Skip to main content

Publication, Part of

NHS Breast Screening Programme, England 2020-21

National statistics, Official statistics

National Statistics

Current Chapter

Quality Statement


Quality Statement

Introduction

This page is designed to accompany the main publication report, available on the publication homepage.

This page includes contextual information to aid understanding and presentation of the data including the methods used to compile the statistics and other background information readers may find useful.

Where appendices are referred to in this quality statement, they are available in a separate document, also published on the main homepage.

The main report presents information about the NHS Breast Screening Programme in England in 2020-21 as well as key statistics from the previous ten years1.

It includes statistics on women invited for breast screening, coverage, uptake of invitations, outcomes of screening and cancers detected.

The statistics in the main report are used to inform policy and to monitor the quality and effectiveness of screening services.

Please note: The NHS Breast Screening Programme statistics for 2020-21 were significantly impacted by disruption from Covid-19. See Appendix G for details. The data collection and validation processes were completed as normal, therefore, the quality and reliability of the statics published in this report are of a comparable standard to previous years.

 

Data sources

The statistics presented in this publication are derived from information that is routinely collected by NHS breast screening services for the operation of the screening programme, including for quality assurance and performance management purposes.

Information is collected on the following NHS Digital Korner Collection (KC) returns:

  • KC62 – Information on invitations, uptake and outcomes from all 78 breast screening units (BSUs) across England.
  • KC63 – Information on the population coverage of the programme from all 151 Upper Tier Local Authorities (LAs).

Data have been provided annually since 1988-89 through the KC62 and since 1994-95 through the KC63.

The data from the KC62 and KC63 are received in aggregate form each year by NHS Digital.

The NHS Data Model and Dictionary Service contain more information on the KC62 and KC63 central return data sets including guidance on content, completion and definitions. Links to the returns are given below:

KC62

http://www.datadictionary.nhs.uk/data_dictionary/messages/centr al_return_data_sets/data_sets/nhs_breast_screening_programm e_central_return_data_set_(kc62)_fr.asp?shownav=1

KC63

http://www.datadictionary.nhs.uk/data_dictionary/messages/centr al_return_data_sets/data_sets/nhs_breast_screening_programm e_central_return_data_set_(kc63)_fr.asp?shownav=1

Further information on the underlying sources of information is available in NHS Digital’s list of administrative sources.

 

Changes were made to the KC62 and KC63 central return data sets (effective from 1 October 2013) to take account of recent changes in policy that have been implemented within the programme.

The policy changes relate to the extension of the screening age range and the screening of women who are at higher risk (see Appendix A in the separate Appendices document). The KC63 return was also amended to enable the collection of data at LA level following changes in the NHS structure (see section on the ‘Impact of NHS Reorganisation’ for more information).

The data from each KC return are collected at the end of each financial year via the Breast Screening Information System.

The KC62 data comes to NHS Digital via a bespoke NHS Digital data collection system to which all Breast Screening Units (BSUs) submit. At source validations are carried out before submissions can be finalised.

The KC63 data comes from the BS select system from which aggregate LA level reports are produced.

Following the national breast screening incident some of the age parameters utilised within the programme were hard coded. This inevitably has had an impact on the age profile of the programme.

The Screening Quality Assurance Services (SQASs) are responsible for quality assuring the screening programme including the KC62 and KC63 data sets before final submission.

Further validation and quality assurance checks are carried out at NHS Digital as part of the publication process.

Regional quality assurance (QA) Managers at the SQASs are asked to check some of the tables produced for publication by NHS Digital as part of the validation process.

For more information on this process, please see Annex A which is a copy of a briefing paper compiled by PHE detailing the collection and assurance of breast screening data.

 

Methods used to compile the Statistics

NHS Digital validates and analyses the KC62 and KC63 data using automated processes developed in SQL3 and SAS4 as well as spreadsheets (Microsoft Excel).

Most of the figures presented in the report and data tables are in the form of simple counts, percentages (rounded to one decimal place) or rates (e.g. number of women with cancer detected per 1,000 women screened).

Due to rounding, the sum of percentages in some tables will not equal 100%.

Definitions and formulae detailing how the statistics used in the report are calculated are given in Appendix B.

Relevance

Appendix E gives details of who uses the statistics in this publication and what they use them for.

 


Accuracy and reliability

The KC62 and KC63 are established collections based on complete data, i.e. not a sample.

All validation queries that were raised through NHS Digital’s data validation processes were resolved satisfactorily.

Data quality issues

The number of women never screened in the 75 and above age group could not be accurately calculated in 2018-19. The number never screened for women outside the routine screening age range is calculated using data from two sources, NHAIS and Breast Screening (BS) Select. Differences in the way these two systems process data for women aged 75 and over meant an accurate reconciliation could not be completed.

KC63 data for the 75 and over age group is therefore no longer included in this publication series.

Calculations based on small numbers

Within the data tables there are rates per 1,000 women screened presented, some of which are based on small numbers and the level at which the rates are presented (BSU) can vary enormously in size, with the smallest unit screening just over 3,000 women aged 50-<71 in 2020-21 and the largest over 33,000 (see data table 14). Rates for smaller units, are therefore often based on a very small number of cases. This is especially true for prevalent5 screening, which accounts for only 12.9% of women aged 50-<71 screened.

The figures in this report represent all women screened in the reporting period, but uncertainty may still arise as a result of natural or random variation, with statistics involving small numbers being most susceptible.

Where statistics in data tables 14 and 15 have been highlighted with an asterisk, caution should be exercised when using them for comparative purposes, either from year to year, or across reporting units.

The asterisks have been applied when the number of occurrences used to make the calculation is less than 25 (e.g. where the number of small cancers used in the calculation of a small cancer detection rate is less than 25).

Occurrences of less than 25 approximate to a Relative Standard Error (RSE) value of more than 20%6.

The RSE is the standard error expressed as a percentage of the measure itself and is used to identify the level of reliability in statistics when the potential for random variation is taken into account.

The higher the RSE, the less confidence there is in the reliability of the statistic.

 

False positive and false negative results

Screening tests are not 100% accurate and in any screening programme there may be some false positive results and some false negative results.

False positives

Some people with a positive screening test result do not actually have the condition being screened for and are said to have a ‘false positive’ result.

In breast screening, false positives refer to cases where the mammogram appears abnormal but after further investigation no cancer is detected. All women whose mammograms appear abnormal are offered further tests until a definitive diagnosis is reached. These diagnostic tests are a routine part of the screening programme.

In 2020-21:

  • 4.0% (47,198) of screened women aged 45 and over had an abnormal result and were therefore referred for assessment (see data table 7a).
  • 77.1% (36,358) of these women were found not to have breast cancer (see data table 8).
  • Based on this, the percentage of all women screened who returned a false positive result was 3.1%7.

False negatives

Some people with a negative screening test result do actually have the condition being screened for. These people are said to have a ‘false negative’ result.

In breast screening, false negatives are identified when a woman is diagnosed with cancer and a review of her previous screening results shows that an abnormality was present which, in retrospect, may have been the cancer.

Cancers can be diagnosed at any time following a screening episode with a negative outcome. Those diagnosed between scheduled screening episodes are known as interval cancers. Interval cancers can be divided into:

  • those that were false negative
  • those that are new (i.e. those that have developed since the last screening episode)
  • those that are not visible on mammography.

There is no generally accepted or expected level of false negatives in the programme, although there are targets to minimise interval cancer rates which will include false negative cancers. Such measurements require long term follow up and are beyond the scope of this bulletin.

Further information on interval cancers is available.

The Independent Breast Screening Review8 published in October 2012 contains more information on false positives and false negatives.

 

Timeliness and punctuality

The breast screening data are made available annually as soon as possible after they have been compiled and validated.

The time delay in publishing the statistics is because the data returns are produced six months after the year end (sufficient time to allow most screening episodes to be completed and outcomes to be recorded).

The statistics published in this report reflect data submitted as of 10th November 2021. At the time of publication (24 February 2022), no amendments to these data had been received.


Accessibility and clarity

The data fields are published in the Data Tables which are available as Excel and CSV files from the publication webpage.

An interactive data dashboard is provided as part of the data resources for this publication. The dashboard has been developed in software called Microsoft Power BI and is designed to make data more meaningful by allowing local, regional and national comparisons over time. 

This includes coverage statistics for women aged 53 to <71 years presented by LA, and uptake statistics for women aged 50 to <71 years presented by BSU.

Coherence and comparability

NHS Digital maintains awareness of changes that may impact on the data through regular communication with NHS Cancer Screening Programmes and the Department of Health and Social Care (DHSC).

Time series

For key statistics, the report presents a 10-year time series where possible. For other statistics, figures for the current year are compared with the previous year, or occasionally with figures from 5 years ago.

The changes in policy, described in Appendix A, need to be taken into account when considering trend data.

Local and regional comparisons

The statistics are presented at a national, regional and local level.

Local level statistics are presented by Upper Tier Local Authority (LA), region (see ‘Impact of NHS Re-organisation’ section) and Breast Screening Unit (BSU).

At a regional level, LA (KC63) data are aggregated up to eight regions with a sub-regional breakdown for the South (showing the South East and South West).

Data from BSUs (KC62) are aggregated to the PHE’s eight reporting regions with sub-regional breakdowns for North East, Yorkshire and the Humber (showing North East and Yorkshire and the Humber) and the South (showing the South East and South West).

Occasionally, BSUs and LAs will move between reporting regions, these are reported under the relevant data table(s).

LAs are assigned to regions based on region of responsibility within the programme. Therefore, although these generally align with the standard LA regions, there are some differences, and this should be considered when comparing regional level LA data with other datasets.

A summary of the changes is shown in the table below.

 

 


Age-extension outside of the research trial

Nine BSUs which screened all women aged 47 to 49 years stopped this practice over the period July-December 2016.

They were screening these women because for operational reasons they could not participate in the age extension trial.

As PHE advises on evidence-based practice for operating screening programmes, screening this cohort of women was ceased as the age extension trial is a randomised control trial (RCT) established to answer whether screening this age group of women is effective.

Following this decision, there was a large fall in the number of women aged 45-49 in these areas who were invited and screened.

The impact of this change can be seen in both the regional and national activity figures in 2019-20.

The research trial stopped recruitment in March 2020. 

 

 

Round length impact on service activity

Screening round length is the interval between the date of a woman’s previous screening mammogram and the date of her next first offered appointment. Screening services are required to invite their eligible screening population within 36 months (three years) of a previous screen.

Due to a variety of factors, including workforce shortages, equipment breakdown and GP re-organisations, around 20 to 25% of services fail to invite 90% of their population within the 36 month target. Where women are not invited during this period there will be an adverse impact on the coverage rate; the rate is based on the eligible population which, by definition, would include these women. It should be noted these women will still receive an invite, however, it will be received outside of the 36 month target.

These factors may cause fluctuations in year on year changes in numbers invited and screened.

From April 2013, LAs took on new responsibilities for public health.

The statistics in this report are therefore presented by LA. LA data was published in this report for the first time in 2012-13 as experimental statistics.

LA coverage statistics for previous years are published as part of the Public Health Outcomes Framework (PHOF) and are available at: https://fingertips.phe.org.uk/profile/public-health-outcomesframework

Although the data source is the same, LA PHOF figures differ slightly to LA figures from the KC63, with slightly more women both eligible and screened identified in the KC63 than in the PHOF.

Coverage at a national level is not affected and only a very small number of LAs show a difference between the two datasets.

The two datasets are run at different points, the PHOF figures being automatically extracted and the KC63 being submitted following QA checks by the BS Select team and PHE. There are a number of possible reasons for the differences in counts of eligible population.

The two different run dates could allow for remedial actions taken by screening manager or more common actions, such as delayed registrations/deductions, amendments to registration/deduction dates, changes to registered postcodes, corrections to date of birth and corrections to registered gender.

Updates to LA mapping files might also assign women to LAs that were previously ‘unknown’.

See Appendix G for a summary of how the programme was affected by disruption from Covid-19 in 2020-21

 

Comparisons with other countries

Some of the statistics in this report can be compared with other UK countries - see sections on coverage, uptake and cancers detected in the main report for more information.


Performance cost and respondent burden

The publication is based on information that has been routinely collected by the programme as part of the performance management of the breast screening organisations.

All data collections used in this publication are subject to assurance by the Data Standards Assurance Service (DSAS) on behalf of the Data Coordination Board (DCB) (previously known as the Burden Advice and Assessment Service (BAAS) procedure).

This is to ensure that data collections do not duplicate other collections, minimise the cost to all parties and have a specific use for the data collected.

Information on all DCB standards and collections are available.

 

Confidentiality, transparency and security

The standard NHS Digital security and confidentiality policies have been applied in the production of these statistics.

An annual risk assessment is undertaken prior to publication which addresses any potential issues around disclosure.

No disclosure issues were identified in relation to this publication and no disclosure controls have been applied.

The eligible populations in two LAs are relatively small and in these instances their data have been combined and reported under other LAs.

Data for Isles of Scilly are reported under Cornwall and City of London are reported under Hackney.

Statistics in this report are therefore presented by 149 LAs, two of which include another small LA.

 

Data revisions

Where any data are re-submitted post-publication, NHS Digital will assess whether the resubmitted data has a significant impact on England-level data.

Where this is the case, the affected Excel tables will be reissued.

Any other related documents, such as interactive reports, will also be updated.

Where the impact to England level data is not significant, footnotes will be made to the affected Excel tables but not to the PDF report.

 

Exclusions

Data from the high-risk screening programme has been published in this bulletin from 2013-14 to 2015-16.

Their status had been designated as experimental statistics due to the lack of nationwide implementation of the programme and limitations of the computer system to provide complete statistics.

Data for 2016-17 is not available because the system was being developed during this time period.

Since 2017-18 this data was collected using a new and more comprehensive collection system.

Their status continues to be designated as experimental statistics whilst the quality of the data from the new collection system is being assessed.

KC63 data for ages 75 and above was not available so is not included in this report.

 

 

In 2020-21, the NHS Breast Screening Programme was significantly affected by disruption from Covid-19, however, the data collection and validation processes were completed as normal. See Appendix G for details.

Footnotes

1. Since 2004-05 this bulletin has been published by NHS Digital. Previous editions published by the Department of Health, are available. 

2. NHAIS

3. Structured Query Language (SQL) is a programming language designed for managing data in relational database management systems.

4. Statistical Analysis System (SAS) is an integrated system of software products which enables functions such as data management, statistical analysis and quality improvement

5. Prevalent screening refers to women being screened for the first time within the breast screening programme. In this statistical bulletin, prevalent screening figures relate to first invitations for routine screening and routine invitations to previous non-attendees.

6. An RSE of 20% or more is often used by the Office for National Statistics (ONS) to advise where figures should be treated with caution.

7. Calculated from the KC62 data set as: (Total number of women referred for assessment – Total number where outcome of assessment is cancer) / Total number of women screened * 100

8. The Independent UK Panel on Breast Cancer Screening (2012), “The Benefits and Harms of Breast Cancer Screening: An Independent Review ”


Last edited: 24 February 2022 9:27 am