Current AI applications are advanced in their analytical power and precision to optimise health care logistics and support health care professionals in clinical decision-making. A primary challenge for AI applications, however, is that their use of patient data will change existing relationships between and among patients, health care staff, clinicians, and health care systems.
In this blog, part of our Public Involvement Blog Series, we hear from Sudhir Shah on his experiences of working on an ARC North Thames study which brought patient and public voices on this topic to the fore. This blog is based on a conversation between Sudhir Shah, William Lammons and Lucy Thompson.
Centering Public Perceptions on Translating AI Into Clinical Practice: Patient and Public Involvement and Engagement (PPIE) Consultation Focus Group Study is the result of a team of ARC North Thames researchers and public collaborators. It aimed to bring the voices of patients and the wider public into the design, development, implementation, and embedding of AI applications in health care, through a series of qualitative public and patient focus-groups. A total of 17 public collaborators representing 7 NIHR ARCs across England participated in 1 of 3 web-based focus group discussions where they were encouraged to voice their understandings, experiences, and perceptions of AI applications in health care.
The next step was “thematic analysis” of discussion transcripts - using code to eke out 3 primary themes – which included recommendations on PPIE for AI in health care - and 7 corresponding inductive subthemes. One might expect that public collaboration in this project had ended by this point, but this was not the case. Sudhir Shah, one of ARC North Thames’ longest serving Research Advisory Panel (RAP) members, and Jamie Hunter, a public advisor with ARC North West Coast’s Care and Informatics research theme, also worked on thematic analysis.
PPIE as personal impact
Sudhir felt “excited to learn about how to get involved in this, because coding work is mostly done by the researchers, not by the PPIE. I felt we were becoming more like peer researchers.” Sudhir’s involvement as a co-researcher on this project meant a better, public informed output at the end. But what we enjoyed about our conversation with him, is the way he talked animatedly about what working on the project had given him. He said:
“Milou [one of our co-authors] was a professional researcher and she helped us with how to do coding work, she supported us to develop an appropriate coding method to analyse the transcript and scoring method.
We looked at the advantages of implementing AI and involve PPIE in the recommendations. We got to 3 main domains, and from those we went back to the transcript again and highlighted paragraphs or sentences which would match these domains. From that coding process we got 7 further subdomains.
I began to understand the research questions properly by doing this analysis. It helped me to articulate the research question better, and this was important for us as co-researchers as we could approach the work more effectively.
I now realise how time consuming [it] is to do this data analysis. It’s not easy, because some of the transcripts were not clear. You've got to have a very good comprehension of all the transcripts, and read between the lines to get proper understanding of what PPIE meant in these contexts.”
This shows the ways in which Sudhir gained new skills through his involvement in the project, which he used to apply a different lens to research data. His involvement also helped to demystify the tricky, individual steps a researcher must go through to create an output.
Ultimately, Sudhir also gained a new appreciation of the end-to-end research process and the diverse ways in which PPIE colleagues should be encouraged to contribute. He continues:
“I started with personal interest, but I got to see it all the way to the finish – the journal article – that we went through the full research cycle together has helped me. We did something outside the box because PPIE people, we mostly provide constructive comments on research papers. But here we have produced a research paper ourselves in a journal. So that was an exciting part of it.”
In addition to understanding the end-to-end research process, Sudhir’s experience has also helped him to build skills he is using in different projects:
“It's also helped me with my other project outside ARC. I did some coding work there also for an East London group within mental health and we looked at clinicians and the patient interviews from the data that was collected, and [the] transcript was produced just like what we did with the PPIE colleagues.”
Seeing the whole person
A big part of William Lammons’ approach to his role as PPIE Lead at ARC North Thames is seeing the whole person:
“When Sudhir came to me and expressed interest in this project, I wanted to facilitate this. It’s important for me in my work that you see the person in front of you not only as someone with a general interest in applied health research or lived experience of a certain condition, but as the whole person – their career, their skills, their motivations, and passion. I knew from our work together that Sudhir had a professional background in computer science and engineering, and I assumed this study might resonate with those prior experiences.
Now, I’m touched – he describes the qualitative interview coding process perfectly. He has really got the gist of the way technical-meets-personal in this work. Overall, it is fun to hear him reflect on this experience, and that he is now able to provide such insight into how he and we can improve.”