BHB Invest: How Technology Will Change the Face of Behavioral Health in the Next Five Years

The Impact of Digital Therapeutics, AI and Interoperability on Client Care, Outcomes and Reimbursement

This article is sponsored by Kipu Health. This article is based on a Behavioral Health Business presentation with Ben Dittman, Founder of Avea Solutions. The presentation took place on October 12, 2022 during the BHB Invest Conference in Chicago. The article below has been edited for length and clarity.

Ben Dittman: About 10 years ago, I got into the behavioral health space and founded a company called Avea Solutions. We were a revenue cycle management software, specifically designed for substance use disorder facilities. Built from the ground up, we brought on about 380 facilities, supported 27 billing companies all pushing a couple of billion dollars’ worth of claims throughout the platform. A year ago, we joined forces with Kipu, the largest EHR within the space.


I was excited about that because one of the ethos of Kipu, and the new leadership, was being open and integrated. Traditionally, revenue cycle management was considered at the end of the patient journey. I’m really excited about how together, we’re bringing the revenue cycle to the beginning with technology integration.

Right now, we are in the middle of a technology revolution. I am approached on a quarterly basis by different companies that want to integrate with us and I only see that growing. We’re starting to see artificial intelligence (AI) and machine learning in almost all the products and everybody talks about interoperability. I’ll discuss what that looks like, and revenue opportunities as well. Finally, we’ll go into analytics and the ROI of some of these new technologies.

Emerging AI technologies


Dittman: One area where we’re seeing more adoption is digital therapeutics. There are apps like Pear Therapeutics, which is FDA-approved, and the outcomes are pretty amazing. They have a ton of outcomes measurement and alumni management tools. We’re starting to see some really neat devices with remote patient monitoring as well.

These digital therapeutics integrate within the clinical workflow and while there’s a learning curve, clinicians are successfully helping patients adapt to these technologies.

Another area seeing a lot of growth is natural language processing where the computer starts to go and pick up text, translates that text or unstructured data into other data, and makes inferences. We’re seeing that right now in clinical documentation. We have a group, for instance, that overlays our telehealth visits and it records the session. In real-time it’s converting that into clinical documentation and supporting the clinician and their clinical notes. It’s also looking at what the payers want and highlighting different areas and saying, “Does this make sense? Do we need to change the verbiage on this?”

Right now, language processing can recognize about four different individuals. It doesn’t work great in a group session where you have an intensive outpatient program (IOP) session and there’s 15 people, but if you have anything over telehealth, the technology can identify each person. Also, if it’s a small group of about four different people, it can decipher what each person says and build out the clinical documentation.

We’re also seeing language processing on the utilization review (UR) side. It can go through and analyze the clinical notes and come up with appropriate recommendations for calls with payers. We’re seeing an increase in links of authorizations as a result. Computer-assisted coding goes through and analyzes the entire clinical record and determines if there’s additional codes that can be built out. It will analyze the entire clinical record, give recommendations, give the right codes and the right modifiers.

Machine learning is going through and analyzing claims that are underpaid, which claims we need to work, what’s the timeliness of each claim, which ones should we focus on first and which ones can we appeal. We’re starting to see a lot of AI chatbots that will actually go out and check eligibility, determine estimated insurance value and request payment, all within the website.

Supporting alumni on the go

Dittman: There are some exciting developments in the alumni space that allow for better engagement and outcomes. For example, apps like Videra allow alumni to check in via a video-recorded session. This is more efficient because it allows them not to have to wait till an alumni coordinator can schedule a call – they can do it when they’re waiting in line to pick up their kid from soccer practice. It will record the check-in and analyze the mood of the person by face recognition and listen to the person’s voice to determine if he/she is using depressive language or showing sign of suicidality.

It will also analyze the rate of speech to determine if the rate has increased, decreased, or if it’s slurred. If anything concerning registers, it sends a response back to the alumni coordinator and the information all flows back into the medical record.

Wearables are also changing how providers check in with alumni. For example, RAE Health is a wearable device that will sense cravings post-discharge or even within the facility, it will alert the clinicians that there’s cravings happening.

Another interesting technology is a low-frequency Bluetooth device that is attached to a pill bottle, and it detects the difference between twisting off, rattling, shaking or dropping. From a medical adherence standpoint, that’s huge. It’s not going to say if that person took the pill, but if you’re looking or dealing with adolescents post-discharge, you want that medication adherence when it comes to something like attention deficit and hyperactivity disorder (ADHD) medication. If they’re not taking it, the chance of relapse is higher.

There’s also MyClearStep, which is a scale for eating disorders. The scale doesn’t have numbers, it connects to WiFi, and immediately rolls back into the medical record and informs the clinicians.

New technologies lead to better Interoperability

Dittman: With interoperability, right now everybody wants to say that they’re integrated, and I would say there are very few integrations that are right. Everyone checks the box, but really it doesn’t flow into the clinical workflow. Providers need to think about the patient flow through different applications. Ultimately, you never want to have to onboard a patient in a separate application. When you do the admissions process, those patient demographics should automatically flow to that third-party application, build that patient profile and kick off the onboarding process.

It’s amazing all the different technology that’s coming out right now and the impact that it’s having. The other piece is, when you are looking at these third-party applications and integrating with any EHR or any platform, you really want to think about what’s also being fed back into the system: the alerts, biometrics, reports. I’d say probably out of every group we start to work with on an integration, about seven of them fail, and they fail because they’re not willing to work within the clinical workflow. They want clinicians going out into another site. They want a separate login to happen which is more difficult for them and if they’re not bought in, patients won’t adopt the technology.

The other thing is being able to tie a primary key from pre-admission. As a patient admits, you want to be able to look at all that patient demographics, but then be able to tie it to every single analytics piece that’s coming out of these different platforms. This is going to be huge for value-based reimbursement negotiations in the future.

If a scheduling change happens, is that app being notified with that scheduling change, because you may want one of these apps to engage post-treatment. What you want is basically a quarantine process. If somebody updates something in the clinical record, you don’t necessarily want it to update in the billing record. If somebody changes the name from Bob to Robert and your billing team has already updated it to maybe a different legal name, you don’t want that to change going forward.

We believe in bidirectional communication with purpose and being able to quarantine changes so that it doesn’t impact revenue down the road. Then one of the biggest things that we’re seeing right now is data storage. Everybody wants to be able to pull their own data out, own their own data, put them in data lakes, cubes, be able to create reports.

Technology’s impact on reimbursement and outcomes

Dittman: We’re starting to see more and more reimbursement impact when it comes to digital therapeutics and remote monitoring devices. We have 20 states currently reimbursing remote patient monitoring for Medicaid programs and we’re starting to see a lot more changes here. We’re also starting to see that they’re not grouping behavioral health in with everyone else, which is great.

One thing that most people don’t think of when they talk about outcomes measurement is the impact that it has on UR. We have started to see that we can present information in that authorization process when they call the payers and say, “Based off of this biometrics data, we need additional authorized days. Based on this ASAM report that we got from ERPHealth, we recommend that the patients stay in this level of care.”

There’s additional insurance revenue and then really coming into value-based reimbursement. There are a lot of great value-based reimbursement negotiations going on with the payers and there is a ton that’s not working right. That’s why it’s key to have the right data to show up with the right information. You can also come up with a great value-based reimbursement agreement, but the payers are not structured to manage those payments.

We have seen quite a few groups out there that think they’ll increase revenue by 25-30%. On paper, it might appear that way but during the claims processing, the financial systems on the back end and the checks aren’t coming through. We talked to a group recently that had over $1.2 million in old accounts receivable (AR) because of this process and it was a small group that needed that money to keep operations open.

Then really, there’s a huge push in analytics. Benchmarking is what we talk about within our organization. Benchmarking is hard because you have a lot of unstructured data and everybody does things differently and there’s no standardized code sets. We’ve been starting to use machine learning and AI to go through and standardize that unstructured data, but when you have 2,800 facilities in this space, we really believe that it’s our mission to be able to show benchmarking across all of our patients. It’s an opt-in process, where if you share your data, then you can see other people’s data.

Everybody right now is managing multiple dashboards. They have marketing analytics over here, reimbursement analytics over there. We really need to be able to tie pre-admission through post-discharge and show the cost of a patient, cost of delivery, the impact that we’re having. Then finally, with predictive analytics, this is where I get really excited, because I truly think it’s going to impact client care UR days.

With pre-admission, we can now look at the claims data that the payers are working off of and see if it’s not what is being delivered in the facilities. Many facilities are delivering care that is at a higher level than what they’re billing out. If somebody’s authorized for four days of detox, but the client needs five, they’re not going back and fighting for that fifth day, they’re using a residential treatment center (RTC) day, that often ends up being longer than one day.

That being said, we start to look at predictive analytics, based off of this insurance company, this diagnosis, we can anticipate how many days that that client’s going to receive from an authorized days standpoint. That is also not consistent by facility. We analyze by case manager, you can have substantially different numbers of authorized days based on who is calling into that payer. You should be tracking it as a facility who’s calling in, what’s the average authorized days, who they’re talking to. On the front end, you should be able to predict exactly how long their insurance company’s going to be able to authorize.

Emerging technologies, such as AI, have been shown to solve a Rubik’s cube within 1.2 seconds. When you look at behavioral health, you can imagine how helpful that will be to providers. I look forward to the efficiencies and outcomes improvements we’ll all experience as a result of an interoperable technology ecosystem.

Kipu Health uses CRM to accelerate patient intake and supports engagement throughout the entire lifecycle—from pre-admissions to alumni status and everything in between. To learn more, visit:

Companies featured in this article: