Exploring perceptions about the potential deployment of Artificial Intelligence within the English NHS Diabetic Eye Screening Programme

Problem

The English NHS Diabetic Eye Screening Programme (DESP) performs approximately 2.2 million eye screening appointments each year, generating over 10 million retinal images that are graded for the presence or severity of diabetic retinopathy. Currently, human graders perform this task within the DESP; however, this is expensive and time-consuming, and their expertise could be applied elsewhere.Previous research has shown that Artificial Intelligence (AI) systems can identify images with no disease from those with retinopathy as equally well as human graders, and could help significantly reduce the workload for human graders.There is a need to examine the perceptions, concerns and views of AI-assisted eye-screening among people living with diabetes and NHS staff, if AI were to be introduced into the DESP. This will help to identify factors that may influence acceptance of this technology by NHS staff and people living with diabetes.

Approach

The aim of this PPIE work is to create two surveys to explore the perceptions, concerns and views among people living with diabetes and NHS staff about the implementation of AI within the DESP.Potential questions for each survey were initially developed from recently published research which had used surveys to investigate perceptions and acceptance of AI in healthcare. Staff working at the North East London NHS DESP and people living with diabetes from the local community were then invited to participate in focus groups to co-design two separate online surveys to examine their perceptions, concerns and views on potential implications of AI within the DESP.

Findings

Each survey was developed through an iterative process, with focus group discussions and post meeting feedback informing question development, domains captured and the overall format and structure of each survey. The two surveys are comparable in content, and the survey for NHS staff additionally examines issues with workforce planning.Surveys will be piloted at the North East London NHS DESP followed by qualitative semi-structured interviews to assess accessibility and usability, and to validate the surveys. Following review, the surveys will be distributed via other NHS DESP sites and via relevant charities to ensure greater participation and representation from a diverse population. Differences in survey responses by population subgroups, including by age and ethnicity, will be examined.

Consequences

The process of survey development has highlighted the importance of patient and public involvement in research. These questionnaires will identify factors that may influence acceptance of AI within the NHS DESP by staff and people living with diabetes, which will assist planning outreach activities to facilitate implementation of this technology. The findings may also highlight the potential for greater use of AI within healthcare, and also impact on screening approaches within the NHS.