Identifying Cancer Recurrence within Patient Care Pathways across Linked National Clinical Datasets

The London School of Hygiene and Tropical Medicine (LSHTM) requires access to National Disease Registration Service (NDRS) data for the purposes of the following research project:

Identifying Cancer Recurrence within Patient Care Pathways across Linked National Clinical Datasets

The following is a summary of the key aims of the research project being undertaken:

The aim of this research study is to develop and validate methods to phenotype cancer recurrence after curative treatment for bowel cancer in linked national clinical datasets and to assess how well the methods extend to other cancer sites.

Key objectives include:

1. Construct care and outcomes pathways of cancer patients across datasets, from diagnosis and treatment to

subsequent investigations and treatments for recurrence (Workpackages 1 & 2)

2. Develop four indicators of the presence and timing of bowel cancer recurrence, one using clinical rule-based

methods, one using statistical modelling and two using machine learning (ML). (Workpackage 3)

3. Validate the four indicators, including using clinical adjudication for a subset of patients (Workpackage 4)

4. Demonstrate the clinical use of the optimal indicator (Workpackage 5)

5. Assess how well the optimal indicator extends to other cancers (Workpackage)

In support of these aims, LSHTM requests access to the following NDRS Datasets:

• NDRS National Radiotherapy Dataset (RTDS)

• NDRS Systemic Anti-cancer Dataset (SACT)

• NDRS Linked Hospital Episode Statistics (HES)- Admitted Patient Care, Outpatient and Accident and Emergency Subsets

• NDRS Cancer Waiting Times

• NDRS Linked Diagnostic Imaging Dataset (DIDS)

• NDRS National Cancer Patient Experience Survey (CPES)

• NDRS Cancer Pathways

• NDRS Somatic Molecular Testing Data

• NDRS Cancer Registrations- including fields relating to deaths.

• NDRS Rapid Cancer Registration Data (RCRD)

The above-listed datasets are required because the research relies on being able to identify in the data the details of all types of healthcare activity in order to distinguish activity for cancer recurrence from activity for diagnoses, treatment and surveillance.

The level of data accessed will be pseudonymised.

Data is required for all of England to ensure that methods are developed in a fully representative cohort of patients and that the recurrence indicator is not sensitive to any idiosyncrasies of coding practices between healthcare providers.

LSHTM is the sole data controller who also processes the data for the purposes described within this Agreement. LSHTM is responsible for ensuring that the data will only be processed for the purposes described above.

Alongside substantive employees of LSHTM, data will be accessed by no more than four PhD students affiliated with LSHTM. Data will only be accessed by students where the processing is in support of the purposes outlined within this Agreement. In line with University’s policies, students who access data are required to undertake training in data protection and information security. Students processing the data will be under the supervision of a substantive employee of LSHTM.

Individuals employed by the Christie NHS Foundation Trust and Leeds Teaching Hospitals NHS Trust may provide advice to the study team. However, they will play no role in determining the purposes and means of processing. These individuals will not process any data disseminated under this Agreement.

LSHTM rely on Article 6 (1) (e) of the UK GDPR as the lawful basis of processing - "processing is necessary for the performance of a task carried out in the public interest or in the exercise of official authority vested in the controller". This is justified and falls under the official functions outlined within the University’s Royal Charter.

LSHTM rely on Article 9 (2) (j) as the legal basis for processing under the UK GDPR – “processing is necessary for archiving purposes in the public interest, scientific or historical research purposes or statistical purposes in accordance with Article 89(1) based on Union or Member State law which shall be proportionate to the aim pursued, respect the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject”. The basis in law is the University’s Royal Charter.

The National Institute of Health and Care Research (NIHR) funds this project, funding is in place until April 2025.

A PPI-focused Study Steering Committee (SSC) has been integral to the design and development from the outset. The SSC will meet twice per year to guide the design and delivery of the project, representing key NHS, data provider and clinician stakeholders and including a PPI and a charity representative for each cancer site. Additionally, the committee will be key in overseeing the planning and delivery of the outputs of the project. There are three confirmed PPI representatives in addition to the PPI co-applicant, representatives from Bowel Cancer UK, Breast Cancer Now, Prostate Cancer UK and NHS England.

No data will flow to the NDRS in support of this request.

The NDRS will provide the relevant records from the following datasets:

• NDRS National Radiotherapy Dataset (RTDS)

• NDRS Systemic Anti-cancer Dataset (SACT)

• NDRS Linked Hospital Episode Statistics (HES)- Admitted Patient Care, Outpatient and Accident and Emergency Subsets

• NDRS Cancer Waiting Times

• NDRS Linked Diagnostic Imaging Dataset (DIDS)

• NDRS National Cancer Patient Experience Survey (CPES)

• NDRS Cancer Pathways

• NDRS Somatic Molecular Testing Data

• NDRS Cancer Registrations- including fields relating to deaths.

• NDRS Rapid Cancer Registration Data (RCRD)

In line with the UK General Data Protection Regulation (GDPR) principle of Data Minimisation, the data will be minimised as follows:

• Data will be restricted to the pertaining C codes for; Oesophago-gastric cancers, Colorectal Cancer, Lung Cancers, Female Breast Cancers, Prostate Cancer, Head and Neck Cancers, Secondary cancers other.

• Data will be limited to patients who were 18 or older upon a diagnosis occurring between 2013 to the latest available data.

• For Cancer Registrations and NDRS Linked HES APC data will be further minimised to two years before diagnosis to the latest available.

Data disseminated under this Agreement will not be linked to any other data already held by the controller.

The data will contain no direct identifying items but will include a unique person ID which can be used to identify the same individual across the datasets requested. There will be no requirement or attempt to re-identify individuals from the pseudonymised data.

Following the receipt of the data, the data will not be transferred to any other location. The data will not leave England at any time, and will not be accessed outside this area.

The data will be stored on servers at LSHTM. All backups will be located onsite at LSHTM.

Data will be accessed onsite at the premises of LSHTM, or by authorised personnel via remote access. The data will remain on the servers at LSHTM at all times. The data will remain on the servers at LSHTM at all times and personnel are prohibited from copying/ downloading data to local devices. The study uses role-based access to the data, which means that only staff involved in the study work can be granted access to the strictly necessary information.

Access is restricted to substantive employees of LSHTM and PhD students affiliated with LSHTM. In line with University’s policies, all staff and students accessing data must complete mandatory data protection and information security training. These individuals will process the data for the purposes described above.

The outputs to date and further expected outputs are as follows:

> Publication of methods and validation work in peer-reviewed articles. The study team have one paper on development and validation of metastatic cancer indicators which has been accepted for publication in the journal BMC Cancer subject to minor revision. A paper on development and validation of the recurrence indicators is due for submission for publication by end of February 2025. A further article on statistical models for identifying recurrence is in preparation and due to be submitted for publication by April 2025. Articles on the clinical use of indicators for bowel cancer are planned for May-July 2025 and articles on recurrence indicators in other cancers in July to November 2025.

> The PPI representatives are closely linked to Bowel Cancer UK, Prostate Cancer UK and Breast Cancer Now. The first paper which has been accepted subject to minor revisions includes a section on the patient view of the findings. The PPI representatives will use their connections to publicise the projects' findings. These publications will be patient summaries of any peer-reviewed publications.

> The study team expect to be able to publish all algorithms on repositories for code sharing such as the Health Data Research UK (HDRUK) Innovation Gateway. This work will begin once papers are published.

> The study team will provide a research report to the study funder, NIHR, on the findings and conclusions of all work packages. Any such information will be inclusive of any recommendations for practice and will include a lay-accessible summary of the work undertaken.

> The study team have presented the findings of the work on metastatic cancer indicators and at conferences (CRUK Data-driven cancer research conference 2024 and Association of Coloproctologists of Great Britain and Ireland Annual Meeting 2024). The team will present the findings of current and upcoming work at methodological and clinical conferences.

The outputs will not contain NDRS data, and will only contain aggregated information with small numbers suppressed in line with relevant suppression rules.

The findings of this research study are expected to contribute to evidence-based decision-making for policymakers, local decision-makers (i.e. doctors), as well as patients, to inform best practices to improve the care, treatment and experience of healthcare users relevant to the subject matter of the study.

The methods the study aims to develop have the potential to allow cancer recurrence to be routinely identified

in cancer registries and national cancer audits by the end of the research project (March 2025). Knowledge of the date of recurrence will open up a large number of research opportunities (which in turn may provide a benefit to the provision of health and social care in England). Examples include:

• It may enable research into patterns of initial care which may be associated with recurrence, allowing studies with shorter follow-up than those using mortality as their key outcome.

• Making cancer recurrence indicators routinely available could allow better monitoring of healthcare providers, leading to improved care.

• It may permit further research into the care and outcomes of patients whose cancer has recurred.

• Recurrence information from routine health data could become available as an outcome in cancer clinical trials, decreasing the need for patient follow-up, making clinical trials more feasible and reducing costs.

The results of this research have the potential to enable these benefits, initially for bowel cancer before extending across other cancer sites including breast, prostate, oesophago-gastric, lung and head and neck cancers. Together these cancers make up 40% of cancer diagnoses, amounting to over 120,000 new diagnoses per year in England alone.

The research has the potential to benefit a huge number of research studies and audit outputs in the future as this study has the potential to allow cancer recurrence to become measurable in routine data. The resulting algorithms could give more representative estimates of cancer recurrence or recurrence-free survival for patients according to disease stage and physical fitness and other specific patient characteristics, treatments, and characteristics of providers of cancer services. An important contribution is that it has the potential to enhance clinical trials that typically have limited applicability in ‘real-world’ settings due to the limited ability to have long-term follow-up.

The study team also expect NHS England and regional commissioners of cancer services to be aided in various ways through the availability of better information on cancer recurrence facilitated by the study’s algorithms. In particular, cancer recurrence is a key element in models evaluating the cost-effectiveness of cancer treatments. More accurate national and regional information on the number of cancer patients who experience a recurrence will guide resource planning. The cancer recurrence indicators can be used to compare outcomes between treatments and providers, with subsequent implications for performance management and quality improvement.

The work to date has led to important new and robust algorithms for identifying metastatic cancer and cancer recurrence in national clinical datasets. The new methods have been internally validated and demonstrate excellent performance. The peer-reviewed articles include the full details of the algorithms so that they can be used widely by other analysts of national cancer data. Once the articles are published we will publish the algorithms on repositories such as the Health Data Research UK (HDRUK) Innovation Gateway.