Public Engagement with Data Linkage and Data Governance – Rapid fire presentations

Authors: Kirsteen C Campbell (1,2), Sarah Chave (3), Rosie Hill (3), Mhairi Docherty (3), Michael Gregg (3), Emma Turner (4,5), Katharine Evans (4,5), Jacqui Oakley (4,5), Stela McLachlan (1,2), Alex Bailey (6), Cassie Smith (7), Rebecca Whitehorn (1,2), Lidis Garbovan (1,2), Robin Flaig (1,2), Andy Boyd (4,5,7)

1 UK Longitudinal Linkage Collaboration, The University of Edinburgh, Edinburgh, UK, 2Usher Institute, The University of Edinburgh, Edinburgh, UK, 3UK LLC Public Advisory Group, UK Longitudinal Linkage Collaboration, The University of Edinburgh, Edinburgh, UK, 4UK Longitudinal Linkage Collaboration, University of Bristol, Bristol, UK, 5Population Health Sciences, Bristol Medical School, University of Bristol, Bristol, UK, 6Medical Research Council Regulatory Support Centre, Edinburgh, UK, 7Health Data Research UK, London, UK

Abstract

Introduction
Legislation around Longitudinal Population Study (LPS) data can be complex and confusing. Therefore, it may prove difficult for LPS participants to fully understand who is allowed to access their data, for what purposes, under what circumstances and their rights. Of particular importance are the UK GDPR (General Data Protection Regulations), Common Law Duty of Confidentiality, Digital Economy Act (DEA) 2017 and Section 251 of the NHS Act 2016. Our vision was to work with our Public Advisory Group (PAG) to produce transparency materials for LPS participants to more easily understand LPS data and the law.

Methods
We surveyed public contributors from UK Longitudinal Linkage Collaboration (UK LLC), Health Data Research UK and DATAMIND to benchmark current understanding. Together, our PAG and legal experts worked to produce an infographic series understandable by a public audience, presented using clear and accessible methods and in a range of language options. Through regular rounds of discussion and feedback, we focussed on the practical implications for LPS participants.

Results
The results of the survey (n=56) suggested there were low levels of awareness of the laws and legal principles. This informed our key messages, highlighting what LPS participants should be aware of and what to expect from their LPS. Sharing these materials continues. They have been distributed to UK LLC Partner LPS to share with their participants, are free to download from the UK LLC website and this work has resulted in a publication in the International Journal of Population Data Science.

Conclusions
Our transparency materials support LPS participants to understand the legal basis for study data use, and their rights. General awareness of the laws and legal principles involved in the use of longitudinal data is low and we recommend data providers should raise awareness in clear and accessible ways.

Authors: Shayda Kashef-Tomlins (1), Zoe Dale, Tom Curran, Dan Clay

1 ADR UK (Administrative Data Research UK)

Abstract

Public involvement and engagement is increasingly recognised as essential to trustworthy and impactful data research. Yet misconceptions persist that members of the public need specialist knowledge or research experience to contribute meaningfully to discussions about data, statistics and governance. Drawing on the experience of the ADR England Public Insights Panel, this presentation will challenge these assumptions and share practical lessons from establishing and running a diverse public panel focused on administrative data research.

The ADR England Public Insights Panel brings together members of the public from a range of backgrounds, ages and lived experiences across England to provide input on data-driven research, public communication, data governance and public trust. Since its launch in 2024, the panel has considered a wide variety of topics, from research design and public engagement strategies to language, accessibility and the communication of complex data concepts. The breadth of subjects discussed demonstrates the value of involving public contributors throughout the research lifecycle, rather than limiting engagement to dissemination or consultation exercises.

Through case studies, reflections and testimonials from panel members, we will explore how inclusive recruitment, accessible facilitation and ongoing support have enabled meaningful participation from people with no prior research background. We will discuss what we have learned from working with researchers who have presented to the panel, including how modes and approaches to communicating in an engaging way to a general public audience.

The session will also offer practical advice for organisations considering establishing public panels of their own, as well as for researchers seeking to engage effectively with existing panels. By sharing successes, challenges and lessons learned, we will demonstrate how diverse public contributors can engage critically and constructively with complex data research issues, and why meaningful involvement depends less on technical expertise and more on creating the right conditions for participation, dialogue and mutual learning.

Authors: Naomi Clements-Brod (1) Rudolf Cardinal, Jan Speechley (2), Jan Davies (2), Shelina Gofur (2), Robyn Frances (2)

1 DATAMIND/University of Cambridge, 2 Public Contributor

Abstract

The NHS collects mental health data that can be valuable for research but people often worry about how this data is used, especially if commercial organisations (not just clinicians or academic researchers) want to use it. It is particularly important to engage the public on this topic because patients do not have to give explicit consent for their routinely collected mental health data to be used for research. There is substantial public sensitivity and wariness around commercial use of routinely collected health data. Data about mental health is considered particularly sensitive and potentially stigmatising.

Patients have a right to understand and choose how their routinely collected health data is used for research. Everyone who works with this data must understand and operate according to the laws and the public’s expectations to create and maintain trust in research with mental health data and its outcomes.

The DATAMIND Lived Experience Advisory Group (LEAG) (https://datamind.org.uk/patients-and-public/lived-experience-advisory-group/), co-developed a new set of guidelines (https://doi.org/10.3389/fpsyt.2026.1760116) providing advice and expectations to industry and the NHS about what is acceptable and desirable when it comes to sharing mental health data with commercial organisations from a patient/public perspective.

Funded by PEDRI’s From Standards to Impact programme, we co-created a public-friendly website and animation to increase transparency about this research to a wider public audience.
This work will provide clarity about:

• how collaborations with industry should work and
• what choices people have about how their data is used.

By involving people with lived experience throughout the project, we ensured the resources were accessible, relevant and grounded in public concerns, helping to strengthen trust and transparency around the use of mental health data.
In this presentation, we’ll share our learnings from the co-creation process and embedding PEDRI’s Good Practice Standards in this work.

Authors: Andrew Steele (1)

1 NHS Golden Jubilee, University of Glasgow

Abstract

Patient and Public Involvement and Engagement (PPIE) is essential to trustworthy data research, but getting diverse input is hard. Traditional panels are limited by recruitment, geography, and cost, so they often miss the communities whose data are actually being linked and studied.

To help, we built Panelyze, an AI platform that supports existing PPIE rather than replacing it. It works in five steps: (1) it builds synthetic personas and panels using census data and real-world patient experience; (2) it reads and analyses a research proposal; (3) it simulates the panel’s discussion; (4) it produces co-production materials such as Plain English Summaries and infographics; and (5) it generates the Panelyze Score, which rates a proposal against the UK Standards for Public Involvement.

Because the synthetic panels match census proportions, and can over-sample demogrpahics or targeted lived experiences, Panelyze brings in voices that are often not heard in person. Testing it on the CAPRIE-2 cardiovascular proposal surfaced real barriers, such as losing wages to attend clinic visits, and checked communications channels and terminology choices like drug names. This gives research teams a low-risk way to test how inclusive their plans are before involving real people and PPIE groups.

For data-linkage work, Panelyze runs fully offline inside Trusted Research Environments and safe havens, so no data leaves the secure space.

Early results and conclusions have been recently published at https://www.frontiersin.org/journals/digital-health/articles/10.3389/fdgth.2026.1771729/full.

We will share what the approach can do, its limitations, and how it can make later human engagement more inclusive and better prepared.

Authors: Joachim Tan (1), Ania Zylbersztejn, Pia Hardelid

1 UCL Great Ormond Street Institute of Child Health

Abstract

Linking data which are routinely collected for administrative purposes, such as hospital and education records, is increasingly used by researchers to understand children’s health, needs and development over time. Linked data can generate population-level evidence, support earlier interventions, and improve outcomes, without the necessity nor expense of new data collection.

Due to the confidential and sensitive nature of such data, their use for research is subject to stringent controls. Permission must be obtained from research ethics committees, confidentiality advisors and data guardians, all of whom want assurance that studies are carried out ethically, lawfully and safely to protect people’s rights and privacy. They often want to know what children and young people think about data linkage, but instead of repeatedly asking them similar questions for every study, we aimed to co-develop general guidance around common themes that could serve as future reference for all.

In 2025, our team met three times with GenerationR Young Persons’ Advisory Groups (YPAGs) from Great Ormond Street Hospital and Liverpool, listened to their feedback and co-produced a guide setting out principles that reflect children’s and young people’s views about when it is acceptable to link their data for research.

Building on previous research on public attitudes to data sharing, the YPAGs discussed and expanded on three core ideas that: (1) research should be for everyone’s good; (2) data should be kept safe and private; and (3) researchers should be open and honest in conducting and communicating research. Through this project, we gained a deeper appreciation of the rules and protections around people’s data, especially from the perspective of individuals whose information is being used. Overall, participants expressed support for data linkage when it is used to benefit everyone, and researchers should strive to be conscientious, respectful and transparent when doing so.

Authors: Rebecca Goulding (1), Gail Davidge, Warren Del-Pinto, Sarah Markham (2), Brian McMillan, Goran Nenadic

1 The University of Manchester, 2 Public Contributor

Abstract

Data and Analytics Research Environments (DARE) UK funded catalyst projects to drive innovation in the field of sensitive data research, including FORTRESS: Federated, Open and Reliable Trusted Research Environments for Synthetic Textual Healthcare Records. We focused on synthetic free-text healthcare data. Synthetic data are data generated to resemble real data (using statistical or machine learning methods) without being about actual people to minimise risk to privacy. Most synthetic healthcare data are structured. Our goal is to design and create a model that can generate synthetic free-text healthcare data, and to develop frameworks to assess the validity of such data.

We assembled a group of 12 people with experience of being involved in healthcare- and/or data-related research. The group was diverse in terms of age, gender and ethnicity. We held three meetings (two online, one in-person) to discuss what we would need to do to build trust in this field, communicate about it with the general population and get more members of the public involved in this type of research.

The first session was introductory and included a draft glossary of terms. The second and third sessions focused on what is important or concerning about how synthetic data may be generated, shared and used. We discussed: transparency in how the model is designed and created, consent to use real data to teach the model and validate synthetic data, the quality of synthetic data and what they could be used for, the safety of synthetic data and the potential for bias and misuse (for example, if synthetic data do not reflect under-served populations).

These discussions have shaped our current research and future plans. This presentation is an opportunity to share our approach to public involvement and invite discussion about best practices for engaging the public in conversations about this type of research.