The Newcastle upon Tyne Hospitals NHS FT - Quarterly HES & ONS extracts (2024/25 application)

Object for Processing:

Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) requires access to NHS England data for the purpose of the following research programme:

The Newcastle upon Tyne Hospitals NHS FT - Quarterly HES & ONS extracts (2023/2024 application)

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is one of the country's top performing trusts in terms of the number of research studies it has supported. Northern Medical Physics and Clinical Engineering (NMPCE) is a Clinical Directorate within NUTH, with research interests and a track record going back 40 years in the development and evaluation of health technology interventions including medical devices, diagnostics, and interventional procedures.

NMPCE will use Hospital Episode Statistics (HES), Emergency Care Dataset (ECDS) and Civil Registration mortality data to support its programme of research which includes studying:

i) uptake and outcomes of novel medical technologies (medical devices, diagnostics, medicines, therapies), and interventional procedures to assess peri-procedural, short-term and long-term safety and efficacy

ii) associations between comorbidities, diagnoses, and hospital admissions

iii) patient pathways through the healthcare system.

This is a not-for-profit research programme which uses large pseudonymised data sets to achieve the aims described. The outputs of the programme are intended to inform national guidance, decision-making, research recommendations, study designs, grant applications to national bodies.

The lawful basis for processing is under Article 6 (1)(e) of the UK GDPR – “processing, as necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller”. The justification is as follows:

NUTH are a public authority conducting transparent observational research and publication of healthcare outcomes from health technologies (devices, diagnostics, medicines, therapies) and interventional procedures in the NHS. Civil Registration mortality data informs overall death rates (any cause) for the patient cohort selected for analysis. Out of hospital deaths and cause of death are not reported through HES, therefore the combined data are more robust for analysis than either source alone. NUTH cannot reasonably achieve the above in a different way. NHS providers have administrative systems which feed into HES; therefore this represents the only national routine dataset of NHS activity in secondary and tertiary care. HES may record in-hospital deaths, but Civil Registration mortality is the only legal record of cause of death, independent of place of death. The pseudonymised data analysis methods developed by NMPCE (non-identifiable patient data) are the least intrusive means of achieving the purpose of NHS healthcare outcomes analysis and publication, in the public interest.

The lawful basis for processing special category data (health data) is under Article 9 (2)(j) of the UK GDPR. The justification is as follows:

Not all the projects are intended to generate generalisable new knowledge for academic publication. NMPCE outcomes research programme also includes processing for reasons of public health, including adoption of medical technologies and outcomes of interventional procedures, to ensure high standards of quality and safety, as described in this application.

Moral and ethical issues and risk of potential harm to the public from the data processing methods, have been considered through active engagement with patient and public involvement (PPI) groups in the Trust and NUTH. Patients and their representatives have told NUTH that they positively endorse and demand such use of their data to benefit medical research and the NHS, as described in this agreement.

The data requested will achieve the aim identified by permitting a timely response from NUTH, to questions of safety and efficacy of medical devices and interventional procedures in current NHS practice.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is the sole Controller and joint processor with Newcastle University. Only substantively employed NuTH and Newcastle University staff will process the data.

NMPCE has used HES data since 2011 in support of its research and evaluation work for several stakeholders. Funding from stakeholders is used to support NMPCE staff time for health technology evaluation, which may or may not include analysis of data from NHS England. However, NMPCE exercises overall responsibility and control over the purposes and means of the processing personal data received from NHS England. NMPCE determines whether HES data should be used, defines the query, conducts data cleaning, and data analysis. In doing so the NMPCE fulfils the role of sole controller by completing the following tasks:

- Decision to process the personal data

- Decision on outcome of processing

- Decision on what datafields and datasets to process

- Decision on cohort selection

- Exercising professional judgement in the processing of personal data

- Have complete autonomy as to how the personal data is processed

Since 2011, NUTH has hosted one of the current National Institute for Health and Care Excellence (NICE) External Assessment Centres (EAC). The NICE EAC is based in the NMPCE Directorate at the Freeman Hospital in Newcastle (staff are substantially employed by NuTH), and has a subcontract in place with Newcastle University for the provision of literature searching, statistical and health economic expertise. The work commissioned by NICE evaluates health technologies and interventional procedures and answers specific research questions raised by the NICE Medical Technologies Advisory Committee (MTAC), Diagnostics Advisory Committee (DAC), Interventional Procedures Advisory Committee (IPAC) and other NICE programmes. NICE had provided a commissioner letter for the External Assessment Centre within NMPCE to support this agreement, which extends to 2025.

NUTH was formerly an HSCIC / NHS Digital approved HES Business Objects (HESBO) and HES Data Interrogation System (HDIS) user. The NMPCE analysts within NUTH are trained and were approved to process episode level pseudonymised HES data for the research purposes described in this DARS agreement, for NHS Health and Social Care system, patient and public benefit. This agreement is to renew and continue the quarterly managed extract service of pseudonymised HES (APC, CC, OP, AE/ECDS) and Civil Registration data from NHS England, which will permit NMPCE to continue delivering the described research objectives for an additional twelve months.

The nature of the work described in this agreement is a wide programme of collaborative research as described above, including long term follow up. The detailed scope of data analysis is determined on a project by project basis, but always within the overarching terms of the Data Sharing Agreement with NHS England.

The data subjects are also determined on a project by project basis, following the convention of defining Population, Intervention, Comparator, Outcomes (PICO) and setting (APC / A&E etc) for every intended analysis project. A protocol for pseudonymised matching is drafted before work commences, where the HES and Civil Registration mortality codes to match the required PICO and setting are determined in an iterative process. Data cleaning rules are applied to reach the final agreed and matched set of data subjects, which may be for a single-armed, or comparative analysis, using control and cohort groups.

Detailed information about the purpose of the projects covered by this agreement is provided below, including the data required, why these data are required and justification for the datasets requested, the level of data (pseudonymised), the number of years requested and the geographical spread of the data requested.

NUTH confirm that there are no alternative, less intrusive ways of achieving the purpose, and details of the efforts taken to minimise the data required are also described below:

Pseudonymised data from the requested datasets are required by NMPCE for the following purposes / activities:

i) to determine the number of patients with a particular disease requiring NHS care, the number being treated with a health technology across England and identification of key providers of care, (e.g. to inform NICE of scope/coverage/uptake of guidance, to inform sample size determination and to target study invitations to relevant organisations. An example where this type of activity led to a successful publication was, NMPCE's cross-sectional Recurrent Respiratory Papillomatosis (RRP) survey which determined the number of patients diagnosed with the condition, the proportion receiving hospital treatment and the breakdown of interventions used in this population (Donne et al. Clinical Otolaryngology 2016). The survey found at least 16 different interventions in use in the NHS for RRP and limited evidence on their comparative safety and efficacy. This work led to a successful application for NIHR Research for Patient Benefit grant funding and the establishment of a national data collection on the existing Airway Intervention Register (AIR) platform for analysis to determine the safest and most effective intervention(s) for the condition, which will benefit current and future RRP patients.),

ii) to determine patient demographics receiving healthcare interventions (e.g. common comorbidities, comorbidity score index, age distribution, gender),

iii) to determine in-hospital outcomes (e.g. efficacy and safety, length of stay) of health care interventions,

iv) to conduct active surveillance/longitudinal analysis to determine long-term outcomes (e.g. efficacy, safety and in-hospital mortality) of healthcare interventions. For example the team undertook an 8 year analysis of almost 100,000 women with surgical insertion of mesh implants for the treatment of stress urinary incontinence which led to a peer reviewed publication (Keltie et al. Scientific Reports. 2017). The retention of HES data up to a maximum of 20 years will permit further, longer term follow-up analyses of these women, to establish the as-yet unknown longevity of mesh implants and any change in efficacy and complication rates as they age,

v) to compare patient outcomes (including in-hospital and longitudinal outcomes) between healthcare interventions (e.g. through propensity matching of cohorts based on comorbidities) to determine relative efficacy and safety. The retention of HES data up to a maximum of 20 years will permit future analyses of paediatric RRP patients in the AIR registry, to determine which treatments have enduring efficacy outcomes into adulthood,

vi) to determine hospital resource usage, including prior to and following healthcare interventions (e.g. patient pathway, number and indication of outpatient attendances, subsequent hospital admission, further interventions, accident and emergency attendances, in-hospital deaths, HRG codes/costs). An example where the team used HES data in this way is the published study of procedural and short-term efficacy of bronchial thermoplasty (Burn et al., 2016),

vii) to determine national coverage of and encourage data submission to national clinical or procedural registries (e.g. identification of trusts not contributing data to registries, data completeness of registry in terms of total number of procedures not recorded in registry, data completeness of registry in terms of complications reported using anonymised matching techniques based on for example treating trust, procedure date, gender, age datafields). The study undertook this type of active surveillance for the Airway Intervention Registry, using the former HDIS system. This is necessary to understand whether the registry data can be considered representative of the total population of interest, and to measure the success of efforts to improve data quality and completeness.

viii) to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching technique based on for example treating trust, sex, procedure date, gender, age data fields),

ix) to conduct exploratory analysis to determine whether outcomes (e.g. complications, diagnoses, in-hospital death) are associated with patient characteristics or setting. An example where the team used the data in this way was in a published analysis of the effect of hospital centre volume on iliac artery stenting outcomes (Goode et al., 2013),

x) study of outcomes and resource usage before and after an event (e.g. new device/procedure) to determine impact.

The most significant issue with the use of HES data alone in healthcare outcomes research is missing information on out-of-hospital deaths (date and cause of death) - noting that from recent publication that 50% of all deaths occur out of hospital. This risks overestimating the efficacy and safety of medical devices and interventional procedures in the aggregated results reported to NICE by NMPCE, to inform national NICE Guidance.

Hence, in addition to the above ten NMPCE purposes / activities, pseudonymised mortality data are required by NMPCE for the following purposes / activities:

xi) to add to the above longitudinal analysis in iv), to determine the long-term outcome of all-cause mortality, requiring both in-hospital (HES) and out-of-hospital mortality data of healthcare interventions,

xii) to add out-of-hospital mortality data to v) above, in comparative patient outcomes analysis,

xiii) to add out-of-hospital mortality data to vi) above, in analysing hospital resource usage prior to and following healthcare interventions

xiv) to add out-of-hospital mortality data to viii) above, to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching techniques based on for example treating trust, procedure date, gender and age data fields),

xv) to add out-of-hospital mortality data to ix) above, to conduct exploratory analysis to determine whether outcomes, including out-of-hospital deaths, are associated with patient characteristics or setting (e.g. age, gender, comorbidities, treating hospital).

In terms of data minimisation, due to the variety of analyses undertaken by NMPCE (different specialties, time durations of interest, research questions posed) the data and data subjects will vary by project. However, NUTH have considered the data fields requested from each data set (mortality, APC, CC, OP, ECDS) and each have been minimised accordingly. As a previous HESBO and HDIS user, NMPCE has used a specific subset of HES and Mortality data fields to inform evaluations of safety and efficacy for given healthcare interventions. For example: admission and discharge information, all diagnoses and procedure codes (4 character), episode and spell information (to build up an entire patient admission), treating organization, referring organization, patient age and gender.

National data are required due to the breadth of healthcare intervention (health technologies and interventional procedures) development and evaluation conducted (i.e. any condition, any procedure, any population). Access to HES data for the whole of England is also particularly important for NMPCE when considering interventions for rare diseases (affecting less than 5 in 10,000 of the general population), novel innovative procedures/devices/medicines coming into use, and for longitudinal analysis where patients may attend a number of different hospitals (e.g. treatment received in specialised treatment centre followed by follow-up care in local hospital). Furthermore, by using data from across the whole of England (rather than a specific localised region), NMPCE can assure that outcomes are generalisable and reflect current practices used across the country to inform national NICE Guidance. For NMPCE to conduct activities for NICE purposes, some of the research objectives, mean access is required to HES/mortality data from all of England.

No identifiable information is requested. NMPCE requires only a unique patient pseudonym that links across each data set: mortality, HES APC, OP, ECDS and CC to determine hospital resource usage and long-term outcomes for each patient.

The amount of mortality data requested has also been minimised by restricting by year to only those patients who have been admitted as inpatients (from 2007/08 onwards), or attended outpatients, emergency care, or critical care departments (from 2013/14 onwards). The frequency of mortality data requested is restricted to quarterly, to ensure safety signals are detected in a timely manner.

HES, ECDS and mortality data will not be used for any external commercial purposes not outlined in this agreement. These data will not be provided in record level form to any third party. These data will not be used for direct marketing. NUTH determines which data are processed for each project/research question and ensures the legal basis for requesting, storing and processing these data are met, therefore NUTH is the data controller and also processes the data for this study. No other organisations process the data for this purpose.

The programme of outcomes research using administrative and registry data currently has funding from the NIHR, Academic Health Science Network - North East and North Cumbria (AHSN NENC) and NICE. NMPCE has a track record of attracting research funding, and intends to seek future funding for their outcomes research programme, from public funding bodies, charities and organisations such as:

- NIHR projects (Invention for Innovation (i4i), Research for Patient Benefit (RfPB), Health Technology Assessment, Research Capability Funding, Newcastle Biomedical Research Centre;

- NIHR infrastructure funding (NIHR Newcastle IVD Co-operative, NIHR Innovation Observatory);

- NICE (the External Assessment Centre, EAC, described below);

- Academic Health Science Network (AHSN) North East and North Cumbria (NENC);

- Engineering and Physical Sciences Research Council (EPSRC);

- Charitable funding (Wellcome Trust, Innovate UK, British Heart Foundation, Newcastle Healthcare Charity);

- Professional societies supporting registry development;

- Commercial companies, which could include medical devices and diagnostics manufacturers.

NMPCE will produce an output in the form of a publication outlining an individual project's findings which, subject to acceptance, will be published in peer-reviewed academic or clinical journals. Additional outputs include grant application to funding bodies, conference presentations, summaries shared with public/patient representative groups and summaries shared on dedicated study websites (where applicable) and internal reports to NICE and other potential funders listed in the objectives for processing.

All outputs will contain only data that are aggregated with small numbers suppressed, in line with the HES analysis guide.

The dissemination and communication approach for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. Examples are described below.

Dissemination of results / outputs.

NMPCE will share aggregated information with NICE, to contribute to the production of national Medical Technologies, Interventional Procedure or Diagnostic Guidance, inform other NICE programmes, and thereby inform healthcare users, health care providers and commissioners of the safety and efficacy, national adoption and key providers of health technologies. An output will be an internal report outlining an individual project's findings submitted to NICE with small number suppression in line with the HES analysis guide.

Communication of results / outputs

One specific NIHR-funded project in progress is the Research for Patient Benefit award PB-PG-0416-20037 - Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP). Recruitment to the study in England will be compared (via routine monthly surveillance) to the number of RRP procedures being reported to the Hospital Episode Statistics (HES). This will enable NMPCE to invite new NHS hospitals/trusts to contribute to the study and also contact registered NHS hospitals/trusts to encourage them to enter data identified in HES to the registry where they have not done so. NMPCE have found that this method of feedback to NHS organisations (via Information Governance teams and clinical coding managers) has increased data entry to previous online databases as well as improving clinical coding accuracy for some interventional procedures.

Exploitation of results / outputs

NMPCE engages with clinical coders in acute trusts across England prior to analysis of HES data, but also provides feedback to clinical coders to demonstrate the external uses of coded hospital data, with the aim to continually improve data quality at source. All publications using HES/mortality data will reference NHS England and cite a copyright statement. NMPCE strives to make all peer-reviewed publications available in open access form in order to maximise the availability of information to health care users, health care providers and general members of the public. For example, NUTH published the analysis code (written in R) which analysed HES APC data to determine safety of mesh implants for stress urinary incontinence. Sharing methodology (not data) with the public will provide guidance to other researchers investigating other interventions in how to clean and analyse a large sample of HES data to identify complications; (Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015). The journal choice for any peer-reviewed publication will depend on the medical device intervention/interventional procedure/medical condition of interest (e.g. cardiac, ENT, oncology) and focus of the research question posed (e.g. methodology, patient safety, national policy).

Target dates to produce the outputs for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. In general, NMPCE projects involving medical device intervention/interventional procedure development or evaluation arise with clinical need, with timeframes governed by suitable funding calls and duration of grant funding.

In the specific example of the NIHR-funded Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP) project, a rolling monthly extract from the AIR will be compared with the latest quarterly HES extract from NHS England in active surveillance of coverage of the registry (total number of RRP cases treated in England versus total number entered in the registry). The target date for recruitment completion was 31st August 2022. A publication with provisional title of “Current practice, safety and efficacy of interventions for Recurrent Respiratory Papillomatosis: evidence from a UK registry” has been submitted (09/02/2023) and under review at the Clinical Otolaryngology journal.

In the specific case of NMPCE projects commissioned by NICE, each individual project is issued with a timeline of dates for completion of project milestones. Most NICE projects require peer-reviewed publication of results within 12 months of project completion. NMPCE strives to share analysis code (where possible ) for transparency and reproducibility purposes for future researchers.

NICE will consider evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

i) NICE IPG434: Radiofrequency cold ablation for respiratory papillomatosis (published November 2012 - update required).

ii) NICE Helium plasma coagulation for treating Respiratory Papillomatosis – monitored (awaiting publication of further evidence)

Updated Outputs 04/03/2024.

Current DSA has only been active since October 2023, and the data disseminated in late December therefore no published outputs and benefits yet.

Outputs from previous iteration:

NMPCE and NICE and NHS England worked collaboratively on a project looking at data linkage methodology (manuscript available freely online: “Comparison of identifiable and non-identifiable data linkage: health technology assessment of MitraClip using registry, administrative and mortality datasets” BMJ Health Care Inform. 2021; 28(1): e100223). NMPCE is currently working with NHS England on following longitudinal outcomes from topics exiting Cancer Drug Fund using data from HES.

Examples of peer-reviewed publications (available in the public domain) resulting from the analysis of HES data previously accessed via HDIS by NMPCE for NICE purposes, with details of the impact these have achieved are as follows:

ia) Keltie et al. Paediatric tonsillectomy in England: a cohort study of clinical practice and outcomes using Hospital Episode Statistics data (2008-2019). Clin Otolaryngol. 2021; 46(3): 552-561.

IMPACT: Analysed 318,453 paediatric tonsillectomies conducted in NHS hospitals in England between 2008 and 2019, demonstrating increase in complications.

ib) Powell et al. Coblation intracapsular tonsillectomy: A cohort study of NHS practice in England using Hospital Episode Statistics. Clin Otolaryngol. 2022; 47(3): 471-7

IMPACT: Analysed a total of 5525 intracapsular coblation procedures from 4 ENT centres in England.

This has triggered an IPG150 review at NICE, and grant applications to NIHR in order to further investigate why the safety profile of tonsillectomy has changed over time.

ii) Page et al. Prostatic urethral life (UroLift): a real-world analysis of outcomes using hospital episodes statistics. BMC Urol. 2021; 21(1): 55.

IMPACT: Demonstrates uptake of UroLift in England after UroLift was added to the Innovation Technology Tariff (ITT) in April 2017 and selected as a Rapid Uptake Product by the Accelerated Access Collaborative in 2018. This real-world analysis shows that UroLift implantation can be delivered safely in a day-case setting with minimal morbidity. However, hospital resource usage for catheterization and emergency hospital attendance in the first 30 days was substantial, and 12% required re-treatment at 2 years.

iii) Powell et al. Balloon dilatation for paediatric airway stenosis: Evidence from the UK Airway Intervention Registry. Clin Otolaryngol. 2020; 45(3): 334-341.

IMPACT: Conducted pseudonymised linkage to determine registry coverage, and also to determine longitudinal outcomes (further respiratory surgery, death).

iv) Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015.

IMPACT: Complications from surgical mesh procedures have led to legal cases against manufacturers worldwide and to national inquiries about their safety. This was one of the largest studies so far looking at adverse event rates of these procedures, and has significantly added to the understanding of the likelihood of adverse events from mesh procedures in a UK NHS setting. This means that women considering mesh procedures in future will have more accurate safety information available to them, so they can make a more informed choice. This publication and its analysis code was published in an open-access form to share both methodology and findings widely. NICE has also updated its NG123 guidance and produced linked Patient Decision Aids for women considering mesh surgery.

v) Burn et al. Procedural and short-term safety of bronchial thermoplasty in clinical practice: evidence from a national registry and Hospital Episode Statistics. J Asthma. 2017; 54(8): 872-879.

IMPACT: Bronchial thermoplasty is a novel treatment for severe asthma, however its mode of action and target patient population were poorly defined, with little evidence available on efficacy and safety in terms of quality and quantity. This study presents procedural and short-term safety evidence from routine UK clinical practice.

vi) Donne et al. Prevalence and management of recurrent respiratory papillomatosis (RRP) in the UK: cross-sectional study. Clinical Otolaryngology. 2017; 42(1):86-91.

Supported by Sims et al. Our experience in developing and operating the Airway Intervention Registry for Recurrent Respiratory Papillomatosis (AIR-RRP): national data collection. NIHR Open. 2023; 2: 22.

IMPACT: Recurrent respiratory papillomatosis is a relatively rare condition, but has a significant impact on patients and their families as it requires repeated anaesthetic procedures to achieve symptomatic control. This study utilised a cross-sectional survey of ENT consultants and analysis of HES data to provide, for the first time, an estimate for the prevalence of recurrent respiratory papillomatosis in the UK. This prevalence informed the sample size for an NIHR Research for Patient Benefit grant application - 3 year award granted with extension due to suspended recruitment to non-COVID 19 studies. Additional funding gained from NICE to support an additional 12 months (end 31st August 2022). Identification of key providers of RRP care in the UK, enabled NMPCE to invite specific acute trusts to contribute to further research in RRP patients, and contribute to future NICE guidance (in update of IPG 434).

vii) Goode et al. Effect of procedure volume on outcomes after iliac artery angioplasty and stenting. British Journal of Surgery. 2013; 100:1189-1196.

IMPACT: This study was able to use HES data to determine if there was an association between outcomes of endovascular iliac artery intervention and centre volume, describing methodology which could be applied to any interventional in-hospital procedure which would benefit future analysis of HES data.

viii) Patrick et al. Monitoring the use and outcomes of new devices and procedures: how does coding affect what Hospital Episode Statistics contribute? Lessons from 12 emerging procedures 2006-10. Journal of Public Health. 2012; 35(1): 132-138.

IMPACT: This study provided evidence that routinely collected HES data have the potential to support quality improvements and evidence-based commissioning of devices and procedures in the national health service.

ix). Keltie et al. Identifying complications of interventional procedures from UK routine healthcare databases: a systematic search for methods using clinical codes. BMC Medical Research Methodology 2014; 14:126.

IMPACT: This study summarised methods of analysing complications of interventional procedures from routine healthcare databases in the UK published in the literature. This study was published in open-access form to share learnings with other researchers and clinical coding teams across the UK to encourage high quality research from HES.

x) Powell J et al. National cohort study of health care resource after pediatric tracheostomy. JAMA Pediatr. 2022; 176 (8): 817-819.

IMPACT: This study demonstrated the significant healthcare utilisation and cost burden associated with hospital readmissions in children with tracheostomies. In England service specifications exist for certain specialist conditions or interventions. Service specifications are clearly defined standards of care expected from organisations funded by NHS England to provide specialised care. Paediatric tracheostomy is not currently included in this list, meaning that the levels of care and support are only defined on a local level. This study recommended the development of a paediatric tracheostomy service specifications, including extensive multidisciplinary support, to help reduce healthcare utilisation in the patient group.

xi) Veeratterapillay R et al. Infection after ureteroscopy for ureteric stones: analysis of 71,305 cases in the Hospital Episode Statistics database. BJU Int. 2023; 131 (1): 109-115.

IMPACT: This study represented the largest series evaluating infectious complications after ureteroscopic stone treatment. The procedure was found to be safe, with low inpatient infective complication and critical care admission rates. However the study highlighted the hospital after care resources required, which were previously unknown.

These publications inform clinical practice in the NHS, and can contribute to national NICE guidance. For example, NICE has incorporated evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

e.g. NICE IPG635: Bronchial thermoplasty for severe asthma (originally published January 2012 - updated December 2018; as a method of active surveillance of safety signals following the intervention.

e.g. NICE IPG425: Endoscopic balloon dilatation for subglottic or tracheal stenosis (originally published April 2012, and replaced with IPG719 in March 2022).

NMPCE has also supported Medical Technology evaluation of two devices (Danis stent, SpaceOAR, MAGECRods) through surveillance of HES to determine coverage of use across in England; this give a direct measure of health technology uptake.

NMPCE has used data from HES to support the drafting of grant applications applicable to a number of NIHR funding streams (including: Digital technologies, HTA, Ambitious data-enabled trials, health services and public health research studies, RfPB) focusing on health outcomes. Research themes for these grant applications have included (but are not limited to):

- safety and efficacy of tonsillectomy,

- safety and efficacy of adenoidectomy,

- longitudinal study to investigate the uptake and safety of benign prostate hyperplasia (BPH) procedures,

- longitudinal study to investigate the association between kidney transplantation and urological cancers,

- longitudinal study of stroke patients and pneumonia outcomes,

- longitudinal study of patients with laryngeal cancer.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is one of the country's top performing trusts in terms of the number of research studies it has supported. Northern Medical Physics and Clinical Engineering (NMPCE) is a Clinical Directorate within NUTH, with research interests and a track record going back 40 years in the development and evaluation of health technology interventions including medical devices, diagnostics, and interventional procedures.

NMPCE will use Hospital Episode Statistics (HES), Emergency Care Dataset (ECDS) and Civil Registration mortality data to support its programme of research which includes studying:

i) uptake and outcomes of novel medical technologies (medical devices, diagnostics, medicines, therapies), and interventional procedures to assess peri-procedural, short-term and long-term safety and efficacy

ii) associations between comorbidities, diagnoses, and hospital admissions

iii) patient pathways through the healthcare system.

This is a not-for-profit research programme which uses large pseudonymised data sets to achieve the aims described. The outputs of the programme are intended to inform national guidance, decision-making, research recommendations, study designs, grant applications to national bodies.

The lawful basis for processing is under Article 6 (1)(e) of the UK GDPR – “processing, as necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller”. The justification is as follows:

NUTH are a public authority conducting transparent observational research and publication of healthcare outcomes from health technologies (devices, diagnostics, medicines, therapies) and interventional procedures in the NHS. Civil Registration mortality data informs overall death rates (any cause) for the patient cohort selected for analysis. Out of hospital deaths and cause of death are not reported through HES, therefore the combined data are more robust for analysis than either source alone. NUTH cannot reasonably achieve the above in a different way. NHS providers have administrative systems which feed into HES; therefore this represents the only national routine dataset of NHS activity in secondary and tertiary care. HES may record in-hospital deaths, but Civil Registration mortality is the only legal record of cause of death, independent of place of death. The pseudonymised data analysis methods developed by NMPCE (non-identifiable patient data) are the least intrusive means of achieving the purpose of NHS healthcare outcomes analysis and publication, in the public interest.

The lawful basis for processing special category data (health data) is under Article 9 (2)(j) of the UK GDPR. The justification is as follows:

Not all the projects are intended to generate generalisable new knowledge for academic publication. NMPCE outcomes research programme also includes processing for reasons of public health, including adoption of medical technologies and outcomes of interventional procedures, to ensure high standards of quality and safety, as described in this application.

Moral and ethical issues and risk of potential harm to the public from the data processing methods, have been considered through active engagement with patient and public involvement (PPI) groups in the Trust and NUTH. Patients and their representatives have told NUTH that they positively endorse and demand such use of their data to benefit medical research and the NHS, as described in this agreement.

The data requested will achieve the aim identified by permitting a timely response from NUTH, to questions of safety and efficacy of medical devices and interventional procedures in current NHS practice.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is the sole Controller and also processes the data for this study. No other organisations process the data for this purpose.

NMPCE has used HES data since 2011 in support of its research and evaluation work for several stakeholders. Funding from stakeholders is used to support NMPCE staff time for health technology evaluation, which may or may not include analysis of data from NHS England. However, NMPCE exercises overall responsibility and control over the purposes and means of the processing personal data received from NHS England. NMPCE determines whether HES data should be used, defines the query, conducts data cleaning, and data analysis. In doing so the NMPCE fulfils the role of sole controller by completing the following tasks:

- Decision to process the personal data

- Decision on outcome of processing

- Decision on what datafields and datasets to process

- Decision on cohort selection

- Exercising professional judgement in the processing of personal data

- Have complete autonomy as to how the personal data is processed

Since 2011, NUTH has hosted one of the current National Institute for Health and Care Excellence (NICE) External Assessment Centres (EAC). The NICE EAC is based in the NMPCE Directorate at the Freeman Hospital in Newcastle. The work commissioned by NICE evaluates health technologies and interventional procedures and answers specific research questions raised by the NICE Medical Technologies Advisory Committee (MTAC), Diagnostics Advisory Committee (DAC), Interventional Procedures Advisory Committee (IPAC) and other NICE programmes. NICE had provided a commissioner letter for the External Assessment Centre within NMPCE to support this agreement, which extends to 2025.

NUTH was formerly an HSCIC / NHS Digital approved HES Business Objects (HESBO) and HES Data Interrogation System (HDIS) user. The NMPCE analysts within NUTH are trained and were approved to process episode level pseudonymised HES data for the research purposes described in this DARS agreement, for NHS Health and Social Care system, patient and public benefit. This agreement is to renew and continue the quarterly managed extract service of pseudonymised HES (APC, CC, OP, AE/ECDS) and Civil Registration data from NHS England, which will permit NMPCE to continue delivering the described research objectives for an additional twelve months.

The nature of the work described in this agreement is a wide programme of collaborative research as described above, including long term follow up. The detailed scope of data analysis is determined on a project by project basis, but always within the overarching terms of the Data Sharing Agreement with NHS England.

The data subjects are also determined on a project by project basis, following the convention of defining Population, Intervention, Comparator, Outcomes (PICO) and setting (APC / A&E etc) for every intended analysis project. A protocol for pseudonymised matching is drafted before work commences, where the HES and Civil Registration mortality codes to match the required PICO and setting are determined in an iterative process. Data cleaning rules are applied to reach the final agreed and matched set of data subjects, which may be for a single-armed, or comparative analysis, using control and cohort groups.

Detailed information about the purpose of the projects covered by this agreement is provided below, including the data required, why these data are required and justification for the datasets requested, the level of data (pseudonymised), the number of years requested and the geographical spread of the data requested.

NUTH confirm that there are no alternative, less intrusive ways of achieving the purpose, and details of the efforts taken to minimise the data required are also described below:

Pseudonymised data from the requested datasets are required by NMPCE for the following purposes / activities:

i) to determine the number of patients with a particular disease requiring NHS care, the number being treated with a health technology across England and identification of key providers of care, (e.g. to inform NICE of scope/coverage/uptake of guidance, to inform sample size determination and to target study invitations to relevant organisations. An example where this type of activity led to a successful publication was, NMPCE's cross-sectional Recurrent Respiratory Papillomatosis (RRP) survey which determined the number of patients diagnosed with the condition, the proportion receiving hospital treatment and the breakdown of interventions used in this population (Donne et al. Clinical Otolaryngology 2016). The survey found at least 16 different interventions in use in the NHS for RRP and limited evidence on their comparative safety and efficacy. This work led to a successful application for NIHR Research for Patient Benefit grant funding and the establishment of a national data collection on the existing Airway Intervention Register (AIR) platform for analysis to determine the safest and most effective intervention(s) for the condition, which will benefit current and future RRP patients.),

ii) to determine patient demographics receiving healthcare interventions (e.g. common comorbidities, comorbidity score index, age distribution, gender),

iii) to determine in-hospital outcomes (e.g. efficacy and safety, length of stay) of health care interventions,

iv) to conduct active surveillance/longitudinal analysis to determine long-term outcomes (e.g. efficacy, safety and in-hospital mortality) of healthcare interventions. For example the team undertook an 8 year analysis of almost 100,000 women with surgical insertion of mesh implants for the treatment of stress urinary incontinence which led to a peer reviewed publication (Keltie et al. Scientific Reports. 2017). The retention of HES data up to a maximum of 20 years will permit further, longer term follow-up analyses of these women, to establish the as-yet unknown longevity of mesh implants and any change in efficacy and complication rates as they age,

v) to compare patient outcomes (including in-hospital and longitudinal outcomes) between healthcare interventions (e.g. through propensity matching of cohorts based on comorbidities) to determine relative efficacy and safety. The retention of HES data up to a maximum of 20 years will permit future analyses of paediatric RRP patients in the AIR registry, to determine which treatments have enduring efficacy outcomes into adulthood,

vi) to determine hospital resource usage, including prior to and following healthcare interventions (e.g. patient pathway, number and indication of outpatient attendances, subsequent hospital admission, further interventions, accident and emergency attendances, in-hospital deaths, HRG codes/costs). An example where the team used HES data in this way is the published study of procedural and short-term efficacy of bronchial thermoplasty (Burn et al., 2016),

vii) to determine national coverage of and encourage data submission to national clinical or procedural registries (e.g. identification of trusts not contributing data to registries, data completeness of registry in terms of total number of procedures not recorded in registry, data completeness of registry in terms of complications reported using anonymised matching techniques based on for example treating trust, procedure date, gender, age datafields). The study undertook this type of active surveillance for the Airway Intervention Registry, using the former HDIS system. This is necessary to understand whether the registry data can be considered representative of the total population of interest, and to measure the success of efforts to improve data quality and completeness.

viii) to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching technique based on for example treating trust, sex, procedure date, gender, age data fields),

ix) to conduct exploratory analysis to determine whether outcomes (e.g. complications, diagnoses, in-hospital death) are associated with patient characteristics or setting. An example where the team used the data in this way was in a published analysis of the effect of hospital centre volume on iliac artery stenting outcomes (Goode et al., 2013),

x) study of outcomes and resource usage before and after an event (e.g. new device/procedure) to determine impact.

The most significant issue with the use of HES data alone in healthcare outcomes research is missing information on out-of-hospital deaths (date and cause of death) - noting that from recent publication that 50% of all deaths occur out of hospital. This risks overestimating the efficacy and safety of medical devices and interventional procedures in the aggregated results reported to NICE by NMPCE, to inform national NICE Guidance.

Hence, in addition to the above ten NMPCE purposes / activities, pseudonymised mortality data are required by NMPCE for the following purposes / activities:

xi) to add to the above longitudinal analysis in iv), to determine the long-term outcome of all-cause mortality, requiring both in-hospital (HES) and out-of-hospital mortality data of healthcare interventions,

xii) to add out-of-hospital mortality data to v) above, in comparative patient outcomes analysis,

xiii) to add out-of-hospital mortality data to vi) above, in analysing hospital resource usage prior to and following healthcare interventions

xiv) to add out-of-hospital mortality data to viii) above, to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching techniques based on for example treating trust, procedure date, gender and age data fields),

xv) to add out-of-hospital mortality data to ix) above, to conduct exploratory analysis to determine whether outcomes, including out-of-hospital deaths, are associated with patient characteristics or setting (e.g. age, gender, comorbidities, treating hospital).

In terms of data minimisation, due to the variety of analyses undertaken by NMPCE (different specialties, time durations of interest, research questions posed) the data and data subjects will vary by project. However, NUTH have considered the data fields requested from each data set (mortality, APC, CC, OP, ECDS) and each have been minimised accordingly. As a previous HESBO and HDIS user, NMPCE has used a specific subset of HES and Mortality data fields to inform evaluations of safety and efficacy for given healthcare interventions. For example: admission and discharge information, all diagnoses and procedure codes (4 character), episode and spell information (to build up an entire patient admission), treating organization, referring organization, patient age and gender.

National data are required due to the breadth of healthcare intervention (health technologies and interventional procedures) development and evaluation conducted (i.e. any condition, any procedure, any population). Access to HES data for the whole of England is also particularly important for NMPCE when considering interventions for rare diseases (affecting less than 5 in 10,000 of the general population), novel innovative procedures/devices/medicines coming into use, and for longitudinal analysis where patients may attend a number of different hospitals (e.g. treatment received in specialised treatment centre followed by follow-up care in local hospital). Furthermore, by using data from across the whole of England (rather than a specific localised region), NMPCE can assure that outcomes are generalisable and reflect current practices used across the country to inform national NICE Guidance. For NMPCE to conduct activities for NICE purposes, some of the research objectives, mean access is required to HES/mortality data from all of England.

No identifiable information is requested. NMPCE requires only a unique patient pseudonym that links across each data set: mortality, HES APC, OP, ECDS and CC to determine hospital resource usage and long-term outcomes for each patient.

The amount of mortality data requested has also been minimised by restricting by year to only those patients who have been admitted as inpatients (from 2007/08 onwards), or attended outpatients, emergency care, or critical care departments (from 2013/14 onwards). The frequency of mortality data requested is restricted to quarterly, to ensure safety signals are detected in a timely manner.

HES, ECDS and mortality data will not be used for any external commercial purposes not outlined in this agreement. These data will not be provided in record level form to any third party. These data will not be used for direct marketing. NUTH determines which data are processed for each project/research question and ensures the legal basis for requesting, storing and processing these data are met, therefore NUTH is the data controller and also processes the data for this study. No other organisations process the data for this purpose.

The programme of outcomes research using administrative and registry data currently has funding from the NIHR, Academic Health Science Network - North East and North Cumbria (AHSN NENC) and NICE. NMPCE has a track record of attracting research funding, and intends to seek future funding for their outcomes research programme, from public funding bodies, charities and organisations such as:

- NIHR projects (Invention for Innovation (i4i), Research for Patient Benefit (RfPB), Health Technology Assessment, Research Capability Funding, Newcastle Biomedical Research Centre;

- NIHR infrastructure funding (NIHR Newcastle IVD Co-operative, NIHR Innovation Observatory);

- NICE (the External Assessment Centre, EAC, described below);

- Academic Health Science Network (AHSN) North East and North Cumbria (NENC);

- Engineering and Physical Sciences Research Council (EPSRC);

- Charitable funding (Wellcome Trust, Innovate UK, British Heart Foundation, Newcastle Healthcare Charity);

- Professional societies supporting registry development;

- Commercial companies, which could include medical devices and diagnostics manufacturers.

NMPCE will produce an output in the form of a publication outlining an individual project's findings which, subject to acceptance, will be published in peer-reviewed academic or clinical journals. Additional outputs include grant application to funding bodies, conference presentations, summaries shared with public/patient representative groups and summaries shared on dedicated study websites (where applicable) and internal reports to NICE and other potential funders listed in the objectives for processing.

All outputs will contain only data that are aggregated with small numbers suppressed, in line with the HES analysis guide.

The dissemination and communication approach for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. Examples are described below.

Dissemination of results / outputs.

NMPCE will share aggregated information with NICE, to contribute to the production of national Medical Technologies, Interventional Procedure or Diagnostic Guidance, inform other NICE programmes, and thereby inform healthcare users, health care providers and commissioners of the safety and efficacy, national adoption and key providers of health technologies. An output will be an internal report outlining an individual project's findings submitted to NICE with small number suppression in line with the HES analysis guide.

Communication of results / outputs

One specific NIHR-funded project in progress is the Research for Patient Benefit award PB-PG-0416-20037 - Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP). Recruitment to the study in England will be compared (via routine monthly surveillance) to the number of RRP procedures being reported to the Hospital Episode Statistics (HES). This will enable NMPCE to invite new NHS hospitals/trusts to contribute to the study and also contact registered NHS hospitals/trusts to encourage them to enter data identified in HES to the registry where they have not done so. NMPCE have found that this method of feedback to NHS organisations (via Information Governance teams and clinical coding managers) has increased data entry to previous online databases as well as improving clinical coding accuracy for some interventional procedures.

Exploitation of results / outputs

NMPCE engages with clinical coders in acute trusts across England prior to analysis of HES data, but also provides feedback to clinical coders to demonstrate the external uses of coded hospital data, with the aim to continually improve data quality at source. All publications using HES/mortality data will reference NHS England and cite a copyright statement. NMPCE strives to make all peer-reviewed publications available in open access form in order to maximise the availability of information to health care users, health care providers and general members of the public. For example, NUTH published the analysis code (written in R) which analysed HES APC data to determine safety of mesh implants for stress urinary incontinence. Sharing methodology (not data) with the public will provide guidance to other researchers investigating other interventions in how to clean and analyse a large sample of HES data to identify complications; (Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015). The journal choice for any peer-reviewed publication will depend on the medical device intervention/interventional procedure/medical condition of interest (e.g. cardiac, ENT, oncology) and focus of the research question posed (e.g. methodology, patient safety, national policy).

Target dates to produce the outputs for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. In general, NMPCE projects involving medical device intervention/interventional procedure development or evaluation arise with clinical need, with timeframes governed by suitable funding calls and duration of grant funding.

In the specific example of the NIHR-funded Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP) project, a rolling monthly extract from the AIR will be compared with the latest quarterly HES extract from NHS England in active surveillance of coverage of the registry (total number of RRP cases treated in England versus total number entered in the registry). The target date for recruitment completion was 31st August 2022. A publication with provisional title of “Current practice, safety and efficacy of interventions for Recurrent Respiratory Papillomatosis: evidence from a UK registry” has been submitted (09/02/2023) and under review at the Clinical Otolaryngology journal.

In the specific case of NMPCE projects commissioned by NICE, each individual project is issued with a timeline of dates for completion of project milestones. Most NICE projects require peer-reviewed publication of results within 12 months of project completion. NMPCE strives to share analysis code (where possible ) for transparency and reproducibility purposes for future researchers.

NICE will consider evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

i) NICE IPG434: Radiofrequency cold ablation for respiratory papillomatosis (published November 2012 - update required).

ii) NICE Helium plasma coagulation for treating Respiratory Papillomatosis – monitored (awaiting publication of further evidence)

NMPCE and NICE and NHS England worked collaboratively on a project looking at data linkage methodology (manuscript available freely online: “Comparison of identifiable and non-identifiable data linkage: health technology assessment of MitraClip using registry, administrative and mortality datasets” BMJ Health Care Inform. 2021; 28(1): e100223). NMPCE is currently working with NHS England on following longitudinal outcomes from topics exiting Cancer Drug Fund using data from HES.

Examples of peer-reviewed publications (available in the public domain) resulting from the analysis of HES data previously accessed via HDIS by NMPCE for NICE purposes, with details of the impact these have achieved are as follows:

ia) Keltie et al. Paediatric tonsillectomy in England: a cohort study of clinical practice and outcomes using Hospital Episode Statistics data (2008-2019). Clin Otolaryngol. 2021; 46(3): 552-561.

IMPACT: Analysed 318,453 paediatric tonsillectomies conducted in NHS hospitals in England between 2008 and 2019, demonstrating increase in complications.

ib) Powell et al. Coblation intracapsular tonsillectomy: A cohort study of NHS practice in England using Hospital Episode Statistics. Clin Otolaryngol. 2022; 47(3): 471-7

IMPACT: Analysed a total of 5525 intracapsular coblation procedures from 4 ENT centres in England.

This has triggered an IPG150 review at NICE, and grant applications to NIHR in order to further investigate why the safety profile of tonsillectomy has changed over time.

ii) Page et al. Prostatic urethral life (UroLift): a real-world analysis of outcomes using hospital episodes statistics. BMC Urol. 2021; 21(1): 55.

IMPACT: Demonstrates uptake of UroLift in England after UroLift was added to the Innovation Technology Tariff (ITT) in April 2017 and selected as a Rapid Uptake Product by the Accelerated Access Collaborative in 2018. This real-world analysis shows that UroLift implantation can be delivered safely in a day-case setting with minimal morbidity. However, hospital resource usage for catheterization and emergency hospital attendance in the first 30 days was substantial, and 12% required re-treatment at 2 years.

iii) Powell et al. Balloon dilatation for paediatric airway stenosis: Evidence from the UK Airway Intervention Registry. Clin Otolaryngol. 2020; 45(3): 334-341.

IMPACT: Conducted pseudonymised linkage to determine registry coverage, and also to determine longitudinal outcomes (further respiratory surgery, death).

iv) Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015.

IMPACT: Complications from surgical mesh procedures have led to legal cases against manufacturers worldwide and to national inquiries about their safety. This was one of the largest studies so far looking at adverse event rates of these procedures, and has significantly added to the understanding of the likelihood of adverse events from mesh procedures in a UK NHS setting. This means that women considering mesh procedures in future will have more accurate safety information available to them, so they can make a more informed choice. This publication and its analysis code was published in an open-access form to share both methodology and findings widely. NICE has also updated its NG123 guidance and produced linked Patient Decision Aids for women considering mesh surgery.

v) Burn et al. Procedural and short-term safety of bronchial thermoplasty in clinical practice: evidence from a national registry and Hospital Episode Statistics. J Asthma. 2017; 54(8): 872-879.

IMPACT: Bronchial thermoplasty is a novel treatment for severe asthma, however its mode of action and target patient population were poorly defined, with little evidence available on efficacy and safety in terms of quality and quantity. This study presents procedural and short-term safety evidence from routine UK clinical practice.

vi) Donne et al. Prevalence and management of recurrent respiratory papillomatosis (RRP) in the UK: cross-sectional study. Clinical Otolaryngology. 2017; 42(1):86-91.

Supported by Sims et al. Our experience in developing and operating the Airway Intervention Registry for Recurrent Respiratory Papillomatosis (AIR-RRP): national data collection. NIHR Open. 2023; 2: 22.

IMPACT: Recurrent respiratory papillomatosis is a relatively rare condition, but has a significant impact on patients and their families as it requires repeated anaesthetic procedures to achieve symptomatic control. This study utilised a cross-sectional survey of ENT consultants and analysis of HES data to provide, for the first time, an estimate for the prevalence of recurrent respiratory papillomatosis in the UK. This prevalence informed the sample size for an NIHR Research for Patient Benefit grant application - 3 year award granted with extension due to suspended recruitment to non-COVID 19 studies. Additional funding gained from NICE to support an additional 12 months (end 31st August 2022). Identification of key providers of RRP care in the UK, enabled NMPCE to invite specific acute trusts to contribute to further research in RRP patients, and contribute to future NICE guidance (in update of IPG 434).

vii) Goode et al. Effect of procedure volume on outcomes after iliac artery angioplasty and stenting. British Journal of Surgery. 2013; 100:1189-1196.

IMPACT: This study was able to use HES data to determine if there was an association between outcomes of endovascular iliac artery intervention and centre volume, describing methodology which could be applied to any interventional in-hospital procedure which would benefit future analysis of HES data.

viii) Patrick et al. Monitoring the use and outcomes of new devices and procedures: how does coding affect what Hospital Episode Statistics contribute? Lessons from 12 emerging procedures 2006-10. Journal of Public Health. 2012; 35(1): 132-138.

IMPACT: This study provided evidence that routinely collected HES data have the potential to support quality improvements and evidence-based commissioning of devices and procedures in the national health service.

ix). Keltie et al. Identifying complications of interventional procedures from UK routine healthcare databases: a systematic search for methods using clinical codes. BMC Medical Research Methodology 2014; 14:126.

IMPACT: This study summarised methods of analysing complications of interventional procedures from routine healthcare databases in the UK published in the literature. This study was published in open-access form to share learnings with other researchers and clinical coding teams across the UK to encourage high quality research from HES.

x) Powell J et al. National cohort study of health care resource after pediatric tracheostomy. JAMA Pediatr. 2022; 176 (8): 817-819.

IMPACT: This study demonstrated the significant healthcare utilisation and cost burden associated with hospital readmissions in children with tracheostomies. In England service specifications exist for certain specialist conditions or interventions. Service specifications are clearly defined standards of care expected from organisations funded by NHS England to provide specialised care. Paediatric tracheostomy is not currently included in this list, meaning that the levels of care and support are only defined on a local level. This study recommended the development of a paediatric tracheostomy service specifications, including extensive multidisciplinary support, to help reduce healthcare utilisation in the patient group.

xi) Veeratterapillay R et al. Infection after ureteroscopy for ureteric stones: analysis of 71,305 cases in the Hospital Episode Statistics database. BJU Int. 2023; 131 (1): 109-115.

IMPACT: This study represented the largest series evaluating infectious complications after ureteroscopic stone treatment. The procedure was found to be safe, with low inpatient infective complication and critical care admission rates. However the study highlighted the hospital after care resources required, which were previously unknown.

These publications inform clinical practice in the NHS, and can contribute to national NICE guidance. For example, NICE has incorporated evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

e.g. NICE IPG635: Bronchial thermoplasty for severe asthma (originally published January 2012 - updated December 2018; as a method of active surveillance of safety signals following the intervention.

e.g. NICE IPG425: Endoscopic balloon dilatation for subglottic or tracheal stenosis (originally published April 2012, and replaced with IPG719 in March 2022).

NMPCE has also supported Medical Technology evaluation of two devices (Danis stent, SpaceOAR, MAGECRods) through surveillance of HES to determine coverage of use across in England; this give a direct measure of health technology uptake.

NMPCE has used data from HES to support the drafting of grant applications applicable to a number of NIHR funding streams (including: Digital technologies, HTA, Ambitious data-enabled trials, health services and public health research studies, RfPB) focusing on health outcomes. Research themes for these grant applications have included (but are not limited to):

- safety and efficacy of tonsillectomy,

- safety and efficacy of adenoidectomy,

- longitudinal study to investigate the uptake and safety of benign prostate hyperplasia (BPH) procedures,

- longitudinal study to investigate the association between kidney transplantation and urological cancers,

- longitudinal study of stroke patients and pneumonia outcomes,

- longitudinal study of patients with laryngeal cancer.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is one of the country's top performing trusts in terms of the number of research studies it has supported, with 501 studies delivered by the trust in 2018/2019. Northern Medical Physics and Clinical Engineering (NMPCE) is a Clinical Directorate within NUTH, with research interests and a track record going back 40 years in the development and evaluation of health technology interventions including medical devices, diagnostics, and interventional procedures.

NMPCE will use Hospital Episode Statistics (HES), Emergency Care Dataset (ECDS) and Civil Registration mortality data to support its programme of research which includes studying:

i) uptake and outcomes of novel medical technologies (medical devices, diagnostics, medicines, therapies), and interventional procedures to assess peri-procedural, short-term and long-term safety and efficacy

ii) associations between comorbidities, diagnoses, and hospital admissions

iii) patient pathways through the healthcare system.

This is a not-for-profit research programme which uses large pseudonymised data sets to achieve the aims described. The outputs of the programme are intended to inform national guidance, decision-making, research recommendations, study designs, grant applications to national bodies (a comprehensive list of outputs is described in 5c).

The justification for Article 6(1)(e) processing, as necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller is summarised briefly as follows:

NUTH are a public authority conducting transparent observational research and publication of healthcare outcomes from health technologies (devices, diagnostics, medicines, therapies) and interventional procedures in the NHS. Civil Registration mortality data informs overall death rates (any cause) for the patient cohort selected for analysis. Out of hospital deaths and cause of death are not reported through HES, therefore the combined data are more robust for analysis than either source alone. NUTH cannot reasonably achieve the above in a different way. NHS providers have administrative systems which feed into HES; therefore this represents the only national routine dataset of NHS activity in secondary and tertiary care. HES may record in-hospital deaths, but Civil Registration mortality is the only legal record of cause of death, independent of place of death. The pseudonymised data analysis methods developed by NMPCE (non-identifiable patient data) are the least intrusive means of achieving the purpose of NHS healthcare outcomes analysis and publication, in the public interest.

The public interest justification for Article 9(2)(j) – for public health purposes is summarised briefly as follows:

Not all the projects are intended to generate generalisable new knowledge for academic publication. NMPCE outcomes research programme also includes processing for reasons of public health, including adoption of medical technologies and outcomes of interventional procedures, to ensure high standards of quality and safety, as described in this application.

Moral and ethical issues and risk of potential harm to the public from the data processing methods, have been considered through active engagement with patient and public involvement (PPI) groups in the Trust and NUTH. Patients and their representatives have told NUTH that they positively endorse and demand such use of their data to benefit medical research and the NHS, as described in this agreement.

The data requested will achieve the aim identified by permitting a timely response from NUTH, to questions of safety and efficacy of medical devices and interventional procedures in current NHS practice.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is the sole Data Controller and also processes the data for this study. No other organisations process the data for this purpose.

NMPCE has used HES data since 2011 in support of its research and evaluation work for several stakeholders. Funding from stakeholders is used to support NMPCE staff time for health technology evaluation, which may or may not include analysis of data from NHS Digital. However, NMPCE exercises overall responsibility and control over the purposes and means of the processing personal data received from NHS Digital. NMPCE determines whether HES data should be used, defines the query, conducts data cleaning, and data analysis. In doing so the NMPCE fulfils the role of sole data controller (according to the ICO definition) by completing the following tasks:

- Decision to process the personal data

- Decision on outcome of processing

- Decision on what datafields and datasets to process

- Decision on cohort selection

- Exercising professional judgement in the processing of personal data

- Have complete autonomy as to how the personal data is processed

Since 2011, NUTH has hosted one of the current National Institute for Health and Care Excellence (NICE) External Assessment Centres (EAC). The NICE EAC is based in the NMPCE Directorate at the Freeman Hospital in Newcastle. The work commissioned by NICE evaluates health technologies and interventional procedures and answers specific research questions raised by the NICE Medical Technologies Advisory Committee (MTAC), Diagnostics Advisory Committee (DAC), Interventional Procedures Advisory Committee (IPAC) and other NICE programmes. NICE had provided a commissioner letter for the External Assessment Centre within NMPCE to support this agreement, which extends to 2025.

NUTH was formerly an HSCIC / NHS Digital approved HES Business Objects (HESBO) and HES Data Interrogation System (HDIS) user. The NMPCE analysts within NUTH are trained and were approved to process episode level pseudonymised HES data for the research purposes described in this DARS agreement, for NHS Health and Social Care system, patient and public benefit. This agreement is to renew and continue the quarterly managed extract service of pseudonymised HES (APC, CC, OP, AE/ECDS) and Civil Registration data from NHS Digital, which will permit NMPCE to continue delivering the described research objectives for an additional twelve months.

The nature of the work described in this agreement is a wide programme of collaborative research as described above, including long term follow up. The detailed scope of data analysis is determined on a project by project basis, but always within the overarching terms of the Data Sharing Agreement with NHS Digital.

The data subjects are also determined on a project by project basis, following the convention of defining Population, Intervention, Comparator, Outcomes (PICO) and setting (APC / A&E etc) for every intended analysis project. A protocol for pseudonymised matching is drafted before work commences, where the HES and Civil Registration mortality codes to match the required PICO and setting are determined in an iterative process. Data cleaning rules are applied to reach the final agreed and matched set of data subjects, which may be for a single-armed, or comparative analysis, using control and cohort groups.

Detailed information about the purpose of the projects covered by this agreement is provided below, including the data required, why these data are required and justification for the datasets requested, the level of data (pseudonymised), the number of years requested and the geographical spread of the data requested.

NUTH confirm that there are no alternative, less intrusive ways of achieving the purpose, and details of the efforts taken to minimise the data required are also described below:

Pseudonymised data from the requested datasets are required by NMPCE for the following purposes / activities:

i) to determine the number of patients with a particular disease requiring NHS care, the number being treated with a health technology across England and identification of key providers of care, (e.g. to inform NICE of scope/coverage/uptake of guidance, to inform sample size determination and to target study invitations to relevant organisations. An example where this type of activity led to a successful publication was, NMPCE's cross-sectional Recurrent Respiratory Papillomatosis (RRP) survey which determined the number of patients diagnosed with the condition, the proportion receiving hospital treatment and the breakdown of interventions used in this population (Donne et al. Clinical Otolaryngology 2016). The survey found at least 16 different interventions in use in the NHS for RRP and limited evidence on their comparative safety and efficacy. This work led to a successful application for NIHR Research for Patient Benefit grant funding and the establishment of a national data collection on the existing Airway Intervention Register (AIR) platform for analysis to determine the safest and most effective intervention(s) for the condition, which will benefit current and future RRP patients.),

ii) to determine patient demographics receiving healthcare interventions (e.g. common comorbidities, comorbidity score index, age distribution, gender),

iii) to determine in-hospital outcomes (e.g. efficacy and safety, length of stay) of health care interventions,

iv) to conduct active surveillance/longitudinal analysis to determine long-term outcomes (e.g. efficacy, safety and in-hospital mortality) of healthcare interventions. For example we undertook an 8 year analysis of almost 100,000 women with surgical insertion of mesh implants for the treatment of stress urinary incontinence which led to a peer reviewed publication (Keltie et al. Scientific Reports. 2017). The retention of HES data up to a maximum of 20 years will permit further, longer term follow-up analyses of these women, to establish the as-yet unknown longevity of mesh implants and any change in efficacy and complication rates as they age,

v) to compare patient outcomes (including in-hospital and longitudinal outcomes) between healthcare interventions (e.g. through propensity matching of cohorts based on comorbidities) to determine relative efficacy and safety. The retention of HES data up to a maximum of 20 years will permit future analyses of paediatric RRP patients in the AIR registry, to determine which treatments have enduring efficacy outcomes into adulthood,

vi) to determine hospital resource usage, including prior to and following healthcare interventions (e.g. patient pathway, number and indication of outpatient attendances, subsequent hospital admission, further interventions, accident and emergency attendances, in-hospital deaths, HRG codes/costs). An example where we used HES data in this way is the published study of procedural and short-term efficacy of bronchial thermoplasty (Burn et al., 2016),

vii) to determine national coverage of and encourage data submission to national clinical or procedural registries (e.g. identification of trusts not contributing data to registries, data completeness of registry in terms of total number of procedures not recorded in registry, data completeness of registry in terms of complications reported using anonymised matching techniques based on for example treating trust, procedure date, gender, age datafields). We undertook this type of active surveillance for the Airway Intervention Registry, using the former HDIS system. This is necessary to understand whether the registry data can be considered representative of the total population of interest, and to measure the success of efforts to improve data quality and completeness.

viii) to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching technique based on for example treating trust, sex, procedure date, gender, age data fields),

ix) to conduct exploratory analysis to determine whether outcomes (e.g. complications, diagnoses, in-hospital death) are associated with patient characteristics or setting. An example where we used the data in this way was in a published analysis of the effect of hospital centre volume on iliac artery stenting outcomes (Goode et al., 2013),

x) study of outcomes and resource usage before and after an event (e.g. new device/procedure) to determine impact.

The most significant issue with the use of HES data alone in healthcare outcomes research is missing information on out-of-hospital deaths (date and cause of death) - noting that from recent publication that 50% of all deaths occur out of hospital. This risks overestimating the efficacy and safety of medical devices and interventional procedures in the aggregated results reported to NICE by NMPCE, to inform national NICE Guidance.

Hence, in addition to the above ten NMPCE purposes / activities, pseudonymised mortality data are required by NMPCE for the following purposes / activities:

xi) to add to the above longitudinal analysis in iv), to determine the long-term outcome of all-cause mortality, requiring both in-hospital (HES) and out-of-hospital mortality data of healthcare interventions,

xii) to add out-of-hospital mortality data to v) above, in comparative patient outcomes analysis,

xiii) to add out-of-hospital mortality data to vi) above, in analysing hospital resource usage prior to and following healthcare interventions

xiv) to add out-of-hospital mortality data to viii) above, to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching techniques based on for example treating trust, procedure date, gender and age data fields),

xv) to add out-of-hospital mortality data to ix) above, to conduct exploratory analysis to determine whether outcomes, including out-of-hospital deaths, are associated with patient characteristics or setting (e.g. age, gender, comorbidities, treating hospital).

In terms of data minimisation, due to the variety of analyses undertaken by NMPCE (different specialties, time durations of interest, research questions posed) the data and data subjects will vary by project. However, NUTH have considered the data fields requested from each data set (mortality, APC, CC, OP, ECDS) and each have been minimised accordingly. As a previous HESBO and HDIS user, NMPCE has used a specific subset of HES and Mortality data fields to inform evaluations of safety and efficacy for given healthcare interventions. For example: admission and discharge information, all diagnoses and procedure codes (4 character), episode and spell information (to build up an entire patient admission), treating organization, referring organization, patient age and gender.

National data are required due to the breadth of healthcare intervention (health technologies and interventional procedures) development and evaluation conducted (i.e. any condition, any procedure, any population). Access to HES data for the whole of England is also particularly important for NMPCE when considering interventions for rare diseases (affecting less than 5 in 10,000 of the general population), novel innovative procedures/devices/medicines coming into use, and for longitudinal analysis where patients may attend a number of different hospitals (e.g. treatment received in specialised treatment centre followed by follow-up care in local hospital). Furthermore, by using data from across the whole of England (rather than a specific localised region), NMPCE can assure that outcomes are generalisable and reflect current practices used across the country to inform national NICE Guidance. For NMPCE to conduct activities for NICE purposes, some of the “Objectives for processing” (described above) require access to HES/mortality data from all of England.

No identifiable information is requested. NMPCE requires only a unique patient pseudonym that links across each data set: mortality, HES APC, OP, ECDS and CC to determine hospital resource usage and long-term outcomes for each patient.

The amount of mortality data requested has also been minimised by restricting by year to only those patients who have been admitted as inpatients (from 2007/08 onwards), or attended outpatients, emergency care, or critical care departments (from 2013/14 onwards). The frequency of mortality data requested is restricted to quarterly, to ensure safety signals are detected in a timely manner.

HES, ECDS and mortality data will not be used for any external commercial purposes not outlined in this agreement. These data will not be provided in record level form to any third party. These data will not be used for direct marketing. NUTH determines which data are processed for each project/research question and ensures the legal basis for requesting, storing and processing these data are met, therefore NUTH is the data controller and also processes the data for this study. No other organisations process the data for this purpose.

The programme of outcomes research using administrative and registry data currently has funding from the NIHR, Academic Health Science Network - North East and North Cumbria (AHSN NENC) and NICE. NMPCE has a track record of attracting research funding, and intends to seek future funding for their outcomes research programme, from public funding bodies, charities and organisations such as:

- NIHR projects (Invention for Innovation (i4i), Research for Patient Benefit (RfPB), Health Technology Assessment, Research Capability Funding, Newcastle Biomedical Research Centre;

- NIHR infrastructure funding (NIHR Newcastle IVD Co-operative, NIHR Innovation Observatory);

- NICE (the External Assessment Centre, EAC, described below);

- Academic Health Science Network (AHSN) North East and North Cumbria (NENC);

- Engineering and Physical Sciences Research Council (EPSRC);

- Charitable funding (Wellcome Trust, Innovate UK, British Heart Foundation, Newcastle Healthcare Charity);

- Professional societies supporting registry development;

- Commercial companies, which could include medical devices and diagnostics manufacturers.

NMPCE will produce an output in the form of a publication outlining an individual project's findings which, subject to acceptance, will be published in peer-reviewed academic or clinical journals. Additional outputs include grant application to funding bodies, conference presentations, summaries shared with public/patient representative groups and summaries shared on dedicated study websites (where applicable) and internal reports to NICE and other potential funders listed in section 5a.

All outputs will contain only data that are aggregated with small numbers suppressed, in line with the HES analysis guide.

The dissemination and communication approach for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. Examples are described below.

Dissemination of results / outputs.

NMPCE will share aggregated information with NICE, to contribute to the production of national Medical Technologies, Interventional Procedure or Diagnostic Guidance, inform other NICE programmes, and thereby inform healthcare users, health care providers and commissioners of the safety and efficacy, national adoption and key providers of health technologies. An output will be an internal report outlining an individual project's findings submitted to NICE with small number suppression in line with the HES analysis guide.

Communication of results / outputs

One specific NIHR-funded project in progress is the Research for Patient Benefit award PB-PG-0416-20037 - Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP). Recruitment to the study in England will be compared (via routine monthly surveillance) to the number of RRP procedures being reported to the Hospital Episode Statistics (HES). This will enable NMPCE to invite new NHS hospitals/trusts to contribute to the study and also contact registered NHS hospitals/trusts to encourage them to enter data identified in HES to the registry where they have not done so. NMPCE have found that this method of feedback to NHS organisations (via Information Governance teams and clinical coding managers) has increased data entry to previous online databases as well as improving clinical coding accuracy for some interventional procedures.

Exploitation of results / outputs

NMPCE engages with clinical coders in acute trusts across England prior to analysis of HES data, but also provides feedback to clinical coders to demonstrate the external uses of coded hospital data, with the aim to continually improve data quality at source. All publications using HES/mortality data will reference NHS Digital and cite a copyright statement. NMPCE strives to make all peer-reviewed publications available in open access form in order to maximise the availability of information to health care users, health care providers and general members of the public. For example, NUTH published the analysis code (written in R) which analysed HES APC data to determine safety of mesh implants for stress urinary incontinence. Sharing methodology (not data) with the public will provide guidance to other researchers investigating other interventions in how to clean and analyse a large sample of HES data to identify complications; (Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015). The journal choice for any peer-reviewed publication will depend on the medical device intervention/interventional procedure/medical condition of interest (e.g. cardiac, ENT, oncology) and focus of the research question posed (e.g. methodology, patient safety, national policy).

Target dates to produce the outputs for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. In general, NMPCE projects involving medical device intervention/interventional procedure development or evaluation arise with clinical need, with timeframes governed by suitable funding calls and duration of grant funding.

In the specific example of the NIHR-funded Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP) project, a rolling monthly extract from the AIR will be compared with the latest quarterly HES extract from NHS Digital in active surveillance of coverage of the registry (total number of RRP cases treated in England versus total number entered in the registry). The target date for recruitment completion is currently 31st August 2022. A publication with provisional title of “Current practice, safety and efficacy of interventions for Recurrent Respiratory Papillomatosis: evidence from a UK registry” is planned for late 2022/early 2023.

In the specific case of NMPCE projects commissioned by NICE, each individual project is issued with a timeline of dates for completion of project milestones. Most NICE projects require peer-reviewed publication of results within 12 months of project completion. NMPCE strives to share analysis code (where possible and journal permitting) for transparency and reproducibility purposes for future researchers.

NICE will consider evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

i) NICE IPG434: Radiofrequency cold ablation for respiratory papillomatosis (published November 2012 - update required).

ii) NICE Helium plasma coagulation for treating Respiratory Papillomatosis – monitored (awaiting publication of further evidence)

Examples of peer-reviewed publications (available in the public domain) resulting from the analysis of HES data previously accessed via HDIS by NMPCE for NICE purposes, for example:

Specific outputs achieved:

NMPCE and NICE and NHS Digital worked collaboratively on a project looking at data linkage methodology (manuscript available freely online: “Comparison of identifiable and non-identifiable data linkage: health technology assessment of MitraClip using registry, administrative and mortality datasets” BMJ Health Care Inform. 2021; 28(1): e100223).

Examples of peer-reviewed publications (available in the public domain) resulting from the analysis of HES data previously accessed via HDIS by NMPCE for NICE purposes, with details of the impact these have achieved are as follows:

ia) Keltie et al. Paediatric tonsillectomy in England: a cohort study of clinical practice and outcomes using Hospital Episode Statistics data (2008-2019). Clin Otolaryngol. 2021; 46(3): 552-561.

IMPACT: Analysed 318,453 paediatric tonsillectomies conducted in NHS hospitals in England between 2008 and 2019, demonstrating increase in complications.

ib) Powell et al. Coblation intracapsular tonsillectomy: A cohort study of NHS practice in England using Hospital Episode Statistics. Clin Otolaryngol. 2022; 47(3): 471-7

IMPACT: Analysed a total of 5525 intracapsular coblation procedures from 4 ENT centres in England.

This has triggered an IPG150 review at NICE, and grant applications to NIHR in order to further investigate why the safety profile of tonsillectomy has changed over time.

ii) Page et al. Prostatic urethral life (UroLift): a real-world analysis of outcomes using hospital episodes statistics. BMC Urol. 2021; 21(1): 55.

IMPACT: Demonstrates uptake of UroLift in England after UroLift was added to the Innovation Technology Tariff (ITT) in April 2017 and selected as a Rapid Uptake Product by the Accelerated Access Collaborative in 2018. This real-world analysis shows that UroLift implantation can be delivered safely in a day-case setting with minimal morbidity. However, hospital resource usage for catheterization and emergency hospital attendance in the first 30 days was substantial, and 12% required re-treatment at 2 years.

iii) Powell et al. Balloon dilatation for paediatric airway stenosis: Evidence from the UK Airway Intervention Registry. Clin Otolaryngol. 2020; 45(3): 334-341.

IMPACT: Conducted pseudonymised linkage to determine registry coverage, and also to determine longitudinal outcomes (further respiratory surgery, death). This additional evidence has triggered an update to NICE Interventional Procedure Guidance on Endoscopic balloon dilation for subglottic or tracheal stenosis (IPG719, published March 2022)

iv) Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015.

IMPACT: Complications from surgical mesh procedures have led to legal cases against manufacturers worldwide and to national inquiries about their safety. This was one of the largest studies so far looking at adverse event rates of these procedures, and has significantly added to the understanding of the likelihood of adverse events from mesh procedures in a UK NHS setting. This means that women considering mesh procedures in future will have more accurate safety information available to them, so they can make a more informed choice. This publication and its analysis code was published in an open-access form to share both methodology and findings widely. NICE has also updated its NG123 guidance and produced linked Patient Decision Aids for women considering mesh surgery.

v) Burn et al. Procedural and short-term safety of bronchial thermoplasty in clinical practice: evidence from a national registry and Hospital Episode Statistics. J Asthma. 2017; 54(8): 872-879.

IMPACT: Bronchial thermoplasty is a novel treatment for severe asthma, however its mode of action and target patient population were poorly defined, with little evidence available on efficacy and safety in terms of quality and quantity. This study presents procedural and short-term safety evidence from routine UK clinical practice.

vi) Donne et al. Prevalence and management of recurrent respiratory papillomatosis (RRP) in the UK: cross-sectional study. Clinical Otolaryngology. 2017; 42(1):86-91.

IMPACT: Recurrent respiratory papillomatosis is a relatively rare condition, but has a significant impact on patients and their families as it requires repeated anaesthetic procedures to achieve symptomatic control. This study utilised a cross-sectional survey of ENT consultants and analysis of HES data to provide, for the first time, an estimate for the prevalence of recurrent respiratory papillomatosis in the UK. This prevalence informed the sample size for an NIHR Research for Patient Benefit grant application - 3 year award granted with extension due to suspended recruitment to non-COVID 19 studies. Additional funding gained from NICE to support an additional 12 months (end 31st August 2022). Identification of key providers of RRP care in the UK, enabled NMPCE to invite specific acute trusts to contribute to further research in RRP patients, and contribute to future NICE guidance (in update of IPG 434).

vii) Goode et al. Effect of procedure volume on outcomes after iliac artery angioplasty and stenting. British Journal of Surgery. 2013; 100:1189-1196.

IMPACT: This study was able to use HES data to determine if there was an association between outcomes of endovascular iliac artery intervention and centre volume, describing methodology which could be applied to any interventional in-hospital procedure which would benefit future analysis of HES data.

viii) Patrick et al. Monitoring the use and outcomes of new devices and procedures: how does coding affect what Hospital Episode Statistics contribute? Lessons from 12 emerging procedures 2006-10. Journal of Public Health. 2012; 35(1): 132-138.

IMPACT: This study provided evidence that routinely collected HES data have the potential to support quality improvements and evidence-based commissioning of devices and procedures in the national health service.

ix). Keltie et al. Identifying complications of interventional procedures from UK routine healthcare databases: a systematic search for methods using clinical codes. BMC Medical Research Methodology 2014; 14:126.

IMPACT: This study summarised methods of analysing complications of interventional procedures from routine healthcare databases in the UK published in the literature. This study was published in open-access form to share learnings with other researchers and clinical coding teams across the UK to encourage high quality research from HES.

These publications inform clinical practice in the NHS, and can contribute to national NICE guidance.

NMPCE has also supported Medical Technology evaluation of two devices (Danis stent, SpaceOAR, MAGECRods) through surveillance of HES to determine coverage of use across in England; this give a direct measure of health technology uptake. In addition, NICE has incorporated evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance: e.g. NICE IPG635: Bronchial thermoplasty for severe asthma (originally published January 2012 - updated December 2018; as a method of active surveillance of safety signals following the intervention.

NMPCE has used data from HES to support the drafting of grant applications applicable to a number of NIHR funding streams (including: Digital technologies, HTA, Ambitious data-enabled trials, health services and public health research studies, RfPB) focusing on health outcomes. Research themes for these grant applications have included (but are not limited to):

- safety and efficacy of tonsillectomy,

- longitudinal study to investigate the uptake and safety of UroLift procedures,

- longitudinal study to investigate the association between kidney transplantation and urological cancers,

- short-term evaluation of ureteroscopy and infection outcomes,

- longitudinal study of stroke patients and pneumonia outcomes.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is one of the country's top performing trusts in terms of the number of research studies it has supported, with 501 studies delivered by the trust in 2018/2019. Northern Medical Physics and Clinical Engineering (NMPCE) is a Clinical Directorate within NUTH, with research interests and a track record going back 40 years in the development and evaluation of health technology interventions including medical devices, diagnostics, and interventional procedures.

NMPCE will use Hospital Episode Statistics (HES), Emergency Care Dataset (ECDS) and Civil Registration mortality data to support its programme of research which includes studying:

i) uptake and outcomes of novel medical technologies (medical devices, diagnostics, medicines, therapies), and interventional procedures to assess peri-procedural, short-term and long-term safety and efficacy

ii) associations between comorbidities, diagnoses, and hospital admissions

iii) patient pathways through the healthcare system.

This is a not-for-profit research programme which uses large pseudonymised data sets to achieve the aims described. The outputs of the programme are intended to inform national guidance, decision-making, research recommendations, study designs, grant applications to national bodies (a comprehensive list of outputs is described in 5c).

The justification for Article 6(1)(e) processing, as necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller is summarised briefly as follows:

NUTH are a public authority conducting transparent observational research and publication of healthcare outcomes from health technologies (devices, diagnostics, medicines, therapies) and interventional procedures in the NHS. Civil Registration mortality data informs overall death rates (any cause) for the patient cohort selected for analysis. Out of hospital deaths and cause of death are not reported through HES, therefore the combined data are more robust for analysis than either source alone. NUTH cannot reasonably achieve the above in a different way. NHS providers have administrative systems which feed into HES; therefore this represents the only national routine dataset of NHS activity in secondary and tertiary care. HES may record in-hospital deaths, but Civil Registration mortality is the only legal record of cause of death, independent of place of death. The pseudonymised data analysis methods developed by NMPCE (non-identifiable patient data) are the least intrusive means of achieving the purpose of NHS healthcare outcomes analysis and publication, in the public interest.

The public interest justification for Article 9(2)(j) – for public health purposes is summarised briefly as follows:

Not all the projects are intended to generate generalisable new knowledge for academic publication. NMPCE outcomes research programme also includes processing for reasons of public health, including adoption of medical technologies and outcomes of interventional procedures, to ensure high standards of quality and safety, as described in this application.

Moral and ethical issues and risk of potential harm to the public from the data processing methods, have been considered through active engagement with patient and public involvement (PPI) groups in the Trust and NUTH, Patients and their representatives have told NUTH that they positively endorse and demand such use of their data to benefit medical research and the NHS, as described in this agreement.

The data requested will achieve the aim identified by permitting a timely response from NUTH, to questions of safety and efficacy of medical devices and interventional procedures in current NHS practice.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is the sole Data Controller and also processes the data for this study. No other organisations process the data for this purpose.

NMPCE has used HES data since 2011 in support of its research and evaluation work for several stakeholders. Funding from stakeholders is used to support NMPCE staff time for health technology evaluation, which may or may not include analysis of data from NHS Digital. However, NMPCE exercises overall responsibility and control over the purposes and means of the processing personal data received from NHS Digital. NMPCE determines whether HES data should be used, defines the query, conducts data cleaning, and data analysis. In doing so the NMPCE fulfils the role of sole data controller (according to the ICO definition) by completing the following tasks:

- Decision to process the personal data

- Decision on outcome of processing

- Decision on what datafields and datasets to process

- Decision on cohort selection

- Exercising professional judgement in the processing of personal data

- Have complete autonomy as to how the personal data is processed

Since 2011, NUTH has hosted one of five current National Institute for Health and Care Excellence (NICE) External Assessment Centres (EAC). The NICE EAC is based in the NMPCE Directorate at the Freeman Hospital in Newcastle. The work commissioned by NICE evaluates health technologies and interventional procedures and answers specific research questions raised by the NICE Medical Technologies Advisory Committee (MTAC), Diagnostics Advisory Committee (DAC), Interventional Procedures Advisory Committee (IPAC) and other NICE programmes. NICE had provided a commissioner letter for the External Assessment Centre within NMPCE to support this agreement, which extends to 2021.

NUTH was formerly an HSCIC / NHS Digital approved HES Business Objects (HESBO) and HES Data Interrogation System (HDIS) user. The NMPCE analysts within NUTH are trained and were approved to process episode level pseudonymised HES data for the research purposes described in this DARS application, for NHS Health and Social Care system, patient and public benefit. This agreement is to renew and continue the quarterly managed extract service of pseudonymised HES (APC, CC, OP, AE/ECDS) and Civil Registration data from NHS Digital, which will permit NMPCE to continue delivering the described research objectives for an additional twelve months.

The nature of the work described in this agreement is a wide programme of collaborative research, including long term follow up. The detailed scope of data analysis is determined on a project by project basis, but always within the overarching terms of the Data Sharing Agreement with NHS Digital.

The data subjects are also determined on a project by project basis, following the convention of defining Population, Intervention, Comparator, Outcomes (PICO) and setting (APC / A&E etc) for every intended analysis project. A protocol for pseudonymised matching is drafted before work commences, where the HES and Civil Registration mortality codes to match the required PICO and setting are determined in an iterative process. Data cleaning rules are applied to reach the final agreed and matched set of data subjects, which may be for a single-armed, or comparative analysis, using control and cohort groups.

Detailed information about the purpose of the projects covered by this agreement is provided below, including the data required, why these data are required and justification for the datasets requested, the level of data (pseudonymised), the number of years requested and the geographical spread of the data requested.

NUTH confirm that there are no alternative, less intrusive ways of achieving the purpose, and details of the efforts taken to minimise the data required are also described below:

Pseudonymised data from the requested datasets are required by NMPCE for the following purposes / activities:

i) to determine the number of patients with a particular disease requiring NHS care, the number being treated with a health technology across England and identification of key providers of care, (e.g. to inform NICE of scope/coverage/uptake of guidance, to inform sample size determination and to target study invitations to relevant organisations. For example, NMPCE's cross-sectional Recurrent Respiratory Papillomatosis (RRP) survey determined the number of patients diagnosed with the condition, the proportion receiving hospital treatment and the breakdown of interventions used in this population (Donne et al. Clinical Otolaryngology 2016). The survey found at least 16 different interventions in use in the NHS for RRP and limited evidence on their comparative safety and efficacy. This work led to a successful application for NIHR Research for Patient Benefit grant funding and the establishment of a national data collection on the existing Airway Intervention Register (AIR) platform for analysis to determine the safest and most effective intervention(s) for the condition, which will benefit current and future RRP patients.),

ii) to determine patient demographics receiving healthcare interventions (e.g. common comorbidities, comorbidity score index, age distribution, gender),

iii) to determine in-hospital outcomes (e.g. efficacy and safety, length of stay) of health care interventions,

iv) to conduct active surveillance/longitudinal analysis to determine long-term outcomes (e.g. efficacy, safety and in-hospital mortality) of healthcare interventions (e.g. 8 year analysis of almost 100,000 women with surgical insertion of mesh implants for the treatment of stress urinary incontinence, Keltie et al. Scientific Reports. 2017). The retention of HES data up to a maximum of 20 years will permit further, longer term follow-up analyses of these women, to establish the as-yet unknown longevity of mesh implants and any change in efficacy and complication rates as they age,

v) to compare patient outcomes (including in-hospital and longitudinal outcomes) between healthcare interventions (e.g. through propensity matching of cohorts based on comorbidities) to determine relative efficacy and safety. The retention of HES data up to a maximum of 20 years will permit future analyses of paediatric RRP patients in the AIR registry, to determine which treatments have enduring efficacy outcomes into adulthood,

vi) to determine hospital resource usage, including prior to and following healthcare interventions (e.g. patient pathway, number and indication of outpatient attendances, subsequent hospital admission, further interventions, accident and emergency attendances, in-hospital deaths, HRG codes/costs), for example in the published study of procedural and short-term efficacy of bronchial thermoplasty (Burn et al., 2016),

vii) to determine national coverage of and encourage data submission to national clinical or procedural registries (e.g. identification of trusts not contributing data to registries, data completeness of registry in terms of total number of procedures not recorded in registry, data completeness of registry in terms of complications reported using anonymised matching techniques based on for example treating trust, procedure date, gender, age datafields), for example, active surveillance undertaken for the Airway Intervention Registry, using the former HDIS system. This is necessary to understand whether the registry data can be considered representative of the total population of interest, and to measure the success of efforts to improve data quality and completeness.

viii) to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching technique based on for example treating trust, sex, procedure date, gender, age data fields),

ix) to conduct exploratory analysis to determine whether outcomes (e.g. complications, diagnoses, in-hospital death) are associated with patient characteristics or setting (e.g. published analysis of hospital centre volume on iliac artery stenting outcomes),

x) study of outcomes and resource usage before and after an event (e.g. new device/procedure) to determine impact.

The most significant issue with the use of HES data alone in healthcare outcomes research is missing information on out-of-hospital deaths (date and cause of death) - noting that from recent publication that 50% of all deaths occur out of hospital. This risks overestimating the efficacy and safety of medical devices and interventional procedures in the aggregated results reported to NICE by NMPCE, to inform national NICE Guidance.

Hence, in addition to the above ten NMPCE purposes / activities, pseudonymised mortality data are required by NMPCE for the following purposes / activities:

xi) to add to the above longitudinal analysis in iv), to determine the long-term outcome of all-cause mortality, requiring both in-hospital (HES) and out-of-hospital mortality data of healthcare interventions,

xii) to add out-of-hospital mortality data to v) above, in comparative patient outcomes analysis,

xiii) to add out-of-hospital mortality data to vi) above, in analysing hospital resource usage prior to and following healthcare interventions

xiv) to add out-of-hospital mortality data to viii) above, to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching techniques based on for example treating trust, procedure date, gender and age data fields),

xv) to add out-of-hospital mortality data to ix) above, to conduct exploratory analysis to determine whether outcomes, including out-of-hospital deaths, are associated with patient characteristics or setting (e.g. age, gender, comorbidities, treating hospital).

In terms of data minimisation, due to the variety of analyses undertaken by NMPCE (different specialties, time durations of interest, research questions posed) the data and data subjects will vary by project. However, NUTH have considered the data fields requested from each data set (mortality, APC, CC, OP, ECDS) and each have been minimised accordingly. As a previous HESBO and HDIS user, NMPCE has used a specific subset of HES and Mortality data fields to inform evaluations of safety and efficacy for given healthcare interventions. For example: admission and discharge information, all diagnoses and procedure codes (4 character), episode and spell information (to build up an entire patient admission), treating organization, referring organization, patient age and gender.

National data are required due to the breadth of healthcare intervention (health technologies and interventional procedures) development and evaluation conducted (i.e. any condition, any procedure, any population). Access to HES data for the whole of England is also particularly important for NMPCE when considering interventions for rare diseases (affecting less than 5 in 10,000 of the general population), novel innovative procedures/devices/medicines coming into use, and for longitudinal analysis where patients may attend a number of different hospitals (e.g. treatment received in specialised treatment centre followed by follow-up care in local hospital). Furthermore, by using data from across the whole of England (rather than a specific localised region), NMPCE can assure that outcomes are generalisable and reflect current practices used across the country to inform national NICE Guidance. For NMPCE to conduct activities for NICE purposes, some of the “Objectives for processing” (described above) require access to HES/mortality data from all of England.

No identifiable information is requested. NMPCE requires only a unique patient pseudonym that links across each data set: mortality, HES APC, OP, ECDS and CC to determine hospital resource usage and long-term outcomes for each patient.

The amount of mortality data requested has also been minimised by restricting by year to only those patients who have been admitted as inpatients (from 2007/08 onwards), or attended outpatients, emergency care, or critical care departments (from 2013/14 onwards). The frequency of mortality data requested is restricted to quarterly, to ensure safety signals are detected in a timely manner.

HES, ECDS and mortality data will not be used for any external commercial purposes not outlined in this agreement. These data will not be provided in record level form to any third party. These data will not be used for direct marketing. NUTH determines which data are processed for each project/research question and ensures the legal basis for requesting, storing and processing these data are met, therefore NUTH is the data controller and also processes the data for this study. No other organisations process the data for this purpose.

The programme of outcomes research using administrative and registry data currently has funding from the NIHR, Academic Health Science Network - North East and North Cumbria (AHSN NENC) and NICE. NMPCE has a track record of attracting research funding, and intends to seek future funding for their outcomes research programme, from public funding bodies, charities and organisations such as:

- NIHR projects (Invention for Innovation (i4i), Research for Patient Benefit (RfPB), Health Technology Assessment, Research Capability Funding, Newcastle Biomedical Research Centre;

- NIHR infrastructure funding (NIHR Newcastle IVD Co-operative, NIHR Innovation Observatory);

- NICE (the External Assessment Centre, EAC, described below);

- Academic Health Science Network (AHSN) North East and North Cumbria (NENC);

- Engineering and Physical Sciences Research Council (EPSRC);

- Charitable funding (Wellcome Trust, Innovate UK, British Heart Foundation, Newcastle Healthcare Charity);

- Professional societies supporting registry development;

- Commercial companies, which could include medical devices and diagnostics manufacturers.

NMPCE will produce an output in the form of a publication outlining an individual project's findings which, subject to acceptance, will be published in peer-reviewed academic or clinical journals. Additional outputs include grant application to funding bodies, conference presentations, summaries shared with public/patient representative groups and summaries shared on dedicated study websites (where applicable) and internal reports to NICE and other potential funders listed in section 5a.

All outputs will contain only data that are aggregated with small numbers suppressed, in line with the HES analysis guide.

The dissemination and communication approach for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. Examples are described below.

Dissemination of results / outputs.

NMPCE will share aggregated information with NICE, to contribute to the production of national Medical Technologies Guidance or Interventional Procedures Guidance, inform other NICE programmes, and thereby inform healthcare users, health care providers and commissioners of the safety and efficacy, national adoption and key providers of medical device interventions and interventional procedures. An output will be an internal report outlining an individual project's findings submitted to NICE with small number suppression in line with the HES analysis guide.

Communication of results / outputs

One specific NIHR-funded project in progress is the Research for Patient Benefit award PB-PG-0416-20037 - Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP). Recruitment to the study in England will be compared (via routine monthly surveillance) to the number of RRP procedures being reported to the Hospital Episode Statistics (HES). This will enable NMPCE to invite new NHS hospitals/trusts to contribute to the study and also contact registered NHS hospitals/trusts to encourage them to enter data identified in HES to the registry where they have not done so. NMPCE have found that this method of feedback to NHS organisations (via Information Governance teams and clinical coding managers) has increased data entry to previous online databases as well as improving clinical coding accuracy for some interventional procedures.

Exploitation of results / outputs

NMPCE engages with clinical coders in acute trusts across England prior to analysis of HES data, but also provides feedback to clinical coders to demonstrate the external uses of coded hospital data, with the aim to continually improve data quality at source. All publications using HES/mortality data will reference NHS Digital and cite a copyright statement. NMPCE strives to make all peer-reviewed publications available in open access form in order to maximise the availability of information to health care users, health care providers and general members of the public. For example, NUTH published the analysis code (written in R) which analysed HES APC data to determine safety of mesh implants for stress urinary incontinence. Sharing methodology (not data) with the public will provide guidance to other researchers investigating other interventions in how to clean and analyse a large sample of HES data to identify complications; (Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015). The journal choice for any peer-reviewed publication will depend on the medical device intervention/interventional procedure/medical condition of interest (e.g. cardiac, ENT, oncology) and focus of the research question posed (e.g. methodology, patient safety, national policy).

Target dates to produce the outputs for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. In general, NMPCE projects involving medical device intervention/interventional procedure development or evaluation arise with clinical need, with timeframes governed by suitable funding calls and duration of grant funding.

In the specific example of the NIHR-funded Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP) project, a rolling monthly extract from the AIR will be compared with the latest quarterly HES extract from NHS Digital in active surveillance of coverage of the registry (total number of RRP cases treated in England versus total number entered in the registry). The target date for recruitment completion is currently 31st August 2021, with an additional 12 months support gained from NICE (up to 31st August 2022).

In the specific case of NMPCE projects commissioned by NICE, each individual project is issued with a timeline of dates for completion of project milestones. Most NICE projects require peer-reviewed publication of results within 12 months of project completion. NMPCE strives to share analysis code (where possible and journal permitting) for transparency and reproducibility purposes for future researchers.

NICE will consider evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

i) NICE IPG425: Endoscopic balloon dilatation for subglottic or tracheal stenosis (published April 2012 - update undergoing)

ii) NICE IPG434: Radiofrequency cold ablation for respiratory papillomatosis (published November 2012 - update required).

iii) NICE Helium plasma coagulation for treating Respiratory Papillomatosis – monitored (awaiting publication of further evidence)

This agreement does not have any associated EU funding

Previous intended benefits have been yielded and outputs achieved their stated purposes (and thus benefits of processing).

NMPCE and NICE and NHS Digital worked collaboratively on a project looking at data linkage methodology (manuscript available freely online: “Comparison of identifiable and non-identifiable data linkage: health technology assessment of MitraClip using registry, administrative and mortality datasets” BMJ Health Care Inform. 2021; 28(1): e100223).

Examples of peer-reviewed publications (available in the public domain) resulting from the analysis of HES data previously accessed via HDIS by NMPCE for NICE purposes, with details of the impact these have achieved are as follows:

i) Keltie et al. Paediatric tonsillectomy in England: a cohort study of clinical practice and outcomes using Hospital Episode Statistics data (2008-2019). Clin Otolaryngol. 2021; 46(3): 552-561.

IMPACT: Analysed 318,453 paediatric tonsillectomies conducted in NHS hospitals in England between 2008 and 2019, demonstrating increase in complications. This has triggered an IPG review at NICE, and grant applications to NIHR in order to further investigate why the safety profile of tonsillectomy has changed over time.

ii) Page et al. Prostatic urethral life (UroLift): a real-world analysis of outcomes using hospital episodes statistics. BMC Urol. 2021; 21(1): 55.

IMPACT: Demonstrates uptake of UroLift in England after UroLift was added to the Innovation Technology Tariff (ITT) in April 2017 and selected as a Rapid Uptake Product by the Accelerated Access Collaborative in 2018. This real-world analysis shows that UroLift implantation can be delivered safely in a day-case setting with minimal morbidity. However, hospital resource usage for catheterization and emergency hospital attendance in the first 30 days was substantial, and 12% required re-treatment at 2 years.

iii) Powell et al. Balloon dilatation for paediatric airway stenosis: Evidence from the UK Airway Intervention Registry. Clin Otolaryngol. 2020; 45(3): 334-341.

IMPACT: Conducted pseudonymised linkage to determine registry coverage, and also to determine longitudinal outcomes (further respiratory surgery, death). This additional evidence has triggered an update to IPG425 within NICE.

iv) Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015.

IMPACT: Complications from surgical mesh procedures have led to legal cases against manufacturers worldwide and to national inquiries about their safety. This was one of the largest studies so far looking at adverse event rates of these procedures, and has significantly added to the understanding of the likelihood of adverse events from mesh procedures in a UK NHS setting. This means that women considering mesh procedures in future will have more accurate safety information available to them, so they can make a more informed choice. This publication and its analysis code was published in an open-access form to share both methodology and findings widely. NICE has also updated its NG123 guidance and produced linked Patient Decision Aids for women considering mesh surgery.

v) Burn et al. Procedural and short-term safety of bronchial thermoplasty in clinical practice: evidence from a national registry and Hospital Episode Statistics. J Asthma. 2017; 54(8): 872-879.

IMPACT: Bronchial thermoplasty is a novel treatment for severe asthma, however its mode of action and target patient population were poorly defined, with little evidence available on efficacy and safety in terms of quality and quantity. This study presents procedural and short-term safety evidence from routine UK clinical practice.

vi) Donne et al. Prevalence and management of recurrent respiratory papillomatosis (RRP) in the UK: cross-sectional study. Clinical Otolaryngology. 2017; 42(1):86-91.

IMPACT: Recurrent respiratory papillomatosis is a relatively rare condition, but has a significant impact on patients and their families as it requires repeated anaesthetic procedures to achieve symptomatic control. This study utilised a cross-sectional survey of ENT consultants and analysis of HES data to provide, for the first time, an estimate for the prevalence of recurrent respiratory papillomatosis in the UK. This prevalence informed the sample size for an NIHR Research for Patient Benefit grant application - 3 year award granted with extension due to suspended recruitment to non-COVID 19 studies. Additional funding gained from NICE to support an additional 12 months (end 31st August 2022). Identification of key providers of RRP care in the UK, enabled NMPCE to invite specific acute trusts to contribute to further research in RRP patients, and contribute to future NICE guidance (in update of IPG 434).

vii) Goode et al. Effect of procedure volume on outcomes after iliac artery angioplasty and stenting. British Journal of Surgery. 2013; 100:1189-1196.

IMPACT: This study was able to use HES data to determine if there was an association between outcomes of endovascular iliac artery intervention and centre volume, describing methodology which could be applied to any interventional in-hospital procedure which would benefit future analysis of HES data.

viii) Patrick et al. Monitoring the use and outcomes of new devices and procedures: how does coding affect what Hospital Episode Statistics contribute? Lessons from 12 emerging procedures 2006-10. Journal of Public Health. 2012; 35(1): 132-138.

IMPACT: This study provided evidence that routinely collected HES data have the potential to support quality improvements and evidence-based commissioning of devices and procedures in the national health service.

ix). Keltie et al. Identifying complications of interventional procedures from UK routine healthcare databases: a systematic search for methods using clinical codes. BMC Medical Research Methodology 2014; 14:126.

IMPACT: This study summarised methods of analysing complications of interventional procedures from routine healthcare databases in the UK published in the literature. This study was published in open-access form to share learnings with other researchers and clinical coding teams across the UK to encourage high quality research from HES.

NMPCE has also supported Medical Technology evaluation of two devices (SpaceOAR, MAGECRods) through surveillance of HES to determine uptake in England. In addition, NICE has incorporated evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance: e.g. NICE IPG635: Bronchial thermoplasty for severe asthma (originally published January 2012 - updated December 2018.

NMPCE has used data from HES to support the drafting of grant applications applicable to a number of NIHR funding streams (including: Digital technologies, HTA, Ambitious data-enabled trials, health services and public health research studies, RfPB). Research themes for these grant applications have included (but are not limited to):

- safety and efficacy of tonsillectomy,

- longitudinal study to investigate the uptake and safety of UroLift procedures,

- longitudinal study to investigate the association between kidney transplantation and urological cancers,

- short-term evaluation of ureteroscopy and infection outcomes,

- longitudinal study of stroke patients and pneumonia outcomes.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is one of the country's top performing trusts in terms of the number of research studies it has supported, with 501 studies delivered by the trust in 2018/2019. Northern Medical Physics and Clinical Engineering (NMPCE) is a Clinical Directorate within NUTH, with research interests and a track record going back 40 years in the development and evaluation of health technology interventions including medical devices, diagnostics, and interventional procedures.

NMPCE will use Hospital Episode Statistics (HES) and Civil Registration mortality data to support its programme of research which includes studying:

i) uptake and outcomes of novel medical technologies and interventional procedures to assess peri-procedural, short-term and long-term safety and efficacy

ii) associations between comorbidities, diagnoses, and hospital admissions

iii) patient pathways through the healthcare system.

This is a not-for-profit research programme which uses large pseudonymised data sets to achieve the aims described. The outputs of the programme are intended to inform national guidance, decision-making, research recommendations, study designs, grant applications to national bodies (a comprehensive list of outputs is described in 5c).

The justification for Article 6(1)(e) processing, as necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller is summarised briefly as follows:

NUTH are a public authority conducting transparent observational research and publication of healthcare outcomes from interventional procedures in the NHS. Civil Registration mortality data informs overall death rates (any cause) for the patient cohort selected for analysis. Out of hospital deaths and cause of death are not reported through HES, therefore the combined data are more robust for analysis than either source alone. NUTH cannot reasonably achieve the above in a different way. NHS providers have administrative systems which feed into HES; therefore this represents the only national routine dataset of NHS activity in secondary and tertiary care. HES may record in-hospital deaths, but Civil Registration mortality is only legal record of cause of death, independent of place of death. The pseudonymised data analysis methods developed by NMPCE (non-identifiable patient data) are the least intrusive means of achieving the purpose of NHS healthcare outcomes analysis and publication, in the public interest.

The public interest justification for Article 9(2)(j) – for public health purposes is summarised briefly as follows:

Article 9(2)(j) - not all the projects are intended to generate generalisable new knowledge for academic publication. NMPCE outcomes research programme also includes processing for reasons of public health, including adoption of medical technologies and outcomes of interventional procedures, to ensure high standards of quality and safety, as described in this application.

Moral and ethical issues and risk of potential harm to the public from the data processing methods, have been considered through active engagement with patient and public involvement (PPI) groups in the Trust and NUTH, Patients and their representatives have told NUTH that they positively endorse and demand such use of their data to benefit medical research and the NHS, as described in this agreement.

The data requested will achieve the aim identified by permitting a timely response from NUTH, to questions of safety and efficacy of medical devices, diagnostics and interventional procedures in current NHS practice.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is the sole Data Controller and also processes the data for this study. No other organisations process the data for this purpose

Since 2011, NUTH has hosted one of five current National Institute for Health and Care Excellence (NICE) External Assessment Centres (EAC). The NICE EAC is based in the NMPCE Directorate at the Freeman Hospital in Newcastle. The work commissioned by NICE evaluates medical devices and interventional procedures and answers specific research questions raised by the NICE Medical Technologies Advisory Committee (MTAC), Diagnostics Advisory Committee (DAC), Interventional Procedures Advisory Committee (IPAC) and other NICE programmes. NICE had provided a commissioner letter for the External Assessment Centre within NMPCE to support this agreement, which extends to 2021.

NUTH was formerly an HSCIC / NHS Digital approved HES Business Objects (HESBO) and HES Data Interrogation System (HDIS) user. The NMPCE analysts within NUTH are trained and were approved to process episode level pseudonymised HES data for the research purposes described in this DARS application, for NHS Health and Social Care system, patient and public benefit. This agreement is to renew and continue the quarterly managed extract service of pseudonymised HES (APC, CC, OP, AE/ECDS) and Civil Registration data from NHS Digital, which will permit NMPCE to continue delivering the described research objectives for an additional twelve months.

The nature of the work described in this agreement is a wide programme of collaborative research, including long term follow up. The detailed scope of data analysis is determined on a project by project basis, but always within the overarching terms of the Data Sharing Agreement with NHS Digital.

The data subjects are also determined on a project by project basis, following the convention of defining Population, Intervention, Comparator, Outcomes (PICO) and setting (APC / A&E etc) for every intended analysis project. A protocol for pseudonymised matching is drafted before work commences, where the HES and Civil Registration mortality codes to match the required PICO and setting are determined in an iterative process. Data cleaning rules are applied to reach the final agreed and matched set of data subjects, which may be for a single-armed, or comparative analysis, using control and cohort groups.

Detailed information about the purpose of the projects covered by this agreement is provided below, including the data required, why these data are required and justification for the datasets requested, the level of data (pseudonymised), the number of years requested and the geographical spread of the data requested.

NUTH confirm that there are no alternative, less intrusive ways of achieving the purpose, and details of the efforts taken to minimise the data required are also described below:

Pseudonymised data from the HES & Civil Registration mortality datasets are required by NMPCE for the following purposes / activities:

i) to determine the number of patients with a particular disease attending an NHS hospital setting, the number being treated with healthcare interventions across England and identification of key providers of care, (e.g. to inform NICE of scope/coverage/uptake of guidance, to inform sample size determination and to target study invitations to relevant organisations. For example, NMPCE's cross-sectional Recurrent Respiratory Papillomatosis (RRP) survey determined the number of patients diagnosed with the condition, the proportion receiving hospital treatment and the breakdown of interventions used in this population (Donne et al. Clinical Otolaryngology 2016). The survey found at least 16 different interventions in use in the NHS for RRP and limited evidence on their comparative safety and efficacy. This work led to a successful application for NIHR Research for patient benefit grant funding and the establishment of a national data collection on the existing Airway Intervention Register (AIR) platform for analysis to determine the safest and most effective intervention(s) for the condition, which will benefit current and future RRP patients.),

ii) to determine patient demographics receiving healthcare interventions (e.g. common comorbidities, comorbidity score index, age distribution, gender),

iii) to determine in-hospital outcomes (e.g. efficacy and safety, length of stay) of health care interventions,

iv) to conduct active surveillance/longitudinal analysis to determine long-term outcomes (e.g. efficacy, safety and in-hospital mortality) of healthcare interventions (e.g. 8 year analysis of almost 100,000 women with surgical insertion of mesh implants for the treatment of stress urinary incontinence, Keltie et al. Scientific Reports. 2017). The retention of HES data up to a maximum of 20 years will permit further, longer term follow-up analyses of these women, to establish the as-yet unknown longevity of mesh implants and any change in efficacy and complication rates as they age,

v) to compare patient outcomes (including in-hospital and longitudinal outcomes) between healthcare interventions (e.g. through propensity matching of cohorts based on comorbidities) to determine relative efficacy and safety. The retention of HES data up to a maximum of 20 years will permit future analyses of paediatric RRP patients in the AIR registry, to determine which treatments have enduring efficacy outcomes into adulthood,

vi) to determine hospital resource usage, including prior to and following healthcare interventions (e.g. patient pathway, number and indication of outpatient attendances, subsequent hospital admission, further interventions, accident and emergency attendances, in-hospital deaths, HRG codes/costs), for example in the published study of procedural and short-term efficacy of bronchial thermoplasty (Burn et al., 2016),

vii) to determine national coverage of and encourage data submission to national clinical or procedural registries (e.g. identification of trusts not contributing data to registries, data completeness of registry in terms of total number of procedures not recorded in registry, data completeness of registry in terms of complications reported using anonymised matching techniques based on for example treating trust, procedure date, gender, age datafields), for example, active surveillance undertaken for the Airway Intervention Registry, using the former HDIS system. This is necessary to understand whether the registry data can be considered representative of the total population of interest, and to measure the success of efforts to improve data quality and completeness.

viii) to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching technique based on for example treating trust, sex, procedure date, gender, age data fields),

ix) to conduct exploratory analysis to determine whether outcomes (e.g. complications, diagnoses, in-hospital death) are associated with patient characteristics or setting (e.g. published analysis of hospital centre volume on iliac artery stenting outcomes),

x) study of outcomes and resource usage before and after an event (e.g. new device/procedure) to determine impact.

The most significant issue with the use of HES data alone in healthcare outcomes research is missing information on out-of-hospital deaths (date and cause of death) - noting that from recent publication that 50% of all deaths occur out of hospital. This risks overestimating the efficacy and safety of medical devices and interventional procedures in the aggregated results reported to NICE by NMPCE, to inform national NICE Guidance.

Hence, in addition to the above ten NMPCE purposes / activities, pseudonymised mortality data are required by NMPCE for the following purposes / activities:

xi) to add to the above longitudinal analysis in iv), to determine the long-term outcome of all-cause mortality, requiring both in-hospital (HES) and out-of-hospital mortality data of healthcare interventions,

xii) to add out-of-hospital mortality data to v) above, in comparative patient outcomes analysis,

xiii) to add out-of-hospital mortality data to vi) above, in analysing hospital resource usage prior to and following healthcare interventions

xiv) to add out-of-hospital mortality data to viii) above, to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching techniques based on for example treating trust, procedure date, gender and age data fields),

xv) to add out-of-hospital mortality data to ix) above, to conduct exploratory analysis to determine whether outcomes, including out-of-hospital deaths, are associated with patient characteristics or setting (e.g. age, gender, comorbidities, treating hospital).

In terms of data minimisation, due to the variety of analyses undertaken by NMPCE (different specialties, time durations of interest, research questions posed) the data and data subjects will vary by project. However, NUTH have considered the data fields requested from each data set (mortality, APC, CC, OP, AE/ECDS) and each have been minimised accordingly. As a previous HESBO and HDIS user, NMPCE has used a specific subset of HES and Mortality data fields to inform evaluations of safety and efficacy for given healthcare interventions. For example: admission and discharge information, all diagnoses and procedure codes (4 character), episode and spell information (to build up an entire patient admission), treating organization, referring organization, patient age and gender.

National data are required due to the breadth of healthcare intervention (medical devices, diagnostics, interventional procedures) development and evaluation conducted (i.e. any condition, any procedure, any population). Access to HES data for the whole of England is also particularly important for NMPCE when considering interventions for rare diseases (affecting less than 5 in 10,000 of the general population), novel innovative procedures/devices coming into use, and for longitudinal analysis where patients may attend a number of different hospitals (e.g. treatment received in specialised treatment centre followed by follow-up care in local hospital). Furthermore, by using data from across the whole of England (rather than a specific localised region), NMPCE can assure that outcomes are generalisable and reflect current practices used across the country to inform national NICE Guidance. For NMPCE to conduct activities for NICE purposes, some of the “Objectives for processing” (described above) require access to HES/mortality data from all England.

No identifiable information is requested. NMPCE requires only a unique patient pseudonym that links across each data set: mortality, HES APC, OP, emergency care, A&E, CC to determine hospital resource usage and long-term outcomes for each patient.

The amount of mortality data requested has also been minimised by restricting by year to only those patients who have been admitted as inpatients (from 2007/08 onwards), or attended outpatients, emergency care, or critical care departments (from 2013/14 onwards). The frequency of mortality data requested is restricted to quarterly, to ensure safety signals are detected in a timely manner.

HES and mortality data will not be used for any external commercial purposes not outlined in this agreement. These data will not be provided in record level form to any third party. These data will not be used for direct marketing. NUTH determines which data are processed for each project/research question and ensures the legal basis for requesting, storing and processing these data are met, therefore NUTH is the data controller and also processes the data for this study. No other organisations process the data for this purpose.

The programme of outcomes research using administrative and registry data currently has funding from the NIHR, Academic Health Science Network - North East and North Cumbria (AHSN NENC) and NICE. NMPCE has a track record of attracting research funding, and intends to seek future funding for their outcomes research programme, from public funding bodies, charities and organisations such as:

- NIHR projects (Invention for Innovation (i4i), Research for Patient Benefit (RfPB), Health Technology Assessment, Research Capability Funding, Newcastle Biomedical Research Centre;

- NIHR infrastructure funding (NIHR Newcastle IVD Co-operative, NIHR Innovation Observatory);

- NICE (the External Assessment Centre, EAC, described below);

- Academic Health Science Network (AHSN) North East and North Cumbria (NENC);

- Engineering and Physical Sciences Research Council (EPSRC);

- Charitable funding (Wellcome Trust, Innovate UK, British Heart Foundation, Newcastle Healthcare Charity);

- Professional societies supporting registry development;

- Commercial companies, which could include medical devices and diagnostics manufacturers.

NMPCE will produce an output in the form of a publication outlining an individual project's findings which, subject to acceptance, will be published in peer-reviewed academic or clinical journals. Additional outputs include grant application to funding bodies, conference presentations, summaries shared with public/patient representative groups and summaries shared on dedicated study websites (where applicable) and internal reports to NICE and other potential funders listed in section 5a.

All outputs will contain only data that are aggregated with small numbers suppressed, in line with the HES analysis guide.

The dissemination and communication approach for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. Examples are described below.

Dissemination of results / outputs.

NMPCE will share aggregated information with NICE, to contribute to the production of national Medical Technologies Guidance or Interventional Procedures Guidance, inform other NICE programmes, and thereby inform healthcare users, health care providers and commissioners of the safety and efficacy, national adoption and key providers of medical device interventions and interventional procedures. An output will be an internal report outlining an individual project's findings submitted to NICE with small number suppression in line with the HES analysis guide.

Communication of results / outputs

One specific NIHR-funded project in progress is the Research for Patient Benefit award PB-PG-0416-20037 - Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP). Recruitment to the study in England will be compared (via routine monthly surveillance) to the number of RRP procedures being reported to the Hospital Episode Statistics (HES). This will enable NMPCE to invite new NHS hospitals/trusts to contribute to the study and also contact registered NHS hospitals/trusts to encourage them to enter data identified in HES to the registry where they have not done so. NMPCE have found that this method of feedback to NHS organisations (via Information Governance teams and clinical coding managers) has increased data entry to previous online databases as well as improving clinical coding accuracy for some interventional procedures.

Exploitation of results / outputs

NMPCE engages with clinical coders in acute trusts across England prior to analysis of HES data, but also provides feedback to clinical coders to demonstrate the external uses of coded hospital data, with the aim to continually improve data quality at source. All publications using HES/mortality data will reference NHS Digital and cite a copyright statement. NMPCE strives to make all peer-reviewed publications available in open access form in order to maximise the availability of information to health care users, health care providers and general members of the public. For example, NUTH published the analysis code (written in R) which analysed HES APC data to determine safety of mesh implants for stress urinary incontinence. Sharing methodology (not data) with the public will provide guidance to other researchers investigating other interventions in how to clean and analyse a large sample of HES data to identify complications; (Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015). The journal choice for any peer-reviewed publication will depend on the medical device intervention/interventional procedure/medical condition of interest (e.g. cardiac, ENT, oncology) and focus of the research question posed (e.g. methodology, patient safety, national policy).

Target dates to produce the outputs for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. In general, NMPCE projects involving medical device intervention/interventional procedure development or evaluation arise with clinical need, with timeframes governed by suitable funding calls and duration of grant funding.

In the specific example of the NIHR-funded Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP) project, a rolling monthly extract from the AIR will be compared with the latest quarterly HES extract from NHS Digital in active surveillance of coverage of the registry (total number of RRP cases treated in England versus total number entered in the registry). The target date for recruitment completion is currently August 2020.

In the specific case of NMPCE projects commissioned by NICE, each individual project is issued with a timeline of dates for completion of project milestones. Most NICE projects require peer-reviewed publication of results within 12 months of project completion. NMPCE strives to share analysis code (where possible and journal permitting) for transparency and reproducibility purposes for future researchers.

NICE will consider evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

i) NICE IPG425: Endoscopic balloon dilatation for subglottic or tracheal stenosis (published April 2012 - update required)

ii) NICE IPG434: Radiofrequency cold ablation for respiratory papillomatosis (published November 2012 - update required).

This agreement does not have any associated EU funding.

Previous intended benefits have been yielded and outputs achieved their stated purposes (and thus benefits of processing).

NMPCE and NICE and NHS Digital worked collaboratively on a project looking at data linkage methodology (manuscript prepared for academic peer-reviewed journal – awaiting submission “Efficient data linkage to inform health technology assessment using real world data: determining patient outcomes by identifiable and non-identifiable methods using the UK MitraClip registry, Hospital Episode Statistics and mortality datasets”).

Examples of peer-reviewed publications (available in the public domain) resulting from the analysis of HES data previously accessed via HDIS by NMPCE for NICE purposes, with details of the impact these have achieved are as follows:

i) Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015.

IMPACT: Complications from surgical mesh procedures have led to legal cases against manufacturers worldwide and to national inquiries about their safety. This was one of the largest studies so far looking at adverse event rates of these procedures, and has significantly added to the understanding of the likelihood of adverse events from mesh procedures in a UK NHS setting. This means that women considering mesh procedures in future will have more accurate safety information available to them, so they can make a more informed choice. This publication and its analysis code was published in an open-access form to share both methodology and findings widely. NICE have also updated their NG123 guidance and produced linked Patient Decision Aids for women considering mesh surgery.

ii) Burn et al. Procedural and short-term safety of bronchial thermoplasty in clinical practice: evidence from a national registry and Hospital Episode Statistics. J Asthma. 2017; 54(8): 872-879.

IMPACT: Bronchial thermoplasty is a novel treatment for severe asthma, however its mode of action and target patient population were poorly defined, with little evidence available on efficacy and safety in terms of quality and quantity. This study presents procedural and short-term safety evidence from routine UK clinical practice.

iii) Donne et al. Prevalence and management of recurrent respiratory papillomatosis (RRP) in the UK: cross-sectional study. Clinical Otolaryngology. 2017; 42(1):86-91.

IMPACT: Recurrent respiratory papillomatosis is a relatively rare condition, but has a significant impact on patients and their families as it requires repeated anaesthetic procedures to achieve symptomatic control. This study utilised a cross-sectional survey of ENT consultants and analysis of HES data to provide, for the first time, an estimate for the prevalence of recurrent respiratory papillomatosis in the UK. This prevalence informed the sample size for an NIHR Research for Patient Benefit grant application - 3 year award granted (end date 31/08/2020 - with extension expected due to suspended recruitment to non-COVID19 studies). Identification of key providers of RRP care in the UK, enabled NMPCE to invite specific acute trusts to contribute to further research in RRP patients, and contribute to future NICE guidance (in update of IPG 434).

iv) Goode et al. Effect of procedure volume on outcomes after iliac artery angioplasty and stenting. British Journal of Surgery. 2013; 100:1189-1196.

IMPACT: This study was able to use HES data to determine if there was an association between outcomes of endovascular iliac artery intervention and centre volume, describing methodology which could be applied to any interventional in-hospital procedure which would benefit future analysis of HES data.

v) Patrick et al. Monitoring the use and outcomes of new devices and procedures: how does coding affect what Hospital Episode Statistics contribute? Lessons from 12 emerging procedures 2006-10. Journal of Public Health. 2012; 35(1): 132-138.

IMPACT: This study provided evidence that routinely collected HES data have the potential to support quality improvements and evidence-based commissioning of devices and procedures in the national health service.

vi). Identifying complications of interventional procedures from UK routine healthcare databases: a systematic search for methods using clinical codes. BMC Medical Research Methodology 2014; 14:126.

IMPACT: This study summarised methods of analysing complications of interventional procedures from routine healthcare databases in the UK published in the literature. This study was published in open-access form to share learnings with other researchers and clinical coding teams across the UK to encourage high quality research from HES.

NMPCE has also supported Medical Technology evaluation of two devices (SpaceOAR, MAGECRods) through surveillance of HES to determine uptake in England. In addition, NICE has incorporated evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance: e.g. NICE IPG635: Bronchial thermoplasty for severe asthma (originally published January 2012 - updated December 2018.

NMPCE has used data from HES to support the drafting of grant applications applicable to a number of NIHR funding streams (including: Digital technologies, HTA, Ambitious data-enabled trials, health services and public health research studies, RfPB). Research themes for these grant applications have included (but are not limited to):

- safety and efficacy of tonsillectomy,

- longitudinal study to investigate the uptake and safety of UroLift procedures,

- longitudinal study to investigate the association between kidney transplantation and urological cancers,

- short-term evaluation of ureteroscopy and infection outcomes,

- longitudinal study of stroke patients and pneumonia outcomes.

The Newcastle upon Tyne Hospitals NHS Foundation Trust (NUTH) is the country's top performing trust for the number of research studies it has supported, with 550 studies delivered by the trust in 2017/18. NUTH has retained its position at the top of the NIHR Activity Research League Table for 7 years running. Northern Medical Physics and Clinical Engineering (NMPCE) is a Clinical Directorate within NUTH, with research interests and a track record going back 40 years in the development and evaluation of health technology interventions including medical devices, diagnostics and interventional procedures.

NMPCE will use Hospital Episode Statistics (HES) and Civil Registration mortality data to support its programme of research into outcomes of novel medical technologies and interventional procedures. This is a not-for-profit research programme which uses large pseudonymised data sets to assess peri-procedural, short-term and long-term safety and efficacy. The outputs of the programme are research publications, in open-access format wherever possible, which are intended to inform national guidance and decision-making.

The justification for Article 6(1)(e) processing, as necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller is summarised briefly as follows:

NUTH are a public authority conducting transparent observational research and publication of healthcare outcomes from interventional procedures in the NHS. Civil Registration mortality data informs overall death rates (any cause) for the patient cohort selected for analysis. Out of hospital deaths and cause of death are not reported through HES, therefore the combined data are more robust for analysis than either source alone. NUTH cannot reasonably achieve the above in a different way. NHS providers have administrative systems which feed into HES, therefore this represents the only national routine dataset of NHS activity in secondary and tertiary care. HES may record in-hospital deaths, but Civil Registration mortality is only legal record of cause of death, independent of place of death. The pseudonymised data analysis methods developed by NMPCE (non-identifiable patient data) are the least intrusive means of achieving the purpose of NHS healthcare outcomes analysis and publication, in the public interest.

The public interest justification for Article 9(2)(j) – for public health purposes is summarised briefly as follows:

Article 9(2)(j) - not all of the projects are intended to generate generalisable new knowledge for academic publication. NMPCE outcomes research programme also includes processing for reasons of public health, including adoption of medical technologies and outcomes of interventional procedures, to ensure high standards of quality and safety, as described in this application.

Through active engagement with patient and public involvement (PPI) groups in the Trust and NUTH, have identified no moral or ethical issues and no risk of potential harm to the public from the data processing methods. Indeed, patients and their representatives have told NUTH that they positively endorse and demand such use of their data to benefit medical research and the NHS, as described in this agreement.

The data requested will achieve the aim identified by permitting a timely response from NUTH, as the Data Controller and Processor, to questions of safety and efficacy of medical devices, diagnostics and interventional procedures in current NHS practice.

Since 2011, NUTH has hosted one of five current National Institute for Health and Care Excellence (NICE) External Assessment Centres (EAC). The NICE EAC is based in the NMPCE Directorate at the Freeman Hospital in Newcastle. The work commissioned by NICE evaluates medical devices and interventional procedures and answers specific research questions raised by the NICE Medical Technologies Advisory Committee (MTAC), Diagnostics Advisory Committee (DAC), Interventional Procedures Advisory Committee (IPAC) and other NICE programmes. NICE has provided a commissioner letter for the External Assessment Centre within NMPCE to support this agreement.

NUTH was formerly an HSCIC / NHS Digital approved HES Business Objects (HESBO) and HES Data Interrogation System (HDIS) user. The NMPCE analysts within NUTH are trained and were approved to process episode level pseudonymised HES data for the research purposes described in this DARS application, for NHS Health and Social Care system, patient and public benefit. However, in consultation with NHS Digital in 2017/18, NUTH had concluded that the technical differences between the new HDIS2 and the predecessor HDIS systems would prevent NMPCE analysts from conducting their established and published pseudonymised HES data analysis methods. NMPCE therefore submitted a DARS application for a managed quarterly extract (with historical data) which will expire on 14/07/2019. This agreement is therefore to renew and continue the quarterly managed extract service of pseudonymised HES (APC, CC, OP, AE/ECDS) and Civil Registration data from NHS Digital, which will permit NMPCE to continue delivering the described research objectives for an additional twelve months. Additionally, in the previous iteration of the agreement NUTH received month and year of death (from ONS mortality dataset), and within this agreement there is a request for full date (day, month and year of death). This allows the Trust to be consistent in the analysis (as it already obtains full date of death for in-hospital deaths captured in HES APC), and also enables more accurate estimates of death rates (particularly important when conducting analysis and determining 30-day mortality rates compared between interventions).

The nature of the work described in this agreement is, in itself, a wide programme of collaborative research, including long term follow up. The detailed scope of data analysis is determined on a project by project basis, but always within the overarching terms of the Data Sharing Agreement with NHS Digital.

The data subjects are also determined on a project by project basis, following the convention of defining Population, Intervention, Comparator, Outcomes (PICO) and setting (APC / A&E etc) for every intended analysis project. A protocol for pseudonymised matching is drafted before work commences, where the HES and Civil Registration mortality codes to match the required PICO and setting are determined in an iterative process. Data cleaning rules are applied to reach the final agreed and matched set of data subjects, which may be for a single-armed, or comparative analysis, using control and cohort groups.

Detailed information about the purpose of the projects covered by this agreement is provided below, including the data required, why these data are required and justification for the datasets requested, the level of data (pseudonymised), the number of years requested and the geographical spread of the data requested. NUTH confirm that there are no alternative, less intrusive ways of achieving the purpose, and details of the efforts taken to minimise the data required are also described below:

Pseudonymised data from the HES & Civil Registration mortality datasets are required by NMPCE for the following purposes / activities:

i) to determine the number of patients with a particular disease attending an NHS hospital setting, the number being treated with healthcare interventions across England and identification of key providers of care, (e.g. to inform NICE of scope/coverage/uptake of guidance, to inform sample size determination and to target study invitations to relevant organisations. For example NMPCE's cross-sectional Recurrent Respiratory Papillomatosis (RRP)survey determined the number of patients diagnosed with the condition, the proportion receiving hospital treatment and the breakdown of interventions used in this population (Donne et al. Clinical Otolaryngology 2016). The survey found at least 16 different interventions in use in the NHS for RRP and limited evidence on their comparative safety and efficacy. This work led to a successful application for NIHR Research for patient benefit grant funding and the establishment of a national data collection on the existing Airway Intervention Register (AIR) platform for analysis to determine the safest and most effective intervention(s) for the condition, which will benefit current and future RRP patients.),

ii) to determine patient demographics receiving healthcare interventions (e.g. common comorbidities, comorbidity score index, age distribution, gender),

iii) to determine in-hospital outcomes (e.g. efficacy and safety, length of stay) of health care interventions,

iv) to conduct longitudinal analysis to determine long-term outcomes (e.g. efficacy, safety and in-hospital mortality) of healthcare interventions (e.g. 8 year analysis of almost 100,000 women with surgical insertion of mesh implants for the treatment of stress urinary incontinence, Keltie et al. Scientific Reports. 2017). The retention of HES data up to a maximum of 20 years will permit further, longer term follow-up analyses of these women, to establish the as-yet unknown longevity of mesh implants and any change in efficacy and complication rates as they age,

v) to compare patient outcomes (including in-hospital and longitudinal outcomes) between healthcare interventions (e.g. through propensity matching of cohorts based on comorbidities) to determine relative efficacy and safety. The retention of HES data up to a maximum of 20 years will permit future analyses of paediatric RRP patients in the AIR registry, to determine which treatments have enduring efficacy outcomes into adulthood,

vi) to determine hospital resource usage prior to and following healthcare interventions (e.g. number and indication of outpatient attendances, subsequent hospital admission, further interventions, accident and emergency attendances, in-hospital deaths, HRG codes/costs), for example in the published study of procedural and short-term efficacy of bronchial thermoplasty (Burn et al., 2016),

vii) to determine national coverage of and encourage data submission to national clinical or procedural registries (e.g. identification of trusts not contributing data to registries, data completeness of registry in terms of total number of procedures not recorded in registry, data completeness of registry in terms of complications reported using anonymised matching techniques based on for example treating trust, procedure date, gender, age datafields), for example, active surveillance undertaken for the Airway Intervention Registry, using the former HDIS system. This is necessary to understand whether the registry data can be considered representative of the total population of interest, and to measure the success of efforts to improve data quality and completeness.

viii) to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching technique based on for example treating trust, sex, procedure date, gender, age datafields),

ix) to conduct exploratory analysis to determine whether outcomes (e.g. complications, in-hospital death) are associated with patient characteristics or setting (e.g. published analysis of hospital centre volume on iliac artery stenting outcomes),

x) study of outcomes and resource usage before and after an event (e.g. new device/procedure) to determine impact.

The most significant issue with the use of HES data alone in healthcare outcomes research is missing information on out-of-hospital deaths (date and cause of death) - noting that from recent publication that 50% of all deaths occur out of hospital. This risks overestimating the efficacy and safety of medical devices and interventional procedures in the aggregated results reported to NICE by NMPCE, to inform national NICE Guidance.

Hence, in addition to the above ten NMPCE purposes / activities, pseudonymised mortality data are required by NMPCE for the following purposes / activities:

xi) to add to the above longitudinal analysis in iv), to determine the long-term outcome of all-cause mortality, requiring both in-hospital (HES) and out-of-hospital mortality data of healthcare interventions,

xii) to add out-of-hospital mortality data to v) above, in comparative patient outcomes analysis,

xiii) to add out-of-hospital mortality data to vi) above, in analysing hospital resource usage prior to and following healthcare interventions

xiv) to add out-of-hospital mortality data to viii) above, to complement national registry data collection to achieve richer information than can be achieved from a single data set (using anonymised matching technique based on for example treating trust, procedure date, gender, age datafields),

xv) to add out-of-hospital mortality data to ix) above, to conduct exploratory analysis to determine whether outcomes, including out-of-hospital deaths, are associated with patient characteristics or setting (e.g. age, gender, treating hospital).

In terms of data minimisation, due to the variety of analyses undertaken by NMPCE (different specialties, time durations of interest, research questions posed) the data and data subjects will vary by project. However NUTH have considered the data fields requested from each data set (mortality, APC, CC, OP, AE/ECDS) and each have been minimised accordingly. As a previous HESBO and HDIS user, NMPCE has used a specific subset of HES and Mortailty data fields to inform evaluations of safety and efficacy for given healthcare interventions. For example: admission and discharge information, all diagnoses and procedure codes (4 character), episode and spell information (to build up an entire patient admission), treating organization, referring organization, patient age and gender.

National data are required due to the breadth of healthcare intervention (medical devices, diagnostics, interventional procedures) development and evaluation conducted (i.e. any condition, any procedure, any population). Access to HES data for the whole of England is also particularly important for NMPCE when considering interventions for rare diseases (affecting less than 5 in 10,000 of the general population), and for longitudinal analysis where patients may attend a number of different hospitals (e.g. treatment received in specialised treatment centre followed by follow-up care in local hospital). Furthermore, by using data from across the whole of England (rather than a specific localised region), NMPCE can assure that outcomes are generalisable and reflect current practices used across the country in order to inform national NICE Guidance. In order for NMPCE to conduct activities for NICE purposes, some of the “Objectives for processing” (described above) require access to HES/mortality data from all England.

No identifiable information is requested. NMPCE requires only a unique patient pseudonym that links across each data set: mortality, HES APC, OP, emergency care, A&E, CC to determine hospital resource usage and long-term outcomes for each patient.

The amount of mortality data requested has also been minimised by restricting by year to only those patients who have been admitted as inpatients (from 2007/08 onwards), or attended outpatients, emergency care, or critical care departments (from 2013/14 onwards). The frequency of mortality data requested is restricted to quarterly, to ensure safety signals are detected in a timely manner.

HES and mortality data will not be used for any external commercial purposes not outlined in this agreement. These data will not be provided in record level form to any third party. These data will not be used for direct marketing. NUTH determines which data are processed for each project/research question and ensures the legal basis for requesting, storing and processing these data are met, therefore NUTH is the sole Data Controller and sole Data Processor.

The programme of outcomes research using administrative and registry data currently has funding from the NIHR, Academic Health Science Network - North East and North Cumbria (AHSN NENC) and NICE. NMPCE has a track record of attracting research funding, and intends to seek future funding for their outcomes research programme, from public funding bodies, charities and organisations such as:

- NIHR projects (Invention for Innovation (i4i), Research for Patient Benefit (RfPB), Health Technology Assessment, Research Capability Funding, Newcastle Biomedical Research Centre;

- NIHR infrastructure funding (NIHR Newcastle IVD Co-operative, NIHR Innovation Observatory);

- NICE (the External Assessment Centre, EAC, described below);

- Academic Health Science Network (AHSN) North East and North Cumbria (NENC);

- Engineering and Physical Sciences Research Council (EPSRC);

- Charitable funding (Wellcome Trust, Innovate UK, British Heart Foundation, Newcastle Healthcare Charity);

- Professional societies supporting registry development;

- Commercial companies, which could include medical devices and diagnostics manufacturers.

In general, NMPCE will produce an output in the form of a publication outlining an individual project's findings which, subject to acceptance, will be published in peer-reviewed academic or clinical journals. Additional outputs include grant application to funding bodies, conference presentations, summaries shared with public/patient representative groups and summaries shared on dedicated study websites (where applicable) and internal reports to NICE.

All outputs will contain only data that are aggregated with small numbers suppressed, in line with the HES analysis guide.

The dissemination and communication approach for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. Examples are described below.

Dissemination of results / outputs.

NMPCE will share aggregated information with NICE, to contribute to the production of national Medical Technologies Guidance or Interventional Procedures Guidance, inform other NICE programmes, and thereby inform healthcare users, health care providers and commissioners of the safety and efficacy, national adoption and key providers of medical device interventions and interventional procedures. An output will be an internal report outlining an individual project's findings submitted to NICE with small number suppression in line with the HES analysis guide.

Communication of results / outputs

One specific NIHR-funded project in progress is the Research for Patient Benefit award PB-PG-0416-20037 - Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP). Recruitment to the study in England will be compared (via routine monthly surveillance) to the number of RRP procedures being reported to the Hospital Episode Statistics (HES). This will enable NMPCE to invite new NHS hospitals/trusts to contribute to the study and also contact registered NHS hospitals/trusts to encourage them to enter data identified in HES to the registry where they have not done so. NMPCE have found that this method of feedback to NHS organisations (via Information Governance teams and clinical coding managers) has increased data entry to previous online databases as well as improving clinical coding accuracy for some interventional procedures.

Exploitation of results / outputs

NMPCE engages with clinical coders in acute trusts across England prior to analysis of HES data, but also provides feedback to clinical coders to demonstrate the external uses of coded hospital data, with the aim to continually improve data quality at source. All publications using HES/mortality data will reference NHS Digital and cite a copyright statement. NMPCE strives to make all peer-reviewed publications available in open access form in order to maximise the availability of information to health care users, health care providers and general members of the public. For example, NUTH published the analysis code (written in R) which analysed HES APC data to determine safety of mesh implants for stress urinary incontinence. Sharing methodology (not data) with the public will provide guidance to other researchers investigating other interventions in how to clean and analyse a large sample of HES data to identify complications; (Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015). The journal choice for any peer-reviewed publication will depend on the medical device intervention/interventional procedure/medical condition of interest (e.g. cardiac, ENT, oncology) and focus of the research question posed (e.g. methodology, patient safety, national policy).

Target dates for the production of the outputs for different types of analyses varies according to each individual project plan, aligned with the intended results of the research. In general, NMPCE projects involving medical device intervention/interventional procedure development or evaluation arise with clinical need, with timeframes governed by suitable funding calls and duration of grant funding.

In the specific example of the NIHR-funded Airway Intervention Registry (AIR) extension - Recurrent Respiratory Papillomatosis (RRP) project, a rolling monthly extract from the AIR will be compared with the latest quarterly HES extract from NHS Digital in active surveillance of coverage of the registry (total number of RRP cases treated in England versus total number entered in the registry). The target date for recruitment completion is currently August 2020.

In the specific case of NMPCE projects commissioned by NICE, each individual project is issued with a timeline of dates for completion of project milestones. Most NICE projects require peer-reviewed publication of results within 12 months of project completion. NMPCE strives to share analysis code (where possible and journal permitting) for transparency and reproducibility purposes for future researchers.

NICE will consider evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

i) NICE IPG425: Endoscopic balloon dilatation for subglottic or tracheal stenosis (published April 2012 - update required)

ii) NICE IPG434: Radiofrequency cold ablation for respiratory papillomatosis (published November 2012 - update required).

iii) NICE IPG309: Percutaneous mitral valve leaflet repair for mitral regurgitation (published August 2009 - update required).

This agreement does not have any associated EU funding.

Previous intended benefits have been yielded and outputs achieved their stated purposes (and thus benefits of processing).

Examples of peer-reviewed publications (available in the public domain) resulting from the analysis of HES data previously accessed via HDIS by NMPCE for NICE purposes, with details of the impact these have achieved are as follows:

i) Keltie et al. Complications following vaginal mesh procedures for stress urinary incontinence: an 8 year study of 92,246 women. Sci Rep. 2017; 7(1): 12015.

IMPACT: Complications from surgical mesh procedures have led to legal cases against manufacturers worldwide and to national inquiries about their safety. This was one of the largest studies so far looking at adverse event rates of these procedures, and has significantly added to the understanding of the likelihood of adverse events from mesh procedures in a UK NHS setting. This means that women considering mesh procedures in future will have more accurate safety information available to them, so they can make a more informed choice. This publication and its analysis code was published in an open-access form to share both methodology and findings widely. NICE have also updated their NG123 guidance and produced linked Patient Decision Aids for women considering mesh surgery.

ii) Burn et al. Procedural and short-term safety of bronchial thermoplasty in clinical practice: evidence from a national registry and Hospital Episode Statistics. J Asthma. 2017; 54(8): 872-879.

IMPACT: Bronchial thermoplasty is a novel treatment for severe asthma, however its mode of action and target patient population were poorly defined, with little evidence available on efficacy and safety in terms of quality and quantity. This study presents procedural and short-term safety evidence from routine UK clinical practice.

iii) Donne et al. Prevalence and management of recurrent respiratory papillomatosis (RRP) in the UK: cross-sectional study. Clinical Otolaryngology. 2017; 42(1):86-91.

IMPACT: Recurrent respiratory papillomatosis is a relatively rare condition, but has a significant impact on patients and their families as it requires repeated anaesthetic procedures to achieve symptomatic control. This study utilised a cross-sectional survey of ENT consultants and analysis of HES data to provide, for the first time, an estimate for the prevalence of recurrent respiratory papillomatosis in the UK. This prevalence informed the sample size for an NIHR Research for Patient Benefit grant application - 3 year award granted. Identification of key providers of RRP care in the UK, enabled NMPCE to invite specific acute trusts to contribute to further research in RRP patients, and contribute to future NICE guidance (in update of IPG 434).

iv) Goode et al. Effect of procedure volume on outcomes after iliac artery angioplasty and stenting. British Journal of Surgery. 2013; 100:1189-1196.

IMPACT: This study was able to use HES data to determine if there was an association between outcomes of endovascular iliac artery intervention and centre volume, describing methodology which could be applied to any interventional in-hospital procedure which would benefit future analysis of HES data.

v) Patrick et al. Monitoring the use and outcomes of new devices and procedures: how does coding affect what Hospital Episode Statistics contribute? Lessons from 12 emerging procedures 2006-10. Journal of Public Health. 2012; 35(1): 132-138.

IMPACT: This study provided evidence that routinely collected HES data have the potential to support quality improvements and evidence-based commissioning of devices and procedures in the national health service.

vi) Keltie K, Cole H, Arber M, Patrick H, Powell J, Campbell B, Sims AJ. Identifying complications of interventional procedures from UK routine healthcare databases: a systematic search for methods using clinical codes. BMC Medical Research Methodology 2014; 14:126.

IMPACT: This study summarised methods of analysing complications of interventional procedures from routine healthcare databases in the UK published in the literature. This study was published in open-access form to share learnings with other researchers and clinical coding teams across the UK to encourage high quality research from HES.

In addition, NICE has incorporated evidence generated from NMPCE through its analysis of HES/mortality data in order to update the following Interventional Procedures Guidance:

1) NICE IPG635: Bronchial thermoplasty for severe asthma (originally published January 2012 - updated December 2018.

NMPCE has used data from HES to develop support drafting grant applications applicable to the following funding streams:

- NIHR Digital technologies,

- NIHR Call for ambitious data-enabled trials, health services and public health research studies.

Research themes have included (but not limited to):

- safety and efficacy of tonsillectomies in children and adults,

- longitudinal study to investigate the uptake and safety of UroLift procedures,

- longitudinal study to investigate the association between kidney transplantation and urological cancers.

This agreement does not support any current or planned PhD / post graduate research study.

If any of the analyses arising from this dissemination were to contribute to such studies in future an agreement to this agreement would be submitted.