As more data is collected about us and used for multiple purposes, transparency has been used as a way of demonstrating that the systems in place for managing and protecting that data are trustworthy. This is in part because transparency is a core principle in the General Data Protection Regulation, intended to strengthen individuals’ rights over how data about them is used, and improve public accountability. There is also increasing demand coming from patients who want to understand what happens to data about them, including when and why it might not be being used most effectively to help drive research and new treatments. These expectations can be summarised as: say what you do, and do what you say.
It is undoubtedly positive that data custodians and organisations are carefully considering how to be transparent about data use, and to make this meaningful, not tokenistic. Yet it can be unclear what the push towards increasing transparency requires in practice. There is a risk that siloed efforts within organisations could make the whole system of data use appear more fragmented and inadvertently increase people’s confusion and anxiety. This could undermine the intention behind improving transparency across the board.
From soundbite to action
UPD convened a workshop to discuss what a shared understanding of transparency for the use and management of patient data might be. Held in late March 2019, attendees included a patient representative and patient group; data controllers including NHS Digital, Public Health England and the Clinical Practice Research Datalink; data processors such as the Healthcare Quality Improvement Partnership; NHS England, the Information Commissioner’s Office, National Data Guardian’s Office and NHSX among others.
Transparency: simple in principle, difficult in practice
We set up the workshop to try and generate consensus. While everyone agreed on the importance of transparency in the abstract, it was much harder to reach common understandings of what this meant in practice. The same things are explained differently, or different approaches use the same terms, adding to the confusion. Key findings were:
1. Peoples’ understanding of transparency is diverse. We did not achieve a shared, practical understanding.
2. People mostly approach transparency as a concept or value, rather than something to deliver. This means it is not just viewed in terms of legal/policy compliance or box-ticking, but instead as something more cultural and embedded. This again leads to different interpretations of how it can be implemented.
3. While everyone wants to be transparent because it’s “the right thing to do”, people articulate their purpose for being transparent in very different ways.
Transparency is a continuum
Each stakeholder will have different requirements (whether statutory or policy), motivations and capabilities to demonstrate transparency in their processes and practices.
The diversity in people’s understandings of transparency was eye-opening: we had been aware there were different views, but they were stark in discussions. However, people acknowledged that working out what transparency meant in their context was a journey and that it’s important to reflect on how you know when it’s achieved. It’s timely now to think about efforts collectively and not in organisational siloes.
Current challenges
Several elements make it difficult to ensure transparency in practice:
1. Issues with data quality. Patient records can include errors or inaccuracies in things like names and addresses as well as clinical codes and diagnoses. This requires thinking through the practical implications and steps needed, such as capacity and willingness to correct errors when identified.
2. The health data system is complex and changing. Therefore, communications are inevitably complex as systems may appear messy and inconsistent from the outside.
3. There can be a sense of anxiety or resistance to talking openly and honestly. This includes preferring to talk only about good news or plans for future improvements, rather than where we are now.
4. There are no obvious leaders in this space that publics are aware of and who are trusted to deliver information or messages about transparency.
5. Operationally and culturally, transparency is not yet business-as-usual for some organisations. Embedding the values and culture of transparency require dedicated time and resource, which may not be considered a priority.
Different approaches to transparency in practice
To help break down the different approaches to transparency, we’ve pulled together the list below. It covers the different lenses through which to view transparency, grouped into what information is provided and how it is conveyed. The list is based solely on the workshop discussion and a small number of patient insights collected by use MY data.
The below can be used as a prompt when thinking about how to uphold the principle of transparency – and whether your interpretation is the same as your colleagues’.
“WHAT”
Answering the right questions (e.g. content of fair processing notices): information given neatly answers the questions your audience(s) may have. Key elements:
-
You explain what data is being used, by who, why and how it’s looked after.
-
The information is wholly relevant, addresses known concerns and leads to better understanding, not just increased awareness.
-
It is not a PR or marketing exercise to sell your approach, but answers what others may want to know.
Checkable, assessable (e.g. audit trails; feedback on research results): anyone else can easily check what you’ve done. Key elements:
-
There is sufficient information provided for independent scrutiny.
-
Publishing information on the rules you follow and any protocols or decisions you’ve agreed as a result of those rules.
-
If someone only had the information you provide, they could to check whether you’ve adhered to the rules.
Open, interoperable data (e.g. All Trials platform; ClinicalStudyDataRequest.com): open data is the default (where appropriate). Key elements:
-
Technical aspects including: open standards; open APIs; standardised languages such as SNOMED; interoperability standards such as FHIR.
-
Being open about what you make available, including: metadata; statistics or findings derived from the data; evidence/rationale for changes in data use/management.
-
May include details on why certain information is not available, and a rationale for this.
“HOW”
Using the right language (e.g. plain language summaries; visualisations and imagery): The words and images used are clear and understandable. Key elements:
-
Plain language is the default, or language is tailored for the audience and their information and accessibility needs.
-
Language used considers what others are doing. Vocabulary is shared and works together across relevant initiatives to produce an understandable overview of data use.
Accessible (e.g. data release register; minutes of decision meetings): content is easily findable. Key elements:
-
Users don’t need in-depth knowledge of system structures, language and relationships to navigate information.
-
Information needs are thought through, including practicalities such as timeliness, level of detail and ease of navigation to relevant related information.
Organisational values and behaviours (e.g. learning and improvement environment within teams; no-blame culture): Openness and honesty are how we do business around here. Key elements:
-
Acknowledgment when things go wrong and seeking to learn and improve.
-
Time and resource are committed to transparency initiatives.
-
Strong leadership demonstrating importance of transparency.
Next steps
We’re open to hearing from others on whether this way of presenting the different approaches to transparency is helpful, and if it should be developed into something more substantial either from UPD or through convening others to produce guidance.