Conversational AI helps Gaudenzia improve therapist performance, gain efficiencies
Photo: Gaudenzia
Gaudenzia is one of the largest nonprofit drug and alcohol treatment and recovery centers in the country, with 51 facilities in Pennsylvania, Maryland, Delaware and Washington, D.C. It serves 19,000 individuals annually and operates 117 drug and alcohol treatment programs for men and women.
THE PROBLEM
It's no secret that communities have been facing a substance use disorder epidemic. When COVID-19 came along, it only exacerbated the problem. Earlier this year, the CDC estimated there were more than 100,000 overdose deaths in 2021, a record-breaking number.
This increase came at a time when the entire healthcare field, including Gaudenzia, faced debilitating workforce shortages. So, not only was the need for care greater, but there were fewer team members available to take on the workload.
"At the session level, each counseling session with a client requires detailed documentation and clinical notes," said Gaudenzia Director of Outpatient Services Andrew Schmitt, who holds a master's degree in social work from Rutgers University and has more than 20 years of experience in the behavioral healthcare field.
"Our counselors also need to check in with clients on factors like their overall mental health, track their progress, and build a rapport of mutual trust and respect," Schmitt explained. "We really value that personal connection. At the same time, robust, accurate clinical documentation is essential to help us stay on track and accountable as an organization."
Documentation is a time-consuming component of counseling that can take away from the time staff spend interacting with clients one-on-one.
"Another challenge unique to our organization is that we've evolved our treatment model to incorporate evidence-based counseling techniques like cognitive behavioral therapy and motivational enhancement," Schmitt noted.
"From a counseling perspective, it can be challenging to keep track of which modalities clients are more responsive to across sessions, or how closely evidence-based techniques are followed from one session to the next," he continued. "This again makes documentation and having robust clinical notes a crucial component of providing quality care. We wanted to find an effective way to alleviate strain on our team members that still ensured accurate documentation and accountability."
From a supervision standpoint, clinical supervisors have a limited capacity for providing direct supervision and sometimes rely on clinicians to self-report on their progress with clients, he added.
"This can lead to a subjective reporting structure, which is a challenge when it comes to accountability," he said. "This also impacts our counselors' ability to get direct, actionable feedback or insights on the ways in which they can improve their counseling approach."
PROPOSAL
When it came to documentation and clinical notes, the idea of automation presented a solution that could free up counselors to focus on client care. That's the mission at Gaudenzia, so this prospect was exciting, Schmitt said.
"From an employer perspective, we highly value our team members, which includes respecting their time and doing all we can to help prevent burnout," he remarked. "By transcribing recorded client sessions and generating suggestions for clinical notes, automation has the potential to help our team members focus on caring for people in need while helping us provide a healthy work-life balance.
"This was especially important amidst the demands brought on by the pandemic, which we continue to feel today," he added.
So Gaudenzia turned to Eleos, a vendor of conversational artificial intelligence, to create data and insights for its behavioral health practitioners.
"As far as session insights go, the proposal was that Eleos could provide data like counselor-to-client talk ratios, and that it could report on how closely evidence-based techniques are followed throughout each session, and across a range of sessions," Schmitt said.
"This provides an opportunity for direct supervision on-demand and allows clinicians to objectively evaluate and improve their own performance," he continued.
Lastly, there's also the component of clients self-reporting information prior to a session, where they can share how they are doing in terms of their mental health and any challenges they are facing.
"It was proposed that this data could be tracked across sessions and would provide a visual representation of how a client progresses over time, or whether additional treatment interventions would be necessary," Schmitt noted.
MEETING THE CHALLENGE
Many Gaudenzia counselors have started using this technology to record client sessions and benefit from the AI-generated documentation suggestions, as well as session insights that help them keep track of their own performance.
"Something our team members have found especially helpful is the ability to identify keyword clusters that clients use often, as this helps flag certain thought patterns or triggers that can be addressed in session," Schmitt explained.
"Our counselors have found that while it takes several sessions for the AI to learn and adjust to the specific counselor and client, the quality of the suggestions for notes generated from a recorded session improves drastically over time and frees our team members up to spend more time interacting directly with clients," he continued.
This really helps paint a full picture of the counseling relationship and how – or whether – a client is progressing with their treatment, he added.
"We only use the technology in sessions with client consent, and while we anticipated challenges with how comfortable clients would be having their sessions recorded, our team members have found that clients are just as interested in the data as their counselors," he said.
"We also often treat mandated clients who don't necessarily feel they need help or understand why we use certain treatment techniques," he continued. "The ability to bring concrete, objective data into sessions helps clients make connections and better understand their own needs. This can make the treatment process more collaborative and engaging."
In order to implement this technology, the organization had to integrate it with its electronic health record – Netsmart's My Avatar. This took the IT and EHR department working with Eleos and Netsmart to create a solution.
"Luckily, Eleos has a great working relationship with Netsmart that enabled this to be a smoother partnering than expected and resulted in a relatively seamless integration," Schmitt noted.
RESULTS
One of the most notable successes the organization has seen so far has been a 35% increase in the use of evidence-based techniques from team members since implementing the AI technology.
"Our organization has made significant strides in evolving our treatment model to one that is evidence-based and person-centered, but for counselors who have been with us for a long time, it can be a challenge to stay on track and avoid unwittingly falling back on previous techniques," Schmitt noted. "Seeing an increase in fidelity to our new treatment model has been significant.
"We've also noted that at the time of proposal, it took our clinicians approximately 10 minutes to complete session notes," he continued. "With Eleos, it currently takes our clinicians an average of 3.2 minutes to complete their session notes. This adds up to a significant improvement in the timeliness and quality of documentation."
As of November 18, 2022, Gaudenzia will have all of its outpatient programs trained on Eleos with a refined onboarding approach. The training process currently involves Eleos staff, Gaudenzia program staff, and IT and EHR teams.
"This helps us create a coordinated and integrated training approach that meets the needs of our counselors and any other team members who need to interface with the technology," he said.
ADVICE FOR OTHERS
"It's essential for any organization considering implementing this kind of technology to take an honest inventory of current projects and demands on staff," Schmitt advised. "While the results are worth the investment, it does require a significant lift to implement new technologies.
"With AI-related platforms, it also requires a cultural shift among team members," he continued. "I would say organizations should be mindful that there can be a wide range of responses, from those who advocate for the technology to those who are skeptical or even paranoid of the implications of AI-based solutions."
What Gaudenzia found helpful on its end was to provide education and resources through internal messaging, and to acknowledge and applaud team members who successfully use the technology to improve their own performance.
"I would also suggest creating a proactive plan for leadership involvement," he concluded. "When clinical supervisors and others in leadership roles are brought in at the ground level, they can understand the ins and outs of the technology and see the benefits first-hand, which helps create a sense of authentic advocacy and allows us to lead by doing."
Twitter: @SiwickiHealthIT
Email the writer: bsiwicki@himss.org
Healthcare IT News is a HIMSS Media publication.