NHS Digital Data Release Register - reformatted

Guy's And St Thomas' NHS Foundation Trust

Project 1 — DARS-NIC-35216-D6G6M

Opt outs honoured: Y

Sensitive: Sensitive

When: 2016/09 — 2016/11.

Repeats: One-Off

Legal basis: Section 251 approval is in place for the flow of identifiable data

Categories: Identifiable

Datasets:

  • Hospital Episode Statistics Admitted Patient Care

Benefits:

The final report will be fed back to GSTT clinicians and implemented into clinical practice. If the findings show a low relapse rate after abbreviated antibiotic treatment courses, GSTT anticipate these findings will influence UK practice and national/international guidelines. The results will be published in a peer-reviewed journal and will thus be available to reviewers planning future national or international guidance documents. The consequences of a shorter-than-recommended courses of antibiotics are reduced costs, less generation of antimicrobial resistance, and less antibiotic side effects. The patient benefit will be improved survival combined with a shorter hospital stay and fewer hospital clinic follow ups. A target date is difficult to define in this situation but current practice at the Trust site will be immediately implemented based on the evidence.

Outputs:

The aim is to produce a detailed summary of all patients treated over the 10 year period that will demonstrate any relapse after surgery for infective endocarditis in a centre that uses shorter than internationally recommended antibiotic courses. This will include bacteria specific infection and tailored antibiotic regimes used. The final report will be a service evaluation for GSTT that will be fed back to the clinicians and implemented into standard clinical practice. GSTT will also submit findings for publication in high level medical journals (the initial target journal is "Heart") in order to influence national/international guidelines such as NICE. The Cardiovascular Patient Forum have agreed for the Trust to present and discuss their project in September (26th) 2016. Processing of data is now expected to be by the end of August 2016 with subsequent publication of reports and journal submissions. An annual report will be sent to CAG by April 2017. All outputs will only contain aggregated data with small numbers suppressed in line with HES analysis guide. Full and final analysis and identifiable data destruction will be completed within 6 months of receiving the data.

Processing:

GSTT have conducted the following information searches on our cohort of 247 patients: All patients have been assessed for readmission to this Trust within 1 year of surgery and any episodes clarified for IE. All patients deceased within 1 year of surgery have cause of death analysis completed by death certification. All surviving patients’ GPs have been contacted for evidence of admission for IE within one year of surgery to other hospitals – several have been unable to clarify. 40 patients have incomplete follow up information because either: 1. the patient is now deceased 2. has left the UK. 3. the GPs have been unable to confirm accurate health history within one year following surgery. Once GSTT have received the HES data from HSCIC, they will identify any admissions for episodes of endocarditis or possible endocarditis. These will then be investigated by contacting the relevant hospitals' microbiology department for clarification and confirmation of diagnosis and required treatment. Once all necessary information has been obtained, the data will be aggregated. Diagnosis for IE can be complex. It may be necessary for the Trust to contact the Consultant responsible for the care episode to discuss and clarify diagnosis should, and only if, a HES record suggests a possible episode of IE. The patients' sensitive and identifiable data will be required in this situation to ensure the correct patient is discussed. This will ensure all findings are accurate. Once diagnosis is clarified, all identifiable and sensitive data will be anonymised along with the rest of the cohort data. All other admission data will be disregarded and destroyed.

Objectives:

To establish the relapse rate of infective endocarditis within one year, in patients treated with heart valve surgery at the Guys & St Thomas Trust (GSTT) centre over a 10 year period. Infective endocarditis (IE) is a deadly disease associated with high mortality (15 - 30%) and severe complications. It is caused by transient bacteria in the bloodstream attaching on to the endocardium and heart valves leading to destruction of the valve tissue with often fatal consequences. Emergency treatment is required in a specialist centre with intensive and prolonged intravenous antibiotic therapy and often high risk cardiac valve surgery. This is a costly and protracted treatment regime including intensive care and frequent outpatient follow ups. Added to this is the ever increasing concern regarding antimicrobial resistance and antibiotic side effects. Should the patient survive the initial episode and treatment, there remains an elevated risk of relapse. However, because the incidence is low and the early mortality rate high, there is limited evidence to improve the management and randomised trials are almost impossible to conduct. Therefore retrospective studies (such as this audit) can provide valuable new information for meta-analysis. As GSTT are a specialist centre, there is a cohort of patients treated for IE that can provide valuable evidence to improve management of IE. The objective for this project is to evaluate whether the antibiotic regime (which is shorter and therefore less costly than internationally recommended) is as effective at reducing the relapse rate of infective endocarditis in patients treated surgically (i.e. undergoing valve replacement surgery) at the centre. This is over a 10 year period. By assessing each patient for confirmed relapse or new episode of IE within the first year of surgery GSTT can provide a robust evaluation. Any relapse or new episode can only be confirmed by hospital admission and clinical diagnosis. Therefore this is key information that is required. GSTT have an identified cohort of 247 individual patients, including information relating to relapse, new episode or death. 40 patients have incomplete information that need to be clarified. As the incidence of endocarditis is a rare occurrence, only a small number of datasets (40) are required to complete this project. The overall cohort is small (@ 250). However, this is considered a significantly large evaluation (most other projects cover only 30-40 individuals) and the absence of 40 datasets will seriously undermine the impact of the analysis and the quality of the results. As GSTT are a tertiary centre, not all patients live locally and further episodes and/or relapses may have been treated at other hospitals elsewhere, of which GSTT are unaware. This includes some patients who are poor compliers with medical care and have no registered GP. Some patients have since died and records of the one year post surgery period are no longer available from their GPs. For surviving patients, some GPs are unable to clarify the one year post operative period. As this is a service evaluation and not a research study there is no patient consent process. Therefore, Section 251 approval has been obtained to request Hospital Episode Statistics (HES) from the Health and Social Care Information Centre (HSCIC). The HES data requested will fill in data gaps for deceased patients and patients with unknown history after treatment. Consultant code is required as diagnosis for IE can be complex. It may be necessary for the Trust to contact the Consultant responsible for the care episode to discuss and clarify diagnosis. This will ensure all findings are accurate.


Project 2 — DARS-NIC-35239-W2W9R

Opt outs honoured: N

Sensitive: Sensitive

When: 2017/12 — 2018/02.

Repeats: One-Off

Legal basis: Informed Patient consent to permit the receipt, processing and release of data by the HSCIC

Categories: Identifiable

Datasets:

  • Office for National Statistics Mortality Data

Benefits:

There are over 3 million people in the UK suffering from COPD. Current costs are £800 million per year to the NHS for those that suffer from COPD, which equates to £1.3 million per 100,000. Exacerbations of COPD represent 1 in 5 hospital admissions with 25% of patients having evidence of respiratory failure at admission. Current evidence demonstrates that an average of 15% of COPD patients will die within 3 months of being admitted with an exacerbation, rising to 30% at 1 year in patients who have had a severe exacerbation requiring respiratory support. The data provided by NHS Digital will allow the study to provide extended mortality data for patients included in the trial to examine for an effect on long term mortality. The outcome of the analysis will provide further information for professionals in this area. Medical specialists will have added information to assist patients in decision making around initiation of HMV. The data will also allow the trial to support commissioners and the trust when considering the importance of changes to service delivery. The dissemination plan will allow these key groups to be targeted. The dissemination plan is to publish in a high ranking general and respiratory medical journal with allied presentations at international and national conferences to ensure key opinion leaders are aware of the results. These opinion leaders can use the information when meeting with their local commissioners to change their clinical practice within UK health care. This analysis is a secondary analysis to a study published in the Journal of the American Medical Association (JAMA – Impact Factor 44) the main study findings are already influencing local policy for the provision of HMV with changes to the patient pathway following an acute exacerbation of COPD. Furthermore the study findings have been incorporated into the recently publish German Thoracic Society guidelines on HMV in COPD demonstrating the effective reach of the dissemination strategy. The data has the potential to assess the mortality benefit of this novel therapy in COPD patients with the most severe disease and high symptom burden. The current evidence from the trial has suggested a benefit to patients in terms of reduced hospital admissions and flare ups of their respiratory disease. This has obvious benefits to both the patient and the health care system (both primary and secondary care). The planned analysis on longterm survival will add weight to the potential patient benefits. The dissemination plan is to publish in a high ranking general or respiratory medical journal with allied presentations at international and national respiratory conferences to ensure key opinion leaders are informed of the trial outcomes and potential benefits to patient care. It is anticipated that the overall study will enhance patient care and wellbeing for patients with end stage COPD and identify those that should receive HMV and therefore, reduce hospital readmissions.

Outputs:

The findings will be published and not be influenced by the funders investment. The dissemination of results are being performed with key respiratory journals and key respiratory conferences as the data will be of great interest to the respiratory physicians with regards to current and future clinical practice. In line with the usual practice within the research group lay summary of the data will be presented to patient participation groups and the Lane Fox Patient Association. All outputs will contain only aggregate level data with small numbers suppressed in line with HES analysis guide. When complete the data will be included within journals and conferences: - Current conferences/presentations to be targeted: British Thoraric Society (BTS) Attendees - Respiratory specialists (consultants & trainees), GPs with interest in respiratory medicine, Respiratory Physiotherapists. Journals to be targeted: Medical journal publication (general or respiratory) Speciality Groups: Medical Grand Round GSTT (Guys and St Thomas Trust) - December 2017. Medical Grand Round is a trust seminar that is held monthly and is open to all the consultants within the trust, medical staff and final year medical students. It is a highly esteemed seminar to present at and the output will demonstrate what research is being undertaken within the trust. This data will not be used to establish a protocol for a clinical trial nor will it be used to show performance of an organisation. The timeline on processing and analysing the data and subsequent publication will be approximately 6 months post receipt of data from NHS Digital.

Processing:

GSTT will provide a list of NHS numbers with dates of birth to NHS Digital. NHS Digital will share back to GSTT Study ID, Date Of Death and Cause of Death. From the data GSTT receive a study identifier will be allocated and this is what will be used for further analysis. The original NHS identifier will then be removed and original transferred data deleted in line with application information. The information needs to be provided in an identified format so it can be match to the rest of the patient data stored within the database from the original trial. This database is de-identified but has a matching document stored in accordance with the original research ethics committee application. GSTT will retrieve cause of death for analysis purposes only. GSTT may use the date of death to determine how soon patients died post participation into the trial. The de-identified death data will be added to a generic database with no personal identifiable data only a study identifier, for example STH34. This will be performed by the identified ONS users. The whole database will then be analysed. Only the named authorised users will have access to the data supplied by NHS Digital. The extracted data will be used to show how many of the participants have died after their participation into the trial and if their cause of death was linked to the pathology they had which made them eligible for the trial initially. The data access will demonstrate mortality numbers for participants with the pathology included into the study. The data which will be shared will be suppressed inline with ONS Disclosure controls. All processing of ONS data is in accordance with standard ONS terms and conditions. Record level data will not be shared with any third party, all those accessing the data will be substantive employees of GSTT. All organisations party to this agreement are required to comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract ie: employees, agents and contractors of the Data Recipient who may have access to that data).

Objectives:

Home Mechanical Ventilation (HMV) is used for patients with symptomatic hypoventilation, they are small devices that can be used at home at night time or during day, dependent upon disease progression. There is a need for a long term study of the effects of HMV in patients with hypercapnic COPD, as there is a considerable morbidity, and mortality with frequent hospital admissions associated with this condition. There are also an increasing number of referrals of patients with hypercapnic COPD to units specialising in ventilatory support for consideration of HMV, yet there is no good evidence for the long term effectiveness of this therapy. For the purposes of this trial HMV is used for patients chronic end stage chronic obstructive pulmonary disease. The hypothesis of this study is that the HMV and home oxygen therapy (HOT) compared to HOT alone will reduce re-admission to hospital in COPD patients who remain persistently hypercapnic following an exacerbation requiring non invasive ventilation (NIV). The primary outcome is to evaluate admission free survival with comparison between HMV and HOT and HOT alone. This data will be analysed by time to event analyses. Duration of hospital admissions will be also analysed. If the trial shows a clinical benefit, it will be important to know if the addition of HMV in patients with hypercapnic COPD represents an optimal use of medical resources. Any additional health benefits have to be judged against the costs of resources required to produce them. The trial will include an analysis of the incremental cost-effectiveness and cost-utility analysis of the two study treatments; HMV with home oxygen and home oxygen therapy alone. Data regarding health economic costs will be obtained on an on-going basis by the research team. The main end-point in the cost-effectiveness analysis will be cost per admission avoided. Costs will include analysis of the equipment costs, maintenance and support costs for the HMV, analysis of medical, nursing and support staff and all hospital admissions. Cost-effectiveness and cost-utility acceptability curves will be analysed on the v13 (1.8.11) 8 basis of a simulation exercise in order to represent the uncertainty surrounding these ratios. Health related quality of life will be measured using a number of validated questionnaires. NHS digital is being asked to supply data on all trial participants who were alive at trial completion and thus the end of intensive trial follow up. Although it is expected that the majority have died some patients may still be alive. The request for data is 2 fold both to confirm health status and to supply, were applicable, date of death and cause of death data. Guys and St Thomas Hospital NHS Trust (GSTT) received hospital admissions data via NHS Trusts directly.


Project 3 — DARS-NIC-44383-L6C0X

Opt outs honoured: N

Sensitive: Non Sensitive

When: 2017/03 — 2017/05.

Repeats: One-Off

Legal basis: Health and Social Care Act 2012

Categories: Anonymised - ICO code compliant

Datasets:

  • Hospital Episode Statistics Admitted Patient Care

Benefits:

Excess alcohol consumption is a growing public health problem, causing 5.3% of deaths worldwide in those aged under 60 years. In the UK, alcohol use is the fourth greatest risk factor for years lived with disability and is second only to tobacco as the leading preventable cause of ill health, costing the NHS £2.7 billion annually, with 78% on hospital based care. Alcohol related hospital admissions have doubled in the last 8 years in England and reducing this burden is a key priority of government public health strategy. It is estimated that 1-2% of attendances to UK A&Es are made by ‘frequent attenders’. Studies show that ‘frequent attenders’ to A&E are also frequent users of other health and social care facilities. There has been a recent call for further research into the predictors of frequent use of healthcare services, supporting the notion that these subgroups are not adequately defined. Alcohol related frequent attenders (ARFAs) are thought to account for 6.7% of frequent attenders. With no singularly defined way of recording and monitoring ARFA hospital admissions/attendances it is difficult to understand the true burden of ARFAs on the NHS. 21 hospitals in England run programmes for ARFAs , with no common method of identifying patients for treatment. By better understanding the characteristics of ARFAs and their patterns of usage of health services through this study, it is hoped it may become possible to identify preventative interventions to avoid further harms to their own health and prior to assimilating high costs to health services. This study will benefit patients and the Trusts within the Kings Health Partners through the development of the risk stratification. This will ensure that ARFAs can receive the specialist treatment that they require, affording them direct health benefits in a setting that is more suited to their needs than in an A&E department, but will contribute to a reduction in hospital admissions with concomitant savings to the NHS.

Outputs:

The proposed project directly informs the design and purpose of services for ARFAs at Kings Health Partners hospitals (Guys and St Thomas', Kings College and South London and the Maudsley NHS Trusts). The analysis of HES will produce a list of characteristics (at aggregated level and will not contain data pertaining to individuals) which are generic to ARFAs eg average age, average level of income deprivation etc which will be used to populate a risk stratification model initially for South London and then nationally. The risk stratification model will then be used to calculate the number of people (ARFAs) who could potentially benefit from accessing ARFA services such as assertive outreach treatment both nationally and locally, informing commissioning of services. This population risk stratification project is part of a wider project to optimise services for ARFAs at KHP, which includes setting up an assertive outreach services (specialist mental health services) to specifically meet the needs of ARFAs. The project is directly supervised by 2 Professors of Addictions at the National Addictions Centre based at KCL, one of whom leads the alcohol strategy for Kings Health Partners hospitals (Guys and St Thomas', Kings College and South London and the Maudsley NHS Trusts). The alcohol strategy steering group includes lay members, service users and other researchers and clinicians working on alcohol-based projects across South London. The principal investigator reports project progress to the alcohol strategy steering group. The principal investigator and Professor/alcohol strategy lead are also part of the ARFA clinical network for South London, which meets every 6 weeks. The group consists of practitioners and clinicians working with ARFAs so provides direct insight in to the day-to-day treatment and issues for this particular patient group. The principal investigator reports project progress to this group. Finally, the principal investigator’s project progress is also monitored on a quarterly basis through King's College London. Project findings are due to be reported November 2018 but due to the close working arrangements with clinicians described above, will inform service design from the outset. The outputs from all of the projects will include peer reviewed papers in academic journals which will be submitted for publication by May 2018. In addition, lay summaries such as newsletters and blogs (on behalf of the South London Academic Health Science Network and Collaboration for Leadership in Applied Health and Care will be produced (March 2017-April 2018). Conference and seminar presentations to academic, policy, professional in the fields of public health and addiction sciences and public audiences will be made between March 2017 and November 2018. All reports and presentations will be produced containing aggregate results with small numbers suppressed that show trends over time, differences across providers, commissioners, geographical areas and by patient subgroups and patient characteristics. The results will contain estimated correlations showing associations between patient outcomes and patient characteristics, hospital, institutional, geographic and environmental factors.   Information about this study and its use of data will be made available to the general public through the South London CLAHRC website.  Outputs will contain only aggregate level data with small numbers suppressed in line with the HES analysis guide.

Processing:

The applicant will be undertaking processing activities for this project. The applicant (Consultant in public health medicine/innovation fellow) is trained in health care analytics and epidemiology and is up to date with NHS information governance training (last update June 2016). The applicant holds an employment contract with Guy’s and St Thomas’ NHS Trust which is part of King’s Health Partners (together with King’s College Hospital NHS Trust, King's College London and South London and the Maudsley NHS Trust). As part of the Guy’s and St Thomas’ employment contract,the applicant holds a King’s Health Partners’ “research passport” to enable her to conduct research from any other of the NHS Trust sites within King’s Health Partners. The applicant will be conducting the proposed research from South London and the Maudsley (SLAM) NHS Trust site, as firstly this is where she is geographically based for work; and secondly, because SLAM currently host HES data, having the necessary IT infrastructure and ability to meet information governance requirements. In terms of the data processing pathway: - The inpatient HES data is downloaded from NHS Digital and stored on South London and the Maudsley NHS Trust’s server. The server is held on-site at SLAM, and access is restricted to named individuals according to SLAM’s security policy. - Storage will be on a storage area network and secured by active directory user group. - Remote access to the database is permitted, but only through terminal Services via secure token (so processing is still carried out on site), and with local printing and downloading disabled. - Only staff who have signed a confidentiality agreement and have received IG training are permitted access. - All access to individual files is recorded, and a sample audited to investigate the existence of any adverse incidents, and ensure that appropriate access has been maintained. - The HES data is imported into STATA SE. Once held in STATA, The applicant will view the data and select a specific cohort for each individual study. Commonly a process will initially take place to define the particular cohort of interest in terms of e.g. individual diagnostic codes or procedure codes. The researchers will use routinely available filter definitions where possible, but may amend these based on the nature of each study’s group of interest. Depending on the research a similar control group may be established. - The applicant then analyses the data, before applying the relevant disclosure controls to any output. Software used will be STATA SE; typically this will involve analysis on several outcome measures, risk adjustment and the construction of control groups. - No record level data would be linked to this dataset, but it may be combined with publically available demographic or geographic data, for example in relation to local Trust performance - Outputs are thus produced which consist of aggregate data (or indicator/statistical data) only with small numbers suppressed in line with the HES analysis guide. The applicant will be the only person who will access the data. They are a substantive employee of Guys and St Thomas' which includes in the employment contract a research passport for the other sites in the Kings Health Partners. The data requested will only be used for the purposes described in this document. South London and Maudsley NHS Foundation Trust will not link the data disseminated by NHS Digital to any other data they may already hold.

Objectives:

The purpose of processing the data is to ascertain "The nature, natural history and characteristics of alcohol-related frequent attenders", by exploring data held within hospital episode statistics to understand this group of patients better including their health and social care needs. The project has 2 specific aims: Aim 1: Explore in a sample of hospital attenders, which medical and socio-demographic characteristics are associated with alcohol-related frequent attendance and different patterns of health service utilisation. Aim 2: Explore costs of health service use by Alcohol Related Frequent Attenders (ARFAs). The two aims are explored by two separate studies outlined below. Study 1: Natural history of ARFAs From national Hospital Episodes Statistics (HES) 2011/12 service use pattern of a pseudonymised cohort of ARFAs during 2015/16, 2014/15, 2013/14, 2012/13, 2011/12. This will yield data on natural history of ARFAs including: co-morbidities (ICD 10 code), mode of admission, length of stay, readmissions, age, gender and geography. The researcher will compare ARFA findings to 3 other groups of patients from 2015/16-2011/12 national HES: non-alcohol-related-frequent attenders, non-alcohol-non-frequent attenders and alcohol-related-non-frequent attenders. Data on the characteristics of frequent attenders and non-frequent attenders will be analysed using STATA SE. A logistic regression approach will be used to explore the variables derived from HES; demographics, diagnosis and attendance frequency. The analysis of HES will produce a list of characteristics which are generic to ARFAs eg average age, average level of income deprivation etc which will be used to populate a risk stratification model initially for South London and then nationally. The risk stratification model will then be used to calculate the number of people (ARFAs) who could potentially benefit from accessing ARFA services such as 'assertive outreach' (specialist mental health services) treatment both nationally and locally Study 2: The cost burden associated with ARFAs Costs of health service usage by the 2015/16 ARFA cohort will be calculated per capita (on the basis of 2015/16 tariffs and occupied beddays) and compared to the costs of non-alcohol related frequent attenders. Total costs of ARFAs will be scaled up to national costs based on epidemiological results from study 1. Costs will be calculated from the health service perspective and will not explore full costs of ARFAs to society. Sensitivity analysis- Using different assumptions and scenarios, how costs vary based on the definition of an ARFA used will be investigated ie comparing the costs to Kings Health Partners (KHP) of ARFAs with varying number of visits per year. Current literature documents multiple ARFA definitions and impact of ARFAs on health services may be important in finalising a definition going forward.


Project 4 — DARS-NIC-91878-Y4M2P

Opt outs honoured: N, Y

Sensitive: Non Sensitive

When: 2017/12 — 2018/02.

Repeats: One-Off

Legal basis: Section 251 approval is in place for the flow of identifiable data

Categories: Anonymised - ICO code compliant, Identifiable

Datasets:

  • Hospital Episode Statistics Admitted Patient Care

Benefits:

Outputs will benefit future patients receiving prosthetic valves, in addition to cardiologists and microbiologists managing the care of patients with infective endocarditis. Two major benefits are anticipated: a) This study will help ascertain whether there has been a rise in cases of infective endocarditis (IE) in patients who have undergone prosthetic valve replacement since UK guidelines were changed (to restrict use of antibiotic prophylaxis) in 2008. This will provide surrogate evidence on the efficacy of antibiotic prophylaxis, and in turn, information on whether the UK currently has the correct policy on use of antibiotic prophylaxis. b) Insight into risk factors and outcome for patients who develop infective endocarditis after transcatheter valve replacement. This is a rapidly growing field, but the incidence, risk factors and outcomes are largely unknown. This is the first study assessing these questions in the UK and potentially would be the largest dataset in the world to date.

Outputs:

To disseminate the results of the study to the wider heart valve disease population, Guys and St Thomas NHS Foundation Trust plan to present the results at major conferences where the key patient representative groups are present (e.g. American College of Cardiology, European Society of Cardiology). Guys & St Thomas NHS Foundation Trust expect that this project will lead to submission of up to 3 manuscripts to high level general or specialist peer reviewed journals (e.g. British Medical Journal, The Lancet, Journal of the American College of Cardiology, Circulation). The Trust will also disseminate results through the British Heart Valve Society and Heart Valve Voice, and plan to present the data at a transcatheter valve patient meeting. Due to journal charges and a lack of central funding the study anticipates that publications will not (initially) be open access, but the study will upload pre-publication manuscripts in open-access repositories. The timescale for submission is 1-2 years following release of the linked datasets. This study is independent of NICE and the findings would be published in the peer-reviewed literature. If there was evidence of increasing rates of IE, the study would notify NICE once the manuscript was accepted for publication. Outputs will contain only aggregate level data with small numbers suppressed in line with the HES analysis guide. This is a non-commercial project and data will not be used for sales and marketing purposes

Processing:

Data capture Guys & St Thomas NHS Foundation Trust will use three NICOR datasets (the National Cardiac Surgery Audit, the UK TAVI registry, and the National Congenital Heart Disease Audit) to identify patients who have undergone transcatheter and surgical valve replacement, and details of their procedure. These patient cohorts will be electronically linked to Hospital Episode Statistics (HES) Admitted Patient Care episodes to identify any patient with a primary diagnosis of “acute or subacute infectious endocarditis” (ICD-10 code I33.0), "endocarditis, valve unspecified" (ICD-10 code I38) or "endocarditis and heart valve disorders in diseases classified elsewhere (ICD-10 code I39)". After linkage of the NICOR datasets to HES, all patient identifiers will be removed with the exception of NHS number for the transcatheter valve intervention with infective endocarditis cohort. For this subgroup (estimated to be approximately 70-140 patients) the study plan to contact the Consultant Cardiologist who performed the operation, in order to request limited data regarding the diagnosis and management which is not available by any other means. Patient populations a) Transcatheter valve implantation (TVI) patients - Patients who have undergone transcatheter aortic valve replacement will be identified from UK TAVI registry. Those who have undergone transcatheter pulmonary valve intervention will be identified from the National Audit of Congenital Heart Disease. b) Surgical valve replacement patients - Patients who have undergone surgical aortic valve implantation will be identified from the NICOR Adult Cardiac Surgery Audit. Those who have undergone surgical pulmonary valve implantation will be identified from the NICOR Adult Cardiac Surgery Audit and the National Congenital Heart Disease Audit. Individual patient data For patients diagnosed with IE after TVI the study has permission from the Confidentiality Advisory Group (CAG) to obtain the patient’s NHS number from NHS Digital. Guys & St Thomas NHS Foundation Trust will then contact the operating consultant to request additional information from the patient’s record. Study Endpoints Primary endpoint/analysis: a) Incidence of IE after TVI and SVR Secondary endpoints/analyses: a) Time to diagnosis of IE from valve intervention b) Patient characteristics in the TVI-IE cohorts (vs those without IE) c) Prosthetic valve characteristics in the TVI-IE cohorts (vs those without IE) d) Procedural factors in the TVI-IE cohorts (vs those without IE) e) Multivariate predictors of IE in patients undergoing TVI f) Survival of patients diagnosed with IE in the TVI and SVR cohorts g) Incidence of IE in patients with a prosthetic heart valve before and after 2008 h) Use of prophylactic antibiotics in TVI-IE (vs those without IE) i) Diagnostic imaging tests/results in those with TVI-IE i) Microbiology (blood culture/valve explant) in patients with IE j) Management of patients diagnosed with TVI-IE k) Analyses above divided by paediatric/adult populations Statistical Analysis Patient characteristics will be reported as categorical or continuous variables as appropriate. Group comparisons will be made using the Student t test or Wilcoxon rank sum test (numerical variables), or the chi-square or Fisher exact test for categorical variables. The study aims to identify independent predictors of IE in the TVI and SVR cohorts using a Cox proportional hazard analysis. Survival of those with IE will be reported using Kaplan-Meier analysis. Guys & St Thomas NHS Foundation Trust will also use Kaplan-Meier curves to show the proportion of valves after TVI and SVR which remain free from infective endocarditis over time. Details of physical security arrangements for data processing: NHSD will provide the NICOR-HES linked datasets, which will contain the NHS numbers for patients within the TVI-IE cohort. These datasets will be downloaded directly from NHSD using secure file transfer. The datasets will be stored on a dedicated laptop (with encrypted hard drive) which when not in active use, will always kept within locked storage in the Department of Cardiology at St Thomas's Hospital, and is not transported outside of the Department. All data will also be encrypted (to minimum standard AES 256), and a BIOS password will be used to make sure data and device can only be accessed by authorised users (see below). There will be no secondary copies of the datasets. The data will not be used on mobile devices. Data analysis The dataset will only be analysed on a dedicated laptop (see above) within the Department of Cardiology at St Thomas's Hospital. This computer will be connected within the Guy's & St Thomas's NHS Foundation Trust maintained network. The data will only be analysed by individuals substantively employed by Guy's and St Thomas' NHS Foundation Trust. Guy's and St Thomas' NHS Foundation Trust will receive two datasets from NHS Digital. The non-TVI group will contain pseudonymised HES data and the TVI group will contain the same HES fields plus NHS number which is identifiable. Once the NHS number has been used to obtain further data from clinicians it will be securely destroyed. From this point the working dataset used for the analysis will be pseudonymised and will only contain the patient's study number and thus no direct patient identifiers. The dataset will never be sent by email or transferred onto memory sticks or external hard drives. Where email is required (for example, with individual patient's consultants to gather clinical data), NHS secure email will be used as the only means of communication. The data will only be used for the purposes described in this document.

Objectives:

Infective endocarditis (IE) is a life-threatening condition caused by infection within the heart, most commonly on a heart valve. IE is much more common in people who have undergone a heart valve replacement. Since 2007 a new keyhole method of heart valve replacement – transcatheter valve implantation (TVI) – has been available, which avoids the need for open heart surgery. TVI is mainly used for patients who are too frail to undergo open heart surgery. There are very few studies examining the risk of IE after valve replacement by TVI. Guys & St Thomas NHS Foundation Trust want to understand how many patients develop this complication, which patients are affected, and to look at how many patients ultimately survive. Guys & St Thomas NHS Foundation Trust also plan to compare the risk of IE after TVI with the risk of IE after conventional open heart surgery. Finally, Guys & St Thomas NHS Foundation Trust will look at whether the number of heart valve replacement patients who develop IE has changed over time. Specifically, the study will look at the numbers of patients with IE before and after 2008, when UK guidelines from NICE advised stopping antibiotics prescriptions for prevention of IE. The study design is a retrospective cohort analysis. Patients who have undergone TVI will be identified from datasets held by the National Institute for Cardiovascular Outcomes Research (NICOR) - the UK TAVI registry (for transcatheter aortic valve intervention), the National Congenital Heart Disease Audit (for transcatheter pulmonary valve intervention), and the Adult Cardiac Surgery Audit (for surgical valve replacement). These NICOR datasets will then be linked to Hospital Episode Statistics to identify patients who have subsequently developed IE. NICOR will send the data identifying these patient cohorts to NHS Digital, who will carry out the linkage. To obtain information about the diagnosis and management of prosthetic valve IE, the research team will contact the consultant who performed the TVI procedure and request further specific information from the patient's record. In summary, the objectives of this study are a) to identify the rates of infective endocarditis (IE) after minimally invasive valve replacement (transcatheter valve implantation; TVI), compared with open surgical valve replacement (SVR) b) to describe the clinical outcomes (e.g. survival) for patients who are diagnosed with IE following TVI or SVR c) to determine which patient groups are at highest risk of IE after prosthetic valve replacement by TVI or SVR d) to identify anatomical or procedural factors which are associated with increased risk of IE after TVI or SVR e) to assess whether the incidence of IE has changed since modification of NICE guidelines on antibiotic prophylaxis. From 2008 onwards NICE advised doctors and dentists not to prescribe routine antibiotic prophylaxis for prevention of infective endocarditis in those with a prosthetic heart valve. The study will analyse whether this guidance has affected the rates of infective endocarditis amongst patients with a prosthetic valve.