AI's Future Role in Virtual Care Management

Artificial Intelligence is everywhere right now. When we look back, I believe we'll reflect on AI as being a revolutionizing force, but one that spawned quite a few distractions along the way. As with any new technology, we're all going to have to make some mistakes and pursue some dead-ends before we find all of the places that this fits.

Healthcare is no exception.

AI isn't new, nor are claims that it is on the precipice of revolutionizing healthcare. If this wave of excitement is going to be anything but hype, it will be because this newest generation of astonishingly capable models have reached a point where they can start to perform tasks that never before could have been put in the hands of AI.

When it comes to Care Management, many are saying that this is the time for AI to make its entry into the list of must-have resources in a care manager's toolbelt.

In this article, we'll lay out some of the ways that Care Management is likely to be impacted by AI in the coming years.

Documentation tasks will decrease, and eventually disappear

This new generation of models is great at summarization. You might have already seen AI assistants joining meetings to take notes, or Google surfacing an AI-generated answer to a question you searched.

Applied to Care Management, this same capability means that having to answer the question "what was discussed on that call?" will soon be a thing of the past. Imagine a world where Care Managers were judged, and where providers could bill, based not off of what got documented properly, but on the true content of their interactions with their patients?

AI's ability to digest, track, and summarize complex conversations means that it can keep track of the facts of a case with the same ease that a Care Manager might talk about them with their patient. AI has the ability to create large, freeform outputs in very specific formats & styles, meaning it can then express those facts that it's capturing and express them in the form of whatever notes payload you need.

That's not to say it's easy to create AI systems that are performant in this respect, but these use cases are attainable, and the healthcare market is awash in various vendors seeking to automate all parts of the medical documentation problem.

Patient Charts will become easier to navigate

Patient health records are a long, complex web of data. In programs such as Remote Patient Monitoring (RPM) and Chronic Condition Management (CCM), profit margins live and die by a very exacting stopwatch. It's not unusual to have a Care Manager spend upwards of 7 minutes prepping for the average patient call. These minutes add up quickly, limiting the number of patients that can be served.

Additionally, it's easy to miss some key piece of data that leads to something being missed - fail to notice a mention of your patient's spouse being sick and you may not know that they'll need a ride to their next appointment.

Generative AI is very good at ingesting & resurfacing insights from complex datasets. Given the right instructions, these models can be used to create a health record that you can have a conversation with. Want to know when your diabetic patient last had an eye exam? These models can tell you that they don't see a corresponding record in the EHR, but that there was a discussion on a call last month mentioning that they received one at an out-of-network clinic.

This sort of AI makes it possible to automate the "hunting and pecking" involved in answering questions about patient's health data. It also means that a care manager can walk into a call with a patient with a summary of the relevant health information for that patient, without the tedious prep time.

Live Guidance Will Mean Fewer Misses

Perhaps the cruelest irony of the Care Management process is that the more complex the patient is, the more likely Care Managers are to miss something. Whether it be forgetting to ask a key question, remember a key fact, or take a key note - the more important it is to get everything right, the greater the chance we get something wrong.

AI opens up the possibility of adding a second qualified Care Manager to each call. It can remind teams of best practices, capture information, and connect Care Managers with resources they can use to improve the care they offer to their patients. All in realtime.

These "live assistants" carry a number of benefits. They ease change management, enabling the latest best practices to be more easily disseminated through an organization. They reduce the threshold to which individual skill is needed to service patients at an expert level, helping newer team members make fewer mistakes. They're a rising tide that lifts all boats.

Monitoring Populations of Patients will be Automated

APCM, and other value-based care programs, are built on a foundation of ensuring that large populations of patients are receiving quality care. Without the underlying fee-for-service structure, caring for those patients can no longer be done effectively or affordable by way prophylactic outreach.

Take the new (TK GPCM1 code) code. With their relatively low reimbursement rate, a Care Management team cannot afford to spend any significant time on the average patient, let alone call each one in a given month. Something needs to be introduced into this equation in order to help our teams scale.

A generative model can quickly monitor large swaths of patient data, identifying the patients that most need help. With enough data, you can get realtime visibility into the needs of any particular patient, ensuring that every phone call your patients receive has a specific purpose, and that patients come away from each interaction with your team feeling like their lives have been made easier.

Algorithmic Patient Engagement Will Deepen and Widen

We've all probably seen chat as a medium through which Generative AI operates. Though much of healthcare will continue to be human-in-the-loop, given the medical importance of the tasks being performed, Chat-driven AI still has its place. Doctors, nurses, and care managers can't be available for patients 24/7, but an AI Chat Agent can.

We've encountered many teams that are already using some sort of programmed patient outreach technology. These deployments are often limited in their scope due the manual nature with which new use cases must be meticulously configured. With Generative AI, this is no longer a requirement.

Though it's possible to be incredibly specific in your instruction, a GenAI agent can scale to encompass many more aspects of patient interactions than could be easily set up and managed manually. It can answer a wider range of questions, it can speak with a human voice, and it can be instructed to take follow-up actions, such as requesting medical documents from other providers or requesting a medication refill.

Taking Action

If you're excited about the potential of this future, we'd love to talk to you. We at Earshot Health are on the forefront of putting these innovations into action. Click the link below and book a free demo of our revolutionary AI Care Management platform.