Abstract P6-08-39: Hereditary cancer risk assessment using a chatbot in women presenting to obstetrics and gynecology practices across the U.S.

Abstract

Abstract Background: Hereditary cancer risk assessment is standard of care in the obstetrics and gynecology (ob/gyn) practice. Historical data indicates 1 in 12 individuals have a family history consistent with hereditary cancer(1), however recent data from a small community practice found that 1 in 4 women met National Comprehensive Cancer Network (NCCN) criteria for genetic testing(2). The aim of this study was to assess the number of women who meet NCCN criteria for genetic testing in a diverse population across the United States (US) by using a computer program that conducts conversations with patients (a.k.a. chatbot) for risk assessment(3). Method: We partnered with Clear Genetics Inc. (San Francisco, CA) to use a HIPAA compliant chatbot for collecting personal and family history of cancer from women in 28 ob/gyn practices across the US. Patients received a text message or email asking them to complete the chatbot five days before their scheduled appointment. Reminders were sent at three days and again one day prior if not completed. After history collection was complete, an algorithm determined if the patient had a known familial variant or met NCCN guidelines for hereditary cancer testing, including BRCA-Related Breast and/or Ovarian Cancer, Lynch, and Polyposis syndromes(3). Incomplete chats, minors under age 18 or those who declined to provide information were excluded. Additionally, some ob/gyn practices elected to exclude pregnant patients. Results: Over 15,000 chatbots were sent to patients. Overall, 65% of patients completed the chatbot and 26% of these patients met NCCN criteria or had a known familial variant(3). 14.5% of completed assessments were excluded (1.2% declined, 1.4% minors and 11.9% pregnant patients). Some patients received the chatbot less than five days before their appointment due to administrative delays at the ob/gyn practice, and 15% of intended recipients did not receive the chatbot due to incorrect contact information. Patients reviewed their experience with an average rating of 4.6 out of 5. Conclusions: A novel chatbot tool was used to collect pertinent cancer history and provide NCCN criteria assessment to identify patients for inherited cancer risk. This study population was nearly four times the size of the most recent study identifying 1 in 4 women met NCCN criteria for genetic testing. In addition, this study was more diverse - examining multiple ob/gyn practices across the US compared to a single community practice. Results of this study were consistent with newly reported data indicating 1 in 4 patients meet NCCN criteria for hereditary cancer testing. There is a need for increased education and tools to help ob/gyn practices identify these patients consistent with ACOG Practice Guideline 634(4), in light of a higher number of patients who need genetic testing.