NHS Digital Data Release Register - reformatted

Bangor University projects

HES data request for SANAD-II trial
NERVES Trial

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

HES data request for SANAD-II trial — DARS-NIC-75079-V7Y7L

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(1) and s261(2)(b)(ii), Health and Social Care Act 2012 s261(1) and s261(2)(b)(ii), Health and Social Care Act 2012 - s261 - 'Other dissemination of information', Health and Social Care Act 2012 s261(2)(b)(ii)

Purposes: No (Academic)

Sensitive: Non Sensitive, and Non-Sensitive

When:DSA runs 2020-03-01 — 2021-02-28 2020.06 — 2020.08.

Access method: One-Off

Data-controller type: BANGOR UNIVERSITY, THE WALTON CENTRE NHS FOUNDATION TRUST, UNIVERSITY OF LIVERPOOL

Sublicensing allowed: No

Datasets:

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

Objectives:

HES data are required for the purpose of the Standard And New Antiepileptic Drug (SANADII) trial.

Epilepsy is a common neurological (brain) disorder. Approximately 2-3% of the population will be given a diagnosis of epilepsy by time they reach 60.

A number of new drugs have been approved to treat epilepsy in the UK in the past few years. These drugs have been shown in research studies to prevent seizures and to be safe. There is as yet, however, no good evidence as to whether they are better at treating epileptic seizures or safer than the standard drugs that have been used for many years.

SANAD-II aims to assess the clinical and cost effectiveness of standard and new antiepileptic drugs to identify which drugs are the most effective and which ones make the best use of NHS resources. SANAD II has two arms, Arm A comparing lamotrigine, levetiracetam and zonisamide in patients with focal onset seizures, and Arm B comparing levetiracetam and valproate in patients with generalised onset seizures or seizures which are difficult to classify. Eligible study participants were aged 5 years or over at the time of randomisation, and have had a diagnosis of newly diagnosed epilepsy. The trial is no longer collecting data from participants. Most participants are still taking antiepileptic drug treatment for epilepsy.

In order to fulfil the objective of the SANADII study relative to cost-effectiveness, Bangor University will be conducting a health economic analysis that will adopt the perspective of the National Health Service (NHS) and Personal Social Services (PSS). The economics data will be based on resource use using entries made in case report forms, and Hospital Episode Statistics data sourced from NHS Digital and devolved equivalents.

Bangor University, The Walton Centre NHS Foundation Trust, and the University of Liverpool are joint data controllers. Bangor University is involved in the decision making and is the only organisation processing NHS Digital data.

The day-to-day running of the trial is being coordinated by the Clinical Trials Research Centre (LCTRC), University of Liverpool. The legal basis for processing falls under GDPR Article 6(1)( e) (public interest) and GDPR Article 9(2)(j) (for research purposes). Processing and dissemination meets public interest criteria as defined by the ICO for good decision-making by public bodies and securing the best use of public resources.

The processing of data for this study is a task of public interest since it will inform the NHS, clinical commissioners and other healthcare decision makers on the most clinical and cost-effective treatment for this condition.

The SANADII trial is a randomised controlled trial funded by the National Institute for Health Research's Health Technology Assessment (NIHR HTA) Programme and is co-sponsored by the University of Liverpool and The Walton Centre for Neurology and Neurosurgery NHS Foundation Trust. NIHR HTA funds independent research on the effectiveness, costs and broader impact of healthcare treatments and tests for those who plan, provide or receive care in the NHS. This funded research serves a variety of key stakeholders including: decision makers in local government, policy-makers including NICE, researchers, NHS health professionals, other NIHR stakeholders, and the general public.

SANADII recruited its first patient on 30/04/2013 and has now recruited 1510 participants from 94 opened sites. The final patient was recruited on 20/07/2017.

In order to undertake the cost-effectiveness analysis, data processing is necessary to cost the treatment arms compared in the study, which comprises an evaluation of the resources used by patients in the SANADII trial. This includes an analysis of all NHS hospital visits, which can be quantified through the provision, by NHS Digital, of pseudonymised data on hospital visits for the SANADII clinical trial participants.

This is a one-off request for Bangor University to receive data on the number and duration of Outpatient, Accident and Emergency and Admitted Patient Care episodes for SANADII trial patients, to cost the separate arms of the SANADII trial and to explore any difference in resource use between the antiepileptic drugs.

Identifying data are only used for the purpose of linking the data for SANADII trial patients and the analysis is undertaken by Bangor University using only pseudonymised data. Data will be pseudonymised by NHS Digital and forwarded to Bangor University, distinguished by patient clinical trial number only. Bangor University researchers will not have access to the 'key' that allows for patient identification.

The data required from NHS Digital is routine Outpatient, Accident and Emergency and Admitted Patient Care data for SANADII trial participants commencing 6 months prior to their entry in the study, to the end of the trial (minimum participation 2 years after randomisation, maximum participation duration 6.5 years after randomisation). The study is providing randomisation start and end dates for each patient to minimise the amount of HES data requested.

Specifically, the data required is the date of each Outpatient visit and Accident and Emergency visit according to Healthcare Resource Groups (HRG), and for Admitted Patient Care, the date and duration of stay, according to HRG, within the specified timeframe.

By measuring hospital resource use, researchers can estimate the overall cost for each patient then generate mean costs with confidence intervals for the trial arms. This information can then be used to inform future NHS treatments for patients with the condition by indicating which arm offers the best value for money. Data is being analysed for total costs. These are estimated from the sum of all Emergency Department, inpatient and outpatient costs derived from the HRGs. In order to represent uncertainty in the total cost, a statistical technique called 'bootstrapping' will be used. Bootstrapping involves taking samples over and over again (10,000 times) from the data to estimate how accurate the mean estimate for the entire population is (i.e. confidence intervals). The results of the work and how it helps the health and social care sector will be published in a HTA report and peer-reviewed journals such as the Lancet or Value in Health.

No patient-level data will be published.

The study requires patient level data so that an accurate costing of resource use per patient can be undertaken. Identifying data is not required for the purposes of this particular task.

In order to provide an accurate estimation of patient NHS hospital resource use, the collection of NHS Digital data was considered the most appropriate route to obtain robust, reliable data on hospital visits. Patient questionnaires have also been completed but are reliant on patient completion as well as patient recall and may introduce biases into the cost-effectiveness analysis results.

Data is restricted to patients who have consented to participate in the SANADII trial. Based on correspondence with NHS Digital in 2017, as part of fair processing, patients were sent a letter in 2018 containing further information on how their personal data would be shared with NHS Digital in order to access their electronic medical records. The letter offered participants the option to opt out of having their data shared with NHS Digital for the outlined purpose.

Only data relevant to identifying the number and duration of Accident and Emergency, Admitted Patient Care and Outpatient visits are required and only for the trial duration.

Outcomes from the trial will be used to inform future clinical practise and a robust economic analysis will support the clinical analysis.

Any results from the study that are shared publicly will be aggregated.

Expected Benefits:

The aim of processing the HES data is to identify which of the two treatment arms of the trial is the most cost-effective method of treating monotherapy with lamotrigine, levetiracetam or zonisamide in patients with untreated focal onset seizures, and monotherapy with levetiracetam or valproate in patients with untreated generalised onset seizures or untreated seizures that are difficult to classify. The results will be considered alongside the evidence of which is the most clinically effective.

Health care professionals and NHS commissioners will be better informed which of the antiepileptic drugs will be the most clinically and cost-effective in terms of patient outcomes, need for hospitalisation or hospital visits, improved care and by extension, which offer the best value for money to the NHS healthcare system. This will be a health care benefit; not only to the NHS in terms of efficacy and best use of resources, but also to patients in terms of clinical outcomes and the global healthcare research community.

NHS Digital data will contribute to the NHS and wider clinical research communities (including decision-makers in local government, policy-makers including NICE, researchers, NHS health professionals, other NIHR stakeholders, and the general public) understanding of the most clinically and cost-effective means of treating epilepsy. It would not be possible to undertake a robust economic evaluation without the NHS Digital HES data.

Benefits are also expected to the NHS through reduced resource use and improved cost efficiencies. The benefits can be measured through changes in epilepsy-related hospital annual submissions via the Patient Linked Information Costing System (PLICS).

Once the results of the study are published, it will up to local decision-makers such as clinical commissioning groups, health-care policy-makers including NICE, researchers, NHS health professionals, other NIHR stakeholders, and the general public to implement the changes.

Outputs:

The plan is to share the results of the cost-effectiveness analysis with key stakeholders, including decision-makers in local government, policy-makers (e.g. NICE), researchers, NHS health professionals, other NIHR stakeholders, and the general public through journals, public reports, presenting at relevant conferences and symposium.

The University of Liverpool and Bangor University will be contributing to the NIHR HTA report underpinning the SANADII study, which will be a wider report on the entire SANADII study published on the NIHR website. One of the sections of this report will be dedicated to health economics and is planned for publication within 12 months of the receipt of data from NHS Digital.

A stand-alone article based primarily on the findings of the cost-effectiveness analysis will be published in a peer reviewed health economics, or other medical journal, that may be relevant to epilepsy e.g. Value in Health or The Lancet within 18 months of receipt of data from NHS Digital.

Presentations relating to the findings of the cost-effectiveness analysis will be given at local, national and international conferences as opportunities arise.

The dissemination activities undertaken concerning the SANADII study are two-fold; to enable the engagement with the scientific and policy-making communities and to ensure that knowledge developed by the research can benefit these communities.

The collaborative SANADII trial team will participate in active communication activities to ensure any information about the project and its results reaches interested groups and civil society and will involve knowledge sharing and dialogue. Planned communication channels to key stakeholders including decision-makers in local government, policy-makers including NICE, researchers, NHS health professionals, other NIHR stakeholders, and the general public will be made through the SANADII website and newsletters. All outputs will be aggregated with small numbers suppressed in line with the HES Analysis Guide.

Processing:

The University of Liverpool Clinical Trials Research Centre (CTRC) will extract the following identifying information from their SANADII patient database and forward this to NHS Digital via secure file transfer:
Adults - Study ID (patient randomisation study number), NHS Number, Date of Birth, Postcode, randomisation date, study completion date
Children - Study ID (patient randomisation study number), Date of Birth, Postcode, randomisation date, study completion date and name.

NHS Digital will link this information to HES Accident and Emergency, Admitted Patient Care and Outpatient data. The study team did not collect the NHS number for children to Liverpool CTRC and do not therefore have this information to pass to NHS Digital for linkage. Liverpool CTRC will send the children's names for linkage purposes. NHS Digital will remove all identifying data items other than Study ID from this dataset and forward the pseudonymised data to the health economics team at Bangor University's Centre for Health Economics and Medicines Evaluation (CHEME) for analysis. Bangor University do not have access to participant NHS number, date of birth, name, or postcode. In order to assess the impact of baseline characteristics Bangor University do have access to age at randomisation and gender. Bangor University will use the NHS Digital data to calculate the patient level costs and then combine those costs with the trial data, which will be extracted to the health economics analysis file. Bangor University will not combine any NHS Digital data with trial data.

There will be no flows of patient-level data from Bangor University. Bangor University researchers will not have access to the 'key' that allows patient identification. No attempts will be made by researchers in Bangor University to obtain the key linking the clinical trial number to any patient identifiers, removing any risk of re-identification. Pseudonymised data will be processed in Bangor University to generate aggregated data pertaining to the cost effectiveness of the two arms in the clinical trial.

Patients will be linked to their arm of the study via their randomisation number. The data is handled by taking the sum of all inpatient and outpatient costs derived from the Healthcare Resource Groups (HRGs) and doing a 10,000 replication bootstrap to generate the confidence intervals for the costs in each arm of the study. The results of the work and how it helps the health and social care sector are published in peer-reviewed journals such as the Lancet or Value in Health.

No patient-level data will be published.

Bangor University will conduct linkage via the HRG codes to obtain HRG costs from publicly available National Tariff registers and, via the patient randomisation study number, will attach additional costs and Quality Adjusted Life Years (QALYs) to a consolidated working file. The multiple line per patient inpatient and outpatient hospital visits in the pseudonymised HES datasets provided by NHS Digital will be costed line-by-line by attaching HRG costs from the most up to date National Tariff to the corresponding HRG in each line of data. The dataset will be compared with pseudonymised data obtained from patient questionnaires.

At this point in time it is not possible to be more specific about which additional cost items and HRG codes will be included as the data has not yet been released; however, any HRGs used in small numbers (<10) will not be identified and subsequently costed under miscellaneous HRGs.

Resource use will be tabulated by trial arm, in order to describe differences in the use of services between randomised groups. Within-trial total costs for each patient will be calculated from the sum of all costs (associated with primary, secondary and community care services, and medication use). Total costs during the course of the trial will be calculated, with summary statistics generated by intervention group.

Costs at baseline, relating to the 6-months preceding randomisation, will be calculated in order to adjust for any baseline difference. This will only relate to secondary care usage. If a hospitalisation is observed for the period subsequent to randomisation, an adjustment may be necessary to apportion costs given that ward costs relate to episodes of care which could start prior to randomisation.

In order to account for any imbalances in important clinical or demographic variables (e.g. age, gender), researchers will implement statistical analysis such as regression analysis to measure the relationship between these demographics, and cost and quality of life. The type of analysis will be confirmed based on the data distributions, in the event that regression analyses are not suitable, generalised linear models will be used as an alternative.

Aggregated data will be stored ready for publishing in peer-reviewed journals, presenting in symposia etc. Commonly used HRGs will be included in the aggregated data. Seldom-used HRGs will not be identified in the aggregated data mitigating any risk of re-identification.

Data received from NHS Digital will be stored on an AES 256 encrypted flash drive. Once costs have been assigned to HRG codes, these will be extracted to the health economics analysis file (containing: resource use and costs derived from trial data, self report quality of life data, age at randomisation, gender, date of primary outcome, type of treatment failure, date of treatment failure, epilepsy syndrome, randomisation number, randomisation date, treatment allocation, number of seizures) , which will be held on secure servers at Bangor University in folders with access restricted to Bangor University analysts in CHEME. At the end of the initial Data Shaing Agreement requests will be made to NHS Digital for archiving the data for a period of 15 years in accordance with clinical trial regulations.

Data processing will only be carried out by substantive employees of Bangor University who have been appropriately trained in data protection and confidentiality.

All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by ͞Personnel͟(as defined within the Data Sharing Framework Contract ie: employees, agents and contractors of the Data Recipient who may have access to that data).

No data will be shared with 3rd parties.

The Data will only be used for the purposes described in this Agreement.

NERVES Trial — DARS-NIC-139741-D9Y1C