NHS Digital Data Release Register - reformatted

GLAXOSMITHKLINE RESEARCH & DEVELOPMENT LIMITED projects

The Extended Salford Lung Study Data Access Project
Investigation of TRELEGY Effectiveness: Usual Practice Design (INTREPID) Exploratory data set

48 data files in total were disseminated unsafely (information about files used safely is missing for TRE/"system access" projects).

The Extended Salford Lung Study Data Access Project — DARS-NIC-115298-L5X4V

Type of data: information not disclosed for TRE projects

Opt outs honoured: No - data flow is not identifiable, Anonymised - ICO Code Compliant, No (Consent (Reasonable Expectation))

Legal basis: Health and Social Care Act 2012 - s261 - 'Other dissemination of information', Health and Social Care Act 2012 - s261(5)(d)

Purposes: Yes (Supplier, Commercial)

Sensitive: Non Sensitive, and Non-Sensitive

When:DSA runs 2021-02-12 — 2022-05-11 2021.04 — 2021.04.

Access method: One-Off

Data-controller type: GLAXOSMITHKLINE RESEARCH & DEVELOPMENT LIMITED

Sublicensing allowed: No

Datasets:

Hospital Episode Statistics Accident and Emergency
Hospital Episode Statistics Admitted Patient Care
Hospital Episode Statistics Outpatients
Hospital Episode Statistics Critical Care
Hospital Episode Statistics Accident and Emergency (HES A and E)
Hospital Episode Statistics Admitted Patient Care (HES APC)
Hospital Episode Statistics Critical Care (HES Critical Care)
Hospital Episode Statistics Outpatients (HES OP)

Objectives:

BACKGROUND
The Extended-SLS is a follow-on study to the Salford Lung Studies (SLS), two landmark effectiveness trials of fluticasone furoate / vilanterol (an inhaled corticosteroid combined with a long-acting-b2-agonist [LABA] in a single inhaler device) in patients with Chronic Obstructive Pulmonary Disease (COPD) (NCT01551758) and asthma (NCT01706198) which ran from March 2012 to December 2016.

The SLS subjects represent patient cohorts that are extremely well-characterised over a short period of their COPD/asthma disease experience. Subjects in the SLS originally consented for information relevant to the study to be shared with the sponsor, GlaxosmithKline Research & Development Limited (referred to in this application as GlaxosmithKline or GSK), and these data were limited to three years prior to randomisation and the twelve-month interventional treatment period. This finite period of data coverage limits the potential to address scientific questions of clinical interest related to long-term COPD/asthma disease progression and associated outcomes (for example, mortality). Broadened access to patients’ data would allow SLS subjects’ entire disease journey to be researched, presenting a rare opportunity to use epidemiological research to improve scientific and clinical understanding of COPD/asthma disease risk, treatment and progression.

The Extended-SLS seeks to broaden capture of SLS patients’ data through the collection of additional subject-level data encompassing past and periodic future (up to 10 years duration from the date of consent) demographic, COPD/asthma risk factors and healthcare-related information from both primary and secondary care, resulting in the creation of the Extended-SLS cohort. Additional patient reported data on early life exposures, impact of disease on sleep, smoking history and other information not typically available from electronic medical records (EMR) will be captured using disease-specific questionnaires. Secondary care data on hospital admissions and attendances available in NHS Digital’s Hospital Episode Statistics (HES) data are crucial to understanding and researching hospitalised exacerbation's of COPD/asthma, comorbidity, long-term safety of COPD and asthma therapies, and healthcare resource utilisation.

The Extended-SLS has three organisations involved in the work but only GSK and Ignite are involved in the processing of the HES being requested in this application:

• GlaxosmithKline (GSK): The sponsors of the original SLS and the Extended-SLS. GSK wish to receive the final pseudonymised cut of study data, including the HES data provided by NHS Digital. GSK are a data controller and a data processor in this study. GSK is an international pharmaceutical company and the relevant GSK party for this agreement is a company registered in England and Wales. GSK produces drugs directly relevant to the ailments being studied and holds the international patent to one of the drugs specifically referred to in this agreement.

• Ignite Data (Ignite): The contracted research company looking after site management, patient consent and data processing. Ignite will have the explicit consent of Participants for bringing together multiple different data sources captured on each individual study volunteer, specifically including the consent of patients to process their HES and mortality data. Note patients provided consent for use of mortality data; however, this application is not requesting ONS mortality data.

• Graphnet Health (Graphnet): The general practitioner (GP) extract provider for the study. Graphnet currently have active data sharing agreements with every GP who is participating in the study for direct care and hold Data Security and Protection (DSP) Toolkit. Study specific agreements exist with each individual GP giving Graphnet permission to extract consented Participant data for this study only. Graphnet will not be in receipt of any HES data provided to Ignite Data by NHS Digital. Graphnet are not directly involved with this application but are a source of the main study data set.

Hospital Episode Statistics will be used to form a longitudinal patient record that GSK wish to build over the lifetime of the study. HES data supplied to Ignite Data and GSK are essential to answer questions in the Extended-SLS relating to:

• Healthcare resource utilisation and costs (HRG); primary care electronic medical records do not accurately and reliably capture complete information about hospital attendances and admissions and therefore cannot inform on the full spectrum of healthcare resource utilisation.
• Severe exacerbation's of COPD and asthma; these are defined, according to global standards, as exacerbation's requiring hospital admission (asthma and COPD) and exacerbation's requiring emergency care (asthma), validation studies demonstrate that severe exacerbation's are not adequately recorded in primary care EMR.
• Frailty and disease severity defined based on prior hospitalisations (all-cause) and comorbidities not managed in primary care.
• Potential, treatment-related adverse events resulting in hospitalisation.
• Impact of treatments and/or disease severity/subtypes on all-cause or COPD- and asthma- related mortality; primary care.
(HRG is analysed using the latest publicly available National Cost Collection:
https://improvement.nhs.uk/resources/national-cost-collection/)

The number of study participants in England in the SLS who have consented onto the study is currently 1,183, with those consenting for NHS Digital sharing their data being 1,121. This is an ongoing study and Ignite Data Limited estimate the cohort to be 1,300 individuals. All participants are 18 years or older. Each study participant has a current diagnosis of COPD or asthma and was previously enrolled on to RCT (Randomised Controlled Trial) known as the Salford Lung Study. Each individual participant was recruited via the SLS study site, which in every case was their local GP. The GP Practices were selected from the Greater Manchester Region based on the fact they had participated in the original Randomised Control Trial (RCT).

GSK and Ignite Data have many questions that can potentially be answered as a result of learnings from the study. Within GSK’s COPD and Asthma disease areas, these data may help to identify other key areas for GSK to focus their medicines development.

By linking NHS Digital’s HES data to the data already held from the SLS study for patients in the Extended-SLS cohort (primary care EHR, disease-specific patient-completed questionnaires, and the SLS trial data), GSK and Ignite Data hope to address a range of research questions. Without access to HES data GSK and Ignite Data will not be able to answer any research questions where data on severe asthma or COPD exacerbations are required as these important adverse outcomes of asthma and COPD can only be reliably identified using discharge data from hospitals. Furthermore, GSK and Ignite Data will not be able to describe the full societal burden of these conditions without access to HES Admitted Patient Care (APC), Accident and Emergency (A&E) and Outpatient (OP) data as these are important sources of health-care resource utilization in asthma and COPD.

Research questions which GSK and Ignite Data cannot fully answer without access to the HES data include:

• Identifying which specific patient and disease traits can be used to guide clinician’s choice of treatment.
• Identifying what impact biomarkers (eosinophils, fibrinogen, etc.) have on (a) the risk of asthma and COPD exacerbations and (b) the risk of pneumonia.
• How patients progress from diagnosis to particular treatments, and if there are factors associated with this progression, e.g. potential clusters of comorbid conditions.
• Identifying if there is a relationship between events pre-COPD or pre-asthma diagnosis (respiratory infections, comorbidities, healthcare-resource utilisation patterns, smoking and body mass index [BMI] dynamic changes) and natural history of COPD and asthma disease progression?.
• Identifying how smoking status modulate or predict disease trajectory in COPD.
• How healthcare resource utilisation differs in patients with and without exacerbations.
• Examining the effect of asthma onset (early vs late onset) on longitudinal healthcare utilisation and asthma management.
• Long term symptom control, adherence and outcomes (e.g. exacerbations) for asthma patients using specific treatments.
• Investigation of sentinel events (e.g. exacerbations, a prescription of antibiotics or prednisolone) trigger a change in management and investigate what the health care is after a sentinel event.

DATA MINIMISATION
Ignite Data Ltd requires HES Critical Care (CC) data to accurately ascertain the level of Healthcare Resource Utilisation (HRU) during hospitalisation. Although these are a subset of admitted patient care episodes, critical care is associated with a higher cost. These data are required for example, to ascertain the number of days of advanced respiratory support for patients ventilated in critical care during an admission for COPD exacerbation.

HES Accident and Emergency (A&E) data is required to understand any serious events which resulted in attendance at the emergency department but did not result in admission to hospital. Without these data the study team would not be able to, for example, ascertain episodes of treatment associated with attendance at A&E for acute exacerbations of Chronic obstructive pulmonary disease (COPD), where the patient is seen in A&E without admission. HES Admitted Patient Care (APC) data is required to ascertain any serious events which resulted in admission to hospital – for example – severe COPD or Asthma exacerbations and other comorbidities; HRU/cost. HES Out Patients (OP) data is required to ascertain secondary care treatment and outcomes, for example, outpatient respiratory clinic appointments where the patient has been referred to the respiratory physician by their GP and associated cost/HRU of this treatment.

The data provided has been linked against each consented participants’ Randomised Control Trial (RCT) data via their unique Study ID. Only this pseudonymised dataset will be provided to GSK by Ignite Data Ltd for analysis. Ignite Data Ltd will destroy the date they hold after processing has been completed. As this data is being linked to an RCT dataset and then viewed prospectively for the next few years, only the years relevant to the study period are being requested. If the number of years are reduced any further it will not be possible to complete any beneficial analysis as the dataset would no longer match the study timeline or provide the depth of information required to study the desired outcomes.

The data has been narrowed by geography due to the fact that the consented patient list only contains patients from specific English research sites. Due to the condition, all patients will live within a reasonable commute of each study site. As the study is a consented and recruited cohort it is not possible to further minimise by demographic information any further. The data cannot be narrow by clinical factors as without data relating to events outside of the main study condition, Ignite Data Ltd and GSK will not be able to understand all cause HRU.

Patient episodes are required to achieve the study purpose. The patients have consented to GSK having access to relevant data in their whole medical record, only then can the study team truly get a picture of their disease history, its treatment and progression historically and during the prospective period of the study. Elective episodes are also required as without data relating to events outside of the main study condition GSK will not be able to understand all cause HRU. Maternity episodes are not required for analysis.

There is a timeframe around the index event required because without the dates relating to events, GSK will not be able to determine whether events occur after the patient’s entry into the study and before study end, therefore during that individual patient’s follow up. The study team will also not be able to determine the pattern or treatment and relevant measures contributing to a developing picture of a patient’s disease progression or management. Further, although record identifiers can be used to link episodes of care into spells, associating spells into super-spells of care (when patients may receive care from more than one hospital or trust) requires dates of admission and discharge.

Not all the fields within the HES APC, A&E and CC dataset are required. Only the specific fields required, which have been reviewed by GSK epidemiologist have been selected.

For linkage, the data will only be linked to the RCT cohort for the analysis set out in this application. As previously described the dataset is minimised by the recruited study cohort and all episodes on the patients are required to determine all cause HRU.

The data is being processed under General Data Protection Regulation Article 6 (1) (f) Legitimate interests as this is for the benefit of science and public health. Processing the NHS Digital data will enable us to address specific research questions relating to asthma and COPD including advances in general knowledge of COPD and asthma disease risk and treatment that may alter how clinicians care for their patients and raise awareness of the burden of these on patients and the wider health system. Insight into the longer-term burden of COPD and Asthma (which are common chronic respiratory conditions which place a high burden on patients and society) would be beneficial to the wider medical science community including medicines developers and healthcare providers.

As special category data, the data is further being processed under article 9(2)(j) (processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes). This is because the consented data will be used to enhance the potential outcomes and improve care, which is in the public interest.

Expected Benefits:

Benefits to the provision of health care
The Extended SLS Study will be one of the key resources used to help researchers:
• Get a better understanding of who is at risk of developing Asthma and COPD and why the progression of the disease varies from person to person;
• Explore the anatomy of the diseases to help develop new medicines and enable more accurate diagnosis
• Look into how existing drugs which are used to treat other conditions might help to treat the progression of Asthma and COPD and improve symptoms.
The Extended-SLS will enable GSK to conduct methodologically-sound research that answers questions important to COPD and asthma Participants, providing further clarity to the scientific and clinical communities around COPD/asthma disease risk, treatment and progression.
The data requested will enhance the analysis of who is at risk of developing Asthma and COPD and allow GSK and the broader scientific respiratory community to answer questions which will ultimately be used to benefit the management of Asthma and COPD patients.

Processing of the HES secondary care data to address specific research questions relating to asthma and COPD using the Extended SLS cohort will have broad public benefit, including advances in general knowledge of COPD and asthma disease risk and treatment that may alter how clinicians care for their patients and raise awareness of the burden of these on patients and the wider health system. For example, analysis of the Extended-SLS data may improve our understanding of groups of patients at higher risk of developing asthma and COPD who may benefit from increased monitoring, or provide additional information on specific groups of asthma and COPD patients who are more likely to experience severe exacerbations and may therefore require alternative disease management.

More generally, creation of the Extended SLS cohort will demonstrate that extension studies of RCTs are feasible and that important research questions can be addressed using data that has already been collected for other purposes (e.g. NHS Digital HES). Thus facilitating GPs participation in research without onerous and time consuming completion of case report forms which require patient data to be extracted from patient records into a research database.

Insight into the longer-term burden of COPD and Asthma (which are common chronic respiratory conditions which place a high burden on patients and society) would be beneficial to the wider medical science community including medicines developers and healthcare providers. We might learn how to better manage these conditions. Patients consenting into this study have an expectation that processing their data may lead to improved healthcare for them and others. Furthermore, sourcing data from EHR held at medical practices and in centralised healthcare records allows an investigational site within the NHS to participate in medical research but with a much-reduced resource cost to the site. Reducing the burden of research at sites would help to increase the amount of medical research that could be conducted in the UK with obvious benefit to both the public and the healthcare system.

Outputs:

Outputs for the Extended-SLS will include a series of reports, conference abstracts (international and UK conferences including: European Respiratory Society, British Thoracic Society, ISPOR, etc) and published manuscripts in peer-reviewed journals on COPD/asthma disease risk, treatment and progression.

GSK is committed to sharing results of impactful research with the wider scientific community and all research evaluating a medicine, providing important scientific knowledge, and/or which has relevance for patient care is published externally (and on our GSK Study Register: https://www.gsk-clinicalstudyregister.com/) in line with their policies.
The first publication and piece of research they intend to conduct using the Extended-SLS cohort is a descriptive analysis , comparing participants in the Extended-SLS with those in the original SLS. The analysis using SLS trial data, Extended-SLS GP data and Extended-SLS data from the disease-specific questionnaires has been completed and is being developed as a manuscript. An abstract on the Extended-SLS study, including results from this initial analysis was presented at the European Respiratory Society congress in September 2020. Once HES data are available, the Extended-SLS cohort will be further described using HES data. Additional outputs of the Extended-SLS, including reports, presentations, manuscripts, etc. will be determined once recruitment into the Extended-SLS study has completed. Three studies that are proposed to begin from 2021 will explore specific questions relating to asthma control, moderate exacerbations of asthma, and use of ‘maintenance and reliever therapy’ regimens in asthma patients in Ext -SLS, with conference abstracts and publications planned.

The audiences for outputs of the wider Extended-SLS will predominantly include researchers, scientists, and clinicians. GSK plan to disseminate information to Extended SLS participants throughout the lifecycle of the study via newsletters to GP sites. Information will need to be disseminated via GPs as GSK does not have access to participants names, addresses or email.

Any published results would contain data which has been aggregated with small number suppression applied as per the HES Analysis Guide to ensure privacy is maintained. All data in outputs will be published in-line with GSK policies.

In accordance with GSK’s internal policy, all trial results will be shared, regardless of whether they reflect positively or negatively on GSK’s medicines. GSK posted information about the EX-SLS study on a publicly accessible register (ClinicalTrials.Gov) before it started and will update it with a result summary after the study is finished. GSK seek the publication of all results of all clinical trials in peer-reviewed scientific journals and the EX-SLS study and any relevant result found as part of the data processing activities in this application will be no different.

Within a year of data receipt, and during the lifetime of the study, there will be multiple outputs: GSK's aim is to provide updates to annual scientific conferences/congresses as the study progresses and to answer further disease specific questions generated by the research community inside and external to GSK.

The outputs of this study aim to support the below listed benefits by looking at the range of benefits for COPD and asthma patients and HCPs as below:
- Identifying what other factors are associated with diagnosis of early COPD. This will increase understanding of groups at higher risk for developing COPD and who may benefit from increased monitoring/ alternative treatment paradigm or disease management
- Identifying what triggers a COPD diagnosis, e.g. are there particular events that lead to a initial diagnosis of COPD? This will improve understanding of presenting events which often suggest a COPD diagnosis; improved patient diagnosis.

- To aim to describe patient progression from diagnosis to therapy with a long-acting bronchodilator (LABD ) to triple therapy (comprising an ICS, a LABA and a long-acting muscarinic antagonist [LAMA]), and describe factors associated with this progression, including the role of potential clusters of comorbid conditions. This will mean patient treatment pathways can assist in identifying subgroups of patients which may be over/under treated based upon their symptoms and COPD profile.

Processing:

DATA SUMMARY:
Pseudonymised* record-level historic Hospital Episode Statistics data extracts linked to a consented cohort of approximately 1,300 extended Salford Lung Study participants. Each patient has provided explicit consent for their personal information identifiers [Participant Surname, Forename, Date of Birth, NHS Number, Address and Post Code] to be provided to NHS Digital.

• Hospital Episode Statistics (HES) Admitted Patient Care (APC) 2009/10 – 2019/20
• HES Accident and Emergency (A&E) 2009/10 – 2019/20 M12
• HES Out Patients (OP) for the periods 2009/10 – 2019/20
• HES Critical Care (CC) for the periods 2009/10 – 2019/20

As per the consent documentation, the study team can only hold 10 full years of prospective data from the date of consent.

METHODOLOGY
1. [Outside of this agreement scope] Cohort data originated from Graphnet, collected from participating GPs, and sent to Ignite Data Limited.

2. Ignite Data Limited apply GSK Study ID to the cohort and the cohort identifiers [Participant Surname, Forename, Date of Birth, NHS Number, Address and Post Code] sent to NHS Digital via Secure Electronic File Transfer (SEFT). It is estimated that the cohort will be 1,300 individuals.

3. NHS Digital will receive the cohort and link to the below periods of HES data. Identifiers will be removed, and pseudonymised data returned to Ignite Data Ltd via Secure Electronic File Transfer (SEFT).

4. Ignite Data Ltd will link NHS Digital Pseudonymised data to further pseudonymised RCT participant data via the GSK Study ID, and pass to GSK for further analysis.

*Pseudonymisation is defined as using a code (GSK Study ID) to replace any directly identifiable personal or sensitive personal information that GSK has no legal requirement, or wish, to hold. The use of this code is to allow GSK to link an individual’s data from different sources over time whilst minimising risk incurred by holding unnecessary Personal and Sensitive Personal information.

The pseudonymised NHS Digital HES data will be linked to pseudonymised RCT data obtained during a patient’s participation using the GSK Study ID. This will enhance GSK and Ignite Data’s understanding of a patient’s Healthcare Resource Utilisation during their study participation.

IGNITE will have access to the record level data to complete this work and this has been made clear to the participant in the Patient Information Sheet (PIS) and Informed Consent Form (ICF) the participants have completed. IGNITE will destroy the data once data processing has been complete in-line with GDPR and Data Protection Legislation.

IGNITE shall transfer the data onto a server within their Microsoft Azure datacentre environment. This environment is heavily access controlled and is only accessible via trained and authorised IGNITE staff. This access control is managed in-line with ISO27001 policy. To safeguard any sensitive data being processed staff are trained on a regular basis. Yearly ISO27001 training is compulsory and is reviewed by external audit yearly, NHS DSP Toolkit is also tied into this regime along with GDPR training. Data accessed within the IGNITE environment shall be exclusively by IGNITE employees. The final audited data is sent via encrypted transfer to GSK where it is then stored behind a VPN controlled firewall on a UK based server with further access right control in-line with GSK’s NHS DSP Toolkit policy.

The data will not be linked or compared (matched) with other data sets not detailed within this agreement. There will be no attempts to try and re-identify or re link to identifiable record level patient data.

The data received by GSK and Ignite Data will not be used for any purpose other than to meet objectives as stated in this Data Sharing Agreement and will not be shared with any other third party or organisation other than in the form of aggregated level data with small number suppression applied as per the HES Analysis Guide.

Microsoft Ltd provide Azure Backup Storage Services for IGNITE Data Limited and are therefore listed as a data processor. They supply support to the system, but do not access data. Therefore, any access to the data held under this agreement would be considered a breach of the agreement. This includes granting of access to the database[s] containing the data.

HES and ECDS DISCLOSURE CONTROL / SMALL NUMBER SUPPRESSION
In order to protect patient confidentiality, when presenting results calculated from HES record level data, outputs will contain only aggregate level data with small numbers suppressed in line with HES Analysis Guide. When publishing HES data, you must make sure that:
• cell values from 1 to 7 are suppressed at a local level to prevent possible identification of individuals from small counts within the table.
• Zeros (0) do not need to be suppressed.
• All other counts will be rounded to the nearest 5.
Data will not be made available to any third parties other than those specified except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide.

Investigation of TRELEGY Effectiveness: Usual Practice Design (INTREPID) Exploratory data set — DARS-NIC-297783-V4P6H