Examining the healthcare inequalities in breast cancer screening during COVID-19

COVID-19 has had an unprecedented impact upon breast cancer screening services. In addition to the backlog of almost 1 million missed mammograms since March 2020, public health measures have impacted on the functioning of screening hubs. To overcome these stresses, breast screening services have moved from inviting women to a pre-specified mammogram (timed invitations) to an invitation to book an appointment (open invitations), in order to increase the efficiency of screening hubs.

Existing studies have shown that women from Non-White backgrounds are up to 63% less likely to attend screening mammograms compared to their White British counterparts. In addition, women from more deprived areas and those with more medical co-morbidities status are less likely to attend breast screening invitations. These studies have all been conducted with closed, as opposed to, the novel open invitation structure. There are, however, concerns that this open model may exacerbate these existing inequalities by introducing further logistical barriers. For example, studies have highlighted that people from ethnic minorities report significant practical barriers to booking appointments.

NHS England's service evaluation has further highlighted these concerns, with the existing uptake of invitations in London almost 10% lower than in 2019. Moreover, they have shown a negative correlation between uptake of open invitations and Index of Multiple Deprivation (IMD).

Currently, screening services in London are utilising both open and closed invitations, and therefore, the true effect of the new invitations on individual patients is unclear. Furthermore, the effect upon ethnic minorities, those with clinical comorbidities or those who were identified as clinically extremely vulnerable is not known, as this is not collected by screening hubs. Understanding the impact on these groups is essential, as they already constitute low uptake populations and are at risk of this inequality widening.

Understanding this impact is integral to allow services to:

(1) target resources on potentially low attendance groups,

(2) amend invitations to meet the needs of the local populations

(3) maintain pre-COVID breast cancer screening uptake, and

(4) aim to maintain screening levels sufficient for the screening programme as a whole to be of benefit.

This work is of significant public health concern and is led by NHS England. The analysis constitutes a service evaluation of the existing invitation framework introduced as a response to COVID-19 in London. The primary aim of which is to ascertain whether the new open invitation type exacerbates healthcare inequalities and affects screening uptake. No existing data on this work has been published, given the recent introduction of such measures. The results from this work will feed into NHS England. NHS England are commissioning this work to determine the impact of COVID upon screening inequalities within the breast screening programme. The NHSE screening lead is leading this evaluation (and this data request), as this will inform the future COVID recovery of the breast screening programme in London specifically:

1) the impact of invitation types on inequalities

2) whether to change policy to improve attendance (which remains low)

3) inform national practice.

This service evaluation is designed to determine the impact upon healthcare inequalities of a new invitation type to breast cancer screening. In order to minimise data requested only pertinent variables such as co-morbidities shown to have an impact from the literature, ethnicity, Index of Multiple Deprivation and COVID diagnosis flags/vaccine indicators have been selected. These variables have been shown to impact on either 1) patient-level attendance or 2) screening service-level resources.

The Data Controller (NHS England) confirms that the purpose of the work outlined in this agreement is for service evaluation and not research. The primary project aim is to determine the impact of open invitations on attendance to the NHS Breast Screening Programme in London during the COVID-19 recovery phase. This will be achieved through the completion of three project objectives:

1) which factors predict overall attendance at Breast cancer screening appointments during the secondary phases of the COVID-19 pandemic,

2) do the factors in (1) differ between open and timed invitation-types,

3) do the populations that do not attend each invitation type differ significantly, to potentially represent a significant healthcare inequality.

Secondary aims will involve examining the spatial accessibility of screening services/hubs in London during the pandemic. This will examine how the effect of changing loco-regional COVID19 positive cases, the type of screening service (mobile versus fixed) and hospital activity, impacts

open the perceived accessibility of individual screening services in London.

NHS England will provide a cohort to NHS Digital to include patients in London asked to screen between August 2020 and July 2021. Imperial College London estimate approximately 1.2 million patients will be invited to screen over this period. The recovery of the health and screening service following its emergency footing where several emergency measures have been taken, continues with screening services in England. Therefore, Imperial College London will be undertaking a further extraction in Autumn 2021 (approx. September/October 2021) to see how dynamic environmental factors (lockdown/Tier* responses) also impact upon the drivers to uptake the new versus old invitation to breast screening, providing legislation permits.

*Whilst the Tiering system is not so relevant now, the environmental factors such as local rates of COVID and impact on local health systems still are relevant. The screening service is predominantly undertaken within fixed centres based within hospitals, and the impact of local rates (freely available to the public) is an important co-variate in understanding attendance.

To evaluate the ongoing impact of Covid-19 upon breast screening and the inequalities caused by open invitations the study team hope to undertake a further data drops for this cohort using GPES data for Pandemic Planning and Research (GDPPR) in the event that COPI Legislation is extended beyond end of March 2022.

In this agreement, NHS England are the sole Data Controller and Imperial College London will be undertaking the data processing. Funding for this service evaluation is from NHS England. Agreements between Imperial College London and NHS England are in place for this activity. No other NHS (including NHS Improvement) or academic institution is involved in this evaluation work and therefore this agreement.

DATA MINIMISATION

Imperial College London intend to use a data set formed from the linkage of a cohort provided by NHS England to NHS Digital's GPES data for Pandemic Planning and Research (GDPPR). No additional data sets are necessary and therefore this is the minimum data required.

The existing NHS England cohort incorporates NHS Numbers and as the screening service does not collect co-morbidity or ethnicity data directly, a linkage is required. As the cohort Imperial College London intend to be linked with GDPPR is a breast cancer screening population in London - the area of evaluation for NHS England - the data will be filtered by:

- geography (London),

- age (50 to 70 years) and

- gender (female).

The variables of interest include co-morbidities, however these may be provided as cluster-level diagnoses to minimise the data provided. The fields requested have been minimised to provide the information required. Of particular interest to this evaluation are the COVID-codes and co-morbidity status of individuals which may impact upon their healthcare engagement during the COVID recovery[a] period.

[a] The 'COVID-19 recovery phrase' refers to the recovery of the health service following its emergency footing, where several emergency measures were taken in the interest of public safety. These include the halting of screening programmes and changes to invitation systems.

The fields requested in the GDPPR data set have been minimised for only information required. Of particular interest are the COVID-codes and co-morbidity status of individuals as determined by fields within the Charlson Co-Morbidity Index (heart failure, myocardial infarction, Chronic obstructive pulmonary disease (COPD), dementia, peripheral artery disease, kidney disease 3 to 5, liver disease, haematological cancer, solid cancer and immuno-suppressive states). The Charlson Co-Morbidity Index is a validated measure of co-morbidity burden, and is prognostic to survival. Data on healthcare encounters are not required, as this is determined by the individual's episode with the breast cancer screening service, which will be included in the cohort provided. The study team only require diagnoses codes, COVID-codes, IMD and ethnicity correct at the time of the screening appointment date.

The study team do not require data on healthcare encounters, as this is determined by the individual's episode with the breast cancer screening service, which will be included in the cohort provided. They therefore only require diagnoses codes, COVID-codes, IMD and ethnicity correct at the time of the screening appointment date.

The study team requires an extraction based on one cohort for the period August 2020 to March 2021 which includes the current cohort invited by the new open invitation type, and a comparative group from the year previously, and then a further cohort submission in August/September 2021 to review the the ongoing recovery response to the pandemic. The cohort will be submitted by NHS England. Two drops of a record-level pseudonymised GDDPR dataset will be sent to the Big Data Analytical Unit (BDAU) at Imperial College London, who will undertake evaluation on behalf of NHS England.

LAWFUL BASIS FOR PROCESSING

The lawful basis for processing data under GDPR has been reviewed and been assessed as acceptable. NHS England, as the Data Controller, process data under Article 6(1)(e): "processing is necessary for the performance of a task in the public interest or in the exercise of official authority vested in the controller" as they are a Public Authority. The Data Protection Act 2018 s7(1)(a) defines ‘public bodies’ for the purpose of the GDPR as “a public authority as defined by the Freedom of Information Act 2000”.

The Data Protection Act 2018 defines 'public authority' as that defined under the Freedom of Information (FOI) Act 2000. The FOI Act was amended by the Health and Social Care Act 2012 schedule 5 paragraph 99(b) to include the NHS Commissioning Board (also known as NHS England) as a named public authority.

The Health and Social Care Act 2012 section 23 13E(1) states that the NHS Commissioning Board 'must exercise its functions with a view to securing continuous improvement in the quality of services provided to individuals for or in connection with (a) the prevention, diagnosis or treatment of illness, or (b) the protection or improvement of public health.

Additionally, NHS England process Special Category Health Data under 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" as the data are required for statistical purposes in the public interest - meeting the conditions in the DPA 2018 Schedule 1 Part 1 (4) - which GDPR Recital 52(2) determines is an appropriate derogation from the prohibition on processing special categories of personal data.

Use of Control of Patient Information (COPI) Regulation - Identifiable information flowing to NHS Digital:

Regulation 3 (3)(b) confirms who is able to process patient information which includes: ‘persons employed or engaged for the purposes of the health service.' The Data Controller for this work is NHS England. NHS England have commissioned Imperial College London (ICL) to deliver this piece of work and ICL are therefore the Data Processors in this agreement. NHS England would be considered as falling under the regulation 3(3).

METHODOLOGY

- NHS England provide a filtered cohort of 1,200,000 individuals from the London area via a Secure Electronic File Transfer account (SEFT). The cohort will contain the following identifiers:

1. Pseudo-Study ID (a cryptic identifier created in the Breast Screening service)

2. NHS Number

3. Date of Birth

- NHS Digital will link this cohort of individuals to the GDPPR data set, and extract a data set using the fields requested. NHS Digital will then remove all identifiers from the extract (except the study ID which is a key placed on encounters at the screening service which cannot be used to re-identify the individual outside the hub)

- NHS Digital will then send the resultant pseudonymised record level data extract via SEFT to Imperial College London’s Big Data Analytical Unit (BDAU) for their analysis.

Two drops of data:

Drop 1 - Upon approval of agreement, a cohort of 1,200,000 individuals will be submitted for the period August 2020 - March 2021 (dissemination in August 2021).

Drop 2 - Approx. August/September 2021, a new cohort (estimated numbers are unknown as this is a new invitation system, but smaller than the 1st cohort) will be submitted for the period April 2021 - end August 2021 (dissemination by approx. October 2021).

NHS England will provide a Study ID (a cryptic identifier created in the Breast Screening service placed on encounters at the screening service which cannot be used to re-identify the individual outside the hub).

The resultant pseudonymised dataset will be transferred to the secure data environment at Imperial College London for processing, called the Big Data Analytical Unit (BDAU). To analyse the data, the study team will be using statistics to examine the effect of the new invitation type, which was developed due to the restrictions of COVID-19, on whether individuals attend their appointment. To understand this effect properly and fairly, it is also important to account for other factors that are known to be associated with lower attendance, for example increased deprivation level. By accounting for all these factors statistically (called multivariate regression) the study team will be able to determine the true impact of the new-invitation type.

No attempt to re-identify individuals will be undertaken.

GDPPR data will be used for the specific purpose stated in this agreement. Due to the use of a cohort, it is not currently feasible to utilise the Trusted Research Environment (TRE), and therefore an extract from NHS Digital to Imperial College London is required.

Although the GDPPR data is pseudonymised, the data is treated as confidential patient data due to this being a restricted data set that is collected only for the purpose of COVID-19 management.

Data processing of pseudonymised data will be undertaken by Imperial College's Big Data and Analytical Unit Secure Environment (BDAU SE), which will act as the secure storage and processing location. Data access is strictly controlled by the Imperial College's Big Data and Analytical Unit (BDAU) with stringent procedures including dataset registration process, limiting access to the data to the minimum numbers of the research team required and BDAU approved staff access facilities. Data analysis will only be undertaken through the BDAU. The data will only be used for the purpose outlined in this Data Sharing Agreement. Imperial College London staff are bound by all policies and regulations as substantive employees of the College.

All data will be stored and analysed within the Imperial College's Big Data and Analytical Unit Secure Environment (BDAU SE). The BDAU SE is located in Imperial College London's Data Centre.

Data Processing is only carried out by substantive employees of the data processor (ICL) and who have been appropriately trained in data protection and confidentiality. All personnel will abide by Imperial College London's as well as national data privacy regulation. Imperial’s Big Data and Analytical Unit Secure Environment (BDAU SE) is a secure research environment, providing a standard operating/access model, secure data storage and processing environment and analysis software. It is ISO 27001 certified and also compliant with NHS Digital’s Data Security and Protection Toolkit. The BDAU SE is located in Imperial College London’s Data Centre and can be accessed remotely by users using multi-factor authentication once the user registration process is completed. Data will only be provided once the appropriate dataset registration process is completed. All data files and directories will be encrypted using advances encryptions standards (Encryption is the process where data is encoded for privacy and a key is needed by the data owner to access the encoded data).

Virtus SDC Ltd are not considered a Data Processor. Virtus are a data centre co-location provider (providing data centre space), not a cloud services provider. Imperial College rents space in Virtus data centres to host BDAU equipment. All data processing are performed by BDAU’s staff. Virtus provide physical security, power and environmental controls. Virtus have a number of quality credentials including ISO27001, certificate copies of which can be accessed or downloaded on the following page: https://virtusdatacentres.com/why-virtus/quality-credentials.

GDPPR Disclosure Controls / Suppression Rules

Whilst there are no specific GDPPR disclosure controls, outputs for public consumption should follow the Government (ONS) Statistical Service disclosure controls.  NHS Digital require that users review and follow the disclosure control guidelines as set out within the HES Analysis Guide. Some, but not all requirements are outlined below:

Disclosure control only needs to be applied to values relating to individuals. No rounding or suppression is required for values not relating to individuals, such as a count of providers.

No small number suppression is required for national totals.

For any sub national geographies e.g. NHS Commissioning Region / Government Office Region or smaller, then the following apply:

• Zeroes can be shown.

• Values between 1-7 to be displayed as “*”.

• Any other numbers rounded to nearest 5.

• Percentages calculated from rounded values

The following outputs are aimed to be produced from this work:

1. A report from NHS England regarding the impact of open invitations on attendance and healthcare inequalities to guide health policy. The primary output aims to be a report made by NHS England regarding the impact of open invitations on attendance and healthcare inequalities to guide health policy. This report will primarily be made available for NHS Breast Screening Hubs in London, and across the country. It is hoped that this will enable these officials to alter their practice in direct response to the result found from this evaluation. The output report hopes to also be made available to the public who remain important stakeholders in screening services and wider health policy makers who will be able to utilise our findings to alter national practice and inform COVID-19 recovery efforts in other regions.

2. Publications in peer-reviewed scientific journals targeted to the British Medical Journal. Estimated in late 2021, these publications aim to detail the association of varying invitation type and socio-demographics with screening attendance.

3. Presentations in late 2021 which hope to be delivered at international conferences to clinicians, public health workers and experts in preventative care. In addition, Imperial College London intend to present this work directly to screening hubs and public stakeholders to inform practice and delivery of screening regionally. This will be achieved through collaboration with NHS England.

Imperial College London intend to also use non-traditional dissemination through social media and blog posts directed by both NHS England's and Imperial College London's communications department. This will allow the findings to reach relevant stakeholders including patients, clinicians and the broader interested audience.

The work is being conducted through the Patient Safety Translational Research Centre which is one of 3 centres in the UK, as the findings from this work are foreseen to translate into clinical practice. All processes will be compliant with Imperial College London's policy on data publication and dissemination.

All outputs will contain only aggregate anonymous summary data with small numbers suppressed in line with the HES Analysis guide.

No raw data will be transferred outside the BDAU SE and neither the data nor the outputs will be used for insurance, marketing or commercial purposes.

Open invitations have been introduced as part of the COVID-19 recovery programme, aimed at reducing the backlog of breast cancer screening appointments. The unintended consequences of introducing these measures upon attendance during the COVID era, and populations who traditionally have low-uptake of screening, is unknown.

This evaluation, led by NHS England, aims at determining this impact with a view to re-evaluate the current screening service practice in London. This is of prime public health concern, and disseminating the findings crucial, to prevent potentially exacerbating healthcare inequalities by continuing to employ current measures in screening services. In London, approximately 268,000 women were screened in 2018/9, however the potential magnitude is greater given the open invitation model, has now been employed across services in England. The need to disseminate the findings to other regions, screening hubs and public health officials is important to ensure they can re-evaluate local policy, as appropriate. Failure to disseminate the findings could result in lower screening uptake or widening health disparities and outcomes amongst subsets of the population.

The findings from this project are hoped to:

1) inform loco-regional policies regarding the use of open invitations during the COVID recovery

2) provide evidence to revert to timed appointments, if appropriate, and

3) allow screening services to tailor resources to increase attendance amongst low uptake groups including those with COVID diagnoses or shielding.

The analysis hopes to achieve the purpose by determining the effect of invitation type (open versus timed) on attendance. Co-variates known to affect attendance such as ethnicity, socioeconomic status and co-morbid status will be adjusted for in this model. Further, this should give an indication on whether the invitation type disproportionately impacts upon these subgroups. Following this analysis, NHS England aim to implement policy based upon these findings, the outcome of which hope to be be determined through screening hub metrics (e.g. uptake and coverage rates).

The work will also inform a PhD student who is substantively employed at Imperial College London, investigating the use of data analytics on the breast screening programme in London which remains in COVID-19 recovery phase. The results from this hope to be used to inform current practice in screening hubs and help guide the national policy on COVID-19 recovery in screening. This is particularly important as the prospect of a third wave approaches and screening services need to adequately prepare. None of the results or methods will be used for/published in any PhD or research work. The involvement of a PhD student is only to facilitate data access and provide infrastructural support as a member of the study team and share their experience and does not affect the purpose of this work as an evaluation.

Yielded Benefits is not a requirement for new applications.