Use GPT-J Prompt Generation with RoBERTa for NER Models on Diagnosis Extraction of Periodontal Diagnosis from Electronic Dental Records

Chuang, Yao-Shun, Jiang, Xiaoqian, Lee, Chun-Teh, Brandon, Ryan, Tran, Duong, Tokede, Oluwabunmi, Walji, Muhammad F.

arXiv.org Artificial Intelligence 

The extent is indicated by the percentage of teeth affected by periodontitis at the identified stage. Grading depends on the risk of disease progression associated with the history of disease progression, local and systemic factors. Despite the introduction of new diagnostic terms for periodontal diseases, dental care providers might not be acquainted with them due to the complexity of this new system. This results in clinical documentation lacking accurate and structured diagnosis, or in some cases, no diagnosis being recorded. Inadequate periodontal diagnoses poses a significant threat to patient care quality. An accurate diagnosis is key to the provision of appropriate patient care, outcome assessment and quality improvement efforts. This, in turn, may hinder future care providers from evaluating the patient's condition precisely and providing optimal treatment. Electronic dental records (EDR) have become widely adopted in dental care, providing an opportunity to address the issue of missing diagnoses. EDRs include comprehensive information on a patient's history, clinical examination, diagnosis, treatment, and prognosis