The Telepsychiatry Advantage: Episode 5

3 Ways AI is Making Psychiatry Better

Hello everyone, and welcome back to the Telepsychiatry Advantage. My name is Doctor Edward Kaftarian and I’m the chairman and CEO of Orbit Health Telepsychiatry. 

Today we’re going to take a look at something that may feel scary, but is also a great opportunity. I’m talking, of course, about artificial intelligence or AI. 

 

The Luddite Lesson

Now this is a topic that’s on everyone’s minds, obviously. And whenever I hear discussions about AI, I hear half the people really excited about it and half of us really terrified. And I get it. Technology can be intimidating, it can be scary because there are lots of implications to new technology. And this fear is not new. In fact, it reminds me of a story in the early 19th century at the dawn of the Industrial Revolution. At the time, there was a group of highly skilled English textile workers and weavers, and they were called the Luddites. And for generations they were the undisputed masters of their craft. But then automation came, and there was a loom that was invented, and suddenly the artisans were terrified. They looked at the new machinery and believed it was going to make their hard earned skills completely obsolete, and it would destroy their livelihoods.

So what did they do? They formed a radical faction. They called themselves, as I said, the Luddites, and they started breaking into factories in the dead of night, literally smashing the machines with sledgehammers, hoping that they could stop the future from arriving. But this didn’t work. And history has taught us a hard lesson. Smashing the machines never works. The Industrial Revolution happened anyway, and it’s good that it happened. The weavers who ultimately survived and thrived were the ones who put down the sledgehammers and learned how to operate the new looms. So the best way to survive a technology revolution is to embrace it. 

As psychiatrists, AI is our automated loom. We can’t afford to be modern day Luddites bearing our heads in the sand, and we need to understand the technology so that we can help navigate it. We can help guide it into the future so it can be used responsibly. If we leave AI entirely in the hands of Silicon Valley engineers, we’re going to lose the clinical nuance, empathy, and ethical boundaries that make our field effective. And we need to take a seat at the table. We need to ensure that our specialty stays relevant.

We also need to protect the integrity of the doctor patient relationship, reduce our own burnout, and ultimately drive better patient care outcomes. And we have to own this transition. We have to be part of this change. We have to embrace AI but also guide it. 

Now let me set the record straight right now. I don’t think that AI is going to replace psychiatrists. I think that people need that human connection and there still is a role for clinical judgment, and this can’t all be substituted by a machine. 

You know, for example, if you’re on the phone and you want help, you often will hear a teleprompter and you will be now steered toward AI. And it’s very frustrating, and it’s actually not good for business, for businesses to always push people to AI and to have a machine talk to them. But AI can be a fundamental tool and part fundamental, part of delivering care better if we integrate it and incorporate it. 

Let me be clear. The AI is not going to be the practitioner. It’s going to be the copilot. And in telepsychiatry, where we rely heavily on digital interfaces, AI is uniquely positioned to solve some of our biggest systemic headaches. 

 

The AI Copilot: Scribes & EMR Burnout

So let’s go over three ways that AI is actively influencing the future of tele psychiatry. 

Number one ambient AI scribes. 

So if you ask any psychiatrist what causes them the most burnout, the answer is almost universally the same. The electronic medical record documentation. It takes hours, and it pulls our eyes away from patient care and the paperwork of it all. And that’s not something that’s unique to telepsychiatry that’s common in all of medicine. The paperwork is the part that the doctors are most frustrated with and don’t like. And I’ve seen in my career telepsychiatrist done really poorly because of bad technical systems like clunky systems, clunky EMR, and gradually this has improved to the point at which telepsychiatry can be done well, easily with the right tools.

And right now, we’re seeing the adoption of AI tools that make things a lot easier for the telepsychiatrist. So there are tools that can securely listen to the telepsychiatry session in real time, and pick out the clinically relevant details and automatically generate a highly accurate, structured psychiatric progress note. It drafts the history of the present illness, the mental status exam, the assessment, and the plan, all while you can maintain unbroken eye contact with the patient through the camera.

And with Orbit Health, we do have some of these tools that are available. But I want to give you a caveat, it has to be done securely, and there are rules around it that need to be maintained and abided by. So you simply review the note once the AI tool has generated the note. You can make some minor edits. You make sure that it’s correct, clinically accurate, and its good judgment, and then you’re done. It saves you time. It also allows you to spend more time talking to the patient, less time charting. And so I believe that these AI scribes are going to be a bigger part of telepsychiatry and health care in general, and they’re going to integrate with the EMRs so that the job of the doctor is going to be a lot more pleasant. So it’s going to make our jobs better, not worse. 

 

The Digital Biomarker: Natural Language Processing

Number two diagnosis. 

So the diagnosis, the copilot. And now we’re talking about natural language processing and biomarkers. So let me explain. Psychiatry can be a subjective field. We don’t have blood tests for depression or x-rays for schizophrenia. We rely on things like our observation of behavior, the affects, speech. You know the drill. But AI is starting to provide us with more objective and measurable data. And researchers are using natural language processing, or NLP to analyze actual speech patterns of patients during virtual sessions. Studies show that AI can detect minute changes in vocal tone, cadence, and word choice that might indicate the early onset of a manic episode, or a slip into a psychosis, sometimes before the patient and the clinician fully realize it. 

So, for example, AI can analyze speech coherence and semantic density, fancy word, to show an accuracy in predicting the transition to clinical psychosis in high risk individuals. So a lot of this research is ongoing and in the works. But we do know that, you know, as psychiatrists, we do know that speech is really important. And the changes in the tone of the speech can signify the beginning of a worsening episode of depression or mania or psychosis. So imagine having an AI assistant running quietly in the background of your telepsychiatry platform, flagging you on your screen. The patient’s speech velocity has increased 15% from baseline, so consider assessing for mania. And then now we are going to pay attention more to potential signs and symptoms of mania. You might ask the patient different questions based on some of this feedback. And that’s going to be fantastic. In fact, I think that’s going to be a lot of fun to be working with the machine in real time to obtain the best information possible and act on it with a great assessment and plan. 

 

Precision Triage: Solving Systemic Headaches

Number three operational efficiency. 

So I’ve spent years working at the intersection of psychiatry and technology, partnering with psychiatrists, county behavioral health departments, and community mental health systems across the country and specifically in California. When you’re dealing with the public sector and community settings, the sheer volume of patients can be really overwhelming, and triaging patients accurately is a massive logistical challenge. You want to know which patients require the most attention at the right time. AI is stepping in to solve this by analyzing patient intake forms, past EMR data, and self-reported symptoms to intelligently route patients to the right level of care.

For example, is this a routine medication follow up, or is this an acute crisis that requires immediate evaluation by a very experienced telepsychiatrist? So by using AI to streamline these operational workflows, we see tele psychiatry can be done really well, workflows that work, clinicians feeling supported, patients getting consistent, high quality care at the right time. And it ensures that our most valuable resource, our time and clinical expertise, is deployed at exactly the right time.

 

Conclusion: A Seat at the Table

The goal of The Telepsychiatry Advantage is simple: raise the standard for clinicians, for systems, and ultimately for patients. And embracing AI isn’t about cutting corners or replacing psychiatrists. It’s about elevating our practice. It handles the administrative heavy lifting so that we can focus on the human element of healing. And I think in the immediate future, we’re going to see a I improving our ability to triage patients and to operationally streamline things.

So whether you’re already practicing remotely, thinking about starting or building telepsychiatry programs in your system, I’m glad you’re here. The future is bright and it’s augmented. We shouldn’t worry about AI replacing us. What we should do is spend time getting involved not only in embracing the AI, but also guiding it and helping make sure that patient care is quality.

Part of guiding it is going to be speaking up when we’re talking about regulatory changes and guidelines and rules and regulations around AI and how it interfaces not only with psychiatry, but also with all of healthcare. And that’s going to be really important that we have a seat at the table so that it’s not just Silicon Valley that is dictating our future, but it’s our experience, our care about the patient, our empathy and the human connection that’s all going to be integrated with AI to provide the best patient experience possible. 

So thank you very much for joining me today. I hope you got a lot out of this episode, and I encourage you to embrace AI, but keep things at a very high standard. Thank you very much and we’ll see you next time.