Modelling impact of interruptions to cancer screening with COVID ( ODR2021_016 )

Text removed is struck through; text added is underlined. Unchanged paragraphs are summarised rather than repeated.

Datasets: + NDRS Rapid Cancer Registrations

Objective for processing

Data for this study has previously been shared when the data were controlled and managed by Public Health England (PHE). PHE facilitated data release via its Office of Data Release service (ODR). ODR was responsible for providing a common governance framework for responding to requests to access PHE data for secondary purposes, including service improvement, surveillance and ethically approved research. All requests to access data were reviewed by the ODR and were subject to strict confidentiality provisions. The responsibility for the management of the National Disease Registration Service of which the National Cancer Registration and Analysis Service is a part, transferred from PHE to NHS Digital (Now NHS England) on 1st October 2021. University College London (UCL) requires access to NHS England data for the purpose of the following research project: Modelling impact of interruptions to cancer screening with COVID. The following is a summary of the aims of the research project provided by UCL: [6 paragraphs unchanged] The following NHS England Data will be accessed: • NDRS Linked Cancer Registration • NDRS Rapid Cancer Registrations The level of the Data will be pseudonymised. The Data will be minimised as follows: • Limited to both male and female patients aged 18 or over at diagnosis from all ethnicities. • Limited to the following geographic areas: England • Limited to conditions relevant to the study identified by specific ICD or OPCS codes; o 1. Lung (C34x) o 2. Colorectal (C18x-C20x) o 3. Prostate (C61x) o 4. Breast (C50x, D05x) o 5. Pancreatic (C25x) o 6. Oesophagus (C15x) o 7. Liver (C22x) o 8. Bladder (C67x) o 9. Kidney (C64x) o 10. Ovarian (C56x) For NDRS Rapid Cancer Registrations the Data will also be limited to patients diagnosed between 01/01/2018 to 30/05/2021 as well as the bullet points mentioned above. If a patient has more than one tumour then take the tumour diagnosed first. For NDRS Linked Cancer Registrations the Data will also be limited to patients diagnosed between 01/01/2013 to 31/12/2017 as well as the bullet points mentioned above. If a patient has more than one tumour then take the tumour diagnosed first. UCL is the research sponsor and the controller as the organisation responsible for ensuring that the Data will only be processed for the purpose described above. The lawful basis for processing personal data under the UK GDPR is: Article 6(1)(e) - 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. The lawful basis for processing special category data under the UK GDPR is: Article 9(2)(j) - 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 essence of the right to data protection and provide for suitable and specific measures to safeguard the fundamental rights and the interests of the data subject. This processing is in the public interest because the resulting research could support policy to mitigate the impact of any delays on cancer outcomes both from COVID and any future emergency. The modelling could also inform policy decisions regarding acceptable targets for w There is no specific funding of this research. No other organisation will be processing data including providing IT support, IT hosting services and IT back up services. Data will only by accessed by substantive employees of the organisation named as controller or processor. This work has not involved members of the public or patients. It was initially conceived within the context of the initial emergency setting of the pandemic response.

Processing activities

Design of study: Mathematical modelling. Study population: All adults aged 18 and over in England with one or more of lung, colorectal, prostate, breast, pancreatic, oesophageal, liver, bladder, kidney, or ovarian cancers. Statistical analysis: This analysis will have three overarching stages: 1) Stage progression. We will analyse retrospective cancer registry data to derive the time taken for each cancer type being investigated to progress between stages. No data will flow to NHS England for the purposes of this Data Sharing Agreement (DSA). We will apply for access through Public Health England's (PHE) Office for Data Release to Cancer Registry data for all patients aged 18 or older in England diagnosed with one of: lung, colorectal, prostate, breast, pancreatic, oesophageal, liver, bladder, kidney, or ovarian cancer diagnosed between 01/01/2013 and 31/12/2017.We will subsequently apply for the Public Health England’s rapid cancer registration dataset providing data from 2018 up to the present. NHS England will provide the relevant records from the NDRS Linked Cancer Registrations and the NDRS Rapid Cancer Registrations datasets to UCL. The Data will contain no direct identifying data items but will contain a unique person ID which can be used to link the Data with other record level data already held by the recipient. Using depersonalised data on age at diagnosis, the year their cancer was diagnosed, the type of cancer, its stage at diagnosis, and year of death (if applicable), we will use a Markov multistate model to enable us to estimate the time taken for each cancer to progress between stages. We will then apply the resulting stage transition estimates to incidence and stage-at-diagnosis data for each cancer to derive the number of cancers expected at localised / advanced stages at different periods of time, under alternative lengths of delays to cancer services.2) Modelling breast cancer screening. We will use a multistate model taking into account the natural history of breast cancer to derive the probability of a cancer being detected by screening or clinically with different periods of disruption to screening programmes. We will apply this probability to a decision analytic model that uses a life-table approach to understand the impact on cancer outcomes. We will analyse alternative catch-up screening strategies to identify that which best mitigates the disruption on breast cancer mortality and life-years gained.3) Estimating impacts of delays and demand for cancer services We will use the estimated number of cancers at different stages to analyse the impact of delays to diagnostic and treatment services on cancer outcomes and demand for services in England. Treatment parameters by stage will be obtained from PHE's Cancer Registry data. All other parameters for the models will be from aggregated anonymous sources, for example those released under an Open Government License, or peer-reviewed literature. The Data will not be transferred to any other location. We will first develop a continuous time Markov multistate model to describe the progression of cancer through the following states: healthy, localised cancer, advanced cancer, and dead1,2. Using these probabilities, and incidence of cancer by stage, we will estimate the expected number of additional advanced cancer diagnoses and the expected number of localised and advanced cancers at different periods of time. To analyse the impact of disruption to the breast cancer screening programme and alternative catch-up screening strategies that could be used when re-starting the programme, the following methods will be applied:1.We will use a multistate model of the natural history of breast cancer in the preclinical phase to derive the probabilities of detecting cancer by screening or clinically (i.e. interval cancers) following different time periods of disruptions to screening services. 2.Using a life-table approach3,4, accounting for the sojourn time of breast cancer by age, stage, and subtype, and using the derived probabilities for screen-detection and clinical diagnoses, we will model the impact of a suspension of screening services and of the backlog on interval cancer diagnoses and subsequent life years lost. We will consider different catch-up scenarios and identify the scenario that gives the fewest interval cancers and the least loss of life years. 3.We will consider delays in screening of between one month and one year and will liaise with PHE screening regarding alternative re-starting strategies that are under consideration. Finally, to analyse the impact on cancer outcomes and on predicted demand for cancer diagnostic, assessment and treatment services we will apply data on diagnostic and treatment modalities by cancer stage to estimate demand for services, and aggregate data on 1-year and 5-year survival to estimate impact on survival and mortality. The Data will be stored on servers at UCL. Our statistical analysis plan has been chosen as it encompasses robust methods with which the study team has experience for predicting medium and long-term cancer outcomes. There are two major caveats to the quality of the rapid cancer registration data for our analyses: missing variables, and data inaccuracies. For the purposes of our study, we have focussed on ten more common cancers, for which missing variables and data inaccuracies in the rapid cancer registration data are less of a problem than in rarer cancers or cancers of unknown primary. In both cases, earlier data are more accurate than the most recently available months, allowing us to take the inaccuracies into account in our modelling. Importantly, in our analyses our focus is on broad TNM stage as early/advanced for cancers as a whole (e.g. breast cancer, rather than breast cancer subdivided by hormone receptor) such that the impact of missing details, for example of cancer subtype, is limited. In addition, these data remain the best possible within the current context and we feel their use is justified given the role our results may be able to have in supporting policy and planning decisions as the pandemic continues to develop. The data will be accessible via secure VPN to the UCL Data Safe Haven, an ISO 27001 accredited secure environment for data storage and analysis that complies with NHS England’s Information Governance Toolkit. Access to the Data Safe Haven involves multifactor authentication and access controls, with lead investigators responsible for regularly ensuring only accredited relevant researchers have access to the data. The Data will be accessed by authorised personnel via remote access. UCL must confirm and provide evidence upon audit by NHS England that access via any remote device complies with the data security obligations within this DSA and the Data Sharing Framework Contract. For remote access: - Remote access will only be from secure locations situated within the territory of use (as further restricted elsewhere within the DSA if so done) stated within this DSA; - Access controls granting users the minimum level of access required are in place; - Remote access is only via secure connections (e.g., VPNs or secure protocols) to protect data; - Multifactor authentication (MFA) is required for remote access; - Device security, including up-to-date software and operating systems, antivirus software, and enabled firewalls are utilised for the remote access; - All remote access is undertaken within the scope of the organisation’s DSPT (or other security arrangements as per this DSA) and complies with the organisation’s remote access policy. The above applies in addition to any condition set out elsewhere within the DSA (e.g. who may carry out processing, and for what purpose). The Data will not leave England/Wales at any time. Access is restricted to employees or agents of UCL who have authorisation from the Principal Investigator. No other organisation is permitted to access the Data. All personnel accessing the Data have been appropriately trained in data protection and confidentiality. The Data will not be linked with any other data outside of this agreement. There will be no requirement and no attempt to reidentify individuals when using the Data. Analysts from UCL will process the Data for the purposes described above.

Expected output

The outputs of this project is modelling to show the impact of different lengths of delays on stage and outcomes at diagnosis. This is relevant for both post-COVID planning, and for cancer health policy more generally. The expected outputs of the processing will be: We are applying for an extension due to personnel changes that have slowed our progress. • Submissions to peer reviewed journals within the next 12-18 months. • Presentations at national and/or international cancer conferences. The outputs will not contain NHS England Data and will only contain aggregated information with small numbers suppressed as appropriate in line with the relevant disclosure rules for the dataset(s) from which the information was derived. The outputs will be communicated to relevant recipients through the following dissemination channels: • Journals with relevant media engagement • Social media • Poster/oral presentations at conferences The target outputs for production and dissemination of the outputs is 12-18 months.

Expected measurable benefits

As previous. The findings of this research study will inform policy related to the impact of delays in screening and treatment on cancer outcomes. This has direct relevance to clinical actions, such as waiting targets. By better understanding the impact of different lengths of delays on different cancers, UCL can improve cancer care and support the design of better management. It is hoped that through publication of findings in appropriate media, the findings of this research will support policymakers and clinicians to improve cancer care. Depending on the results of our analyses, UCL will optimise the potential benefits of the work through contacting relevant charities such as Cancer Research UK, whilst also taking advantage of their clinical and public health networks to maximise dissemination.

Benefits reported

Our UCL's modelling has shown interesting results on the impact of delays, but we they have not yet finished our their analyses. At present, this project is completing analyses into how different lengths of delay in managing different cancers can impact outcomes. This involves detailed modelling of how cancers grow over time. Based on this, UCL can then simulate the impact of delays of different lengths on demand for different types of services. This is because the stage at which a cancer is diagnosed impacts how to treat them and prognosis. UCL expect this to turn into valuable evidence that will support clinical and policy decision making.