Chatbot comparable with counseling for helping patients with genetic testing decisions
Click Here to Manage Email Alerts
The use of chatbot tools may be a valuable part of health management strategies to improve access to cancer genetic services, according to study results.
The randomized BRIDGE trial included 3,073 people (median age, 43.8 years; range, 25-60; 72.9% women; 75% white) who had a primary care visit in the prior 3 years and were eligible for cancer genetic evaluation. No study participants had a cancer history or had received prior cancer genetic testing.
The trial took place between August 2020 and August 2023 at University of Utah Health and NYU Langone Health.
Researchers randomly assigned 1,554 participants to a chatbot intervention through which they received an invitation via a patient portal outreach message to participate in a pretest genetics education chat.
The other 1,519 participants received enhanced standard of care, which included an invitation to schedule a pretesting appointment with a trained genetic counselor.
Completion of pretest cancer genetic services via a genetics education chat with chatbots or genetic counseling appointments, as well as completion of genetic testing, served as the primary outcomes.
Results showed estimated percentage point differences between groups of 2 (95% CI, 1.1 to 5) for completing pretest cancer genetic testing and 1.3 (95% CI, 3.7 to 1.1) for completing genetic testing, demonstrating equivalence in primary outcomes.
“We were trying to determine whether there are ways to automate pieces of the genetic counseling workflow to make genetic counselors’ time more efficient and focused on areas that need their specialized training,” Kimberly A. Kaphingst, ScD, professor in the department of communication at The University of Utah and director of cancer communication research at Huntsman Cancer Institute, told Healio. “Some of the pretest genetics education can be more standardized and potentially automated. We wanted to see if we could supplement the genetic counseling workforce to meet some of this increasing demand.”
Healio spoke with Kaphingst about need to identify patients who might benefit from genetic testing, and how chatbot tools or other approaches may enhance access to and uptake of these services.
Healio: What motivated you to conduct this study?
Kaphingst: There is an increasing demand for cancer genetic testing. More people are meeting the criteria and there aren’t enough genetic counselors. This is particularly true in our region in the rural/frontier areas. In our case, for example, we identified 22,000 patients in two health care systems who could potentially benefit from cancer genetic testing. That’s very difficult given the current genetic counseling workforce.
Healio: How did you conduct the study?
Kaphingst: We developed a clinical decision support algorithm that reviewed or mined the data in the electronic health records in both systems. We identified patients who met specialty National Comprehensive Care Network criteria, and drew random samples of about 3,000 patients. After random assignment, we provided outreach either through a link to the chat or an invitation to call and schedule the genetic counseling appointment. Our study was unique in that it wasn’t referral. It was active outreach to patients.
Healio: What findings stood out to you?
Kaphingst: We tested a hypothesis of equivalence analysis, which means we weren’t testing whether the chatbot was better. We wanted to know if we would essentially see the same service delivery outcomes between the two. Primary outcomes were equivalent between groups. About one-quarter of people in both groups went on to do pretest services, and about 12% to 13% of those who were contacted went on to do genetic testing. This was essentially the same between the two groups. This is important because it shows this is a tool that could be used without affecting people’s ability to get cancer genetic services.
Healio: What barriers do you think prevented individuals in your study from pursuing genetic testing?
Kaphingst: I think it probably wasn’t the right time for some people. Maybe they had competing priorities. Some people didn’t open their MyChart messages. We set reminders and contacted them by phone, but they may never have seen the information.
People also were concerned about the cost. Even though insurance covers most of it, there were still concerns about copays. Perhaps people don’t realize that costs have come down substantially. Once we were able to contact patients and talk with them about cost and insurance coverage, they felt reassured.
Healio: What are your next steps in research?
Kaphingst: We’re planning to start integrating some of the chatbot tools into the genetic counseling workflow here. We are also starting to partner with other health care systems to test the tools in their systems, as well. We think this will allow greater outreach to patients, and hopefully it will advance the goal of increasing access to cancer genetic testing by making the genetic counseling workflow more scalable with automated tools.
References:
- Kaphingst KA, et al. JAMA Netw Open. 2024;doi:10.1001/jamanetworkopen.2024.32143.
- University of Utah Health. Can chatbots help with genetic testing for cancer risk? (press release). Available at: https://healthcare.utah.edu/huntsmancancerinstitute/press-releases/2024/09/can-chatbots-help-genetic-testing-cancer-risk. Published Sept. 9, 2024. Accessed Oct. 9, 2024.
For more information:
Kimberly A. Kaphingst, ScD, can be reached at kim.kaphingst@hci.utah.edu.
Collapse
Healio Interviews
Click Here to Manage Email Alerts