Using AI large language models to assess dental history in systemic conditions

Abstract

Technological advancements, particularly in artificial intelligence (AI), are transforming the field of dentistry. AI—including machine learning (ML) and deep learning (DL)—mimics human cognitive processes to enhance diagnostics, treatment planning, and patient care. This study aimed to develop an AI-driven tool for the more effective and efficient evaluation of patients’ dental histories and to compare the time required between AI-assisted and conventional methods. HistorAI analyzes patient anamnesis forms and generates comprehensive reports. A 22-item anamnesis questionnaire, covering both oral and systemic health, guided the structured prompting of AI models (GPT-4 and Gemini). GPT-4 was integrated via an Application Programming Interface (API) to analyze data, provide treatment suggestions, generate prescriptions, and recommend referrals. Evaluation times and outcomes were compared between AI-assisted and conventional methods using descriptive statistics and independent-samples t-tests, with effect sizes calculated using Cohen’s d, and significance set at p < 0.01. HistorAI successfully evaluated medical and dental histories, identified contraindicated medications and anesthetics, assessed patient complaints, and provided preliminary treatment recommendations. The AI-assisted process significantly reduced the time required to complete dental history assessments compared with conventional methods (p < 0.01). A Cohen’s d of 2.599 indicates a substantially higher efficiency for the AI-assisted group. The AI-powered tool enhanced efficiency and clinical decision-making in dental practice while maintaining clinician oversight. Further clinical validation and careful consideration of ethical implications are essential to ensure the safe and responsible integration of AI into dental workflows.

Ähnliche Arbeiten