Go to content

About the client

The mission of the Health Research Authority (HRA) is to protect and promote the interests of patients and the public in health and social care research. With 200 staff, HRA enables collaboration to streamline the set-up and review of research and provides specialist advice, guidance, and learning. HRA reviews around 6,000 new research studies each year.

About the project

The Streamlined Data-Driven Research Programme (SDDR), funded by the NHSX AI Lab, is an opportunity for HRA to ensure that high-quality Artificial Intelligence (AI) and data-driven technologies are developed and deployed as quickly as possible, to unlock the potential of this technology in the interests of patients, the NHS and the UK economy.

As part of this programme, HRA asked us to help create a decision tool to help applicants in the field of AI work out whether their research is eligible for approval by the HRA and which protocols they need to satisfy to obtain approval. The IRAS website already had a suite of decision tools which lacked cohesion. We were asked to create a solution that could become the blueprint for the entire suite.


  • Provide recommendations to devise an efficient online tool to help AI research applicants navigate the approval and regulations protocols.
  • Review the existing online decision tools to provide consistency in the way all research applicants navigate approval and regulations protocols.
  • Reduce staff time filtering non-eligible research and assisting applicants in navigating approval and regulations protocols.

We conducted user research, designed and built a prototype, and tested and provided our recommendations for the best approach to develop a data decision tool for the IRAS website.

This is a project that aligns with our love of taking complex user journeys and tasks and turning them into things that are easy to use and understand.

Ian Axton
Ian Axton, Studio 24

Information gathering

We undertook a series of stakeholder interviews and information-gathering sessions to understand the suite of tools including:

  • business processes that underpin them
  • technical and design constraints
  • content including language and logical structure
  • structure and wording of the AI decision tool
  • user flows and journeys.

Complex content mapping

The language and user journeys had to be easy to understand. With multiple pathways resulting in recommendations (end nodes) we need to create a design pattern that enables researchers to easily view and amend their responses to questions – this would allow them to alter their path and discover alternate and potentially more efficient outcomes.

Using wireframes for rapid testing

After mapping the user flows we started designing a tool for the user interface. This must be easy to use and meet both user and business needs. 

We designed a ‘low fidelity’ prototype of the decision tool in the form of wireframes. 

Wireframes allow us to explore, test and refine rapidly. The interactive wireframes are published to a test environment where we gave participants task-based exercises to perform. We assessed whether they could achieve the tasks by navigating the wireframe.  We learned from these insights and integrated them into the next iteration of the prototype.

Building a HTML prototype for a more authentic experience

Next, we built a ‘high fidelity’ HTML prototype of the tool so it could be tested in the browser.

There was another round of user testing on the HTML prototype, this time with a larger set of users. A diverse set of people were invited and we gathered feedback on both pain points and helpful features of the prototype. 

User feedback was integrated into the next iteration of the HTML prototype. 

The prototype was submitted to the same round of testing once again. But this time with a different set of users. This reduces the risk of adjusting the prototype to the preferences of a small group of users and avoids potential bias due to emotional attachment.

Presenting our final recommendations

After the user testing and research, we presented our recommendations in a robust written report. Most importantly, our recommendations are evidence-based and informed by our technical expertise.

The report includes:

  • an executive summary
  • a summary of our recommendations
  • a description of research performed and its outcomes
  • relevant research data (as appendices)
  • fully tested prototype

We delivered the blueprint for a well-researched tool to meet HRA user and business needs.