Mental health provider dramatically reduces case review time using AI

Cerebral is a virtual mental health provider with a mission to democratize access to quality care for all. With several full-service monthly subscription plans to choose from, members gain online access to therapists and prescribers.

Visits are done via video chat and members can message their care team directly. Members can have medications delivered directly to their home.

THE PROBLEM

Unfortunately, checking the cards at Cerebral was entirely manual, time-consuming and costly.

The manual internal audit and monitoring processes required hundreds of hours per month from a team of clinical auditors who were responsible for individually identifying notes to audit. To do this, they worked through a process of creating a selected sample of clinicians to review and issues or topics to review for compliance or adherence, ensuring that previous audit findings were not repeated in the environment.

“To get a relevant, selected sample, the manual monthly build required us to go through hundreds of clinicians’ calendars and appointments, identify medical record links and copy them into the EHR, and then find and review the actual session notes,” said Julia Dolan, associate director of behavioral operations at Cerebral.

“The results and observations were then recorded separately in another document, analyzed and summarized in a report for management and the individual clinician. Finally, they were summarized separately for senior management and executive management within the company,” she continued.

In addition to being time-consuming, the process was also prone to human error. The successful completion and review of the providers’ audit results depended on the auditor’s meticulous attention to detail.

“They had to type the clinician’s name and medical record number into a scoring template that was used to document feedback,” Dolan explained. “This, of course, introduced typical audit hazards, such as name mismatches, incorrect links, or the omission of an audit component on a chart.

“As we remain committed to continuous improvement and iteration to ensure high-quality audit feedback, we identified gaps in the manual process and remained diligent and proactive with the resources at our disposal. This is no small task for clinical management,” she added.

PROPOSAL

Supplier Brellium offered an AI-based chart review technology designed to help management of unstructured clinical notes by using artificial intelligence to perform comprehensive and real-time audits. This technology aimed to address a major pain point in healthcare administration: the inefficiencies and errors in clinical documentation due to manual processing.

“The core functionality of the system was to pre-screen all unstructured clinical notes for various types of errors and inconsistencies that could potentially impact the quality of patient care and/or insurance billing, such as client mood, successes and follow-ups, evidence-based interventions, and determining whether treatment goals are measurable and time-bound,” Dolan said.

“The AI ​​is specifically trained to identify clinical errors and spot trends that may require attention,” she continued. “This preemptive analysis also seeks to ensure that billing and/or coding errors can be identified and corrected prior to insurance claims being submitted, significantly reducing the risk of claim denials and the need for costly rework.”

In addition, by automating the review process, the technology promised to reduce the time providers spent on manual audits by 98%, she added. Where traditional methods could typically only thoroughly review a small portion of patient visits, this AI-driven approach allowed Cerebral to expand the percentage of notes it reviews, increasing review coverage while dramatically reducing the labor-intensive task of manual screening — with a 78% reduction in time spent, she said.

“This significant improvement in efficiency has freed up many resources, allowing care providers to focus their time and energy on more client-focused activities and other administrative matters,” she noted.

TOOK UP THE CHALLENGE

In December 2023, Cerebral began using the automated chart review process within its behavioral health department. Cerebral was challenged with processing thousands of therapy session notes each month and needed a scalable system to ensure compliance with its internal quality control and compliance requirements, as well as insurance mandates—without the unmanageable burden of manual reviews.

“Our clinical team worked directly with the supplier to “A custom AI system that integrates Cerebral’s unique session note formats, service timings and specific audit requirements of insurance partners,” Dolan explains.

“After creating individual question sets and parameters for each type of note template, Brellium’s system automatically screens session notes in Cerebral’s EHR and flags any necessary adjustments. This significantly reduces resources spent on administrative audits and frees up time for patient care, while also reducing the risk of financial loss from insurance chargebacks,” she said.

RESULTS

Within six months of implementing the AI ​​technology, Cerebral saw a massive 81% improvement in meeting critical underwriting criteria and reduced time spent on case reviews by 78%, signaling significant progress in operational efficiency.

Additionally, clinicians’ average audit scores improved by 8% between March and June 2024, Cerebral said.

ADVICE FOR OTHERS

“AI-powered chart review has allowed us to extend our documentation review in an objective and operationally efficient way,” Dolan said. “It has allowed Cerebral to more efficiently assess trends and error patterns through precise data feedback that could be missed with manual review.

“This allowed us to more easily identify areas for improvement and better implement additional education across our clinical teams in a meaningful and impactful way to drive continuous improvement,” she continued.

She concluded that the AI-based chart review has made Cerebral’s workflows significantly more efficient and improved the quality of note documentation.

Follow Bill’s HIT reporting on LinkedIn: Bill Siwicki
Send him an email: bsiwicki@himss.org
Healthcare IT News is a publication of HIMSS Media.