Multi-stakeholder Approach for Designing an AI Model to Predict Treatment Adherence

Abstract

Artificial intelligence (AI) can transform healthcare by improving treatment outcomes and reducing associated costs. AI is increasingly being adopted in healthcare. In this regard, one area where AI can have a significant impact is in improving adherence to treatment, which is critical to achieving desired health outcomes. It is well known that poor adherence can lead to treatment failure, disease progression and increased healthcare costs. However, the factors that influence adherence to treatment remain unclear. In this context, this study sought to implement an open innovation methodology based on co-creation to understand the requirements for the development of an AI model to aid in the prediction of treatment adherence. Semi-structured interviews were conducted with eleven stakeholders from four groups: patients, healthcare professionals, data scientists and pharmacists. The needs and requirements received were categorized into four key aspects that were translated into requirements and needs: understanding the nature of the drivers, scope and impact of the problem; identifying data sources; understanding relevant data points; and addressing potential ethical issues.