See your Eleos ROI with our new calculator.

Try Now

I had the privilege of attending the National Council for Mental Wellbeing annual conference, NatCon23, this month. The insights, trends, and dialogues I encountered at the conference were enriching and served as a testament to the dynamic nature of behavioral health.

As I reflect on the myriad of conversations and presentations, one key lesson reverberates above all: the convergence of technology and behavioral health services is not only inevitable, but also absolutely necessary for the evolution of this field.

Cloud-based EHR adoption in behavioral health has lagged behind other areas of care.

The adoption of electronic health records (EHRs) in the behavioral health sector has been considerably slower compared to other healthcare areas. In 2020, only 55% of behavioral health providers had adopted an EHR, compared to 96% of general practice providers and 94% of specialists. (Source: 2020 National Electronic Health Records Survey, which is conducted by the Office of the National Coordinator for Health Information Technology [ONC])

The unstructured nature of behavioral health data makes it difficult to store in a standardized manner.

Multiple factors contribute to this lag, including the nature of behavioral health data—which is often narrative and less structured, making it challenging to standardize within digital records as intended by the HITECH Act of 2009.

Still, EHRs have become an essential part of the overall care delivery process, serving as comprehensive repositories of patient data over time. They capture a vast array of information, including medical history, diagnoses, medications, and treatment plans. And they had a large presence at NatCon23.

Even when data is properly stored, EHRs alone cannot transform information into intelligence.

Yet, while they house an indispensable layer of data, these systems do not inherently provide a layer of intelligence. EHRs record and store information—but they do not analyze, interpret, or make predictions based on that information.

The Donabedian model—a framework for examining health services and evaluating the quality of care—actually provides an excellent analogy for these layers. EHRs, in this context, contribute significantly to the “structure” element by organizing and storing data. However, to influence the “process” and “outcomes” components—both of which involve the delivery and results of care—we must infuse intelligence into the system.

AI is the key to turning raw healthcare data into meaningful healthcare insights—at both the individual and population levels.

This is where the integration of artificial intelligence (AI), augmented intelligence, and machine learning into EHRs becomes crucial. These technologies can:

  • analyze the vast troves of data stored in EHRs,
  • identify patterns,
  • generate insights, and
  • transform raw data into actionable intelligence.

AI is the bridge between pure data and better care quality, helping healthcare providers focus more on their clients, improve their processes, and ultimately, enhance outcomes. (Great discussions with Wes Williams, Anh Kremer, and Ken Lauria on this topic!)

Due in part to its EHR adoption lag, the behavioral health sector has a unique opportunity to “leapfrog” AI adoption. 

The relatively slow uptake of EHRs in this sector has presented an unexpected opportunity for behavioral health providers to become technological trailblazers. This opportunity is due in large part to timing, as behavioral health is starting to embrace digital transformation at the exact time that cloud-first EHRs are becoming more prevalent.

A significant upgrade over traditional on-premise systems, these cloud-based EHRs are (somewhat) more flexible, scalable, and accessible. Most importantly, these systems are more likely to be designed in a way that facilitates interaction with advanced technologies like AI.

Unlike their legacy counterparts—which often require significant resources to extract, process, and use data—cloud-based EHRs can more easily embed AI applications (like Eleos Health), allowing for seamless data exchange.

This makes it easier for providers to leverage AI for data analysis, workflow automation, and decision support. Thus, the late adoption has inadvertently positioned behavioral health providers to leapfrog legacy systems and directly adopt more advanced, AI-compatible technologies—effectively accelerating the sector’s digital transformation.

PREDICTION: By 2030, behavioral health will rank among the top five sectors in healthcare with respect to the adoption of Large Language Models (LLMs) and AI-based solutions.

The effective use of AI in healthcare is predicated on one overarching principle: “It’s the workflow, stupid.”

Many of us remember the legendary phrase, “It’s the economy, stupid.” A similar principle can be applied to the use of AI in healthcare: “It’s the workflow, stupid.”

While the accuracy of AI models is undeniably important, and the datasets on which they are trained is critical, the true value of AI in healthcare extends well beyond its ability to make accurate predictions or diagnoses. Rather, it’s about how AI is embedded within existing workflows—and how it can streamline and automate them—thus enabling healthcare professionals to deliver better care more efficiently. 

It’s about the transformation of cumbersome, time-consuming tasks into seamless, automated processes. Ultimately, AI’s potential to revolutionize healthcare lies in its ability to:

  • augment human expertise,
  • alleviate administrative burden, and
  • enhance clinical decision-making.

This opens the door for optimizing provider workflows, improving care outcomes, and promoting health equity (similar to what we discussed Aimee Prange, Sara Miscannon, and Luke Beasley—and in separate conversation with Tom Morgan). So, just as economic considerations drive political strategies, workflow considerations should drive healthcare AI strategies.

The move to value-based care is picking up speed, putting an even brighter spotlight on AI-driven workflow optimization. 

Reflecting on the insights from NatCon23 and connecting them with Medi-Cal’s upcoming CalAIM payment reform—which goes into effect on July 1, 2023—it becomes evident that optimizing workflows is central to delivering comprehensive, coordinated care.

CalAIM represents a significant step toward value-based care, with its emphasis on enhancing patient outcomes, promoting health equity, and curtailing healthcare costs. Thanks to fascinating insights shared by Josh Ciszek, Kathy McCarthy, Don Taylor, and Gordon Richardson, I learned more about this impending structural change to reimbursements (which will no longer cover travel or documentation time!).

This imminent shift underscores the critical role augmented intelligence can play in clinician workflows (trickling down to community outreach as well, right Amanda Rankin?).

AI can give providers the freedom to focus more on the clients in front of them.

After all, traditional workflows don’t always foster the best care environment or client experience. So, using technologies like augmented intelligence to reimagine those workflows in a way that benefits both providers and clients—thus fostering greater overall value—is critical in light of payment reform. 

Concurrent documentation is one example of this that popped up in several NatCon23 discussions. The concept of concurrent documentation first emerged as a way to address the growing challenge of completing and submitting documentation quickly and accurately. 

However, if we zero in on the core issue and remain open to alternative solutions, it becomes clear that AI can provide a solution to the same issue while also empowering providers to focus more on their clients during sessions.

PREDICTION: Workflow integration will become the number-one factor driving CEOs and CIOs to add an AI-based solution to their IT stack. Accuracy will come in a close second, as unspecialized AI models have become a commodity (versus specialized models designed for specific verticals).

Providers’ appetite for value-based care is on the rise, but they are unsure how to actually achieve it.

The shift toward value-based care in behavioral health is gaining significant momentum, prompted in part by the successful Certified Community Behavioral Health Clinic (CCBHC) model and initiatives like the Illinois Health Practice Alliance. (Loved our discussion around this, David Berkey, Gilbert Lichstein, Fabian Camarena, and Juan Flores!)

These examples represent necessary early steps toward proving the effectiveness and sustainability of value-based care in behavioral health. The CCBHC model has been particularly influential, demonstrating how comprehensive, coordinated care can significantly improve behavioral health outcomes.

Similarly, the IHPA in Illinois has shown how collaboration between payers and providers drives more integrated care delivery and leads to improved patient satisfaction, reduced hospital admissions, and better management of chronic conditions. 

In my conversations with various providers at NatCon23, a common thread surfaced: there is a strong interest in transitioning to value-based care (VBC), but a general uncertainty around how to go about it.

Effective payer-provider collaboration starts with a shared definition of value.

To turn this enthusiasm into tangible action, we need a collaborative approach—one built on a closer partnership and shared vision among providers and payers at both the national and state levels. 

In a previous piece, where I delved into the Donabedian model, I posed a question that could be a potential starting point for these discussions: which outcomes should we track? The answer isn’t straightforward in behavioral health, where value-driven metrics extend beyond just patient-reported outcomes.

To spark the kind of transformation we want at scale, it’s important to consider things like social determinants of health and the sort of whole-person information currently tucked away within patient-clinician dialogues. Mining and analyzing this sort of information will play a pivotal role in how we measure and understand outcomes going forward.

PREDICTION: In 2024, we’ll see the first genuine value-based agreement between major national insurers and community behavioral health providers—one in which outcomes measurement won’t be solely tied to patient-reported surveys.

That’s a wrap! I met a lot of great people and had many stimulating discussions during NatCon2023. I’m so grateful to our amazing Eleos Health team for sparking such meaningful conversations, meetings, and dinners—and participating in one truly unforgettable karaoke session. I can’t wait for NatCon24!

** The writing of this article was augmented by the use of AI—specifically GPT4 and Bard. However, the AI didn’t author the piece entirely from the outset.