Abstract
The development of artificial intelligence (AI) in healthcare is advancing rapidly and playing
an important role in healthcare worldwide. However, real-world implementation remains
challenging in terms of economic value, patient data security, and acceptance by health
professionals. Hence, this study was conducted with three main components: (1) a systematic
review and meta-analysis of global studies evaluating the cost-effectiveness of precision medical
AI (AI-PM); (2) a nationwide online survey of Thai health providers on their use of and perceptions
toward AI in healthcare; and (3) a model-based (decision-tree and Markov model) cost-utility
analysis of community-based screening for pulmonary tuberculosis (TB) in high-risk populations
using AI compared with the current approach.
Results: (1) The literature indicates that, overall, AI-PM tends to be cost-effective or cost-saving,
with positive net monetary benefit (NMB) and gains in quality-adjusted life years (QALYs).
Key drivers of NMB include the unit cost of AI-PM and the extent to which AI-informed decisions
are adhered to in practice. (2) The Thai health-professional survey revealed highly positive views
toward using AI to support clinical decision-making and diagnosis, alongside significant concerns
about patient data security and the quality of physician –patient communication.
(3) The economic evaluation found that AI-based pulmonary TB screening is cost-effective in
Thailand at a willingness-to-pay threshold of 160,000 THB per QALY. In particular, using medical
AI alone to interpret chest radiographs (AI alone) was the most cost-effective option: although it
incurs higher lifetime costs, it delivers better health outcomes than interpretation by general
practitioners alone (GP alone).
Conclusion: AI in healthcare has substantial potential to improve healthcare. To enable
effective and sustainable adoption, however, standardized approaches to cost-effectiveness
evaluation are needed, along with closing reporting gaps, addressing ethical issues such as data
security, and establishing robust policy frameworks for implementation and reimbursement
within health insurance systems.
