OBJECTIVES: Artificial Intelligence (AI) in recent years has been developed to offer solutions to the challenges currently faced by health systems. The adoption of such technologies will require a prior assessment of their health economic value, particularly in countries where health technology assessment bodies exist. There were two key objectives of this research. Firstly, to determine appropriate methods to assess the health economic value of AI tools for diagnostic imaging. Secondly, to identify the key economic value drivers of these interventions in the UK’s NHS.

METHODS: A targeted literature review was conducted to identify previous health economic evaluations of AI interventions. Search strings from a previous review were augmented to capture additional publications. Exclusion criteria were applied to select papers for full review. Next, a panel of eleven NHS clinicians with a variety of experience and expertise were interviewed to elicit where AI tools to support CT and chest X-ray imaging could most benefit the NHS.

RESULTS: The targeted literature review identified 24 evaluations, the majority of these conducted a cost minimisation analysis. No evaluations quantified the value of the systemic benefits that AI tools offer. For AI assisted chest X-ray interpretation, 70% of clinicians agreed that the technology could reduce time and resources to perform audits. It was agreed that an AI tool could accelerate time to clinical decision making and treatment in some indications if the AI could perform alone. Similarly, for AI supported CT imaging clinicians delegated most importance to turn around time improvements or reduction of unnecessary follow ups.

CONCLUSIONS: There is a lack of health economic evidence to assess the value of AI technologies. Furthermore, the evidence that has been published may not capture the true benefit of the technologies as they exclude outcomes relevant for systemic value drivers such as reducing turnaround times.