Dear Ed,
We have recently discovered data quality issues that impact two accredited official statistics releases. I therefore request a temporary suspension of their accredited official statistics status while we investigate and resolve these issues.
Breast cancer screening
I previously wrote to you about data quality issues with cervical cancer screening data provided by an external contractor. Our investigations have since revealed similar problems in the breast cancer screening data for Scotland. These issues relate solely to data provision; the delivery of both screening programmes continues as normal.
When we were preparing the publication for the 23/24 screening programme statistics, our quality assurance indicated that key measures such as uptake and coverage were inaccurately overestimated. Our investigations identified errors in the way the external data provider was classifying historic screening programme attendance. We also found an error with the way that the data provider has been coding attendance status at certain clinics which was recently exacerbated due to changes in the technology used at the point of data capture. These inaccuracies mean that these data cannot meet the quality standards that our users expect and the Code of Practice for Statistics requires. Our investigations suggest these issues affect all statistics published from 2015/16 onwards but with a greater impact evident from 2021 onwards when the additional technology changes occurred.
We have notified users that the planned February 2026 publication will be delayed until these issues are resolved and added notices to earlier releases outlining the data quality concerns. Given the public importance of cancer screening programme performance measures, we are also exploring whether alternative interim data sources might provide users with insights until a longer-term solution can be established, for example the data held by the six screening centres for financial reporting purposes. As a longer-term solution, we are also exploring options with the external data provider to source individual-level patient data that would enable PHS to apply the correct definitions at source and not rely on aggregate returns, but we currently have no indication of how long this might take.
If we can source suitable alternative data on either an interim or longer-term basis, I think it would be appropriate for these to be published as statistics in development until we are content they meet the requirements for official statistics. This will help signal to users that these statistics have undergone a major change in their data source and might not necessarily provide the same information that these statistics have done in the past. We can then work with users as we develop the statistics further to ensure they answer the key questions they have about the breast cancer screening programme.
The IT systems for the breast and cervical screening programmes are nationally managed by NHS National Services Scotland (NSS). In advance of any future procurement exercise for new systems, we are now working with NSS to ensure patient-level data provisioning and quality assurance requirements are clear throughout systems’ lifecycles, aligned to the strategy for renewal of services.
Scottish Health Services Costs (the “Costs Book”)
The Costs Book reports annual expenditure on hospital and primary care by NHSScotland’s 14 territorial boards, the State Hospital and the National Waiting Times Centre at the Golden Jubilee University National Hospital. Its uses include informing Scottish Government’s funding allocation decisions. The Office for National Statistics also use the data in their UK Health Accounts publication.
Issues have arisen impacting both the timeliness and quality of the statistics due to changes to the way that expenditure data is captured locally. Historically, Scotland provided aggregate level costing data whereas the rest of the UK uses the Patient Level Information and Costing System (PLICS) to deliver more granular and insightful costs data. To improve the quality and value of the Scottish data, Boards began to implement PLICS in 2025/26. Five Boards also agreed to apply the new approach to derive their 2024/25 expenditure for the Cost Book, with the remaining Boards continuing to use their existing methods (and transition to using PLICS for future Cost Book submissions). The data quality is therefore impacted by this variation in data sources across the Boards because the 2024/25 data will not be comparable with previous years and the ability to compare costs across Boards will be limited until they are all using the same underlying methodology.
The timeliness has also been impacted. PLICS implementation has taken longer than anticipated and Boards have not been able to meet the submission deadlines originally agreed to align with the planned publication date in March 2026. We have assessed various options to address this, taking account of the extent of data currently missing. In our judgement, the best option is to delay publication until all Boards have submitted data, rather than produce an interim set of estimates with many elements missing. We are working closely with Scottish Government and ONS as key users of this data to ensure that any decisions taken about future publication timings and approaches take account of the potential impact on their work. We will provide a publication date and highlight with users as soon as we have more certainty about the data submissions. The publication will include clear information about the different sources of data used to produce the statistics and any limitations this places on their interpretation.
Next steps
The data underlying both sets of statistics requires significant development to meet the quality requirements of the Code of Practice and enable timely reporting. In my judgement the required temporary changes of their accreditation status to become “official statistics in development” should remain until we are confident we can provide reliable, coherent and timely statistics that meet users’ needs. Compliance reviews of both releases would be welcome in the future, once their development work is complete. We will ensure users and OSR are updated regularly about this development work, the timings of future releases, and are involved in any key decisions.
I am copying this letter to Alistair McAlpine, Scottish Government Chief Statistician; Anita Morrison and Nicola Edge, Heads of Health and Social Care Analysis, Scottish Government; Tasmin Sommerfield, National Clinical Advisor/Deputy Director of Screening, NSS; Alexander Cruikshank, Team leader National Screening Programmes, Scottish Government.
Yours sincerely,
Scott Heald
Director, Data and Digital Innovation
Head of Profession for Statistics