Integrating AI Tools Into Clinical Practice for Genetic Counselors

This is the second part of an interview with Colleen Caleshu, MS, CGC. For the first part, click here.

Artificial intelligence (AI) has become a major topic in medicine in recent years, and the field of genetic counseling is no exception. At the 44th National Society of Genetic Counselors (NSGC) Annual Conference, held November 6 to 10 in Seattle, Washington, Colleen Caleshu, MS, CGC, the senior director of research and real world data at Genome Medical, chaired a session focused on AI in genetic counseling called “Healthcare Administrative AI Tools in Practice: Implementation Insights from Early Adopters.”

At the conference, CGTLive® interviewed Caleshu to get her insight on practical considerations for genetic counselors who may be considering implementing AI tools into their practice. Caleshu emphasized the importance of AI literacy and being transparent with patients, and also discussed the question of whether AI ma eventually replace genetic counselors.

CGTLive: If someone hasn’t integrated AI into their practice, what are some of the things they need to know before doing that?

Colleen Caleshu, MS, CGC: I think if you're interested in integrating AI into your practice, and you're not sure where to start, talking to colleagues who have already done so is a great place, and that doesn't have to be just genetic counselors. It can also be other people at your institution. Many hospitals will have folks whose job it is to help implement these things and to support providers in doing so, and you can learn from them.

There was another great session at the conference. Attendees here can see the recordings and genetic counselors who didn't attend can buy the recordings. It was another AI session that was on AI literacy, a lot of basics in assessing quality. It was a great place to start, too.

How do you talk to patients about the implementation of AI?

Everything we know about AI in medicine is so early, but because it's such a hot topic, the early research studies get a lot of attention. I think we need to be careful to keep doing studies to learn more about that and many other things. But one thing that has come out of early research studies is that people very much want transparency. They want to know. If you write a message to your genetic counselor, and the response was written by AI, they want to know. There's sort of different stances on that right now in what both providers and hospitals are doing, but transparency and consent at this stage of the game is really important. For example, for the solutions that will listen to the discussion between the patient and the clinician and then write the clinic note on it, there are states that require consent if a patient is going to be recorded, and there are states that don't. But above what the law requires at this stage of the game, it seems pretty clear that it's important that we get consent from an ethical and trust and relationship-building standpoint. Patients need to know and have a choice of whether or not that AI solution is being used to record and document their appointment.

The other thing I would actually add, is it's also clear that a subset of patients are very eager for this to be part of their care and want it now, especially as the adoption of these tools in people's personal lives has just taken off. I mean, there's no technology that's been adopted this fast, at least according to some reports. I was actually at a National Academies of Sciences, Engineering, and Medicine workshop on AI and genomics last week, and we had a patient speaker. His son has a genetic disease, and he was really sharing that it really feels like to get quality care right now he wants his son's doctors to be leveraging all the tools that are available available to them, including generative AI tools that might help them do a better job or not miss things they would otherwise miss. And I've seen that sentiment elsewhere. I think we need to be responsive to it as healthcare providers.

Do you think AI will ever replace genetic counselors?

The conversation that we had was really focused on what are sometimes called healthcare administrative tools, or things that help with documentation and scheduling and communication, etc. There's obviously conversations across all different fields these days about AI replacement, and certainly in genetic counseling I think many people think about whether AI tools could replace genetic counselors. I find that to be a very polarizing question. People are either very extremely enthusiastic (often nongenetic counselors) or very worried.

I personally am more of a middle ground person because we know that not enough people have access to genomic medicine. We know from research studies that not everyone who needs a genetic test or genetic risk assessment necessarily needs counseling by a counselor, and we know that many people who do need counseling by a counselor aren't accessing it right now. So I think that we're in a position to lead as genetic counselors on how we develop tools that can meet patient needs when maybe the patient needs are such that a tool is enough, and ways that we can identify the patients who really need counseling by a genetic counselor, so we make sure that they get care with a genetic counselor because of that higher level of need.

This transcript has been edited for clarity.