講演情報
[O21-3]AI Chatbotを活用した遺伝性乳がん卵巣がん症候群拾い上げシステムの臨床研究
○佐藤 杏1, 羽田 恵梨1, 成松 宏人1,2,3 (1.神奈川県立がんセンター 遺伝診療科, 2.神奈川県立がんセンター 臨床研究所 がん予防・情報学部, 3.神奈川県立保健福祉大学 ヘルスイノベーション研究科)
Purpose
To validate the medical accuracy of a chatbot-based preliminary screening system for Hereditary Breast and Ovarian Cancer (HBOC) by using it in the patient population that actually visits hospital.
Methods
We developed a chatbot-based preliminary screening system for HBOC, i.e., the taking of medical and family history and determining whether the patients meet the BRCA1/2 testing criteria of the NCCN guidelines. We conducted a test operation on the patients who consecutively visited Kanagawa Cancer Center.
Results
A total of 11 participants were enrolled between October and December 2020. In a preliminary survey, 6(54.5%) had never heard of chatbot and 7(63.6%) had never used it. All the participants were able to complete the chatbot operation, and the average time required to the operation was 18.0 minutes (±SD5.44). The system’s determination of whether the participants met the criteria based on the information from the chatbot were all consistent with that by medical specialists. We compared the medical histories obtained from the participants by certified genetic counselors (CGC) with those by the chatbot. Of the 11 participants, 3 entered different information from that obtained by CGC. This was caused by forgetting to enter medical history by participants or communication errors with the chatbot. No discrepancies were found between the family histories by CGC with those by the chatbot.
Conclusion
The study demonstrated the medical accuracy of the chatbot-based preliminary screening system for HBOC in actual patients, indicating its usefulness in clinical practice.
To validate the medical accuracy of a chatbot-based preliminary screening system for Hereditary Breast and Ovarian Cancer (HBOC) by using it in the patient population that actually visits hospital.
Methods
We developed a chatbot-based preliminary screening system for HBOC, i.e., the taking of medical and family history and determining whether the patients meet the BRCA1/2 testing criteria of the NCCN guidelines. We conducted a test operation on the patients who consecutively visited Kanagawa Cancer Center.
Results
A total of 11 participants were enrolled between October and December 2020. In a preliminary survey, 6(54.5%) had never heard of chatbot and 7(63.6%) had never used it. All the participants were able to complete the chatbot operation, and the average time required to the operation was 18.0 minutes (±SD5.44). The system’s determination of whether the participants met the criteria based on the information from the chatbot were all consistent with that by medical specialists. We compared the medical histories obtained from the participants by certified genetic counselors (CGC) with those by the chatbot. Of the 11 participants, 3 entered different information from that obtained by CGC. This was caused by forgetting to enter medical history by participants or communication errors with the chatbot. No discrepancies were found between the family histories by CGC with those by the chatbot.
Conclusion
The study demonstrated the medical accuracy of the chatbot-based preliminary screening system for HBOC in actual patients, indicating its usefulness in clinical practice.