IADR Abstract Archives
Individualized Application of AI Combined With Micro-Invasive Caries Treatment was Cost-Effective
Objectives: There is uncertainty around the cost-effectiveness of artificial intelligence (AI) for medical applications. We assessed the value of information towards relevant input parameters (like AI accuracy, costs, population caries risk profile).
Methods: The health intervention (AI) was a trained Convolutional Neural Network to classify bitewing radiographs. We employed a Markov model and microsimulation to quantify the value of information in a mixed public-private-payer perspective in German healthcare. Our health outcome was tooth retention years.
Results: In the base-case scenario (uncertain accuracy, uncertain risk profile of the population, uncertain costs of the AI), AI was more effective (tooth retention for a mean [2.5-97.5%] 62.8 [59.2-65.5] years) and less costly (378 [284-499] Euro) than dentists without AI (60.4 [55.8-64.4] years; 419 [270-593] Euro). The economic value of reducing the uncertainty around the AI’s accuracy or costs was limited, while information on the risk profile of the population was more relevant. In high-risk populations, using an AI followed by micro-invasive therapy (e.g., resin infiltration) was highly cost-effective.
Conclusions: Individualized application of AI for caries detection may optimize cost-effectiveness.
Methods: The health intervention (AI) was a trained Convolutional Neural Network to classify bitewing radiographs. We employed a Markov model and microsimulation to quantify the value of information in a mixed public-private-payer perspective in German healthcare. Our health outcome was tooth retention years.
Results: In the base-case scenario (uncertain accuracy, uncertain risk profile of the population, uncertain costs of the AI), AI was more effective (tooth retention for a mean [2.5-97.5%] 62.8 [59.2-65.5] years) and less costly (378 [284-499] Euro) than dentists without AI (60.4 [55.8-64.4] years; 419 [270-593] Euro). The economic value of reducing the uncertainty around the AI’s accuracy or costs was limited, while information on the risk profile of the population was more relevant. In high-risk populations, using an AI followed by micro-invasive therapy (e.g., resin infiltration) was highly cost-effective.
Conclusions: Individualized application of AI for caries detection may optimize cost-effectiveness.