Patient and public involvement to build trust in artificial intelligence: A framework, tools, and case studies

Abstract

Artificial Intelligence (AI) is increasingly taking on a greater role in healthcare, especially during the current COVID-19 pandemic. However, hype and negative news reports about AI abound. Integrating patient and public involvement (PPI) in healthcare AI projects may help in adoption and acceptance of these technologies.

We argue that AI algorithms should also be co-designed with patients and healthcare workers.

We specifically suggest: 1) including patients with lived experience of the disease, and 2) creating a Research Advisory Group (RAG) and using these group meetings to walk patients and carers through the process of AI model building, starting with simple (e.g. linear) models. These meetings should be repeated to elicit feedback from the stakeholders, explain model predictions and get guidance on model modifications.

We hope that the approach of involving patients, clinicians and data scientists, in a virtuous cycle of co-design, will be used in future AI projects in healthcare. Lived experience of a disease is important in healthcare research. If properly designed, patient and public involvement can lead to better outcomes in health research.

We present case studies and a methodology of how modern data science can be applied to healthcare, where data scientists, clinicians, and patients work together. We have shared a framework, best practices and tools that can be used for engaging with patients and explaining AI to them. The strategy of co-designing with patients can help set more realistic expectations for all stakeholders, since conventional narratives of AI revolve around dystopia or limitless optimism. We hope that AI research in healthcare can be adopted faster if humans slowly build up trust in machines, over repeated and carefully calibrated interactions.