The Intersection of Artificial Intelligence and Health Technology Assessment

Health Technology Assessment (HTA) serves as a vital tool for evaluating the worth and impact of health technologies, offering evidence-based guidance for their adoption and use. Defined as a multidisciplinary, policy-oriented research methodology, HTA uses explicit methods to determine a technology’s value throughout its lifecycle to inform decision-making and promote an equitable, efficient, and high-quality health system.

Artificial intelligence (AI) possesses the capability to significantly enhance HTA processes by improving data collection, analysis, and overall decision-making, thereby increasing the accessibility, quality, and efficiency of healthcare services. Given the growing volume of available data and recent advancements in analytical methodologies, AI has the potential to strengthen the evidence base required for HTA.

This scoping review aimed to explore the opportunities and challenges associated with integrating AI into HTA, with a specific emphasis on economic dimensions. The core objective was to analyze how AI-driven tools and methods can optimize economic evaluation frameworks—such as cost-effectiveness analysis and resource allocation strategies—to support the efficiency and sustainability of health systems.

Methodology

The study utilized the scoping review framework proposed by Arksey and O’Malley, encompassing five essential steps. A systematic search was conducted across three major databases: PubMed, Scopus, and Web of Science, covering literature published between 2000 and 2024. The search queries combined keywords related to AI (e.g., “machine-learning,” “big data,” “deep learning”) and HTA (e.g., “health technology assessment,” “cost-effectiveness analysis,” “economic evaluation”). The initial search yielded 345 results, eventually leading to the inclusion of 38 relevant articles for qualitative synthesis. The synthesized findings focused on the role of AI applications in economic evaluations and were presented through a narrative synthesis approach.

Key Findings: The Transformative Role of AI in HTA

The findings of the review confirm that AI significantly enhances HTA outcomes by driving methodological advancements, improving utility, and fostering healthcare innovation. The results were categorized into themes covering methodological changes, utility, behavioral models, diagnostic innovations, infrastructure, and ethical challenges.

I. Methodological Advancements in Economic Evaluation

AI has the potential to revolutionize HTA decision-making by analyzing vast datasets, including electronic health records (EHRs), medical claims, and real-world data (RWD).

• Cost and Expenditure Estimation: AI enables more accurate estimation of costs and future healthcare expenditures.
• Cost-Effectiveness Analysis (CEA): AI plays a critical role in CEA by helping to predict incremental costs and effectiveness necessary to compute the Incremental Cost-Effectiveness Ratio (ICER).
• Heterogeneity Assessment: Advanced techniques, such as causal forests, leverage individual-level data to assess variability in CEA, allowing policymakers to derive tailored policy rules for specific subpopulations regarding the cost-effectiveness of interventions.
• Modeling Efficiency: Using AI in CEA through discrete event simulation (DES) created accessible software tools (like DESnets) that demonstrated superior performance in expressive capacity, transparency, and computational efficiency compared to traditional methods.
• Downstream Event Evaluation: AI algorithms evaluate treatment-related complications and their associated costs by analyzing billing codes and claims data, which is essential for assessing the financial impact of new medical technologies. AI also helps address uncertainties associated with Next-Generation Sequencing (NGS) data, improving cost-effectiveness evaluations by leveraging genetic data alongside observational datasets.

II. Enhanced Utility and Outcome Metrics

The integration of big data and health information systems (HIS) facilitated by AI provides data-driven insights that support better resource allocation and demonstrate the cost-effectiveness of various interventions.

• Quantifying Health Outcomes: AI enhances the quantification of key population health metrics, including the Value of Statistical Life (VSL), Quality-Adjusted Life Years (QALY), and Disability-Adjusted Life Years (DALYs).
• Real-Time Efficiency: Leveraging real-time analytics demonstrated that AI could significantly reduce mortality rates, increase QALY, and lower overall healthcare costs, particularly in high-cost settings like Intensive Care Units (ICUs).
• Practical Examples: Studies showed that Deep Learning (DL) algorithms used for diabetic retinopathy (DR) screening achieved comparable QALYs and significant cost savings relative to human graders. Furthermore, research evaluating treatments for multidrug-resistant tuberculosis (MDR-TB) found that all-oral regimens provided greater QALYs and longer survival rates, confirming them as a long-term cost-effective option despite higher initial costs.

III. Behavioral Modeling and Policy Support

AI is crucial for developing behavioral models within the HTA framework, helping to understand and predict human behaviors that impact health interventions.

• Policy Simulation: Researchers demonstrated the use of the “What If…” deep learning (DL) algorithm to simulate pandemic scenarios, predicting critical metrics like daily transmission rates (RE) and unemployment rates (UER). These insights enabled policymakers to evaluate the economic impact of health policies and identify optimal strategies through reinforcement learning.
• Integration of Behavior: Researchers emphasized the need to integrate behavioral factors (societal norms, decision-making processes) into health economic models using tools like data mining and agent-based modeling to enhance the efficiency and affordability of public health initiatives.

IV. Infrastructure and Data Management

Robust infrastructure is essential for effective AI application in HTA.

• Data Standardization and Integration: Infrastructures must include mechanisms for data standardization, computational processing power, and secure storage to manage diverse data sources, such as biological, genetic, and electronic health records (EHRs). AI is a critical tool for data integration and linkage, facilitating the merging of disparate data sources like genetic information and EHRs.
• Mitigating Data Quality Issues: Machine Learning (ML) models can address challenges in databases such as data-sharing restrictions, unobserved confounders, and missing data, which frequently compromise data quality in economic analyses.

Challenges and Ethical Considerations

Despite the clear benefits, integrating AI into HTA introduces critical ethical and operational challenges that require deliberate planning and policy development.

• Ethical Concerns: Major challenges include biases, lack of transparency, and accountability. Concerns also extend to data collection and administration, requiring robust ethical frameworks and governance policies to address potential misuse or breaches created by large data volumes.
• Regulatory Deficiencies: AI-based medical devices (MDs) often fail to meet HTA standards due to inadequate evidence from clinical investigations, highlighting the need for regulatory bodies to establish uniform evaluation standards.
• Implementation Hurdles: Other barriers include regulatory hurdles, high IT costs, complex data-sharing agreements, strict adherence to information governance policies, and insufficient experience among stakeholders. The rapid updates of AI technologies necessitate continuous adjustments in HTA frameworks to maintain relevance and accuracy.

Conclusion

AI offers unparalleled opportunities to revolutionize HTA by enabling comprehensive and efficient analyses, fostering innovation, and improving patient outcomes. The primary areas of impact are seen in advanced economic evaluation methodologies (like CEA and ICER calculation), better quantification of utility metrics (QALY, VSL), and sophisticated behavioral modeling for policy simulation.

However, ensuring responsible and sustainable implementation requires the development of robust data management strategies and regulatory frameworks. Future research must prioritize establishing comprehensive frameworks for AI integration, fostering collaboration among diverse stakeholders (academia, regulators, providers, policymakers), and continually improving data quality and accessibility.

Analogy: Integrating AI into HTA is like giving an architect a high-powered supercomputer capable of modeling not just the structure of a building (the technology) but also the intricate flow of traffic, utility costs, and user behavior in real-time, allowing for optimized, cost-effective, and equitable designs before the first brick is laid.

Reference: Ramezani, M., Bakhtiari, A., Daroudi, R., Mobinizadeh, M., Fazaeli, A. A., Olyaeemanesh, A., Rabiee, H. R., Ramezani, M., Mostafavi, H., Sazgarnejad, S., Bordbar, S., & Takian, A. (2025). Applications of artificial intelligence and the challenges in health technology assessment: a scoping review and framework with a focus on economic dimensions. Health Economics Review, 15(46). https://doi.org/10.1186/s13561-025-00645-4.