The Economic Value and Clinical Impact of Artificial Intelligence in Healthcare: A Scoping Literature Review

Artificial Intelligence (AI)-supported healthcare has seen a substantial rise in development in recent years. The health economic impact is a crucial factor in the decision-making process regarding AI adoption. This study aimed to analyze the latest research progress and evidence on the cost-effectiveness and clinical efficiency of healthcare AI software from various perspectives, as well as to identify future opportunities and remaining challenges. A review of global literature was conducted using two key databases, PubMed and Embase, along with other related bibliographic resources. The literature search yielded 1,178 unique articles, of which 31 were included in our analysis. These studies covered a wide variety of clinical use cases and healthcare domains, including disease diagnosis (n=13, 41.9%), risk analysis (n=6, 19.4%), screening or patient triage (n=6, 19.4%), treatment delivery (n=5, 16.1%), and patient engagement or follow-up (n=1, 3.2%). Among the included studies, 24 (77.4%) examined the cost-effectiveness of AI compared to standard human-based practices from the perspectives of patients, healthcare systems, payors, or society. The remaining 7 studies, including 5 clinical trials, concluded that AI can enhance clinical efficiency by shortening labor time or patient journey in the clinic. The findings of this targeted literature review demonstrated that leveraging AI in human decision-making has the potential to improve multilevel health outcomes. However, there is a shortage of prospective health economic studies, particularly long-term evaluations, highlighting the disparity between the rapid progress of AI and its lagging utilization in real-world practices.