Present 

Members 

Helen Boaden (Chair)

Stephen Balchin

Vanessa Cuthill

Colin Godbold

Monica Magadi

Isabel Nisbet

Advisors

Ross Young (Data Protection Officer, UKSA)

Secretariat

Pamela Calderon Ambrossen

Natasha Kong

Matt Short

Simon Whitworth (also presented for items 2 and 5)

In attendance

Rhys Nadin, Data Governance, Legislation & Policy, UKSA (for item 2)

Emma Rourke, Health, Population and Methods, ONS (for item 3)

Will Laffan, Central Policy Secretariat, UKSA (for item 3)

Louisa Blackwell, Population Statistics, ONS (for item 3)

Harriet Beach, Social Statistics Transformation Policy, UKSA (for item 3)

Michael Cole, Population Statistics, ONS (for items 3 and 4)

Alex Mylles, Population Statistics, ONS (for item 4)

Apologies

Rob Bumpstead

Emma Uprichard

1. Minute and matters arising from the previous meeting and correspondence and update to terms of reference NSDEC(23)09

1. The meeting began with introductions for the benefit of the new chair, Helen Boaden.
2.  Colin Godbold declared his interest arising from his role as Chair of the Health Data Research UK Public Advisory Board.
3. Rob Bumpstead and Emma Uprichard gave their apologies.
4. Members of the Committee had approved the minutes from the last meeting in the interim period via correspondence.
5. Matt Short updated the Committee with progress on actions from the January 2023 meeting. All actions were completed or in progress.
6.  Matt Short presented on an update on the Committee’s terms of reference which were amended to reflect the NSDEC’s current working practices. These changes included:
   • An addition to the specific responsibilities of the NSDEC to include advising the National Statistician on the use of NSDEC’s ethical principles by researchers and statisticians
   • A change in wording around the responsibility of the NSDEC to “promote transparency around the use of public data”, which has been changed from “promote transparency around data shares”.
7. The Committee agreed to the changes made to the terms of reference, and made the following comment:
   • The NSDEC noted that UKSA’s ethical principles must remain robust and flexible in a fast-paced landscape, and recommended a review of these ethical principles.
8. Action – The Secretariat to reflect the approved changes in the terms of reference and update the UK Statistics Authority website.
9. Action – The Secretariat to maintain a review of the NSDEC’s ethical principles and present to the NSDEC an updated landscape review of ethical principles in government.

2. Review of ONS data ethics policy and audit review NSDEC(23)10

1. Outcome – the Committee supported this work subject to assurances to points outlined below in section 2.5.
2. Detail of the item and discussion:
3. Simon Whitworth, Head of Data Ethics and Research Accreditation and Rhys Nadin, Data Protection Compliance Manager from the Data Governance, Legislation and Policy team in the UK Statistics Authority (UKSA) presented this item.
4. The paper first outlined the Office for National Statistics (ONS) data ethics policy. It then detailed the audit process for research projects that applied to the NSDEC or utilised the Centre of Applied Data Ethics (CADE) self-assessment tool, which ensures that these projects adhere to the UKSA’s ethical principles. It also detailed various programmes of work of the audit function for the NSDEC to review, which included:
   • The potential escalation procedure for non-compliant research projects
   • The high-level Data Ethics Compliance Review Plan for 2023/24
   • The criteria for selection and plans for the future of Data Ethics Compliance Reviews for 2023/24
5. The NSDEC approved the future programmes of work for the audit function detailed in the paper. These following points were also raised by the NSDEC within the discussion:
   • Recognised that the ONS shows commitment to adhere to ethical principles as evidenced by the increase in the number of projects that apply for NSDEC consideration and/or utilise the CADE self-assessment tool.
   • Asked why there was no central register of ONS research and statistics projects as this may help in revealing the extent of usage of the ethics self-assessment.
   • Asked why the CADE self-assessment tool was not a compulsory part of ONS projects.
   • Recommended that training for the wider ONS to ensure projects adhere to UKSA ethical principles would be useful, especially as it supports the National Data Strategy.
   • Advised that the criteria for projects that are selected for data ethics compliance reviews should also prioritise (1) projects where the research team are new to applications to the NSDEC and utilising the CADE self-assessment tool, (2) projects of a strategic nature for the ONS such as the Integrated Data Service, and (3) multi-agency projects.
   • Emphasised the importance of the Secretariat in developing and communicating a narrative on the importance of ethical consideration and compliance for the ONS.
6. Action – The Secretariat to invite the National Statistician, Sir Ian Diamond to attend a future meeting of NSDEC.
7. Action – The Secretariat to approach senior IDS leadership to present on how the IDS complies with UKSA ethical principles to the NSDEC.
8. Action – The Secretariat to introduce training regarding data ethics compliance in projects to the wider ONS.
9. Action – The Secretariat to develop a narrative or case studies to demonstrate the importance of adhering to data ethics principles and communicate this to the wider ONS.
10. Action – The Data Governance, Legislation and Policy team to undertake the approved future plans of the audit function, subject to the adjustment to the criteria for projects that are selected for data ethics compliance review, as outlined in point 2.5.

3. Update on engagement plans for 2023 recommendation on the future of population and social statistics NSDEC(23)11

1.  Outcome – the Committee supported this work and asked to be kept updated on developments regarding the 2023 Recommendation and its planned public consultation.
2.  Detail of the item and discussion:
3. Emma Rourke, Interim Deputy National Statistician (Health, Population and Methods) from the ONS and Will Laffan, Head of Policy Group from the UKSA presented this item.
4. In 2014, the Government set out its ambition that “censuses after 2021 will be conducted using other sources of data and providing more timely statistical information.” In the 2018 Census White Paper “Help Shape our Future”, the ONS committed to delivering recommendations on the future of social and population statistics – including census arrangements. To provide evidence for the recommendations, the ONS has for several years been conducting research into the viability of using primarily administrative sources to produce estimates about the population, its characteristics and housing.
5. This item was presented in response to a request by the NSDEC at the January 2023 meeting for further information regarding the planned public consultation which would inform the 2023 Recommendation, especially regarding the audience of the consultation and the materials presented to them (Evidence to support the National Statistician’s recommendation to government in 2023 on the future of population and social statistics NSDEC(23)05).
6. The paper provided an overview of the upcoming consultation on the future of population and social statistics, which will feed into the 2023 Recommendation. The paper covered the following areas:
   • the purpose of the consultation and its different audiences
   • the materials that will be presented, with a focus on the intentions for them with regards to the general public
   • the ONS’s broader work to understand public attitudes towards the linkage and use of administrative data for statistical purposes
7. The NSDEC appreciated the update on plans for a public consultation on the future of population and social statistics for England and Wales, and raised the following points in the discussion:
   • Offered their support and guidance with the development of the 2023 Recommendation, and expressed an interest in being kept updated on its progress.
   • Commented that administrative data cannot be treated as a general term as the public’s attitude towards different classes of administrative data would differ, such as health, criminal and education data. Thus, the Panel advised that in the public consultation, there should be more clarity in communications on what constitutes “administrative data”, especially regarding what classes or categories of data would be included.
   • Advised that there should be clear communication to the public regarding what areas are open for consultation.
   • Recommended that the ONS should treat the audiences of the public consultation with more granularity. For example, academics working in different fields would likely respond differently to the consultation, so it would be helpful to know who will be covered by roundtables, and whether the Royal Statistical Society would be included.
   • Recommended that the ONS consider the public in their capacity as data users, as the public also utilise administrative data such as in research projects, for example tracing community roots.
   • Asked whether a media handling plan was in place. The team confirmed that it was, and in parallel, the ONS is engaging with other data owning government departments, who would need to be bought into the proposal and to gain an understanding of their conditions to make their data available.
8. Action – The 2023 Recommendation team to keep the NSDEC informed of the progress on its public consultation and towards a Recommendation, as this develops.

4. Update on estimating ethnicity via admin data NSDEC(23)12

1. Outcome – The Committee welcomed the presentation and were reassured of the team’s further plans for this project,
2. Detail of the presentation and discussion:
3. Alex Mylles, Data and Research Analyst, ONS, and Michael Cole, Head of Social Statistics Admin First (SSAF), ONS presented this item. This item is a follow-up on the paper “Resolving conflicting ethnicities within administrative data when producing admin-based ethnicity statistics NSDEC(22)01” which was presented to the NSDEC in February 2022.
4. This presentation outlined findings from a series of focus groups exploring the public acceptability of a range of possible methods for dealing with missing and inconsistent ethnicity records in administrative data. The potential methods that were discussed with participants in the focus groups were as follows:
   • Replacing inconsistent data with the most recently recorded ethnicity
   • Replacing missing data with a previously stated ethnicity if the most recent record says “Unknown”
   • If the most recent record is “Any Other Ethnic Group”, replacing it with a previously stated ethnicity
   • Replacing missing data with a previously stated ethnicity if the most recent record says “Not stated” (also known as “Refused”)
   • Using statistical methods to impute an ethnicity where there are no ethnicity records for the person
5. The presentation then detailed the public’s responses to these options and the way in which these views have been incorporated into the team’s workplan. Overall, participants were broadly supportive of the first three approaches detailed above and these are the approaches taken so far in the research on producing admin-based ethnicity statistics. Participants were not supportive of looking beyond a “not stated”/”refusal” (method iv above) and this matched the team’s approach. With regards to statistical imputation, the findings suggested that this may be acceptable to the public, but specific research on a chosen method would be required. The team has confirmed that they currently do not use imputation. The team will do further research on this and will conduct further public acceptability testing if a specific imputation method is proposed.
6. The Committee welcomed this presentation and was assured that the NSDEC’s comments from February 2022 had been considered. The following further comments and observations were made by the NSDEC in the discussion that followed:
   • Emphasised that maintaining the public good is essential, and agreed with the team’s approach to conduct further research on methods and conduct further public acceptability testing regarding the statistical imputation method.
   • Suggested that there is a possibility for future work on the definition of ethnicity and categories of ethnicity, as well as how to operate in a world where people do not declare their ethnicity.
   • Asked whether the 3 methods above will cover most of the missing and inconsistent ethnicity records in administrative data.
   • Acknowledged the team’s plan to publish the findings from this exercise in due course, as per their commitment to keeping users informed.

5. How is ONS managing access to linked data assets to ensure public good aggregate outputs are produced in ethically appropriate ways from them? NSDEC(23)13

1. Outcome – The NSDEC welcomed the presentation and were reassured of how the Digital Economy Act 2017 Research Strand was operationalised.
2. Detail of the presentation and discussion:
3. Simon Whitworth, Head of Data Ethics and Research Accreditation, UKSA presented this item. This item was brought forward from the NSDEC January 2023 meeting (NSDEC(23)04).
4. The presentation first described how the Digital Economy Act 2017 Research strand (DEA) is operationalised to deliver data to the analytical community safely through an agreed framework. It also discussed how the 5 Safes are deployed in the DEA accreditation framework for processors, researchers, and research projects to enable safe and secure data sharing.
5. The Committee welcomed the informative presentation and were assured by the operationalisation of the DEA framework. The following comments were raised in the discussion that followed:
   • Suggested that with regards to the journey for data access through the DEA (displayed in a figure during the presentation), the end point for the journey should be the secure deletion of data once the retention period is reached.

6. Any Other Business

1. The NSDEC discussed the project that was received via correspondence in March 2023 – “COVID-19 and Respiratory Infections Survey (CRIS) NSDEC(23)08”.This project is led by the ONS, and aims to fill the data gap as the future of the UK’s COVID Infection Survey (CIS) is still under discussion. To ensure that the data gap is filled, the COVID-19 Infection and Respiratory Infections Survey (CRIS) will collect survey data from a subset of CIS participants that have agreed to be contacted for follow-up research, filling and providing information on long-COVID, other respiratory infections, absences from work due to respiratory infections.
2. The Committee recognised the public good of this project and gave their approval. The following comments were raised in the discussion:
   • Requested clarification on different ethical consent processes for this survey and ONS surveys in general, and if there are any implications of utilising implied consent instead of explicit consent
3. Action – The Secretariat to arrange a briefing for the NSDEC on the use of consent in surveys.
4. The next NSDEC meeting will be on the 5^th of July 2023.