Cost-effectiveness of AI for caries detection: randomized trial

Area of Science:

• Dental diagnostics
• Artificial intelligence in healthcare
• Health economics

Background:

• Proximal caries detection is crucial for timely intervention.
• AI tools offer potential for improved diagnostic accuracy in dentistry.
• Evaluating the economic impact of AI in clinical practice is essential.

Purpose of the Study:

• To assess the cost-effectiveness of AI-supported proximal caries detection compared to traditional methods.
• To analyze the impact of AI on treatment decisions and long-term patient outcomes.
• To determine the economic viability of AI in dental diagnostics from a healthcare payer perspective.

Main Methods:

• A randomized controlled clustered cross-over trial involving 23 dentists evaluating 20 bitewings with and without AI support.
• Evaluation of detection accuracy (true/false positives/negatives) and assigned treatment decisions (non-invasive, micro-invasive, invasive).
• A Markov simulation model populated with trial data to assess lifetime cost-effectiveness and tooth retention, using a mixed public-private-payer perspective in German healthcare.

Main Results:

• AI-supported detection was significantly more sensitive but resulted in more invasive treatments.
• Both AI and no-AI groups demonstrated identical effectiveness in tooth retention (mean 49 years) and nearly identical costs (approx. 330 Euro).
• Cost-effectiveness remained uncertain, with no significant advantage for AI regardless of willingness-to-pay thresholds.

Conclusions:

• Increased accuracy of AI in caries detection did not translate to superior cost-effectiveness due to a trend towards more invasive treatments.
• The economic benefits of AI in dental diagnostics are limited if treatment decisions are not optimized.
• Integrating AI for both detection and subsequent treatment management could enhance its cost-effectiveness.