would like to use certain SAIL data to work out whether SAIL contains enough microbiology and other data to carry out this work. This IGRP application is to ask for permission
to access specific datasets so that they can do this. Investigating the impact of modelled school travel Dr Sarah Rodgers, Associate Professor 03/12/2014 Approved 25/02/2015 routes on child health using GIS and routine linked in Spatial Epidemiology, Swansea data ‐ Is the relative density of fast food outlets near toUniversity The easy availability of “fast food” from unhealthy food outlets is thought to be linked to obesity rates. This study will investigate whether the high number of unhealthy food outlets on home‐to‐school routes has an schools affecting childhood obesity? impact on childhood obesity. Data on the location of food outlets will be collected and grouped around schools. This dataset will be brought into the SAIL databank using standard procedures and linked to demographic and health data. Using children’s Body Mass Index data, this project will investigate the link between an unhealthy food environment and children who are overweight and obese. Creation of an ICD10 Trauma Mortality Prediction ModelProfessor Ronan Lyons, Professor of The aim of the project is to develop a statistical model that accurately predicts mortality following injury to be used with routinely collected hospital inpatient data. The model would need to encompass every 03/12/2014 Approved 15/02/2015 Public Health, Swansea University combination of physical injury as most injury events resulting in hospitalisation include multiple injured body parts. The purpose behind the research is to provide statistical models that would support efforts to improve trauma services in Wales, the UK and worldwide, and to help evaluate the impact of prevention at the population level. Similar models used in other countries have led to improved survival. An example of the utility of such a model for evaluation of preventive activities would be its ability to help assess whether new cars, that are designed to be safer, result in less serious injuries to occupants and pedestrians. Such a model could be applied routinely to the data in SAIL and used to help evaluate initiatives aimed at improving safety or care following injury, or be used outside SAIL. Some 45,000 people are admitted to Welsh hospitals every year as a result of trauma and there are about 1100 injury related deaths.
Neurological Needs Assessment for Hywel Dda Victoria Edwards, Primary Care Hywel Dda is undertaking a Population Need Assessment as part of the required response to Together for Health – Delivery of Neurological Services. As part of this we would like to obtain data to have a picture of 10/11/2014 Approved 21/11/2015 Locality Development Manager, the numbers of conditions among the population as recorded by general practitioners within the health board. This will be broken down by age category and county of residence. Only some GPs within the health Hywel Dda Health Board board are signed up to SAIL, so we will assess whether the GPs in SAIL are representative of the health board. We will then estimate the numbers of conditions across the whole health board, adjusting for differences between SAIL GPs and all GPs if necessary. Understanding Suicide Clusters tHrough ExploRIng Dr Ann John, Associate Professor in This study will set out to explore the factors and mechanisms that may trigger and maintain a suicide cluster and cause it to subside. To do this an anonymised e‐cohort of individuals who presented to the Princess 06/11/2014 Approved 21/11/2015 Self Harm Behaviours (CHERISH) Public Mental Health, College of of Wales Emergency Department with self‐harm from December 2006 – February 2009 will be created. A descriptive and time trend analysis will be conducted in which the basic demographic and clinical Medicine, Swansea University characteristics of this cohort will be collected. Additionally an electronic anonymised case‐control study will be conducted. A case will be defined as any individual who presented to Princess of Wales Emergency Department with any mention of self‐harm during the cluster (December 2007 to February 2008). Controls will be non‐cluster related Emergency Department attendees for self‐harm (i.e. separated from the cluster by either distance or time). This will allow for an exploration of the characteristics of the population that presented to Emergency Departments for self‐harm during the cluster. Such characteristics include deprivation, previous help‐seeking behaviour, previous episodes of self‐harm, diagnosis and treatment of mental health disorders and educational attainment.
Improving outcomes for older people with urinary tract Dr Harry Ahmed, NIHR Doctoral 06/11/2014 Approved 21/11/2015 infection in primary care Research Fellow, Cardiff University Urinary tract infection (UTI) is the commonest bacterial infection in older people. Major gaps in the evidence base need to be addressed to facilitate better quality of care for this important condition that affects School of Medicine morbidity, mortality and dignity in older people. We will use anonymised data from general practice to (1) estimate the burden of UTI, (2) investigate the effect of different commonly used antibiotic regimens on patient outcomes, and (3) derive a prognostic score to predict the risk of hospitalisation from severe or complicated UTI. Our findings will inform clinical guidelines and improve patient care. Real‐World Data Collection for Type 2 Diabetes Daniel Thayer, Senior Research Type 2 diabetes mellitus (T2DM) is becoming a very common lifelong condition, occurring in parallel with a rise in obesity. Managing blood sugar and cardiovascular disease (CVD) risk factors in patients with T2DM 15/10/2014 Approved 14/01/2015 Mellitus Analyst, SAIL Databank ‐ Swansea is critical for slowing the progression of the disease as well as preventing long‐term vascular complications. Diabetes can be managed through diet, exercise, and medication. The SAIL Databank will be used to get University an overall picture of the care of patients with T2DM, as well as the effects the disease has on their health.
Specific areas of investigation will include: what NHS services and treatment are being used? What health outcomes do these patients use? And, what are the costs to the NHS? In order to answer these questions, patients with diabetes will be matched to similar patients who don't have the disease. This design allows identifying what effects are attributable specifically to diabetes. Patients will be followed over many years, based on the data available in SAIL. Two groups of patients will be identified: those who have data available immediately from diagnosis with T2DM, and those who may have been diagnosed long in the past, giving a picture at the long‐term results of the disease.
Dual Diagnosis problem in Wales Susie Lucas, GP Engagement Lead, For the purposes of this study the definition of dual diagnosis is defined as follows: 29/09/2014 Approved 08/12/2014 College of Medicine, Swansea “An individual who presents with co‐existing mental health (and/or Personality University Disorder) and substance misuse problems (drugs and/or alcohol).”
The intention is to use SAIL to establish how many people are currently recorded as suffering with a Dual Diagnosis of Mental Health and Substance Misuse issues in Wales. Using Primary Care data will help identify a cohort of patients with Dual Diagnosis. The project expects to encounter issues with the coding of both Mental Health and Substance Misuse. This information will be fed back to the ABM Dual Diagnosis
Working Group to inform plans to implement new services, training and new care pathways for this condition. A report will be sent to the GP practices signed up to SAIL to inform them about their practice population
and coding to hopefully improve coding in this area.
The findings will be displayed at Health Board, Locality and Cluster level. It is important to ensure that there are no small numbers displayed within tabulations as this could potentially disclose identity, so suppression and or aggregation techniques will be employed. Also, it is important to ensure that GP practice identity is carefully anonymised in reports.
Whitehall II Study ‐ Clinical Events Aida Sanchez, Whitehall II Data The Whitehall II Clinical Events project is at the core of the Whitehall II Study. It will involve linking with electronic hospital records from England, Scotland and Wales, in order to obtain objective health outcomes 19/09/2014 Approved 08/12/2014 Manager, Whitehall II Study, that will validate and supplement the self‐reported information provided by the study participants. These clinical data will be suitably anonymised and will be used in the UCL premises by the Whitehall II University College London researchers for statistical analysis to study how midlife factors contribute to adverse ageing outcomes such as chronic disease, depression, cognitive impairment and functional health in later life. They will also assess how the adoption of healthy lifestyle at older ages modifies functional health.
Bridgend Falls Hugo van Woerden, Director of 10/09/2014 Approved 15/11/2014 Innovation and Development, Public The report will consider the factors which influence an individual’s risk of a fall, specifically considering the impact of a pilot intervention designed to reduce the number of falls within the Bridgend area. The
Health Wales intervention is a two stage process. Stage one involves giving the patient advice on preventing falls and an assessment of their individual risk. If patients are deemed at high risk they are then offered further falls
prevention strategies. Rates of events such as hospital admissions and primary care use will be compared in the intervention and control group.Factors such as pre‐existing medical complaints and use of certain medications would be considered as part of the investigation. The analysis will be restricted to Bridgend residents that have received the first stage of a pilot intervention. Cost of reduction in Tonsillectomies to NHS Robert McLeod, Welsh Clinical Stricter guidelines, pressure to save money and limited hospital beds have caused a dramatic reduction in Tonsillectomy over the past 10 years. We believe that this reduction in tonsillectomy may have a hidden 10/09/2014 Approved 03/10/2014 Academic Trainee in ENT, NHS and cost to the NHS and wider economy. We aim to investigate if patients who have had a tonsillectomy attend their GP less than those suffering the same symptoms but have not undergone a tonsillectomy. We also Cardiff University believe that these patient may have greater numbers of antibiotic prescriptions and higher numbers of missed days of work/school. Understanding more about the long term impact of tonsillectomy on GP attendanc will allow more informed decisions regarding the future funding and guidelines for tonsillectomy. Evaluation of the Community Pharmacy Common Yvonne Fullwood, Senior Managing The Choose Pharmacy service aims to make the pharmacy the first port of call for patients seeking advice and treatment of common ailments, thereby reducing the number of patients who visit their GP and other 08/09/2014 Approved 16/10/2014 Ailments Service Consultant, ICF GHK health care services for the treatment of these ailments. The SAIL data will be used to help measure whether there is any change in the number of GP consultations for common ailments following the introduction of Choose Pharmacy in the pathfinder areas. If there are changes in the number of GP consultations, the SAIL data will also help to confirm whether these changes would have happened if the Choose Pharmacy service had not been available to patients.
The project will also identify whether there is any duplication of the services provided ‐ ie do patients visit both the pharmacy and the GP for the same common ailment. The data will also help to understand whether: • Patients are more likely to visit Choose Pharmacy rather than the GP for advice and treatment about a particular group of common ailment; and •Whether patients who use Choose Pharmacy only once visit the GP for advice and treatment for subsequent common ailments patients.
Interactive Health Visualisations ‐ Demonstrator Phase Dr Kelly Knock, Data Scientist/Project 01/09/2014 Approved 21/10/2014 Manager, We Predict Ltd. We Predict use complex big and small data to produce actionable insights. In this project, We Predict will adapt and apply their global expertise in predictive analytics to the health sector to benefit people in ABMU. We Predict are a local company who operate in Swansea.
Objectives: • Use sophisticated data linking and analysis methods on SAIL data to describe, evaluate and predict health and healthcare in ABMU. • Present this intelligence in a user friendly, interactive interface for use by lay people, health managers and clinicians to inform planning and evaluation of services. • Produce tables and views from SAIL for analysts to interrogate. • Adapt and apply We Predict expertise in predictive analytics from other sectors to the health sector.
Demonstrator Phase: Primary and secondary care data relating to four priority areas in ABMU (e.g. Obesity & Diabetes, Musculo‐Skeletal, Atrial Fibrillation, Frailty, Unscheduled Admissions and Mental Health ‐ to be confirmed) will be extracted, cleaned and presented to demonstrate population health, service utilisation and predictions for associated environmental risk factors across ABMU. The output of the project will be a user friendly, interactive data visualisation suite containing maps, infographics and charts for the live reporting of key intelligence associated with the agreed upon priority areas within ABMU.
Clinical Object Library ‐ Demonstrator Phase Huw Morgan, Technical Director, GP Currently to get information from SAIL, analysts have to write queries in a computer language called SQL. Often research or work is repeated for example in every research project about diabetes, the analyst has to 01/09/2014 Approved 31/12/2014 Commissioning Solutions Ltd define what combination of clinical codes are to be used to recognise patients with diabetes, and code these into a complex string of computer programming language called a SQL query. The clinical object library (COL) will be a tool available for use within SAIL just as other tools are currently made available (query writing tools or excel). Using the COL a user will be able to select the codes they need for their research question using a graphical interface, and the tool will then compose the appropriate SQL query to identify these records. This means that in the future research can be done using SAIL data without every project needing to have a SQL programmer within the team. Research queries defined using the COL can be saved and used again, or shared with other researchers.
There will be no requirement upon SAIL researchers to use the COL when querying SAIL data – it will be an addition to the variety of tools already available to use if desired. COL does not give any additional access to SAIL data beyond what a researcher will have already applied for and had approved by the IGRP process. Using COL simply removes the need to write SQL queries. Researchers using COL will devise their own data definitions by selecting the appropriate combination of codes as they currently do, but will be able to complete this process more simply and easily.
What part does inclusion on the Palliative Care Susie Lucas, GP Engagement Lead, The primary care team is central to the delivery of high quality end of life care. Primary care teams manage the worry and stress present at the end of life. Having open and honest conversations, where they are 06/08/2014 Approved 16/09/2014 Register and having an Advanced Care Plan in place College of Medicine, Swansea wanted, is essential to giving patients and their carers the time to adjust and make plans for death. This study is investigating how end of life care plan influence patients' place of death. The interest of the study is play in achieving a good death? University also to look at whether or not being on the palliative care register is an appropriate marker for having an end of life care plan in place.
GP Practice Population Profiling Study in supportof the Susie Lucas, GP Engagement Lead, 06/08/2014 Approved 16/09/2014 GP Cluster Network Development Domain of the GMS College of Medicine, Swansea The purpose of the GP Engagement team is to expand the number of practices who are signed up to SAIL and to retain the current GP practices who are signed up to SAIL. The GP Engagement strategy determines Contract University priorities and activities to realise these key objectives. One of the main requirements of the strategy is to provide an incentive for GPs to sign up to SAIL and as part of this work we need to identify research projects with relevance to Primary Care. The pressure of secondary care and reducing unscheduled admissions leave GPs with big challenges that they find difficult to achieve. Providing a tool that could enable GPs to monitor and investigate the number of admissions in secondary care using data linkage from SAIL has been identified as useful by a number of practices. Improving cluster/locality work between practices is now
part of the QOF QP area of the revised GP contract. There is an opportunity for SAIL to facilitate this cluster work by providing linked data around these areas back to GP practices. This information will help them
make informed decisions about work to be addressed within the cluster and will improve GP engagement with SAIL. Inequality and deprivation are key areas that Health Boards and GP practices want to investigate to be able to plan service delivery more effectively. SAIL is already working with Dr Charlotte Jones, Chair of GPC Wales and Dr Tony Calland, Chairman of the BMA’s Medical Ethics Committee to develop a
methodology that will display the deprivation profile of a GP practice population and compare it with other practices in Wales. It is hoped that the practice population profiling work will enable consistency of activity
across Wales and ensure that obstacles / barriers encountered in developing / delivering these areas are highlighted to Health Boards (and from there to Welsh Government). The SAIL Databank represents a potentially invaluable resource for giving background information to support these priorities, as well as evaluating the effectiveness of interventions. Potential areas of research include: measuring existing patterns
of service provision; profiling populations; comparing the outcomes of different treatment options; and evaluating the effectiveness of interventions. This work will be invaluable as a proof of concept of the potential
use of SAIL data and illustrate the value of consistent high quality data. Results will be shown at an all‐Wales level; in addition, individual Practices who are signed up to SAIL will receive a report showing the
results for their own practice, cluster, and Health Board in comparison to all of Wales (for analyses in which this is possible without small numbers). However, direct comparisons of groups at the same level will not be made. Patient pathways for individuals with reported Dr Melanie Jones, Reader, Swansea 14/07/2014 Approved 16/09/2014 disability. University The aim of this project involving SAIL is to show the unique contribution data‐linking can make to the evidence base. By investigating linked survey data (National Survey for Wales) and administrative data (Welsh Demographic Service (WDS), GP event data, Patient Episode Database for Wales (PEDW), and A and E data) we hope to compare the medical history/pathways between individuals who report disability and those who do not. At the same time we will explore the effects disability status has on employment status within Wales. This projects hopes to increase the understanding of the relationships between disability and employment, to further inform policy makers and to create awareness among researchers of the benefits of using linked data. Development and validation of algorithms for detectio Dr Mark D Atkinson, Research Officer, Data can be collected in many ways in different datasets. Data collected by GPs is entered into a number of different computer systems using codes which can either be typed in or entered from menus. GPs are also 13/06/2014 Approved 26/09/2014 of clinical concepts in health datasets. Swansea University encouraged by the Quality Outcomes Framework (QOF) rules to record certain diagnoses or treatments with particular codes for financial reward. These rules change over time. Other datasets, such as diagnoses recorded and operations carried out in hospitals are entered into the computer system by specialist coders after the events have occurred. These coders work according to guideline which again can change over time and vary regionally.
Numerical values (eg; laboratory tests, weights, heights, numbers of alcohol units) need to be checked for units and for unlikely values. Understanding of the data by specialists is important in interpreting it. We hope to develop sets of rules by which data checks can be made and the maximum use made of the data. In order to test these methods, we need to link to alternative sources of data giving information about the same clinical concepts for the same people.
The evaluation of a new coagulation biomarker on Professor Phillip Adrian Evans, A biomarker of coagulation is a blood test that acts as an indicator of blood clotting in healthy individuals, in subjects with disease and is a measure of response to treatment. The Haemostasis Biomedical 09/06/2014 Approved 15/07/2015 outcomes in acute critical illness Professor of Emergency Medicine Research Unit (HBRU), in The College of Medicine, Swansea University, has developed a new biomarker of blood clotting, which is being tested against currently available biomarkers. We will investigate if this Haemostasis, Swansea University biomarker can detect early changes in clotting in critical illness, including stroke, heart disease, sepsis, and cancer. We will further investigate if this biomarker has greater sensitivity and specifity in the diagnosis of a range of critical illnesses and predict disease outcome. Our biomarker has the potential to replace and improve on existing methods for detecting clot structure that involves complicated preparation and expensive instrumentation. Therefore we require assess to the SAIL database so that we can rigorously evaluate the biomarker against patient outcome data that is available on the SAIL database.
The assessment in the variation of angiography and Hugo van Woerden, Director of To compare the primary and secondary care experiences of patients with Coronary Heart Disease for differences which might be associated with variations in emergency and elective revascularisation and 02/05/2014 Approved 09/07/2014 revascularisation procedures in Wales. Innovation and Development, Public angiographies. The investigation will consider factors influencing the provision of revascularisations and angiographies in a cohort of patients with CHD, recruited from GP data, secondary care data and mortality Health Wales data. It follows on from the 2009 report Equity in the provision of angiography and revascularisation in Wales published by the National Public Health Service for Wales and utilises anonymised but linked primary and secondary care data stored in the Secure Anonymised Individual Linked (SAIL) databank at Swansea University. Can primary care data be used to capture blood Rebecca Woodfiekd, Clinical Raised BP is one of the most important treatable causes of stroke; a major cause of death and disability in the UK. Variation in BP ‐ occasional high readings or fluctuation in readings over weeks, months or years ‐ 02/05/2014 Approved 15/07/2014 pressure variability in a large population‐based Research Fellow, Edinburgh may also cause stroke (even if the average BP is normal). BP variation is not currently treated in the UK, but potential treatments already exist. My research will use data from the SAIL databank, a large population cohort? University database, to explore methods of measuring BP variation in large populations. Developing these methods will be informative for large studies, enabling future investigation of the associations between BP variation an stroke. In the long run, this will guide the development of more personalised treatments for stroke prevention. Self harm electronic cohort to further understand Dr Ann John, Associate Professor in 24/04/2014 Approved 10/06/2014 multiple aetiological factors Public Mental Health, College of Self‐harm is an important public health problem concerning a spectrum of behaviours and intents. People who self‐harm are at an increased risk of suicide in both the short and long‐term and assessing their needs Medicine, Swansea University could have important implications for informing suicide prevention strategies. Individuals who self‐harm are typically managed in primary care or seek no help at all. However there is currently no monitoring of self‐ harm beyond emergency department attendance for planning service provision. This study aims to establish an anonymised electronic cohort of those with a diagnosis of self harm across multiple settings (general
practice, hospital) based on routine electronic records from linked datasets available in SAIL. The establishment of the electronic cohort will allow for the monitoring of trends in self‐harm over time, exploration of health contacts across different settings (primary and secondary care) and planning of appropriate service provision. Fracture in the Elderly Multidisciplinary Rehabilitation Nefyn Williams, Associate Director, Trials need to assess the health economics of new treatments being tested, as well as whether or not the treatment is successful. This includes an assessment of change (increase or decrease) in the use of health 16/04/2014 Approved 15/07/2014 Study (FEMuR) NWORTH services by the participant. Traditionally such information is obtained through interview, even though much of the data has already been provided by the participant and is available electronically. This project tests a proposed methodology to reduce patient interview by obtaining relevant data from SAIL and is intended to reduce the patient burden. As part of the FeMUR trial, routinely collected primary care data related to patient
health service use after a hip fracture will be captured from SAIL and the accuracy and completeness of this data will be compared to that gained by the traditional method of patient questionnaire.
Fractures in people with a diagnosis of epilepsy Owen Pickrell, Clinical Lecturer, 07/04/2014 Approved 25/04/2014 Epilepsy is one of the most common neurological conditions affecting around 1% of the population. Fractures are serious, accounting for a significant proportion of emergency hospital attendences and causing pain Wales Epilepsy Research Network and other complications to the individual as well as necessitating periods of time off work and school. It is thought that people with epilepsy have a higher risk of fractures due to injuries during and after seizures
and side effects of anti‐epileptic drugs. This project aims to record the risk of bone fractures in people with epilepsy when compared to people without epilepsy and identify possible factors involved with fractures e.g. certain types of anti‐epileptic drugs. The results of the project would provide useful information in attempting to reduce the risk of fractures in the future.
Utilisation of evidence‐based secondary prevention Hugo van Woerden, Director of Acute Coronary Syndromes (ACS) are a major cause of morbidity and mortality in Wales. Patients surviving an ACS are at risk of developing a series of complications including heart failure, re‐infarction, stroke and 02/04/2014 Approved 23/04/2014 strategies in patients with Acute Coronary Syndromes Innovation and Development, Public death within 5 years. Evidence‐based, guideline recommended treatment has proven to be highly effective in preventing a secondary event however treatment is often underutilised due to under prescribing and lack
Health Wales of patient adherence to treatment. We will be looking at prescribing and risk factor management following an acute event to see if it follows the recommended treatment regime, which will enable us to identify any gaps in treatment provision/use and to assess the effects of these differences. CACI SBRI Challenge ‐ Health Service Improvement Simon Broxton, Principal Consultant, 07/03/2014 Withdrawn/Not Approved N/A through Intelligent, Self Service, Integrated, Evidence CACI Ltd The use of Search Business Analytics enables people to use the same approaches as internet search engines that use ‘natural language’ queries to search and then rapidly filter results using drop—lists and a ‘click Based Analysis and select’ approach on values returned to go from hundreds of millions of records to the results of interest within a matter of seconds.
Through Search Business Analytics, provide technical and non‐technical staff with the ability to quickly and easily search, analyse, report and dashboard strategic needs assessments across linked health and social criteria and share this information with key stakeholders. This will identify areas where resources could be used more effectively to improve health outcome quality and deliver cost savings.
This project is a demonstration project and will use the CACI owned ACORN dataset to prove this concept.
The main delivery mechanism for this solution will be through the configuration and extension of CXAIR by Connexica Ltd, to be held within the secure area of the SAIL databank for the purposes of this project.
The solution will enable the drill down into cohorts of the population on the basis of independent or linked data i.e. health, social care or health and social care. In line with SAIL policy, In any case where the drill‐ down in data would return 5 or less records, the data will not be displayed in any outputs. Additionally, it will be clear to see patterns of current patient journeys through pathways of care across health and social care services, to identify potential earlier interventions and understand the impact of these interventions.
An investigation into the changing epidemiology of Dr Mari Morgan, Clinical Scientist Staphylococcus aureus is a bacterium that is commonly carried on human skin and mucosa. The public are generally aware of the meticillin resistant variety (MRSA) through widespread media coverage. Both the 05/03/2014 Approved 09/04/2014 meticillin‐resistant and sensitive Staphylococcus (Welsh Healthcare Associated sensitive (MSSA) and MRSA varieties can cause serious illness or death. If the bacterium enters the body through broken skin or an implanted medical device,for example an artificial hip, it can cause local or aureus (MRSA/MSSA) bacteraemia in Wales Infection Programme), Public Health widespread infections. Bacteraemia is the term used for infection in blood that may develop into septicaemia, a very serious illness. With widespread preventive efforts in hospitals, the numbers of MRSA infections Wales have steadily fallen in Wales over a number of years. However, there has been no equivalent decrease in MSSA bacteraemia and a recent increase in numbers has been identified and is a growing cause for concern. Public Health Wales have been asked to investigate the risk factors associated with this bacteraemia, in order to inform the revision and development of interventions to reduce the risk of infection. PHW has imported an anonymised set of laboratory data (DataStore) into SAIL to support this work.
To improve the health and well being of the ABMU Andrew Evans, Development Data from the SAIL databank will be extracted onto a SQL Server environment, within the SAIL databank, where the data will be risk stratified and be made available to end users via a number of reports. These reports 24/01/2014 Approved 16/04/2014 popluation through better use of population health Manager, The Sollis Partnership will initially focus on the following areas: information o GIS (Geographic Information System) o Health planning and trend reporting o LSOA comparison o Deprivation o Cohort reporting for specific disease groups Data modelling functionality will also be provided to end users. When data is presented to users, via a secure portal, it will be done at an aggregated level with strict small number confidentiality criteria so that there is no possibility of a patient's confidentiality being compromised. Phase 1 of this project will test the feasibility of the solution and a subset of the population will be used. Phase 2 will extend this to the entire population.
Interactive health visualisations Dr Kerry Bailey, Health Lead, We We Predict use complex big data to produce actionable insights. In this project We Predict will adapt and apply their global expertise in predictive analytics to the health sector to benefit people in ABMU. We Predict 24/01/2014 Approved 16/04/2014 Predict are a local company who operate in Swansea. We Predict will: • Use sophisticated data linking and analysis methods on SAIL data to describe, evaluate and predict health and healthcare in ABMU. • Present this intelligence in a user friendly interactive interface for use by lay people, health managers and clinicians to inform planning and evaluation of services.
• Produce tables and views from SAIL for analysts to interrogate.
Phase 1: data relating to obesity from both primary care and secondary care and the clinical consequences, in this case Diabetes, will be extracted and used in maps and other graphical representations to
demonstrate the health of people across ABMU, the use of health services and the correlation and predictive value of environmental risk factors and impact of services. Output: a user friendly, interactive data visualisation suite using maps, infographics and charts to represent key intelligence around obesity and type 2 diabetes within ABMU.
Communities First James White, Research Fellow, 24/01/2014 Approved 11/04/2014 Cardiff University There is little evidence on the effects of area‐wide regeneration projects on health. This in part is because randomised controlled trials, the ‘gold standard’ of evaluative research, are often not possible. Recently, the
Wanless Report and others have described the opportunities offered by 'natural experiments' to evaluate regeneration activities.
The project aims to investigate whether the Welsh Government's Communities First regeneration programme improved mental health, enhanced social cohesion and reduced health inequalities. We will carry out a natural experiment to evaluate the Communities First programme in Caerphilly county borough, linking information from Caerphilly county borough council on defined areas which did and did not receive regeneration interventions to Caerphilly Health and Social Needs Electronic cohort Study (E‐CATALyST) data collected before and after Communities First in 2001 and 2008. We will link the LSOA intervention
information to E‐CATALyST data in the Secure Anonymous Information Linkage databank and carry out an economic evaluation by quantifying the cost of primary care contacts and hospital admissions
It is expected that the project will produce high quality evidence on the impact of area‐wide regeneration on mental health. This is essential for policy and strategy development to improve mental health and reduce the large inequality in mental health status in Wales. Phase 1 (WP1 SAIL System): Health Impact of Dr Sarah E Rodgers, Associate 24/01/2014 Approved 23/04/2014 Structural Energy Performance Investments in Wales: Professor in Epidemiology, Swansea The SAIL data will be used to retrospectively to evaluate Arbed phase 1, aiming to evaluate the health impacts of ARBED interventions through using routinely‐collected anonymised data held within the SAIL databank. An Evaluation of the Arbed Programme University Data from housing interventions will be linked with various medical records to support the analysis into the health comparison study to evaluate any derived health benefits. Analysis will concentrate on camparison of pre and post intervention data within the study cohort which will also be compared against two other comparator groups, details of each group are as follows: ‐ Study Cohort: 7,460 households subjected to interventions ‐ Comparator Cohort: 13,000 social housing residences ‐ All of Wales Cohort: All low income neighbourhoods in Wales (top 10% of deprived areas in Wales).
Data collected in WP1 will be used to assist an economic evaluation as part of WP4 to determine if the investments can be considered an efficient method of improving public health over and above other intended programme benefits.
ABMU SBRI Challenge, Clinical Object Library Huw Morgan, Technical Director, GP GPCSL will create a 'clinical object library' that describes each patients interaction with health services based on SAIL data, and allows unprecedented ease of access to understanding highly complex clinical 17/01/2014 Approved 16/04/2014 Commissioning Solutions Ltd pathways, interventions and outcomes. By defining the building blocks (or "objects") that make up this journey such as disease, interventions, pathways, services, guidelines (eg NICE) and outcomes it will allow users to rapidly visualise complex services and scenarios. In addition these building blocks will be modular and reusable allowing complex objects to be created by combining existing ones. Having created these objects users will interact with them using an 'Object Library Interface'. This interface uses modern swipe, touch and drag and drop techniques to select and combine objects to visualise current services , understand how interventions lead to outcomes and plan new services around predicting the impact of positive or negative outcomes.
ABMU challenge: Conception and implementation of a Axel Schmidt, Head of Software Gesundheitsforen Leipzig (GFL) will provide an environment in which the analysis can be undertaken; this comprises: a data warehouse, a web portal, and an analytics platform. GFL will build a data warehouse 17/01/2014 Approved 16/04/2014 support tool for health data analysis Development, Gesundheitsforen which supports discrete data marts. Each data mart contains structured and cleansed data from a specific SAIL field. To provide performance and redundancy, these data marts are extendable. This approach will Leipzig GmbH allow the solution to easily scale and retain required performance characteristics. To make these data accessible to the end‐user we will provide an easy to use web application: SpagoBI. SpagoBI is written in Java –it provides a high level of expandability and portability.
The final application will comprise a flexible on‐Demand BI solution containing pre‐written reports and the ability for the user to create custom reports and data‐dashboards from using single blocks down to filtering data to meet the user’s needs.
Health IRIS; how can we improve the health and well Chris Poole, Clinical Science Liaison, 17/01/2014 Approved 16/04/2014 being of our population through the better use of Digital Health Labs The Abertawe Bro Morgannwg University (ABMU) Health Board has commissioned Digital Health Labs, a Welsh software company, to develop flexible web‐based tools allowing rapid processing of health data held population health information? in the SAIL Databank by non‐programmers. In addition to giving a general overview of health in Swansea, Neath, Port Talbot and Bridgend, these tools will enable ABMU health researchers to select any group of patients and gain a detailed understanding of the different ways in which their health is managed. This will help health managers to understand not only the best ways to treat and manage illness but also to
understand factors that may predispose certain people to disease, providing targeted opportunities for keeping those people well.
Cost utilisation of patients on MS Register Ceri J Phillips, Professor of Health 11/12/2013 Approved 16/01/2014 Economics, Swansea Centre for This project is part of the work of the UK MS Register. It will estimate the costs due to people with MS visiting NHS services (such as GP visits, stays in hospital, etc.) and some further costs to society. Information Health Economics, Swansea about health service use will be found in SAIL datasets, and costs will then be calculated using ‘unit‐costs’ available in published sources for this type of work. We will be able to use this information to work out University costs per person with MS. By taking into account the number of times people visit the GP and other healthcare settings, we will also see where the costs are incurred. This will give us a better picture of which services are most used by people with MS, and the associated costs, to add to the evidence for future planning and policy‐making. The EQ5D is a general quality of life questionnaire, widely‐used in health economic
studies, and the MSIS‐29 is an MS‐specific questionnaire used to assess the impact of MS. Both these questionnaires are completed by people with MS as part of the UK MS Register. By linking the EQ5D and MSIS‐29 responses to the SAIL data we will be able to model the questionnaire responses against the SAIL data to relate quality of life and impact of MS to healthcare service use.
Creation of Synthetic Datasets for SAIL Testing and Daniel Thayer, Senior Research 11/12/2013 Approved 16/01/2014 The SAIL Databank uses robust information governance to assure that medical data is not misused, as well as to prevent the disclosure of identity. Key safeguards include review of all projects by the Information Development Analyst, Swansea University Governance Review Panel, requiring all analysis to be conducted within a secure remote desktop environment, limiting access to only the data required for a particular project, and review of all outputs from the gateway. These procedures are essential for responsible use of anonymised health data. However, they also add a large amount of overhead to each project and are restrictive (by design).
Certain types of work do not require access to the genuine data, only "SAIL‐like" data. This includes some training, testing of software tools and developing new methods. By 'SAIL‐like' data we mean data that have been changed so that they resemble the real data, but do not relate to anyone. For example, the columns for certain information e.g. disease codes, ages, prescriptions, etc., would be randomly reordered. In this way, we propose the creation of synthetic versions of all SAIL core datasets by assigning events from the real SAIL datasets at random into a new, scrambled dataset. We will develop testing procedures to verify
that no real information about individuals is disclosed, and that no linkage back to the real datasets is possible. We propose that these synthetic datasets can then be removed from the SAIL Gateway and used for various development and testing purposes. Welsh Study of Mothers and Babies: Prospective Shantini Paranjothy, Clinical Senior 10/12/2013 Approved 16/01/2014 cohort study of renal disease in early childhood. Lecturer, Cardiff University Long term kidney disease is one of the most common illnesses of childhood causing lifelong health needs. Babies born with abnormalities in the kidney are at risk of long‐term kidney disease that may require treatment such as dialysis and kidney transplant as they grow older. Opportunities for early diagnosis and treatment of kidney disease may reduce the severity of illness over time. In some pregnancies, a marker (enlargement of the tubes that collect urine in the kidney, called pelvicalyceal dilatation) is detected in the baby’s kidney at theantenatal ultrasound scan. The meaning of this marker for the future health of the baby (for example the risk of urine infections or hospital admissions for kidney problems) is not known. There are no consistent guidelines about how to treat babies with this condition and so the identification of this marker can lead to anxiety for families and additional tests during pregnancy and after birth. We were funded by NISCHR to study the meaning of markers of uncertain significance (Welsh Study of Mothers and Babies). The study has recruited 21,106 pregnant women, with consent to follow‐up the babies in future studies. There are 150 babies in our study with the kidney marker. We did not find any associations with immediate birth outcomes such as stillbirth, premature birth or genetic problems. We propose to use routinely collected health data to explore whether babies with a marker in one or both kidney(s) detected during pregnancy have more urinary tract infections or hospital admissions for kidney problems during childhood, compared to babies without a marker. This information is vital for pregnant women, families and clinicians to facilitate the development of appropriate clinical guidelines and care pathways during pregnancy and after birth, including the appropriate use of medical interventions and treatment.
Social care and health of older people (SCHOOP): The Alison Porter, Senior Research Officer There is a link between use of social care and use of health care by people aged 65 and over. Appropriate social care can maintain health and delay or avoid the need for health care such as hospitalisation. 02/12/2013 Approved 16/01/2014 relationship between use of social care and use of Swansea University Conversely, a spell in hospital can mark the beginning of a need for social care. health care by people aged 65+ in Wales before and after the implementation of the Chronic Conditions The Welsh Government’s Chronic Conditions Management policy (2008) sets out a plan for improving care in the community and reducing demand for hospital care. Our study will examine the impact of the policy Management policy ‐ an observational study using across Wales on emergency hospitalisations of people aged 65 and over and see how this varies with varying social care input. routine data We will look in more detail at the same age group in Swansea. We will track health and social care use by individuals using linked anonymous data. This will enable us to understand more about how social care interventions may help to reduce hospitalisation, and about how demand on social care may change over time as a result of changes in health care input. The Swansea case study will help to show the feasibility of extending this analysis to the whole of Wales.
GP Practice Population Profile using the SAIL databank Dr Charlotte Jones The population that is registered to a GP practice can vary considerably by age and deprivation which can influence the demand that is placed on a practice. Currently practices are unable to calculate what their 02/12/2013 Approved 22/01/2014 own practice population looks like and how it compares with their peers. It is believed by having this information it will support discussions with the BMA relating to the GP funding formula. Therefore, HIRU is developing a standard report that can be generated for all practices in Wales to facilitate this work.
CALON ‐ Cardiac Ablation: Linking Outcomes for NICE Ruth Poole, Researcher (on behalf of The two main aims of the CALON project are to: 18/11/2013 Approved 05/12/2013 (Pilot project to examine the feasibility and usefulness Principal Investigator Dr Grace • find out more information about how well cardiac ablation procedures work and how safe they are. of data linkage between relevant data sources in order Carolan‐ Rees) Cedar • assess the value of using data linkage, a method of linking patient records, as a method for healthcare research. to capture health and social outcomes). Abnormal heart rhythms can cause very unpleasant symptoms such as chest pain, headache and tiredness. Cardiac ablation is a type of procedure which treats abnormal heart rhythms by destroying sections of tissue in the heart. The National Institute for Health and Care Excellence (NICE) recommend that some cardiac ablation procedures are only used in certain circumstances, as there is currently limited evidence on how well they work and/or how safe they are. We will look at existing health records to try to answer these questions. We will also evaluate the feasibility and usefulness of the method of linking patient records for the purpose of informing NICE guidance.
Investigating outcomes of Biventricular pacemaker Mark Anderson, Consultant This project aims to investigate health outcomes for patients with severe heart failure after having a biventricular (biv) implant. BIV pacemakers cost approximately £15,000 and have shown reduction in the amount 08/11/2013 Approved 05/12/2013 implants Cardiologist, ABM Morriston Hospital healthcare required after having a BIV pacemaker. Outcomes we are interested in looking at are admissions to hospital and length of stay, cardiovascular drug management such as beta blockers/statins and
mortality compared to a control cohort.
Using predicitive models for unscheduled care Dr Kerry Bailey, Epidemiologist Emergency admissions are costly both to the individual and to the Health Service. For some years there has been an attempt to quantify risk and produce risk stratification of patients to improve case management 28/10/2013 Approved 04/11/2013 Consultant in Public Health, Phd and avoid emergency admissions. Since April 2013 GPs in Wales are expected to risk stratify their patients and produce a list of the top 5% most at risk. The current model in Wales, PRISM, combines secondary Student, Swansea University care data with primary care data and although compares well to other combination tools( PARR, SPARRA etc) it is several weeks out of date by the time it is reviewed back in GPs surgery. Using SAIL data I would like to see if we can model a risk tool based purely on primary care data and then compare this to PRISM. The advantages of this would be that this would enable prediction in real time and utilise more detailed primary care data. I also intend to see if there are other factors ( e.g living alone, temperature) which are not currently included in models which may improve the predictive value of the model.
Data linkage in social care: a pilot project Professor Catherine A Robinson, Head Existing data linkage projects in Wales that utilise routinely collected administrative data have foccused on health datasets only. To build the complete picture of service provision there is a need to broaden this to 22/10/2013 Approved 04/11/2013 of School and Professor of Social include social service provision by Local Authorities, and provision of support by third sector organisations. Policy Research, Bangor University The work proposed here is a pilot project to test the feasibility of linking datsets from a local authority, the NHS and third sector organisations. The focus of this work will be on individual level data from adults who
are referred to social services in order to avoid admission to hospital or to facilitate their discharge from hospital. The data linkage will include data from statutory and third sector organisations and services which provide interventions and support in community settings.
We will use the tried and tested Secure Anonymised Information Linkage (SAIL) system (Swansea University) which allows large routine datasets to be presented for research links at the level of the individual, but with all individuals remaining anonymous.
Part of the feasibility study will be to assess the range and quality of the data available in each of the relevant organisations and to better understand the utility of data linkage across statutory and third sector organisations, in particular, to inform future service organisation and delivery. Identifying the true cost of asthma in the UK: Dr Gwyneth Davies, Senior Clinical Identifying the true cost of asthma in the UK 22/10/2013 Approved 18/11/2013 secondary analyses of national stand‐alone and Lecturer and Respiratory Physician, This work will provide the Wales’ most detailed, comprehensive and up‐to‐date estimates of the true prevalence and costs from asthma. Asthma is now one of the commonest long‐term conditions in the UK and it is linked databases in England, Northern Ireland, College of Medicine, Swansea therefore important to understand the costs of healthcare, social care and societal costs in order to inform decisions on care provision and planning. Scotland and University Wales Welsh data will be integrated into a state‐of‐ the‐art overview for the UK and its member countries of how many people have asthma and the full cost of this. We will analyse a range of relevant data sources, including national health surveys, GP data, prescribing, emergency care, hospital and mortality data to estimate asthma frequency and to what extent the health service is used by people with asthma. We will need to look at new ways to link and model data and will also flag up important evidence gaps and provide guidance on the relative importance and costs of filling these.
2014 update of the Global Burden of Disease: UK Professor Ronan Lyons, Professor of Policy makers increasingly use measurements of the health of the population to judge the effectiveness of policies. The Global Burden of Diseases was initiated in 1990 to calculate the burden of diseases, risk 22/10/2013 Approved 04/11/2013 Public Health, Swansea University factors and injuries using a new metric ‐ the Disability Adjusted Life Year (DALY). DALYs combine a measure of premature mortality with a measure of the impact of disability. The results of the GBD studies are used b policy makers to study the effectiveness of health policies. I have attached a Lancet paper showing the latest results for the UK amongst rich countries. There are a number of inaccuracies in the latest approach ‐ mainly from insufficient data on how common diseases are and also in the impact these conditions have on peoples lives. This planned latest update of the GBD project will use improved measures of disease frequency for the UK and separately for Wales. We are applying to use the linked data in SAIL so that these calculations will be as accurate as they possibly can be.
Further costs of head trauma in a small Swansea Ceri J Phillips, Professor of Health 07/10/2013 Approved 04/11/2013 The analysis will estimate the costs of head trauma patients for the NHS and personal social services (eg A&E attendances, inpatient admissions, length of stay, etc). Data on the use of the health service resources cohort Economics, Swansea Centre for will be collected using SAIL extractions based on a detailed data items table. Costs will then be calculated using unit‐costs estimated from current published sources. Costs per patient will also be calculated as well Health Economics, Swansea as frequency of service use multiplied by the relevant unit cost (also derived from current published sources) to produce a total cost per patient, with a segregation of costs in relation to the where they were incurred. University This would aid in the identification of the services most utilised by the patients, which would in turn add to the evidence for future policy making.
Pilot study of patients with eating disorders Jacinta O.A Tan, Clinical Associate Eating disorders have the highest mortality of any mental disorder. Research suggests that only a minority of people with eating disorders receive help from mental health services. Current service data in Wales 07/10/2013 Approved 02/04/2014 Professor, Swansea University suggest referrals to secondary care are highly variable across regions as compared to projections based on population demographics.
At present, we lack evidence of: how well eating disorders are recognised and diagnosed at the primary care level in Wales; what happens to patients with eating disorders after recognition at primary care, and how many or few enter Tier 2‐3 care; and what happens to patients with eating disorders after entering Tier 2‐3; and what inpatient admissions occur for these patients. Finally, we do not know what the trends have been over time. These are all concerns which need systematic exploration. This pilot research is intended to develop hypotheses for a large scale research proposal using SAIL to investigate treatment across Wales for eating disorders.
Epilepsy ‐ factors associated with A&E attendance Dr Owen Pickrell, Clinical Lecturer , Epilepsy is one of the most common neurological conditions affecting around 1% of the population. People with epilepsy account for 2‐3% of all emergency medicine department (Accident and Emergency) 30/09/2013 Approved 29/10/2013 Wales Epilepsy Research Network attendances. Whilst some of these attendances are appropriate‐many of these emergency department admissions in people with epilepsy might be prevented by appropriate escalation of primary care/outpatient management. The project aims to identify factors associated with emergency attendances of people with epilepsy and identify any potential interventions that may reduce emergency department presentations in the future. Routine data extraction for CAP Professor Richard Martin, Professor of Whilst the NHS has a limited amount of money the cost of treatments is an important comparison to make; termed the cost effectiveness. Would introducing PSA testing be cost effective compared to current NHS 11/09/2013 Approved 18/10/2013 Clinical Epidemiology, School of practice? In other words, is PSA testing effective, allowing men to be cured (saves more lives), and do treatment costs compare favourably to current NHS practice, where men who request a PSA test receive Social and Community Medicine, standardised information about the potential benefits and harms of PSA testing. University of Bristol When comparing the cost effectiveness of treatments it is important that information is collected on all people to ensure researchers arrive at the right answer. This involves finding out about their health and any other treatments they may have had. The NHS collects a wealth of routine information for administrative purposes when people access healthcare (such as the type of operation they have and the amount of time
spent in hospital). Using these routine data will allow us to include all individuals in the trial in the cost effectiveness comparison.
Impact of Care and Repair interventions on health Sarah Rodgers, Lecturer in Spatial The work carried out by Care and Repair Cymru ensures that the homes of older and disabled people are safe and enables individuals to live as independently as possible in their own communities. The types of work 01/09/2013 Approved 18/10/2013 servise utilisation and health outcomes Epidemiology, Swansea University carried out to create safer homes includes installing handrails, fitting new locks and fitting specially adapted bathrooms. This project aims to determine whether the work carried out by Care and Repair reduce or delay clients accessing Health and Social Care Services and also to determine if the cost of the work carried out by Care and Repair to create a safer home is cheaper than a client accessing Health and Social Care services. Comparisons will be made between those who received Care and Repair work and those who did not. By creating safer homes it is hoped that this would prevent adverse health outcomes therefore reducin the burden on Health and Social Care services. Using health record linkage in the UK Millennium Ronan Lyons, Professor of Public UK birth cohort studies have made important contributions to understanding of disease and health through early childhood and into adult life. The UK Millennium Cohort Study was set up to understand the health and 20/08/2013 Approved 18/10/2013
Cohort Study to investigate childhood obesity, asthma Health, Swansea University social and economic circumstances of British children born in the new century. It includes children from disadvantaged social circumstances, ethnic minorities and those living in all four UK countries. We plan to infections and injury add to the information already collected on these children at ages 8 months, 3, 5, 7 and 11 years by obtaining information recorded electronically about their use of hospital services as either inpatients or outpatient This will help us answer important questions about children’s health, particularly concerning children who become obese or overweight in the preschool years, who have developed asthma or other allergic
disorders, infections or serious injuries, by 11 years of age. We will also compare survey data with health records to understand more about the quality, completeness and consistency of recording of a child's ethnic group in hospital health records (not possible for Wales as ethnicity is not included in PEDW data supplied to SAIL). Finally we will determine the feasibility and value of linking cohort members to their
electronic primary care records (Welsh children only). It is important to note that records will only be linked where there is explicit consent to do so
We have over 1700 Welsh MCS participants with consent for health record linkage from those with parental responsibility. We propose to link their MCS data with the electronic health data held within SAIL. Further this, we intend to apply to have English and Scottish anonymised medical records for consented MCS children to be imported into the SAIL Databank. The SAIL approach to split file linkage would be used by HSCIC
(England) and ISD (Scotland) to link to hospital inpatient records. We intend to use the SAIL Databank and Gateway as a resource to provide secure privacy‐protecting access to investigators to analyse linked MCS
survey (n= 18,000 excluding N. Ireland) /health records for the purposes detailed in sections 4b and 5b. An exploration of bringing Energy Performance Lowri Wyn Francis, Research and The main aim of the project is to investigate the usefulness and sensibility of linking EPC data into an anonymised dataset. The project will be a feasibility study as it will not be possible to know the size of any 29/07/2013 Approved 18/10/2013 Certificate data into an anonymised dataset. Information Professional Trainee, cohorts that can be flagged up until the linkage has happened. There are records of EPC data for 13,109 households in Gwynedd, therefore it is anticipated that the potential population size for the study will be Gwynedd Council approximately 28,840 individuals. The study will explore many different aspects of linking the EPC data into an anonymised dataset. Firstly, how effective can the address matching be in the anonymisation process, and will report on the success of linkage. Having anonymised properties, it would be possible to flag up the family groups living in the properties at the time of certification, therefore a correlation could be identified between the number of rooms and family size and structure. Another aspect that could be explored is migration by looking at the frequency of change of occupancy by housing type. A comparison betwee EPC certification and housing improvements data from the Homes Energy Efficiency Database (HEED) will enable the researcher to identify how up to date the EPC's stay for the 10 years they are valid. An associatio between the energy efficiency rating of the home and occupants health could also be explored once the data is linked. This could be done by creating anonymised subsets of residents of the houses who are having acute respiratory episodes, and comparing them to housing type, energy efficiency and overcrowding.
An e‐cohort study of trends in the incidence of Ann John, Senior Lecturer, College of Research has suggested a change in the way GP's are recording depression in both adults and children. The recording of symptoms is increasing overtime with a parellel decrease in new diagnoses. Similar trends 29/07/2013 Approved 16/10/2013 recorded anxiety and depression symptoms and Medicine, Swansea University have been found with regard to anxiety diagnosis/symptoms in adults, however research has yet to be done with children and adolescents. Research has also shown a steady increase in antidepressant prescibing diagnoses and antidepressant prescription in children in both adults and children. The purpose of this research is to examine time trends in the incidence and sociodemographic variation in GP recorded symptoms and diagnoses of depression and anxiety in children and young people and adolescents. We will also examine time trends in antidepressant prescribing for children and adolescents. To do this we will create a retrospective e‐cohort of children and adolescents using routinely collected data available in SAIL. Data will be collected on new incidence of depression/anxiety symptoms/diagnoses and antidepressant prescribing for each year from 2003 to 2011.
Educational attainment at key stage 1 of children with Cathy White, Consultation This study aims to investigate the effect of epilepsy and epilepsy treatment on educational attainment of children at key stage 1 (age 7 years). We aim to investigate both whether there is an effect of the use of anti‐ 11/07/2013 Approved 18/10/2013 epilepsy, and children of mothers with epilepsy. Paediatrician, ABM Morriston Hospitalconvulsants and education in children who have epilepsy at key stage 1 (primary exposure) as well as seperately investigating if anti‐convulsants or lack of anti‐convulsants being taken by pregnant epileptic women (secondary exposure) has an effect on their childs' educational attainment.
Smoke‐free legislation and fetal, infant, and child Jasper v Been, Research Fellow, 19/06/2013 Approved 12/07/2013 health in Wales: interrupted time series analysis. Centre for Population Health Sciences, The University of Tobacco use kills over five million people each year worldwide. In addition, exposure to other people’s tobacco smoke (‘second‐hand smoking’; SHS) is estimated to kill another 600,000 people, including 165,000 Edinburgh children. To protect people from harmful effects of SHS the World Health Organization (WHO) recommends that smoking be banned in public places. Among other positive effects, research now shows that deaths from cardiovascular disease among adults and hospital admissions for asthma among children are reduced following smoking bans in different countries. Despite this, only 10% of the world’s population is currently protected by such bans. A better understanding of the range of effects that smoking bans may have can help policy makers worldwide to make well‐informed decisions about the implementation and expansion of policies to further protect people from SHS. The current project aims to investigate whether the smoking ban that was introduced in Wales in 2007 has led to improvements of infant and child health. We will evaluate changes in a range of adverse health outcomes that are known to be affected by SHS: stillbirth and infant death, congenital anomalies, preterm birth, low birth weight, and respiratory diseases in childhood including asthma and airway infections. For this purpose we will use data on all of these outcomes from the Wales Electronic Cohort for Children (WECC) that includes information about all children that lived in Wales during the study period (1998‐2011). We will investigate how trends in these outcomes have been changing over time and whether particular changes have occurred following the introduction of the smoking ban. Secure access to WECC will be provided through the Health Information Research Unit at Swansea University, ensuring that no single person can be identified from the database. We will present our findings to colleagues at international congresses and describe our findings in a scientific publication that can be publicly accessed to ensure its availability to relevant audiences.
Cost of wound management to the NHS Ceri J Phillips, Professor of Health The analysis will estimate the costs of wound management for the NHS and personal social services (eg A&E attendances, inpatient admissions, length of stay, etc). Data on the use of the health service resources 03/06/2013 Approved 18/07/2013 Economics, Swansea Centre for will be collected using SAIL extractions based on a detailed data items table. Costs will then be calculated using unit‐costs estimated from current published sources. Costs per patient will also be calculated as well
Health Economics, Swansea as frequency of service use multiplied by the relevant unit cost (also derived from current published sources) to produce a total cost per patient, with a segregation of costs in relation to the where they were incurred University and the potential opportunity for cost shifting. Long term outcomes in patients with Graves’ Dr Onyebuchi Okosieme, Consultant 03/06/2013 Approved 18/07/2013 disease Endocrinologist, Prince Charles Graves hyperthyroidism is a condition characterised by thyroid gland overactivity and affects 05‐1.0% of the population. Most sufferers are women, typically in their thirties to fifties. Three established treatment Hospital, Cwm Taf Health Board modalities are available for patients with Graves' disease, namely, (i) antithyroid drugs, (ii) radioactive iodine, and (iii) surgery (thyroidectomy). Untreated disease is associated with considerable morbidity including heart rhythm disorders, heart attacks, heart failure, strokes, and osteoporosis (thinning of the bones). Many of these complications occur long after the disease has been treated and there is evidence that patients
who have been treated for Graves' disease suffer a higher mortality from cardiovascular disease than matched counterparts in the population. However, it is not known whether the treatment of Graves' disease has any relationship to the development of long term complications. This knowledge will be crucial in guiding clinical decisions on therapeutic choices The study will address this by analysing the long term outcomes in a cohort of patients with Graves' disease who have been treated in the three different ways. Patients who have been treated with Graves' disease from 1998‐2010 would be identified through a hospital laboratory
database. The clinical and demographic details of these patients will be imported into the SAIL database using a secure anonymisation process. Analysis of patient outcomes in comparison to age and gender matched controls will be undertaken through the SAIL database and its linked datasets. Impact of the NICE guideline recommending cessation Mark Anderson, Consultant Infective endocarditis (IE) is a rare condition with significant morbidity and mortality. It may arise following bacteraemia in a patient with a predisposing cardiac lesion. In an attempt to prevent this disease, over the 23/05/2013 Approved 26/07/2013 of antibiotic prophylaxis for prevention of infective Cardiologist, ABM Morriston Hospital past 50 years, at‐risk patients have been given antibiotic prophylaxis before dental and certain non‐dental interventional procedures. Recent NICE guidlines suggest cessation of antibiotic prophylaxis in dental endocarditis practices, and this study aims to determine the effects of these guidlines on incidence of infective endocarditis. Linking spatial accessibility of services to treatment Dr Richard Fry, Senior Research Access to health care has long been associated with health inequalities in Wales. Using the e‐GIS (environmental geographic information system) and the SAIL databank we propose to test a methodology in which 23/05/2013 Approved 26/07/2013 of chronic health conditions Officer, HIRU we can anonymously link residential level accessibility scores to an individual’s health records.
By using a GIS system we estimate a number of accessibility measures from residential addresses to health care providers (eg GP surgeries and clinics). So one measure may be the direct distance between the two points, another may use roads, and another may use footpaths. These calculated measures are saved in a file which is processed by NWIS to securely match and anonymise the residential addresses to a RALF (Residential Anonymised Linking Field (RALF) This process follows the same security protocols that are used in the matching and anonymisation of individuals. The data are then loaded into SAIL, with a further encryption of the RALF, enabling the potential to link the eGIS measures to individuals, via the information stored in the anonymised Welsh Demographic Service dataset. This will be done for a regional area within South Wales.
We can then find outcome measures of interest for the patient, depending on the disease condition being investigated. These can then be linked to the measures of distance to the GP surgery and other health providers to see the effect of these accessibility issues on health outcomes. We.will then compare these with a range of socioeconomic factors. Thus we will estimate the potential health impacts of differences in spatial accessibility to care services.
Colorectal polyp risk factors‐Bowel Screening Wales Professor Frank Dunstan, Professor of In Wales in 2010, colorectal cancer was the second most common cancer in men and third most common in women. The Bowel Screening Wales (BSW) programme began in 2008, aiming to reduce mortality from 15/04/2013 Approved 14/06/2013 Medical Statistics, Insitute of Primary bowel cancer by 15% by 2020. Screening is available for men and women aged 60‐74. Initial screening involves detection of faecal occult blood (FOB) in bowel stools. Sometimes FOB can be due to bowel polyps Care and Public Health, School of or cancer. Most bowel polyps are benign but some (e.g. adenomas) are pre‐cancerous. Participants who test positive for FOB are offered colonoscopy. Colonoscopy aims to detect colorectal cancer and there is the Medicine, Cardiff University opportunity to detect and remove polyps/adenomas before progression to cancer. This study aims to 1) investigate whether sleep disruption is associated with an increased risk of colorectal polyps/adenomas and 2) research modifiable risk factors for primary prevention programmes and secondary prevention of recurrence of screen‐detected polyps. Data from an existing Bowel Screening Wales dataset (2008‐09) will be linked with further BSW data and GP records via SAIL to estimate the odds ratio of colorectal polyps (and cancer if possible) due to an association with factors such as sleep and metabolic syndrome.
Investigating modifiable risk factors for Escherichia Dr Mari Morgan, Clinical Scientist The increase in E .coli bacteraemia particularly in the elderly in Wales is a growing cause for concern. Public Health Wales have been asked to investigate the risk factors associated with E. coli bacteraemia, in 07/03/2013 Approved 13/06/2013 coli bacteraemias in the Welsh population (Welsh Healthcare Associated order to devise population level interventions to reduce the burden of E. coli. Infection Programme), Public Health Wales Understanding the risk profiles and outcomes of Dr Phillip Freeman, WCAT Clinical 27/02/2013 Approved 07/05/2013 Welsh patients with a diagnosis of acute coronary Fellow in Cardiology, Cardiff Broadly speaking there are two major types of heart attack. The first is the type were a plaque (narrowing) in the coronary artery (the vessel supplying blood to the heart muscle itself) ruptures and develops a clot that syndrome. University completely blocks the artery. These patients become very unwell and require immediate medical attention that will unblock the artery usually with a procedure called a coronary angiogram and angioplasty (removin the clot and restoring flow). Unfortunately a significant proportion of these patents die before they can get to hospital (30‐40%). The second (and more common) type of heart attack is where the plaque ruptures but no clot forms. These patients have a brief episode of pain and then recover with minimal damage to the heart muscle (we call them small heart attacks or NSTEACS). NSTEACS are often seen prior to large heart attacks that are described above; and this is the main reasons we aggressively treat these patients, in order to stop them having a large heart attack which they may not survive or which may leave them severely debilitated.
We know that a number of clinical markers predict the risk of patients with NSTEACS, and we also know the higher a patients risk the more likely they are to gain benefit from invasive therapies (stent procedures or bypass surgery). We also know that sometimes patients are not suitable for these invasive therapies because the risk of the procedure outweighs the benefit of the procedure.
It is well documented that without a formal risk scoring system the accuracy of doctors at predicting both risk of heart attack and risk of procedure is very poor. We want to compare our population of formally risk assessed patients (who have had invasive therapy) to the population of patients in South Wales who have suffered a heart attack and who have not been referred (or risk assessed) for invasive therapy. We also
want to see how well other important medical conditions are treated in both groups (high blood pressure, cholesterol and diabetes).
We hope that this audit will enable us to identify any groups of patients who may receive considerable benefit from invasive therapies but who are not currently referred for this because of inadequate or inaccurate assessment. We also want to identify how well we are treating both groups of patients in the context of conditions that also add to their future risk of cardiovascular complications (high blood pressure, high
cholesterol and diabetes).
Assessing the costs of care settings for older people Dr Paul Nash, Director MSc in Ageing 29/01/2013 Approved 26/03/2013 in Wales Studies, Centre for Innovative Ageing, Using retrospective longitudinal information to compare and contrast the utilisation of health and social services and their associated costs by those living in: Extra Care; in their own homes; in sheltered housing; Swansea University and in residential care settings will provide a clear evidential base. Making use of local authority "supporting people and adult social care" information as well as information contained within the SAIL database at Swansea University a pragmatic cohort of older people will be
identified from Swansea and their use of health and social care services monitored over a given time period. It is hoped that SAIL will provide access to information on individual use of health related services such as hospital admissions, medication usage and GP Events while social services departments will be able to provide access to information on individual use of services including assessment care management, day services, homecare provision, and respite. Similarly, supporting people services will be able to identify whether any housing related support is being provided. Methodology: A researcher has been appointed through Swansea University to work with partners already identified within Swansea Local Authority. With LA guidance, the researcher has started to collect the service use and will utilise published unit cost data pertaining to those suitable participants/patients.Through data linkage the use of NHS services pre and post social care services delivery will be used to estimate a complete
economic costing across health and social care services. Data Linkage: Proof of concept for Cardiff Domestic Lisa Megraw, ICT & Data Supervisor, 24/01/2013 Approved 25/03/2013 and Sexual Violence Abuse (DSVA) Data Partnership Cardiff Women's Aid Victims of domestic violence and sexual abuse in the Cardiff area may approach a number of statutory health, local authority, voluntary organisations, or the police for help. These agencies form collaborative groups to ensure that information is exchanged to ensure the best possible help is available. This is highly sensitive information normally shared on a face to face basis by suitably appointed and authorised personnel, but currently there is no way to collate the information across these organisations to carry out general research on the underlying causes of these problems or the effectiveness of the various interventions. In order to explore the possiblility of developing a linked dataset, this is a pilot study to process a sample of data from one of the partner organisations(Cardiff Womens Aid) into SAIL. This sample, which will include about 500 cases, is likely to include people who may move home frequently, particularly around the time they present to service providers, and in some cases may supply false information.The first aim of this pilot study is to see if there is a high enough matching rate to make linkage viable. Then we aim to demonstrate to the Cardiff partner organisations the security of the SAIL data handling process, and the value of linking datasets in an anonymised environment. This is to help these agencies reach agreement about setting up a larger data sharing project.
Reducing clinical risk in the management of alcoholic Dr Michael Jose, Consultant ‐ 14/01/2013 Approved 26/03/2013 re‐attenders in Emergency Departments Emergency Medicine, Aneurin Bevan This study is for an MD thesis in the College of Medicine, Swansea University, which is investigating the clinical management and clinical risks for alcohol‐related re‐attenders at Emergency Departments in Swansea Health Board and Newport. The study patients are 74 from Morriston Hospital Swansea and 97 from the Royal Gwent Hospital, Newport
The SAIL data will enable us to investigate how mortality rates and hospital admission rates (for alcohol‐related causes including liver cirrhosis, acute and chronic pancreatitis, upper gastrointestinal bleeding, etc) among the study patients compare with those in the corresponding age‐group and sex matched general population. This will help enable the project to assess the degree of clinical risk in the study patients.
Dynamic Wales Electronic Cohort for Children (D‐ Dr Melanie Hyatt, Research Officer, 04/12/2012 Approved 25/03/2013 WECC) Centre for Health Information, Research has shown that poor education is associated with ill‐health. We propose a study to improve child health and education through the use of routinely‐collected anonymised data held within the SAIL Research and Evaluation (CHIRAL), databank. We will convert the Wales Electronic Cohort for Children study into a dynamic e‐cohort (D‐WECC) capable of tracking individuals across datasets monitoring life events including health and education. We Swansea University will include new births and flag migration. We will create novel links from D‐WECC to a traditional cohort: the Millennium Cohort Study (MCS), to explore the effect that additional variables, such as socio‐economic status and perceptions of the local environment, has on the well‐being of children in Wales who overlap between the two cohorts. In this way we will be able to assess the robustness of using each data source alone compared with the increased power of using a hybrid design. Results will aid in assessing the ability of using routine data to identify health inequalities of children in Wales. This study offers a unique opportunity to inform policy and intervention studies to ensure limited resources are most effectively used to improve health and well‐being for the most disadvantaged children. Quantifying the burden of electrical burn injuries Professor Ronan Lyons, Professor in This project aims to quantify the burden of electrical burn injuries, by identifying electrical burn injury patients in the SAIl database and linking their health records together anonymously. The burden of electrical bu 22/11/2012 Approved 04/01/2013 Public Health, Swansea University injuries in Wales and the UK will be expressed in terms of health service use, disability, fatality, and direct medical costs generated as a result of electrical burns. By assessing the burden of electrical burn injuries, this study hopes to raise the profile of electrical burns amongst the public and policy makers, and direct funding to support intervention/prevention activities.
MCS Health Record Linkage Project for Welsh Professor Lucinda Platt, Professor of The Millennium Cohort Study is a multi‐disciplinary survey of over 19,000 children born in 2000‐2001 from across the UK who are followed over time. The sample includes an oversample of children from Wales. The 19/11/2012 Approved 04/01/2013 participants Sociology, Director & Principal children have been surveyed at ages 9 months, 3, 5 and 7 years old, with data collected from parents and, from age 3, through direct measures from the children themselves. By linking the Welsh children's survey Investigator of the Millenium Cohort data to their hospital records the project will be able to provide an important and valuable research resource for the research community, which will enhance understanding of children's health and wellbeing, health Study, Institute of Education, inequalities and the consequences of hospitalisation and for subsequent outcomes. HIRU will be supplying anonymised electronic health records to the MSC research project of participants that have given informed University College London consent.
UK Biobank: Welsh healthcare data linkage and Professor Ronan Lyons, Professor of The study will use anonymised data linkage to bring together lifestyle, environmental and genetic information from Welsh resident volunteers in the Biobank project, and their continuing history of health service 07/11/2012 Approved 07/01/2013 informatics support. Public Health, Centre for Health utilisation already accumulating in SAIL. This will allow the relationships between the biobank measured factors and the development of diseases to be examined by researchers who will be looking for patterns that
Information, Research and Evaluation might indicate associations between these factors and disease. Particpants from the Biobank project have been linked in a previous project, which already received IGRP approval 0091. This study will update those (CHIRAL), School of Medicine, records and add newly recruited participants. Swansea University
Suicide Information Database‐Cymru (SID‐Cymru) Ann John, Clinical Senior Lecturer in 18/10/2012 Approved 04/01/2013 Public Mental Health, College of Medicine Swansea University The information available on people who complete suicide in Wales is limited. This acts as a barrier to our understanding of their underlying problems and the effectiveness of efforts at preventing this tragic event. We know there are high levels of mental health disorders, substance misuse, previous self harm and contact with services. However much of this knowledge is based on small studies. More is required, to understand the health and wider social circumstances (short and long term) of those who complete suicide, in order to inform local and national policies.
This study will look at possible risk factors for completed suicide from the whole Welsh population, whether known or unknown to mental health services. We will explore factors such as visits to the general practitioner, medication given, self harm, general hospital admissions, medical history and employment. It will highlight areas for intervention to prevent suicide in the general population of Wales.
We will use the Swansea University Health Informatics Research Unit’s (HIRU)’s, Secure Anonymised Information Linkage (SAIL) database which links anonymised health and social services data on the population of Wales. We aim to establish the Suicide Information Database‐ Wales (SID‐Cymru) to provide a central repository for information on all probable suicide deaths in Wales in order to support policy and prevention.
Population feasibility for transcranial magnetic Huw Morgan, Technical Director, GP 27/09/2012 Approved 04/01/2013 stimulation (TMS) Commissioning Solutions Ltd This project will provide background summary information on the population of Wales who have depression, such as the average age of the patients, the proportion of males and females, the severity of the disease and typical patient pathways and outcomes for these patients. The information will be used by Neuradaptix to assess the feasibility of the application of TMS therapy for patients with depression.
SAFER 2 (Support and Assessment for Fall Emergency Helen Snooks, Professor of Health 30/08/2012 Approved 03/10/2012 Referrals). Care of older People who Fall: Evaluation of Services Research, Swansea the clinical and cost‐effectiveness of new protocols to University Sail data will be used to measure the outcomes of a structured trial in which some patients, who have fallen and been attended by an emergency ambulance, will receive the current standard treatment, whilst othe allow paramedics to assess and refer older people to will be processed according to a new protocol that allows referral by the ambulance crews to different services. The rates of further falls, subsequent 999 calls, A&E attendances for falls, fall related injuries, hospital appropriate community based services admissions, quality of life, 'fear of falling', patient satisfaction and death will be compared and summarised for a period of time following the initial 999 event. Comparisons will be made between the group receiving standard treatment and those receiving the new treatment. Where possible this will be achieved in Wales by anonymising the study particpants into SAIL so that their routinely collected data can be compared. A simultaneous process in England will be processed by the NHS Information Centre. Social Deprivation and Epilepsy in Wales Owen Pickrell, Clinical Research Epilepsy is a common disease affecting around 1% of the general population. Epilepsy can possibly decrease quality of life due in part to factors such as the inability to drive and gain suitable employment. This 30/08/2012 Approved 03/10/2012 Fellow, Wales Epilepsy Research project aims to investigate whether there is a link between social deprivation and the number of people with epilepsy. If any relationship is found it would highlight important areas for future spending decisions on Network epilepsy healthcare and research. Mortality in Childhood Epilepsy Dr Cathy White, Consultant Paediatric The aim of the study is to use the SAIL database to identify a cohort of childhood epilepsy from the GP database, and group diagnoses by idiopathic and symptomatic epilepsy with the aid of co morbidity diagnoses. 30/08/2012 Approved 03/10/2012 Neurologist, ABM ULHB ‐ Paediatrics The project will also involve accessing the ONS Mortality database to access death certificate information, and will also involve accessing the Address Registry database to gather deprivation information on the Morriston Hospital cohort. Mapping dermatology service demands (proxy for Fiona Kinghom, Consultant in Public Cardiff and Vale UHB plan to enhance and further develop current dermatology service provision in Cardiff and Vale, to offer patients improved information and quicker access to high quality services, with reduced 28/08/2012 Approved 03/10/2012 population needs) Health, Public Health Wales/Cardiff waiting times. A new service model has now been developed incorporating a number of new elements which will improve patient experience, enhance primary care skills, embrace IT innovation and foster and Vale UHB partnership working between primary and secondary care clinicians . However, to ensure that the UHB has adequate resources to deliver the new service it needs to have an understanding of current demand for dermatological care. A number of data options have been considered and the best, most practical option is to use primary care data using read codes that capture most dermatological conditions. These read codes are a source of robust compulsory information, and are used as a monitoring tool within the GMS contract. These data will help the UHB to provide a baseline of primary care activity (in part as a proxy for population need), on which future change in service use can be measured. Journey mapping of patients with multiple chronic Sarah Lowe, Senior Research Officer, Anonymised cohorts of people who had an emergency admission to hospital for: heart disease, chronic respiratory conditions, prostate, breast, lung and colorectal cancers, diabetes or stroke between 2007 and 2009 22/08/2012 Approved 03/10/2012 conditions Knowledge and Analytical Services, will be linked across primarycare, secondary care, and Accident and Emergency care, to captureevents beforeand afterthe emergency admission. Groups will be established according to what othercomplicating Department for Strategic Planning, chronic conditions were also present,and the numberand type of health service interactions and totallength of stay in hospital, will be summarised forthe period2003 to 2011. The three groups of conditions that Finance and Performance, Welsh have requiredthe most health service supportoverall will be examined forpatternsof variation in the care, to see if ‘good practice’ events, leading to better outcomes (e.g. shorterlength of stay orless returnsto Government hospital) can be observed. As a demonstrationprojectthe reportingwill suggest hypotheses forfurtherexplorationand illustrate to WG colleagues some of the issues thInformatics Research Unit’s (HIRU)’s, Secure Anonymised I Prescribing trends in epilepsy Owen Pickrell, Clinical Research 17/08/2012 Approved 03/10/2012 Fellow, Wales Epilepsy Research Epilepsy is a common disease affecting around 1% of the general population. The majority of patients with epilepsy will have to take lifelong medication. It is very important therefore for doctors to prescribe the most Network appropiate medication for the patient. Over the last 10 years there have been several large drug trials in patients with epilepsy which have resulted in new guidelines for doctors when prescribing medications for people with epilepsy. We want to investigate trends in prescribing patterns for anti‐epileptic medications in Wales to see if they follow the guidelines. If any differences are identified then this would be an important area for future work as it is very important that people with epilepsy in Wales get the best possible care by being prescribed the most appropiate medication for their condition. Almirral COPD Research Huw Morgan, Technical Director, GP This project will provide background summary information on the population of Wales who have COPD, such as the average age of the patients, the proportion of males and females, the severity of the disease 17/08/2012 Approved 03/10/2012 Commissioning Solutions Ltd measured by a standard scale known as GOLD,(based on lung function measurements), smoking status, and the drug treatments currently being used to treat these patients. The information will be used by Almirall when presenting the potential impact of their new treatment (drug) to the All Wales Medicine Strategy Group.
Long‐term health sequelae of cryptosporidiosis Angharad Davies, Clinical Senior Cryptosporidium is a protozoan parasite and in the UK it is the commonest protozoal cause of acute gastroenteritis, with between 3000 and 6000 laboratory‐confirmed cases reported annually. Two species, 16/07/2012 Approved 17/08/2012 Lecturer in Microbiology, Swansea Cryptosporidium parvum and Cryptosporidium hominis, account for most of these cases. Disease is commoner in young children but it can affect healthy people of any age. It is a particular problem in the profoundly University immunocompromised.Little is known about the long term health effects of Cryptosporidium infection but there is evidence to suggest that, rather like other bacterial, causes of gastroenteritis, infection may have long‐ term sequelae. It is these sequelae which we would like to investigate with the help of the SAIL project. The project involves anonymised linkage of data within SAIL to create a case‐control study to see if infection results in higher rates of persistent symptoms and diagnoses of IBS The UK MS Register Dr Kerina Jones, Senior Research 18/06/2012 Approved 17/08/2012 Fellow in Health Informatics, MS is a complex neurodegenerative condition affecting an estimated 85,000 ‐ 100,000 people in the UK. There is a lack of information about MS to support research and service planning. The MS Society has Swansea University commissioned a project to develop a prototype MS Register in pilot sites that could be scaled‐up to become a national Register. The prototype collects data from; neurology clinics in 5 pilot sites (2 in England, 1 in Wales, 1 in N Ireland and 1 in Scotland); directly from people with MS as they complete questionnaires via a secure internet site; and from routine data such as GP and hospital records. With consent of the people with MS and ethical approval we are able to link these data at the individual level and then remove personal details to create an anonymous MS Register as a rich resource for research and service planning. Provision of quality improvement toolkits to primary Martin Holloway, Senior Information 30/05/2012 Approved 03/10/2012 care Analyst, Public Health Wales PCQIS provides a national platform for the dissemination of quality improvement products specifically for primary care. PCQIS is advised by a national steering group which includes representation from, GPC Wales,
RCGP, The Deanery,etc. This group outlines our work plan. There is a large range of products in our catalogue and general information found on our web pages.
http://www.wales.nhs.uk/sitesplus/888/page/45127
This project will test out the Read code selections in our new products and those currently under development. The outputs may, modify this selection or the enquiry logic, providing evidence of general data quality and coding habits. This will help to provide appropriate advice relating to the recording of clinical interventions and to help accurately audit the quality of services to patients. We would like to apply for a generic
application so that we could test our products as and when they are developed without making multiple applications for essentially the same output
Data Visualisation Proof of Concept David Ford, Co‐Director Health Dealing with vast amounts of data is a challenge and a new technique called Data Visualisation offers the ability to view such data in a different way, which could potentially highlight interesting aspects of the data 01/05/2012 Approved 18/06/2012 Information Research Unit (HIRU), that are not already known. This project is a proof of concept to understand the process required to work with the SAIL data and the data visualisation tools. To do this a subset of SAIL data will be identified for a Swansea University group of randomly selected individuals aged 50 yrs + which will list all of their key health events for 12 months. These data will be displayed using Data Visualisation techniques to see if any underlying patterns of care are highlighted. Sudden death and epilepsy in Wales Rhys Thomas, Clinical Lecturer in Status epilepticus (SE) is a life‐threatening complication of epilepsy and a medical emergency. Within SAIL the aim is to identify patients presenting with SE and then comparing mortality rates between those who 24/04/2012 Approved 18/06/2012
neurology, Wales Epilepsy Research had a previous epilepsy diagnosis to those who didn't. This will be done by obtaining epilepsy as SE diagnosis information from primary care data and then linking them to the ONS mortality data. A mortality Network prevalence for the two cohorts can then be compared.
Nature of death will be an important factor with regards to mortality rate ‐ we are particularly interested in people who have Sudden Unexpected Death in Epilepsy (SUDEP) coded as the cause of death, as we will want to compare the prevalence of this between the two cohorts. For some patients an episode of SE is their first presentation of epilepsy and it would be interesting to confirm how many of those presenting with SE went on to develop epilepsy . Research into the UK Tertiary Healthcare Referral Victoria Norman, Managing Director, a)For a list of conditions which might lead to a referral to tertiary services, we will examine a cohort of cases of these condtions within the GP data in SAIL to establish the level of the condion in the population, and 03/04/2012 Approved 16/05/2012 Market Health Focus Referrals (HFR) what proportion of this group is has actually been recorded as being referred to such services. For example the numbers with diabetes will be used to establish the potential size of the population that could be referred to an exercise referral scheme, and examine how many actually are recorded as being referred. We will then validating the baseline referral information using secondary data from NHS England (the Qresearch Database at The University of Nottingham)
Fforestfach fire incident public health impact follow‐up Huw Brunt, Consultant in On 16th June 2011, public health agencies in Wales were notified of a fire at an unoccupied warehouse on the Fforestfach Industrial Estate, Swansea. The fire generated a dense, dark smoke plume over the three 03/04/2012 Approved 15/05/2012 Environmental Health Protection, week period for which it burned with fine particulate matter (PM10) being the emission of primary public health concern. It is considered plausible that those living up to 2km from the incident scene (estimated Public Health Wales population 24,812) were exposed elevated levels of PM10 and may have suffered considerable anxiety. This project aims to assess whether there has been any measurable health impacts of this fire on population health in the 12 month period following the fire. Weight change associated with anti‐epileptic drugs Owen Pickrell, Clinical Research 03/04/2012 Approved 15/05/2012 Fellow, Wales Epilepsy Research Epilepsy is a common disease affecting around 1% of the general population. The majority of patients with epilepsy will have to take lifelong medication. It is very important therefore for patients and doctors to fully Network understand any potential side effects that these medications may cause. It is known that several anti‐epileptic drugs (AEDs) can cause weight change. There is a current lack of evidence however as to how much the most common AEDs can change body weight. This study aims to collect data on the change in body weight caused by five of the most common AEDs. This would be done by looking back at anonymised data from GP practices. The results would enable patients to make a more informed choice when discussing the choice of AED with their doctor. Testing the usability and accuracy of active case Huw Morgan, Technical Director, GP The concept of active case finding is to assist clinicians and case managers in identifying patients that may be more suitable for intervention and programmes of care. The objective of these programmes is to 03/04/2012 Approved 18/06/2012 finding models and medicines optimisation Commissioning Solutions Ltd improve outcomes and efficiencies. Risk stratification is commonly used to predict patients at risk, for example those likely to be admitted. Gaps in care identifies possible clinical interventions based on techniques to assess efficiency of identification and established guidelines such as those published by NICE and WHO. Medicines optimisation recommends changes to prescribing that may lead to better outcomes or efficiencies. GP Commissioning have management of patient interventions.esting established a tool (HINT) that is able to combine these methods and wishes to explore whether this provides more accurate and useful recommendations for intervention. Anonymous SAIL data from primary and secondary care contains the required content and volume to achieve this. Evaluation of the Flying Start initiative Sarah Lowe, Senior Research Officer, Flying start areas will be determined as a list of households translated from the school catchment areas designated as part of the initiative. These households will be processed into SAIL as a set of residential 20/03/2012 Approved 01/06/2012 Knowledge and Analytical Services, anonymised linking fields, which enables the existing health records of 0‐3 year olds living in these areas to be established. When this group of children progress through education, the key stage achievements can
Department for Strategic Planning, be summarised. Changes in achievement, and in health service visits will be measured for the children in years leading up to the baseline date, 01 st April 2009, and annual summaries will be produced from 2009 Finance and Performance, Welsh onwards to establish changes in health service utilistation and educational achievements in successive annual cohorts.Changes in flying start areas will be compared to the control areas to distinguish the initiative Government effects from general changes over time.
Fuel Poverty Sarah Lowe, Senior Research Officer, Home improvements have been carried out for a number of years on the homes of people classified as being in fuel poverty. By anonymising these property details into the SAIL environment, the health service 13/03/2012 Approved 26/04/2012 Knowledge and Analytical Services, utilisation of those people effected can be compared before and after the intervention, and against the general population, to establish if the number or GP visits, illnesses, etc are affected by these home Department for Strategic Planning, improvements to the extent that they might form the basis of a measure of effectiveness. Finance and Performance, Welsh Government
Prognostic factors and outcomes for medical Dr Stephen E Roberts, Reader in 24/02/2012 Approved 15/05/2012 emergency admissions Epidemiology & Public Health, There are often large differences in mortality following admission as an emergency across hospitals in the UK. Although some of this variation is linked to case mix (factors such as age and co‐mrorbidities) the College of Medicine, Swansea reasons for much of the variation in death rates is not known. It is also unclear whether this variation across hospitals extends to other outcome measures including lenths of inpatient stay, major hospital infections University and high rates of readmission. We have designed a large record linkage study across England and Wales to gain important new insight into factors that are, increasingly, thought to have a possible influence on mortality after emergency
hospitalisation. For example, we will establish the extent to which outcomes vary according to whether people are admitted at weekends rather than during the week, according to the size of the hospital, and the specialist services provided by the hospital. We will also establish whether outcomes have changed over time during following reductions in junior doctors' working hours, and determine how they vary according to
the condition the patient is admitted for. The study will provide important new findings that will help inform service provision.
Neath Port Talbot Air Quality Alert Pilot Project Ronan Lyons, Professor of Public 09/02/2012 Approved 18/06/2012 Health, Co director of Health Information Research Unit, College of Routinely collected health service utilisation data will be compared between a recruited cohort of residents with respiratory conditions who have agreed to receive air quality early warnings and a anonymous Medicine, Swansea University controls selected from the general population of asthma sufferers. Comparisons will be made, both before, during, and after a period of time when they will receive early warnings of poor air quality in terms of visits GPs, A&E attendance and hospital inpatients spells. The identities of all those in the study will be anonymised so that researchers will not be able to identify subjects.
Comparison of prescribed incontinentence drugs. Do Huw Morgan, Technical Director, GP 27/01/2012 Approved 06/03/2012 patches lead to more consistant and sustained uptake Commissioning Solutions Ltd Oxybutynin is a common type of medication prescribed for incontinence. This is commonly prescribed over a long period of time. There are recognised risks of side effects when taking this drug, and so often of the medication than tablet forms? patients get switched from one product to another over time. Eventually some patients may stop taking the drug altogether. We wish to explore if the prescribing of the drug by sticky patch reduces these problems, by comparing the prescribing patterns of the drug over time in the GP data routinely presented in SAIL. Optimal methicillin‐resistant Staphylococcus aureus Owen Bodger, Mathematical 18/01/2012 Approved 29/02/2012 (MRSA) response in Wales: a modelling framework for Modeller, Swansea University We would like to find out how length of stay in hospital is related to the prevalence of MRSA in the hospital. For this we would need to know in each of 18 main hospitals in Wales, for each patient diagnosed with healthcare associated infections MRSA the number of days they have stayed in hospital and the number of days untill diagnosis. We can then compare this data to other data we have regarding prevalence rates of MRSA, and use it to determine how well our models describe reality. Our models predict a rise in MRSA levels with longer stay in hospital on average. We would also like to see whether MRSA is spread locally (in a room or ward) or nonlocaly (by transmission through HCW. For this we would need the ward where the patient was at diagnosis and during treatment. A promise has been made to Welsh Healthcare Appointed Infection Programme (WHAIP) to not use the names of hospitals in any report or publication. Due to the need to link PEDW data back to WHAIP data, hospital names would be needed but again the promise would be that they will not be used for any final report or publication. 1 ‐ Unfortunaterly PEDW data does not make available a date of diagnosis field, therefore reporting the number of days until diagnosis will not be possible. The nearest estimate would have to come from the episod start date for each episode in a spell. Only with the start of a new consultant episode will you be able to pick up any new diagnoses in a spell of treatment 2 ‐ Ward level data is also not available in PEDW therefore an exact ward location may not be possible. Specialty of treatment may reveal which ward a patient was admitted to. This would be dependent on the hospital site of treatment and how they admit their patients.
Sian Griffiths, Consultant in Public The need for joint replacements to alleviate pain and dysfunction from hip and knee arthritis is increasing. Hip replacements are very successful but there is some variability in the success of knee replacements. 19/12/2011 Approved 04/03/2012 Using the SAIL databank, to develop a comprehensive Health Medicine, Public Health There has been an enormous growth in knee replacements. The aim of this project is to better understand trends in the need for knee joint replacement in order to improve service planning. This will be undertaken by tool for describing the epidemiology of orthopaedic Wales NHS Trust/Cardiff and Vale anonymised data linkage by studying the patterns of health service utilisation of people who have had joint replacements in the past and of those who might be suitable, such as those attending their general care pathways in order to support improvements in University Health Board practitioners with knee pain orthopaedic service design and quality, including the measurement of patient outcomes. Understanding healthcare utilisation of responders Hilary Fielder, Professor of Screening, We want to link the information routinely collected by the screening programmes to other health service data to gain new insights into reasons why people do or do not take up the offer of screening. We hope the 19/12/2011 Approved 29/02/2012 and non‐responders to the bowel screening Cardiff University results will give us ideas about how to improve both the screening programme and primary care to improve early detection of cancer. programme in Wales before and after invitation to screening in order to develop new strategies to Through improved understanding of the differences in co‐morbidities and health service usage we are also looking to inform professional training to support improved uptake and therefore to reduce inequities in improve early detection either through screening or mortality rates. opportunistic testing. Chronic obstructive pulmonary disease (COPD) for Tracy Price, Advanced Public Health 19/12/2011 Approved 29/02/2012 patients registered with Abertawe Bro Morgannwg GP Intelligence Analyst, Public Health GP practices in the ABM UHB area have been grouped into community networks. The Public Health Wales Observatory have been asked to determine COPD prevalence (using GP data), hospital admission rates, practices Wales NHS Trust mortality rates and COPD outcome measures (e.g. flu vaccination uptake) for these community networks.
Modelling risk pathways in coronary heart disease William King, WCAT Clinical Fellow It is well known that there exists an inequality in the rate of people dying as a result of coronary heart disease in Wales according to the level of deprivation of the population. Those individuals who are most 01/11/2011 Approved 12/12/2011 management: simulation of interventions to reduce in Public Health Medicine, Cardiff deprived have a substantially greater rate of death from coronary heart disease compared to those who are less deprived. It is known that at least some of this difference relates to differences in the risk factors inequality in coronary heart disease outcomes University between the groups (for example it is known that those who are more deprived tend to smoke more), and there is some historical evidence that more deprived individuals receive less healthcare for the condition in situations where they have an equivalent need for healthcare. The primary aim of this project is to understand the extent to which these different possible causes are influencing the differences in death rates between more and less deprived groups. In order to answer this question, the intention is to use the anonymised data held on people in Wales relating to their attendances in primary care, their admissions to hospital, and deaths from coronary heart diseas or from another cause, to build a computer simulation model. By modelling the data we can investigate the extent to which the observed difference in mortality relates to components of healthcare provision or differences in risk factors between deprivation groups. As a second stage of the analysis, there is potential to use the simulation model to examine the impact of different possible interventions (for example inviting a particular age group to be screened for risk factors for
developing the condition) on the inequality of mortality rates between the different deprivation groups. The contribution of HIRU to this project will be in the provision of access to data so that analysis can be carried out to provide aggregate parameters for the model. Extraction of the data will require linking of general practice data held in SAIL, PEDW data, and mortality data. There will be no use of individual or practice level data in the project beyond carrying out analysis to produce risk models of progression through the coronary heart disease pathway. Such risk model information will be aggregate, and as a consequence it will not possible to identify any individual from it.
Predictive risk stratification: impact on care for people Helen Snooks, Professor of Health The PRISMATIC research study is looking at the impact of a predictive tool available to general practices. Commissioned by the Welsh Government, the Prism tool assesses patients' risk of having an emergency 01/11/2011 Approved 26/04/2012 with or at risk of chronic conditions Services Research, Swansea hospital admission in the following 12 months, providing a score out of 100 for every patient (with 100 being the highest possible risk). The tool collates data from 37 primary care, hospital care and demographic University variables to provide a secure web‐based view for approved GP practice staff. These include the number of medications that a patient has received, the number of outpatient appointments over the past year, and the patient's age. Prism has been available and used in a number of practices across Wales since 2009, and appropriate governance and data security measures have been put in place by the NHS Wales Informatics
Service. Prism is set to be made available in the ABMU area in late 2011, and the research team have received funding from the Department of Health, to describe the introduction of the predictive model and to estimate its effects on the care of, and resources used, by patients at high estimated risk of emergency hospital admission, and those at lower estimated risk. As part of the study it is planned to collect Prism data from participating GP practices at three different points in time. At the same points in time, patients from participating practies will be sent questoinnaires about their health and health service usage. In order to securely, and anonymously link patient data from the Prism dataset and questionnaire data from consenting patients, we plan to embed the study data into SAIL. The Prism and questionnaire data will be linked using
Anonymous Linking Fields (ALFs) ensuring that the analysis will be undertaken on consolidated, yet anonymised data. Using SAIL also provides an important opportunity to validate data from patient recall with data
recorded in the SAIL databank, in relation to service use for example. Note, however, that this latter validation exercise is not part of the current project plan.
International Meta‐Analysis of Vasovasostomy and Professor Ronan Lyons, Professor of There are plausible biological mechanisms for the theory that vasectomy might cause prostate cancer (see attached paper for more details). The Western Australia Record Linkage Project has investigated this 01/11/2011 Approved 12/12/2011 Prostate Cancer in Vasectomised Men Public Health, CHIRAL, College of relationship with their data and the results were inconclusive due to the small number of vasectomy reversals and the relative lack of observed person‐time at risk in older men where prostate cancer risk becomes Medicine, Swansea University more frequent. As a member of the IHDLN (International Health Data Linkage Network), HIRU would like to explore this relationship with the SAIL data and pool the anonymised data to perform a large international meta‐analysis of vasectomy reversals and prostate cancer in vasectomised men. Risk factor assessment for prevention of premature Professor Julian Halcox, Professor of The aim of our study is to use the powerful SAIL database to explore how often CVD risk factors are NOT formally assessed and managed in patients presenting with heart attacks and angina at a young age. These 29/09/2011 Approved 12/12/2011 CVD in clinical practice Cardiology, Director Cardiovascular data will be used in the future to identify and improve the detection and treatment of those at risk, to assist the development of strategies to reduce the burden of premature CVD as a major health problem in Wales. Research Group Cymru, Cardiff University School of Medicine
GFL AHI Proof of Concept Kevin Curtis, Managing Director, To use routinely collected data available within the SAIL databank covering both primary and secondary care to achieve three things: 08/08/2011 Approved 28/09/2011 Applied Health Intelligence Ltd. 1) Understand how well our existing software and models that originated in Germany will work with Welsh data 2) Stratify the available population data by using a clinical classification system to provide homogeneous groups based upon morbidity and prevalence 3) Accurately predict health events for anonymised individuals using longitudinal data
Real‐world Cholesterol Monitoring and Control: A Ashley Cagar, Associate Project 08/08/2011 Approved 28/09/2011 Unique Study Involving EMR (SAIL) + PRO (MediGuard) Manager, MediGuard/Quintiles MediGuard will identify members with suspected high cholesterol and invite them to participate in the study. Interested subjects will be screened through a web‐based survey tool; eligible and interested subjects will continue forward to complete an on‐line questionnaire that contains questions related to: disease history; most recent cholesterol results; treatment satisfaction; adherence; health resource use; and demographics. Finally, patients will provide identifying information and an Electronic Signature of consent for their anonymised medical data to be utilised. Subjects will be compensated £20 upon receipt of the completed on‐line survey and electronic consent. The resulting MediGuard dataset of consenting participants will be processed as a normal SAIL dataset (i.e., identifiable data will be sent to NWIS to be put through the ALF creation process). Participant response data will be processed into SAIL, and linked to the ALFs provided by NWIS, creating an anonymised linked dataset that includes routinely collected information from GP practice data. The latest available cholesterol tests and cholesterol results will be accessed by the SAIL analyst, to produce the specialised linked subset of data for statistical analysis. Analysis will be performed on the anonymised linked dataset accessed within the SAIL gateway at Swansea University. Results will be disseminated in conference posters/presentations and manuscripts. Care will be taken to ensure sufficient aggregation of results to make potential disclosure of individual level data impossible. As a secondary stage, the result dataset may be examined for the feasibility of further stratification, e.g by demographic group, disease type, etc.
Health impact, and economic value, of Ronan Lyons, Professor of Public This is an NIHR funded evaluation of a natural experiment. The natural experiment is the substantial improvement of housing quality in 9,256 social housing homes in Carmarthenshire over a 5 year period. The 27/07/2011 Approved 28/09/2011 meeting housing quality standards Health, Swansea University evaluation aims to quantify the health and educational benefits of improving housing conditions. Using the standard SAIL methodologies of splitting files and multiple encryptions of individual, household and small area codes means that it is possible to track the health of residents without ever knowing who they are or where they live. This will allow us to quantify the expected benefits of living in homes meeting 21st century quality standards. Potential benefits include fewer deaths, heart attacks, improved mental health, and possibly improved educational attainment in children brought up in warmer homes. Injuries may be increased (unfamiliar lay outs) or decreased (fewer hazards). The health of residents before and after the intervention will be tracked and compared with the general population and a comparator group of social housing in Swansea Change in alcohol outlet density and alcohol‐related David Fone, Professor of Health This is an evaluation of a natural experiment in the density of alcohol outlets. The hypothesis is that reduced accessibility is associated with less consumption and lower rates of alcohol related health problems. It 27/07/2011 Approved 28/09/2011 harm to population health Sciences Research, Cardiff University is essentially an ecological study in which different area based measures of exposure are related to a series of potential health and social effects, including hospital admissions for diagnoses due to alcohol, evening and weekend emergency department attendances, alcohol consumption measured by the Welsh Health Surveys and alcohol or violence related crime obtianed from the police. A list of all premises selling
alcohol between 2005 and 2010 will be used to produce density maps at LSOA level. Network analysis of distance between all households and premises in Wales will be undertaken. All household level data will be anonymised using the RALF system operated by NWIS and all small areas (LSOAs) encrypted using the now standard SAIL approach .This system ensures that there is no risk of identifying any individual or household and now even the small areas are anonymised.
An analysis of the clinical and environmental factors Dr Steve Luzio, Senior Lecturer, Type 1 diabetes is a form of diabetes most commonly diagnosed in childhood that requires insulin injections for life. Careful control of blood sugar levels can reduce the risk of long term problems such as eye and 18/07/2011 Approved 28/09/2011 associated with onset and outcome in childhood type Swansea University kidney disease, but most patients do not manage to achieve this level of control. The Brecon group of paediatricians have kept records of all children newly‐diagnosed with type 1 diabetes in Wales, dating back to 1 diabetes: an epidemiological analysis from the 1995. This group (the Brecon Database) now comprises around 2500 children and young adults. Here we propose to link this dataset anonymously to GP and other health records. By studying records of insulin and
Brecon Cohort blood testing kit use, average sugar levels and clinic attendances, linked to data on adverse outcomes such as hospital admissions,and eye problems, we can begin to understand what it is that makes good control difficult in diabetes and which patients are most at risk. Predicting the impact of HPV vaccination on cervical Dr Shantini Paranjothy, Clinical Senior In Wales and Scotland screening is initiated at age 20 years and women offered the HPV vaccine through the catch‐up programme in Jan 2009 will be eligible for routine cervical screening from Jan 2011. We intend 07/06/2011 Approved 28/09/2011 cancer: the interaction with screening Lecturer, Cardiff University to perform a record linkage study in Wales to ascertain the uptake of screening in vaccinated cohorts. Vaccination status is recorded on the National Community Child Health Database (NCCHD), screening uptake and clinical results are held by Screening Division, Public Health Wales, NHS Trust. The Cervical Screening Wales (CSW) programme database contains records of all women who attended for cervical screening in Wales from 1999 onwards. It contains demographic information, NHS number, dates of cervical smears and results. The NCCHD consists of anonymised records of children who (since 1987) have been born, treated or resident in Wales. Overall data items are 95% complete and these data are used to monitor vaccination uptake. We will use this database to ascertain the cohort of women who were eligible for the catch‐up vaccination campaign and their first cervical smear between 2011 and 2012 inclusive
(age 20‐22 years, year of birth 1990 or later). These data will be record‐linked to the CSW database to provide screening uptake outcome, using the Secure Anonymised Information Linkage (SAIL) research tool.
The NHS administrative register will be used to ascertain any deaths or migration out of Wales during the study period. The Welsh Index of Multiple Deprivation (at LSOA level), as recorded on the NCCHD, and Townsend score (at LSOA level) will be used as a proxy measure for social deprivation.
Chronic obstructive pulmonary disease (COPD) Rhys Gibbon & Tracy Price, GP practices in the ABM UHB area have been grouped into community networks. The Public Health Wales Observatory have been asked to determine COPD prevalence (using GP data), hospital admission rates and 13/05/2011 Approved 08/06/2011 prevalence, hospital admissions and deaths for Advanced Public Health Information mortality rates for these community networks. patients registered with Abertawe Bro Morgannwg Analysts, Public Health Wales NHS health board. Trust Proton pump inhibitors as a risk factor for Sinead Brophy, Senior Lecturer, 13/05/2011 Approved 01/07/2011 campylobacter and other gastroenteitis infections – a Swansea University Currently several epidemiologic studies show that an increased risk of infections in hospitalised patients and in the general population are associated with the use of gastric acid–suppressing drugs. However, cohort study using routine data. conflicting data have been produced on the risk of bacterial infections associated with the use of acid‐suppressing drugs. No research has evaluated the incidence of gastroenteritis (GE) among patients who take PPIs. Further research is required to more fully quantify the risk of enteric infections associated with the use of acid‐suppressing drugs considering the evaluation of the incidence of GE. We thus assessed (addressing the incidence of GE among patients) whether the use of gastric acid–suppressant drugs is associated with an increased risk of bacterial enteral infections.
We would like to examine the incidence rate over a 12 month period before and after the prescription of PPI compared to controls.
To examine affect of mothers health on birth outcome Dr Sinead Brophy, Senior Lecturer, 11/05/2011 Approved 08/08/2011 and health service usage Swansea University We would like to embed the study data into SAIL in order to give long term follow‐up of the study participants. This will include validating data from patient recall and reducing the questionnaire burden by using routine data as much as possible. For example, we would like to examine treatment for depression in pregnancy and the affect this has on the weight and gestional age of the baby and the future health of the infant.
Establishing the prevalence of certain rare Ronan Lyons, Professor of Public Long term conditions that are managed without regular hospital attendance tend to be recorded in GP systems. The condition may be noted by a GP at any point in the history of the patient, so that identification of 21/04/2011 Approved 08/06/2011 neurological conditions as recorded in Primary Care Health, Co director of Health patients requires searching the data over a long time span. The electronic recording of GP practice records has gone from limited recording to substantial recording from the late 1980s to the present day, during whi using the SAIL databank Information Research Unit, College of time governmental initiatives have altered coding priorities. Diagnoses may be subject to general refinement or review over time, and for individual patients as they present to different GPs over time, Patients move Medicine, Swansea University and register with different GPs. GPs may retire and the patient registration is transferred. All these things lead to complications in the analysis of such data. Using a set of hereditary neurological conditions as an example, the study aims to document and address complications with an aim of establishing a robust methodology for completing such investigations.
Multiple Sclerosis Case Ascertainment by querying the Inocencio Maramba, Quantitative An algorithm will be developed to ascertain the cases of Multiple Sclerosis from the records in the GP and Patient Episode Database Wales (PEDW) databases, as well as from the ONS Mortality records for Wales. 16/03/2011 Approved 18/04/2011 SAIL Databank Researcher, HIRU, Swansea The demographic characteristics of the records People With MS (PWMS) found through this method will be summarised and analysed. University An anonymised PWMS cohort will be assembled from the extracted records and will be analysed for the following: • Age and sex distribution • Geographic distribution • Size estimation through the capture‐recapture method • Number of mortalities in the cohort and demographic characteristics of the mortalities • Other analyses such as length of stay in hospital, most common procedures, etc, may also be carried out.
Prevalence and prevention of obesity and other risk Sinead Brophy, Senior Lecturer, This study examines growth rates and attendance at the GP for children who are categorised as high risk for diabetes and cardiovascular disease (CVD) compared to those categorised as low risk. This could help to 16/03/2011 Approved 14/04/2011 factors for CVD among young people with a focus on Swansea University identify early predictors of obesity and risk factors for CVD among children. high risk ethnic minority groups. Creation of a diabetic pregnancy cohort linked to CARI Sue Jordan, Reader, Swansea A number of drugs are widely prescribed in pregnancy however, it is not known whether som eof these drugs may harm the fetus. A variety of complementary approaches, including large multicenter record linkage 16/03/2011 Approved 16/05/2011 and community child health data University studies, are needed to evaluate safety of medicine use in pregnancy. To evaluate safety in relation to teratogenicity (capacity to cause malformations), population‐based congenital anomaly registers, which are already networked across Europe (EUROCAT) with a common database, can provide a cost‐effective mechanism which is as yet under exploited. The enormous population coverage of registers when combined gives sufficient statistical power for the identification of associations between specific drugs and specific malformations. This sub‐project will analyse a linked anonymised database of diabetic and non diabetic pregnancies with health and birth outcomes, in relation to the use of insulin analogs (exposure to which is increasing in the pregnant population), to determine whether such drugs have any adverse outcomes during pregnancy.
EUROmediCAT: Safety of Medication use in Pregnancy Sue Jordan, Reader, Swansea A number of drugs are widely prescribed in pregnancy however, it is not known whether these drugs may harm your unborn baby. A variety of complementary approaches, including large multicenter record linkage 11/01/2011 Approved 25/02/2011 in Relation to Risk of Congenital Malformations. University studies, are needed to evaluate safety of medicine use in pregnancy. To evaluate safety in relation to teratogenicity (capacity to cause malformations), population‐based congenital anomaly registers, which are Framework 7 project no 260598 already networked across Europe (EUROCAT) with a common database, can provide a cost‐effective mechanism which is as yet underexploited. The enormous population coverage of registers when combined gives
sufficient statistical power for the identification of associations between specific drugs and specific malformations. This project will develop and test an efficient pharmacovigilance system for safety of drugs during
pregnancy in relation to teratogenicity by (i) enhancing the information regarding drug exposure in the EUROCAT database, covering a total population of 3 million births 1995‐2010, through linkage to electronic
databases containing prescription information, and by linkage to chronic disease cohorts (ii) analysing the enhanced EUROCAT database in relation to four drug groups of public health concern ‐ new antiepileptics, insulin analogs, SSRI antidepressants, and antiasthmatics ‐ exposure to all of which is increasing in the pregnant population (iii) interrogating health care databases to monitor the effectiveness of drug safety
recommendations and pregnancy prevention programmes through drug utilisation studies, and to provide an exposure profile for pregnant women (iv) (not involving HIRU) conducting a scoping study of the implications for drug safety of growing internet use by pregnant women, in terms of access to safety information about teratogenicity, and access to drugs with teratogenic potential. Do children who move frequently have poorer health Hayley Hutchings, Senior Lecturer, House moves are regarded as a stressful life event. Findings from a previous study 11/01/2011 Approved 13/04/2011 and educational outcomes? School of Medicine, Swansea from Swansea highlighted that children who presented with a burn before the age of University three were more likely to move house before their sixth birthday compared to unburned control children. Another recent study has also suggested that children who move frequently are more likely to commit suicide. There has been little research however that has examined the effect of frequency of house moves on the long term health or educational outcomes of children. The proposed study will examine anonymised child health records within the Wales Electronic Cohort for Children to determine any association between the frequency of house moves and hospital admissions, outpatient visits, vaccination status and educational achievement (teacher assessment at key stage 1), having adjusted for known risk factors for poor health outcomes such as social deprivation. The health and educational outcome data will be obtained from routinely collected datasets and record linked using anonymous Residential Anonymous Linking Fields (RALFs).
The Epidemiology of Open Tibial Fractures in Wales Ian Pallister, Reader in Trauma & Fractures of the tibia (shin bone) are particularly severe injuries when there is an open wound in the overlying skin. Combined surgery from specialised orthopaedic & plastic surgeons is required. This will mean that 04/11/2010 Approved 18/02/2011 Orthopaedics, Morriston patients are often transferred to the hospital with these specialist services. Recently published standards of care aim to improve the outcome for patients by encouraging transfers to occur as soon as possible. Hospital/School of Medicine However, our preliminary data shows that a significant number of patients remain under the care of the admitting hospital, and are only transferred once problems occur, sometimes weeks or months after injury. Presently it is impossible to know how large a proportion of patients sustaining open tibial fractures receive ortho‐plastic care in the intended manner. The long term fate of patients who are never transferred remains unknown. This study will give us accurate numbers of recorded open tibial fractures for South Wales, and will also give us valuable information about the follow‐up needed during the healing process. These data will help us develop the service we provide, and hence improve the care patients receive.
The Epidemiology of Body Mass Index in the Welsh Joanne Demmler, Research Assistant, 19/10/2010 Approved 07/02/2011 SAIL Primary Care Database HIRU, Swansea University Obesity presents a major public health challenge to Wales over the coming decade and has been identified as a high priority area for health improvement by the Welsh Assembly Government. There is relatively little published literature on the epidemiology of obesity, particularly in relation to the Welsh population, and there is particularly poor information on trends in the severely obese (BMI of over 50 kg/m2).
The aim of the study is to examine: The project aims to: 1. Measure the completeness and validity of obesity data in the SAIL database 2. Measure changes in the pattern of severe obesity over time 3. Examine the association between obesity and co‐morbidity 4. assess the effect of obesity medication on patients with a high BMI
The SAIL database contains records from around 120 GP practices and is an idea source of information for such a study.It may be that the quality and completeness of data varies by practice or computer system and this project aims to assess whther the data can be used for this purpose. Outputs will include the number of practices with high quality data and trend data for BMI by age/gender groups.
Developing new risk models to improve targeting of Samantha Turner, Researcher As part of a major initiative to reduce the number of fire related deaths and injuries in the home, FRSs throughout the UK have been offering free Home Fire Safety Checkss. 19/10/2010 Approved 07/02/2011 the Fire and Rescue Services (FRS) Home Fire Safety Assistant/PhD Student, School of Although FRSs in Wales have developed area and household level risk models to target their free HFSC initiative towards higher risk households, many house fires still occur outside of these targeted areas. Check (HFSC) Medicine, Swansea University We propose to work with the FRS to help them improve their current risk models. By linking anonymised data from the FRS to anonymised health, education and environmental data stored within the SAIL databank, we will investigate whether certain health or environmental exposures (for example, having a smoker reside in the household, or people with chronic conditions, or those living alone) are at appreciably higher risk of
house fires. The SAIL models cannot be used by the FRSs in an operational sense due to the anonymisation but if high risk conditions are identified we will be able to advise the FRSs which organisations they need
to work with to help them market HFSCs to those at highest risk.
An e‐cohort study of serious adverse outcomes from Dr Laura Shine, ST4 Old Age There has been concern for some years that antipsychotic drug use in older people with dementia is associated with poor outcomes and increased mortality. A recent Department of Health commissioned review 19/10/2010 Approved 09/02/2011 antipsychotic drug use in older people with dementia Psychiatry, Abertawe Bro Morgannwy highlighted the difficulties, emphasising that these medications are frequently inappropriately prescribed to treat behavioural symptoms of dementia. The report suggests that up to 1,800 extra deaths per year in the in the community and care homes. University Health Board UK are attributable to these medications. This may be a conservative estimate as it is based on an extrapolation of short‐term drug trial data that does not reflect true prescribing patterns.
Our study involves a large‐scale epidemiological review of the relationship between these medications and adverse outcomes for people with dementia. SAIL is an anonymised electronic database that allows
researchers to anonymously identify patient groups and then investigate relevant outcomes. Due to the large amount of patient data held it allows easy access to produce large scale studies. Patients diagnosed with dementia and prescribed antipsychotic medication can be identified and then reviewed over a specified time period to investigate potential tadverse events.
This project will give us invaluable baseline information. If the use of these medications is prevalent and associated with significant adverse effects the next steps would be to develop a large scale intervention project to see whether we could reduce antipsychotic use in this vulnerable patient group in an acceptable manner. Population based ankylosing spondylitis cohort (PAS) Stefan Siebert, Senior Lecturer and 19/10/2010 Approved 07/02/2011 Consultant Rheumatologist, Swansea This proposal is for a population based cohort of people with AS, utilsing disease‐specific data (rheumatology database, radiology imaging) and patient completed data, linked with existing electronic data from University and ABM University Trust clinical, laboratory and administrative systems held by HIRU. All patients with a diagnosis of AS living in Wales will be approached to participate in the cohort and give a well characterised cohort with robust clinical data as its base. Question to be asked of the cohort will include: risk of adverse events such as myocardial infarction associated with AS, cost/burden of AS (cost of AS to the NHS and the patient), factors associated with a severe outcome (i.e. early life factors associated with need for surgery and/or need for anti‐TNF medication). Retinopathy Screening Dr Steve Luzio, Senior Diabetic Retinopathy (DR) (damage to the back of the eye) is a major complication of diabetes, and a leading cause of blindness in the working age population. Screening for DR ensures early detection when treatme 19/10/2010 Approved 13/04/2011 Lecturer/Portfolio Development is most effective. Current recommendations are that screening for diabetic retinopathy should be carried out on an annual basis in all patients with diabetes. In the retinopathy screening database all patients have Manager DRNW, Cardiff University existing eye photographs and basic demographic data (age, sex, type and duration of diabetes). Data available via the HIRU databases will include more information on these patients (e.g. blood pressure, glycaemic control, treatment, etc). From this data we aim to devise a method, using modelling and statistical techniques, to predict patients who will or will not develop diabetic retinopathy. The health economic benefits of stratified screening will also be ascertained. The objective therefore is to determine a safe and more cost effective program of screening for diabetic retinopathy.
Diabetes Cohort Helen Davies, Senior Research Currently there are no published clinical guidelines to help health professionals manage and advise patients with latent autoimmune diabetes in adults (LADA). It is not clear within the literature what happens to the 13/10/2010 Approved 07/02/2011 Officer, Swansea University patient (patient pathway) once they have been classified as having autoimmune diabetes. For example;
(i) What treatments are prescribed in LADA and whether these treatments are adequate in maintaining blood glucose levels. (ii) Treatment by length of time to auto‐antibody testing. (iii) The complications of diabetes incurred in LADA compared with other types of diabetes. (iv) The value of testing Thyroid function and other tests in LADA compared with other types of diabetes. (v) Explore other autoimmune conditions in LADA compared with other types of diabetes.
We would like to examine the patient pathway for autoimmune diabetes with a view to increasing our understanding of this condition and ultimately to inform and improve patient care. This project will involve a da analyst (Mark Atkinson) examining multiple data sources (outlined in section 9s) within health.
SAFER (Support and Assessment for Fall Emergency Helen Snooks, Professor of 13/10/2010 Approved 11/02/2011 Referrals) Trial Healthcare Research, Swansea The project involving HIRU is aimed at following up study patients who have had contact with an emergency department or been admitted as an inpatient, or have died, since their initial 999 call and inclusion in the University study. This will be tracked and linked anonymously through SAIL, and output for analysis in SPSS.
Dermatology service development for Hywel Dda Dr Emma Vesey, Hywel Dda Health The HIRU data will help establish the burden of skin disease in West Wales (Carmarthenshire, Pembrokeshire, Ceredigion). In addition it will be used to determine possible associations between type of skin 04/08/2010 Approved 26/08/2010 Health Board based on a healthcare needs Board disease and healthcare usage and area of residence, age, gender and level of deprivation. This information will then be used to better inform the development of dermatology services and resource allocation for the assessment Hywel Dda area. Risk of hospital admission and cause‐specific Professor David Fone, Professor of Anonymously linking the CHSNS data to SAIL data will allow us to investigate whether common mental disorders are associated with admission to hospital and re‐admissions and cause‐specific mortality. We will 17/06/2010 Approved 16/07/2010 mortality associated with common mental disorder: Health Sciences Research, Cardiff examine whether these associations vary according to individual and neighbourhood social cohesion, household and area income, and socio‐demographic characteristics (e.g. age, sex, socioeconomic status, the Caerphilly Health and Social Needs electronic University educational achievement, employment, housing tenure). We will account for differences in illnesses at baseline (e.g. Physical and Mental Health), diabetes, arthritis, back pain, cardiovascular disease, cancer, cohort study. respiratory diseases) and lifestyle factors (e.g. smoking, alcohol consumption, exercise, diet) in increasing the risk of hospital admission and mortality.
Adiponectin Levels in People with Latent Autoimmune Sinead Brophy, Senior Lecturer, 28/05/2010 Withdrawn/Not Approved N/A Diabetes – A Case Control Study. Swansea University The project team have conducted research investigating the adiponectin levels of patients with LADA matched with type 2 diabetic and non‐diabetic controls. The work has been submitted to a peer reviewed journal
and received a recommendation that the analysis should be controlled for BMI (Body Mass Index) as a cofounder. The project team do not have BMI data for the non‐diabetic controls and would like to explore the potential of using SAIL data to supplement their study data. The project team has a list of all the NHS numbers for all of the participants and will securely send this to HSW for anonymisation. Once anonymised the data will be loaded into SAIL and for each of these anonymous individuals the GP data will be searched for weights, height & BMI measurements. These measurements will then be appended to the original study
dataset (within the SAIL environment) and the statistical analysis will be repeated including the new BMI data as a cofounder.
Using anonymised routine data to Investigate the Mick Dennis, Reader in Old Age In Wales suicide is one of the leading causes of death among young people. There is growing recognition that up to 5% of suicides may occur in clusters (i.e. close together in time/space) in people under 25 years. 28/05/2010 Approved 16/07/2010 temporo‐spatial patterns of suicide and other causes Liaison Psychiatry, School of Our understanding of what triggers a suicide cluster, what influences its continuation and eventual cessation, and, critically, what advice to give to communities experiencing a suicide cluster, is limited. In of death in Wales 2003 to 2008 Medicine, Swansea University 2007/08 a cluster of approximately 20 suicides occurred among young people living in the Bridgend area. In order to establish the context of the Bridgend suicide cluster we need to have a clear picture of the
temporo‐spatial patterns of suicide and other causes of death in Wales for the period before and during the cluster. The routine data on cause of death held in the SAIL data bank will allow us to do this.
MINISAIL: Using routine data to estimate service need Dr Ann John, StR Public Health, Hon Common mental disorders (CMD), such as depressive and anxiety disorders, have serious costs to the individual and society. They account for over 30% of work days lost due to ill health and 20% of visits to the GP. 05/05/2010 Approved 16/07/2010 for Common Mental Disorders in Primary Care Lecturer Public Mental Health, School At any time around 30% of people are affected. of Medicine, Swansea University We have demonstrated in publications from an earlier WORD‐funded project that the Mental Illness Needs Index (MINI), a score based on local area Census data collected every decade, is associated with the
prevalence of CMD. We now plan to look at the relationship between levels of CMD in Wales, assessed using general practice diagnosis codes from the Secure Anonymised Information Linkage (SAIL) project, and a
MINI‐like score (MINISAIL), calculated from data more regularly available than Census data. We will then be able to compare actual levels of CMD currently diagnosed in general practice to levels predicted using MINISAIL scores, based on routine data collected by the SAIL project. This will highlight possible areas of unmet need.
Decisions for the allocation of resources can then be made in a more responsive, open and fair way to reduce variations in services, address differences in access to care and ensure those suffering with these chronic diseases are recognised when planning healthcare. INTEGRIS: Integration of European Injury Statistics Professor Ronan Lyons, Professor of Injuries are an important cause of disability and death in Europe. The European Injury Database (IDB) was implemented in 1999 and currently involves 12 member states. IDB is a standardised reporting system 22/03/2010 Withdrawn/Not Approved N/A Public Health, School of Medicine, implemented in accident and emergency departments in a sample of hospitals. Swansea University Different types of data are available in the IDB and in the routinely collected hospital discharge registers (HDR). Although HDRs may not always have all the desired detail, they are implemented in all countries. INTEGRIS aims to link these two datasets to improve detail on the nature of injuries for admitted patients. This is pilot study towards a more comprehensive Europe‐wide collection of anonymised injury surveillance data to support the development and evaluation of policies to prevent injuries. EU member states have accepted the World Health Organization and European Commision recommendations on improving injury surveillance.
In Wales, PEDW is the hospital discharge register. By linking this to EDDS, the new Emergency Department Data Set, the consequences of injuries in terms of severity and the need for treatment can be better judged. Approximately 10‐20% of hospital admissions are the result of injuries. This use of hospital data will play a valuable part of documenting the consequences of injury for individuals, society and the NHS. It willalso act as a stimulus to the EU and member states to support the development and evaluation of injury prevention policies and initiatives. Before the creation of the SAIL system separate anonymised inpatient and emergency department data were provided to the IDB project. Wales has historically provided the UK contribution to injury surveillance. This proposal requests permission to use SAIL to provide these data. Data from up to three hospitals will be provided in the first instance, following Caldicott Guardian and R+D approval from each relevant health board. Such approvals have already been provided from ABMU HB and are being requested from two others.
Biomedical Informatics in Biomarker Discovery and Professor John White, Professor of This project seeks to develop, and establish the potential of, a secure controlled‐access web‐based informatics platform in the study of biomarkers. There are four research areas in this study, with their focus on biomarkers for: 10/03/2010 Approved 07/05/2010 Validation Reproductive Biology, Swansea Prostate cancer (SD) Barrett’s oesophagus (GJ) Universiy Lung cancer (PL)
Female infertility and endometriosis (JW)
(Initials of authorised researchers as per question 1 above.)
The authorised researchers will use the platform to query SAIL for records of interest to their area of study. By working in collaboration with the Pathology Department at Abertawe BroMorgannwg University Health Board, NHS Pathology staff will action
requests from researchers and retrieve tissue specimens relating to these records from the archive. Tissue specimens each have a unique number that does not relate to an individual, but only to the specimen. This means that anonymity is not
compromised when researchers see specimen numbers. Pathology staff will examine the specimens and analyse Tissue Micro Arrays (TMAs) thus creating a results dataset. Results datasets will be uploaded onto SAIL and linkage will be made to Pathology
data using the tissue specimen number (present in both the research dataset and the Pathology data) as the unique key. Further linkage to additional datasets will be made as required by means of the ALF_E as per usual within SAIL.
If there are additional research datasets of interest, pertaining to particular studies with ethical approval and participant consent (as appropriate), these will be incorporated into SAIL using the standard HIRU split‐file method. In this way, demographic
data (File_1) will be sent to HSW for matching, anonymisation and the assignment of an ALF to each individual, to create the File_3 which comes to HIRU. The experimental results comprise File_2 which come to HIRU directly. The combination of File_2
and File_3 (via a system linking field retained in both parts when the data file is split) creates an anonymised version of the original dataset.
Through these methods, anonymised research data gained through laboratory experiments will be uploaded onto the platform. The platform will provide the facility for researchers to query the SAIL databank for data of interest to these conditions,
along with their research data. This will provide richly enhanced datasets for the study of potential biomarkers and the platform will include software for data analysis. Anonymous linkage at the individual level will operate according to the methods
operating in the SAIL system. The platform will conform to the stringent security arrangements that have been developed as part of the SAIL system.
The Research Assistant will access the data purely for the methodological development of the platform and will not be engaged in producing research results. The four project leads (SD, GJ, PL and JW) will be engaged in the development of the
platform to ensure it meets their needs, but will not have access to any data until the platform is ready to be implemented. Once it is implemented, it will operate on a strict user‐only account basis with a time‐limited access agreement so that the
platform can be evaluated. The four project leads will conduct research in their areas of interest, PG will analyse the tissue samples (TMAs, etc.) and CD will act as the liaison with the Health Board. In future, i.e. after evaluation, further use of the
platform will be subject to project‐level applications to the IGRP, and submissions for research ethics review where applicable.
Environmental and socio‐demographic factors Dr Stephen E Roberts, Senior Lecturer The SAIL data will allow us to investigate the hospitalised incidence (per 100 000 population) and case fatality at 30 days for serious asthma overall across Wales and according to socio‐demographic factors such 04/03/2010 Approved 26/04/2010 associated with the incidence and outcome of serious in Epidemiology & Public Health, as social deprivation quintile, age group, sex, urban v rural residence, background air pollution, size of hospital, and distance to hospital. asthma School of Medicine, Swansea University
Outcome measures from routine administrative data Dr Stephen E Roberts, Senior Lecturer Routine administrative inpatient and mortality held in SAIL will allow us to produce outcomes measures for all acute medical admissions ‐ and for admissions for several subgroup conditions ‐ at the hospital level 04/03/2010 Approved 30/04/2010 in Epidemiology & Public Health, across hospitals in England and Wales during a recent 12 month period (1st June 2008 to 31st May 2009). School of Medicine, Swansea The subgroup conditions are COPD, MI, stroke, pulmonary embolism, DVT, cellulitis, and GI bleeding. University The outcome measures would be length of hospital stay, mortality rates (irespective of cause of death), readmission rates (irrespective of cause of readmission) and recorded c.difficile infection rates. They would be obtained via HES for England and PEDW and NHSAR for Wales. These crude, basic outcome measures will be at the hospital level not at the patient level.
They will be used to investigate correlations with data at hospital level on consultant input and service configuration, obtained via an online survey of acute care hospitals across England and Wales. The data will
be anonymised and no hospitals would be named in any report. There will be no comparisons performed between England and Wales hospitals.
Examining the relationship between antibiotic Rebecca Cannings‐John, Statistician, Where antibiotic prescribing has reduced in general practice, there is concern that it may have fallen too much, putting patients at greater risk from complications from infections. To our knowledge no studies have 01/12/2009 Approved 25/01/2010 prescribing and serious or complicated infections from Cardiff University measured in patients presenting to practices with an respiratory tract infection (RTI), individual exposure to antibiotics and serious or complicated common infections presenting in secondary care as well as general respiratory tract infections: an individual level practice, therefore underetimating the number of complications. analysis. Using individual data arising from general practices, we plan to explore this relationship, with the aim of understanding whether patients who do not receive antibiotics for a RTI have a higher risk of developing a serious or complicated common infection (in general practice or in secondary care). We further plan to explore the impact that other factors may have on the occurrence of a serious or complicated common infection,
such as prior antibiotic use and patient characteristics such as co‐morbidity. Development & Evaluation of Prognostic Models in Dr Paul Lewis, Senior Lecturer Submit dataset for breast cancer cohort to HIRU. Dataset will include patients who have been diagnosed for breast cancer in the 1990’s.Variables in dataset will include patient forename, surname, NHS number, 01/12/2009 Approved 19/02/2010 Breast Cancer Biomedical Informatics, Swansea tumour grade, number of nodes positive, tumour size. This data has been collated and prepared by Dr Paul Lewis’ group. University 2. Request for anonymised data for all patients: date of death and cause of death if available. 3. Apply prognostic models to dataset. 4. Evaluate models for sensitivity and specificity using R The Epidemiology of Harmful Algal Blooms and the Stephanie Hinders, PhD Student, 01/12/2009 Approved 19/02/2010 Future Implications of Climate Change School of Medicine, Swansea The aim of the main project is to firstly asses the current state of the epidemiology of Harmful Algal Blooms in the North Atlantic Waters. This will be achieved by identifying the main species, which cause shellfish University posioning. Secondly, to identify the climatic varibles that affect each of the main toxic species, and to use ecological niche modelling to predict their future status.
There are few published records of shellfish poisoning and we suspect that there is a high level of underreporting. Using the hopsital addmissions and pathology records from the HIRU dataset, we aim to quantify the frequency of incidence of shellfish posioning, and identify which toxins are the most prevlevant. This will enable us to gain a good understanding of how often cases of shellfish posioning occur in Wales and assess the efficiency of shellfish toxin monitoring system.
Prevalence and prevention of obesity and other risk Sinead Brophy, Senior Lecturer, To develop growth curves of weight and height (IOTF BMI measures) for the children surveyed in the main study. These curves can be used to examine how growth in the first year of life predicts the development of 01/12/2009 Approved 26/01/2010 factors for CVD among young people with a focus on Swansea University risk factors for CVD and T2D at age 11‐13. high risk ethnic minority groups.
Augementation of PROBAT study data with SAIL Sue Jordan, Reader, Swansea 11/11/2009 Approved 14/01/2010 University We are proposing to exploit the new research networks established by the Wales Office of Research and Development (WORD) to answer a crucial research question and address a major public health issue: allergic disease in childhood, such as asthma and eczema.
No‐one fully understands why childhood asthma and eczema have become so common. Authorities suggest that environmental changes are responsible. It has been noticed that the increase in these conditions co‐ incided with the introduction of antibiotics. Some research suggests that children who take antibiotics in their first two years are at increased risk of asthma and eczema. To explore the possible links between prescribed antibiotics and these conditions, we propose to examine the medical records, both electronic and paper, in general practice and hospital, of a cohort (group) of children who have joined our existing study. This study is investigating the use of probiotics to prevent eczema. The parents in this study have already given consent for medical records to be examined (a copy of the consent form has been submitted with this application). This work also forms an important part of the piloting and implementation of research studies that use electronic patient records in Wales. These are designed to support research into many aspects of health. This project will allow us to test the data obtained from the records with that obtained from clinical examination by consultants. We aim to determine whether: • When other risk factors are accounted for, there is both a dose‐response relationship and a biologically plausible association between antibiotics prescribed and the incidence of allergic disorders in children at 2 years of age. • The discrepancies between parents’ reports of antibiotic usage, hospital records and anonymised primary care electronic records are sufficiently small to justify collecting data on medication usage from electronic data only.
We also aim to establish a framework for prospective research in medical prescribing using anonymised data.
Is capillary glucose testing associated with improved Dr Jeffrey Stephens, Reader in 11/11/2009 Approved 06/01/2010 HBA1c? Diabetes, Consultant Physician, Controversy exists on the prescription of glucose testing strips to patients with diabetes as the strips are costly and the current management algorithms (NICE, Wales Diabetes Guidelines) are based on HBA1c. The Swansea University Swansea Diabetes Research Group (DRG) is interested in using the SAIL databank to investigate if there is an association between Hb1Ac levels and the number of glucose testing strips prescribed in the six months prior to the Hb1Ac test, for different treatment groups (diet only, those on oral agents only and those receiving insulin, which may be in combination with oral agents).
Folate Augmentation of Treatment Evaluation for Professor Keith Lloyd, Professor of Background 11/11/2009 Approved 14/01/2010 Depression (FolATED): a randomised controlled trial Psychological Medicine, Director of Recruitment to clinical trials can be challenging. The purpose of this study was to test a new method of identifying suitable participants using routinely collected data (housed in the SAIL databank) that would make the Institute of Life Science, Director it easier for practices to identify potential subjects for a clinical trial and consequently reduce their workload, whilst potentially maximising recruitment and reducing costs. of the Mental Health Research Stage 1: Network Cymru, Swansea University Methods Swansea University’s Health Information Research Unit’s Secure Anonymised Information Linkage (SAIL) database of routinely collected health records was interrogated, using Structured Query Language (SQL). The
folATED inclusion/exclusion criteria were translated into read codes which were used in the algorithm to identify suitable anonymous participants. Two independent clinicians rated the eligibility of the potential participants’ identified. Inter‐rater reliability was assessed using the kappa statistic and inter‐class correlation. Results The study population (N=37263) comprised all adults registered at five general practices in Swansea UK. Using the algorithm 867 anonymous potential participants were identified. The sensitivity and specificity results > 0.9 suggested a high degree of accuracy from the algorithm. The inter‐rater reliability results indicated strong agreement between the confirming raters. The Intra Class Correlation Coefficient (Cronbach’s
Alpha) was > 0.9, suggested excellent agreement and Kappa coefficent > 0.8; almost perfect agreement. Conclusions This proof of concept study showed that routinely collected primary care data can be used to identify anonymous potential participants for a pragmatic randomised controlled trial of folate augmentation of
antidepressant therapy for the treatment of depression. Further work will be needed to assess generalisability to other conditions and settings and the inclusion of this approach to support Electronic Enhanced Recruitment (EER).
Stage 2: As the data on SAIL can never be de‐anonymised, the next phase is a pilot project for the translation of the algorithm running on anonymised SAIL data to run on live clinical systems, where the individual physician
can generate a list of potential identifiable participants, with minimal time and effort. Thus this would reduce their workload, with the potential of maximising recruitment. The method ensures confidentiality of Hospitalised prevalence and five year mortality for Dr Stephen E Roberts, Senior Lecturer The SAIL data will allow us to investigate the hospitalised prevalence (per 100,000 population) and long term mortality overall for severe IBD across Wales from 1999 to 2007 and according to socio‐demographic 24/09/2009 Approved 28/09/2009 severe IBD in Wales in Epidemiology & Public Health, factors such as social deprivation quintile, age group, sex, urban v rural residence and distance to hospital. School of Medicine, Swansea University Incidence and case fatality for upper gastrointestinal Dr Stephen E Roberts, Senior Lecturer 21/09/2009 Approved 09/12/2009 (GI) bleed in Wales in Epidemiology & Public Health, The SAIL data will allow us to investigate the hospitalised incidence (per 100,000 population) and case fatality at 30 days overall across Wales and according to socio‐demographic factors such as social School of Medicine, Swansea deprivation quintile, age group, sex, geography, size of hospital, and day of admission (weekday v weekend) for upper GI bleed. The SAIL data will also enable investigation of any trend over time in the incidence University and case fatality for upper GI bleed in Wales from 1999 to 2007. Hospital admissions of people with diabetes in Rhys Williams, Dean of Medicine and How many admissions and how many occupied bed days in hospital are the result of diabetes as a principal diagnosis and how many have diabetes recorded as a subsidiary diagnosis, how valid are the data and 14/08/2009 Approved 22/09/2009 Wales ‐ the size of the problem Professor of Clinical Epidemiology, how have things changed since the 1980's? School of Medicine, Swansea University Non‐conveyance of 999 callers: risk, views and Mohammed Yousef Al‐Sulaiti, PhD Patient clinical records (PCRs) completed by emergency ambulance crews based at three Swansea ambulance stations for patients that they attended but did not convey between 01/07/2006 and 31/12/2006, were 24/07/2009 Approved 22/09/2009 decision making Student, CHIRAL, School of Medicine, collected (n=827) through a manual sorting process within the ambulance service. The PCRs were entered into an electronic database and uploaded to the SAIL databank through a process involving file splitting, Swansea University matching and anonymisation, enabling linkage of each PCR to other health service routine data without identification of individuals. These data relate to GP out of hours’ calls, NHS Direct calls, GP contacts, accident and emergency attendances, hospital admissions, outpatient appointments and deaths. All health service contacts made within six months of the 999 call were identified and time from call to first contact was noted. Measurement of the Global and Wales Burden of Professor Ronan Lyons, Professor of This project is the Welsh/UK component of the Global Burden of Disease international study which aims to improve measurement and understanding of the magnitude and causal factors involved in injury related 03/07/2009 Approved 16/07/2009 Injuries Public Health, Swansea University deaths. It is also an essential part of a Wales Burden of Injuries Study being carried out for the Public Health Wales NHS Trust and the Welsh Assembly Government. Current official statistics in most countries, including the UK, greatly underestimate the impact of injuries on society, as many injury related deaths are coded to unspecified causes or to other diseases. The purpose of the study is to provide better information on the size of the injury problem in the world and Wales to guide policy development on injury prevention. How are pregnancy outcomes affected by intervention Shantini Paranjothy, Senior Clinical 19/05/2009 Approved 17/07/2009 (colposcopy) following abnormal cervical smear? Lecturer and Honorary Consultant in The SAIL anonymised version of the Cervical Screening Wales database will be used to select anonymous individuals that have had a first episode colposcopy and also randomly select 2 controls (women who Public Health Medicine, Cardiff have had negative smears and no history of abnormal smears or colposcopy) match for age and GP practice. Data from April 2001 to March 2003 will be used for this project. These selected individuals will then be University and Screening Services anonymously linked to the National Community Child Health Database and the All Wales Perinatal Survey to identify key measures, such as gestational age, stillbirths and perinatal deaths. From these databases, Wales rates of preterm birth, low birth weight, perinatal mortality and time from intervention to pregnancy outcome can be calculated. The combinationof hypertensionand diabetesmarkedly increasescardiovascular riskand mortalityand bloodpressure reductionis an important Are blood pressure levels taken during a secondary Caroline Brooks, Senior Data Analyst, 19/05/2009 Approved 06/08/2009 care diabetic clinic likely to be higher than when CHIRAL,School of Medicine, Swansea target for patientswithdiabetes.In the UK the NationalInstituteof ClinicalExcellence (NICE) recommend a cut off levelof 140/80for treatment in measured in primary care? University patientswithtype 2 diabetes.Thisrequiresfrequentbloodpressuremonitoringby a reliableand accurate method. Measuringarterial bloodpressure usinga sphygmomanometer oran automatic device isprone to error. For example, itmay be influenced by a numberof variablesrelating to the patient, the observerand the surroundingenvironment.Anexample is‘WhiteCoateffect’ whichis estimatedto occur in approximately 20% of the UK primary care population. It has also been suggested that thiseffect may be more pronounced inthe secondary care hospitalsetting. Thisis particularly important when the patient’s diabetic care is shared between primary and secondary care. In ourpractice itis notunusual for the patient no hospitals would be named in any report. There will be no comparisons performed between England and Wales hospitals. ta usingthe tissuespecimennumber(presentinboththe research dataset and the Pathologydata) asthe uniquekey. Furtherlinkage to additional datasets will be made as required by means of the ALF_E as per usual within SAIL.
If there are addit Use of a patient linked data warehouse to facilitate Caroline Brooks, Senior Data Analyst, 19/05/2009 Approved 06/08/2009 diabetes trial recruitment from primary care CHIRAL,School of Medicine, Swansea Recruitmentof patients into diabetes clinical trials remains poorcomparedto recruitmentfor otherlife threateningconditions such as cancer. Within recentyears therehas been a shift of diabetes care fromthe University secondary to the primarycare setting. Of interestmuch of the recruitmentofparticipants to diabetes (and otherstudies) still occurs within secondary care. This may result in a potentialobstacle to trial recruitment.It may also be argued thatthis is a particularproblem at the presenttime, since many of the currentstudies are looking to recruit treatment‐naivesuboptimally controlledpatients with diabetes orthose freefrom macrovascular or microvascular complications. Therefore, much of the routine care is provided by the general practitioner. The increased use of electronic resources and databases in the routinemanagementof patients’ care provides an opportunity to facilitate participantrecruitment into diabetes (and other) studies. The SAIL (Secure Anonymised InformationLinkage) databank has already been used to estimate the denominatorof the numberof newly diagnosed diabetic adults forthe calculation of prevalence and incidence of latentautoimmune
diabetes in adults (LADA). Our aim was to examine the use the databank as a tool to identify potential participants for two factitious exemplar trial protocols with specific inclusion and exclusion criteria.
WECC Professor Ronan Lyons, Professor of To improve understanding of the factors underlying the development of childhood obesity and inform the development of interventions to prevent or halt the increase in obesity. 15/05/2009 Approved 06/08/2009 Public Health, CHIRAL, School of Medicine, Swansea University PsyCymru Professor Keith Lloyd, Professor of Patient databases such as PEDW and GPD contain information about patient diagnoses stored as ICD‐10 and read codes respectively. We will search these databases for the codes corresponding to psychotic 08/05/2009 Approved 04/08/2009 Psychological Medicine, Swansea illnesses to arrive at a large prevalence cohort of people with Psychosis in Wales. The annonymised linkage field or ALF in each dataset will enable us to link the people with a psychosis read code across all the University datasets. This will provide us with a large amount of data about medication, health outcomes and mortality for all people with psychosis in Wales. It would also provide the potential to incorporate a longitudinal perspective.
Import the datasets of results from consenting individuals and link it to their electronic records on the SAIL databank. This will be done adhering to HIRU's data scoping protocol. Only participants who have consented to this will be used. Address any methodological issues encountered.
3) To scope 2 main research questions:‐ a) Firstly we will investigate the feasiblity of assessing the factors associated with premature death in severe mental illness. b) Secondly we will look at associations between a range of treatment interventions and longer term outcomes.
Psychosis Diagnostic Validation Professor Keith Lloyd, Professor of 08/05/2009 Approved 05/08/2009 Psychological Medicine, Swansea Analysis: We will calculate the concurrent validity of ICD‐10 code diagnoses in PEDW and Read codes in the GP dataset by comparing them with the OPCRIT diagnoses obtained from the case notes (Craddock et al University 1996). The exercise will be carried out at Clydach War Memorial Hospital, Swansea. It would involve two Clinical Studies Officers (CRO) Cymru and Trust employees, based in the School of Medicine, Swansea manually going through a sample of patients case notes and carrying out an exercise using OPCRIT (Operational Criteria for psychotic and affective illnesses). The OPCRIT diagnostic system is a suite of computer programmes that allow data entry and generate psychiatric daignoses based on casenote material. The CROs would then record the results onto two separate encrypted, zipped files; that would be password protected. The data scoping tool protocol devloped by HIRU would be followed. File 1 would contain the Data Owners local identifying code (System ID), NHS number and personal details, such as name, address, postcode etc for each individual, but no clinical detail. They would send this file to Health Solution Wales (HSW) for matching and anonymisation. They would record the system IDs and clinical data (diagnoses ‐ translated into ICD‐10 codes), but no personal identifiers on File 2, and send to Health Informatics Research Unit (HIRU), Centre for Health Information, Research and Evaluation (CHIRAL), School of Medicine, Swanse University. HSW would validate the NHS numbers against the AR register and then attach an Anonymous Linking Field (ALF) by secure encryption, along with a geocode (geographical area). They would then strip the file of demographics and send to HIRU as File 3. HIRU would join File 3 to the clinical data in File 2 to create File 4. Further confidentiality checks (i.e. check for rare conditions) would be undertaken by HIRU and then the ALF would be encrypted again into the ALF_E, which would become File 5. Data would be loaded onto the Secure Annonymised Information Linkage (SAIL) databank according to the file transfer regulations. This new loaded dataset could then be used to compare with the existing HIRU data by matching on the ALF field. This analysis and matching process would be conducted by Joanna McGregor, Research Assistant, Mental Health Team. A comparison would then be made between the diagnosis results from the OPCRIT exercise and the diagnoses recorded in the routinely collected data already on the SAIL databank. We will be comparing diagnoses as dichotomous outcomes using Chi square and Kappa statistics with confidence intervals (Cohen 1960). To detect a statistically significant (P .05) Kappa of 0.90, with 80% power and a proportion of positive diagnoses of 0.7 assuming the two tailed null hypothesis value of Kappa to be 0.50 will require a sample size of 44 subjects for each outcome (Donner & Eliasziw 1992). We will report confidence intervals around the obtained values for Kappa.
This quality check would provide useful information for the development of e‐cohorts for future studies.
A pilot study of health and environmental data linkage Professor Ronan Lyons, Professor of UK Biobank is a major UK medical research initiative, and a registered charity in its own right, with the aim of improving the prevention, diagnosis and treatment of a wide range of serious and life‐threatening 08/04/2009 Approved 14/05/2009 for Welsh UK Biobank participants Public Health, CHIRAL, School of illnesses – including cancer, heart diseases, diabetes, arthritis and forms of dementia. To do this vital work we need your help. We are now recruiting 500,000 people aged 40‐69 from across the country to take part Medicine, Swansea University in this project. UK Biobank's Cardiff assessment centre has been opened and participants have given informed consent that their electronic health records can be used to augment the study data. The UK Biobank is currently investigating the best way to acquire electronic health records. HIRU has developed the SAIL databank that contains anonymised patient level data from routinely collected health and social care providers. This pilot study proposes to investigate the information governance, techinical and logistical issues of supplying anonymised electronic health records to a research project of participants that have given informed
consent. Free School Breakfast Initiative Data Augmentation an Professor Ronan Lyons, Professor of Anonymise components of the existing study data, incorporating this into SAIL and then linking to educational attainment data. Analysing relationships between the study data and educational attainment. 08/04/2009 Approved 14/05/2009 Analysis Public Health, CHIRAL, School of Medicine, Swansea University GeneCIS SAIL data linkage for Grand Challenge Project David Ford, University Director of 08/04/2009 Approved 14/05/2009 Health Informatics Research Labs, Clinical electronic information captured about patients with IBD (Inflammatory Bowel Disease), within gastroenterology clinics at Neath Port Talbot hospital, will be anonymised, securely transferred and loaded into CHIRAL, Swansea University the SAIL databank. These anonymous individual identifiers will be linked to existing data stored on the SAIL databank (general practice, hospital, pathology) producing a complex dataset of the IBD patients medical history. A machine learning algorithm, prepared by the Engineering department in Swansea University, will be installed on one of the secure terminals and will link to this anonymous dataset. The algorithim will
then process the data exploring patterns/clusters within the data that might be predictors to the patients future surgery requirements. Demostrating the potential of this tool.
Community Acquired UTIs and Community prescriptionKerry Bailey, SpR Public Health, ‘Resistant strains’ of bacteria are those that cannot be treated fully by the normal course of antibiotics. There has been a lot of publicity recently about certain types of resistant bacteria, for example ‘MRSA’ or 12/03/2009 Approved 25/03/2009 of Antibiotics – the association for individuals and NPHS and Swansea University ‘Methicillin Resistant Staph. Aureus’. Resistance occurs naturally through ‘mutation’ of bacteria, but can be greatly increased through prescription of antibiotics. family members This study aims to look at one type of common infection, urinary tract infection, or UTI, and then compare the prescriptions of antibiotics for people who have got a laboratory proven resistant bacteria to those who have got a none resistant or ‘sensitive’ strain. Then there will also be a comparison of all of the antibiotic prescriptions of household members in those with resistant strains compared to those who have the
‘sensitive’ type. This will tell us if there is an association between prescriptions to household members and another member of the family developing a resistant infection.
Can routinely collected health related data be used to Azzam Farroha, Research Coordinator 24/06/2008 Approved 20/07/2008 monitor the prevalence of depression and post and Postgraduate Student, Swansea traumatic stress disorders in patients admitted to University hospitals due to burn injuries? Calculating the electoral ward prevalence of asthma in Martin Heaven, Public Health The SAIL data will allow us to count up the number of persons registered by GPs as having asthma on a geographical basis according to where they live. These figures can then be used with total populations for the 17/06/2008 Approved 20/07/2008 the Swansea area using linked Primary Care data Information Manager/HIAT Team same geographical areas (electoral wards), to establish asthma rates for each area. The age groupings will be used to adjust for the effects of different population profiles in the electoral wards. As the ward Leader, Nation boundaries that we will be using are politically defined, and not related to the distribution of the disease, we may apply further statistical techniques to the answer to try and determine if the choice of boundaries has by chance, an impact upon the results. We will then try to determine if the results have been effected by other issues, such as a variation in the classification, by doctors, of persons presenting with symptoms, as being asthma or not.