Yvonne W Leung, Jeremiah So, Avneet Sidhu, Veenaajaa Asokan, Mathew Gancarz, Vishrut Bharatkumar Gajjar, Ankita Patel, Janice M Li, Denis Kwok, Michelle B Nadler, Danielle Cuthbert, Philippe L Benard, Vikaash Kumar, Terry Cheng, Janet Papadakos, Tina Papadakos, Tran Truong, Mike Lovas, Jiahui Wong
{"title":"The Extent to Which Artificial Intelligence Can Help Fulfill Metastatic Breast Cancer Patient Healthcare Needs: A Mixed-Methods Study.","authors":"Yvonne W Leung, Jeremiah So, Avneet Sidhu, Veenaajaa Asokan, Mathew Gancarz, Vishrut Bharatkumar Gajjar, Ankita Patel, Janice M Li, Denis Kwok, Michelle B Nadler, Danielle Cuthbert, Philippe L Benard, Vikaash Kumar, Terry Cheng, Janet Papadakos, Tina Papadakos, Tran Truong, Mike Lovas, Jiahui Wong","doi":"10.3390/curroncol32030145","DOIUrl":null,"url":null,"abstract":"<p><p>The Artificial Intelligence Patient Librarian (AIPL) was designed to meet the psychosocial and supportive care needs of Metastatic Breast Cancer (MBC) patients with HR+/HER2- subtypes. AIPL provides conversational patient education, answers user questions, and offers tailored online resource recommendations. This study, conducted in three phases, assessed AIPL's impact on patients' ability to manage their advanced disease. In Phase 1, educational content was adapted for chatbot delivery, and over 100 credible online resources were annotated using a Convolutional Neural Network (CNN) to drive recommendations. Phase 2 involved 42 participants who completed pre- and post-surveys after using AIPL for two weeks. The surveys measured patient activation using the Patient Activation Measure (PAM) tool and evaluated user experience with the System Usability Scale (SUS). Phase 3 included focus groups to explore user experiences in depth. Of the 42 participants, 36 completed the study, with 10 participating in focus groups. Most participants were aged 40-64. PAM scores showed no significant differences between pre-survey (mean = 59.33, SD = 5.19) and post-survey (mean = 59.22, SD = 6.16), while SUS scores indicated good usability. Thematic analysis revealed four key themes: AIPL offers basic wellness and health guidance, provides limited support for managing relationships, offers limited condition-specific medical information, and is unable to offer hope to patients. Despite showing no impact on the PAM, possibly due to high baseline activation, AIPL demonstrated good usability and met basic information needs, particularly for newly diagnosed MBC patients. Future iterations will incorporate a large language model (LLM) to provide more comprehensive and personalized assistance.</p>","PeriodicalId":11012,"journal":{"name":"Current oncology","volume":"32 3","pages":""},"PeriodicalIF":2.8000,"publicationDate":"2025-03-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11940934/pdf/","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Current oncology","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.3390/curroncol32030145","RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"ONCOLOGY","Score":null,"Total":0}
引用次数: 0
Abstract
The Artificial Intelligence Patient Librarian (AIPL) was designed to meet the psychosocial and supportive care needs of Metastatic Breast Cancer (MBC) patients with HR+/HER2- subtypes. AIPL provides conversational patient education, answers user questions, and offers tailored online resource recommendations. This study, conducted in three phases, assessed AIPL's impact on patients' ability to manage their advanced disease. In Phase 1, educational content was adapted for chatbot delivery, and over 100 credible online resources were annotated using a Convolutional Neural Network (CNN) to drive recommendations. Phase 2 involved 42 participants who completed pre- and post-surveys after using AIPL for two weeks. The surveys measured patient activation using the Patient Activation Measure (PAM) tool and evaluated user experience with the System Usability Scale (SUS). Phase 3 included focus groups to explore user experiences in depth. Of the 42 participants, 36 completed the study, with 10 participating in focus groups. Most participants were aged 40-64. PAM scores showed no significant differences between pre-survey (mean = 59.33, SD = 5.19) and post-survey (mean = 59.22, SD = 6.16), while SUS scores indicated good usability. Thematic analysis revealed four key themes: AIPL offers basic wellness and health guidance, provides limited support for managing relationships, offers limited condition-specific medical information, and is unable to offer hope to patients. Despite showing no impact on the PAM, possibly due to high baseline activation, AIPL demonstrated good usability and met basic information needs, particularly for newly diagnosed MBC patients. Future iterations will incorporate a large language model (LLM) to provide more comprehensive and personalized assistance.
期刊介绍:
Current Oncology is a peer-reviewed, Canadian-based and internationally respected journal. Current Oncology represents a multidisciplinary medium encompassing health care workers in the field of cancer therapy in Canada to report upon and to review progress in the management of this disease.
We encourage submissions from all fields of cancer medicine, including radiation oncology, surgical oncology, medical oncology, pediatric oncology, pathology, and cancer rehabilitation and survivorship. Articles published in the journal typically contain information that is relevant directly to clinical oncology practice, and have clear potential for application to the current or future practice of cancer medicine.