Thursday, April 16, 2026

< + > Standard AI is a Black Box. Here is Why RAAPID Built a Glass One for Risk Adjustment.

The problem with AI in the revenue cycle is transparency. It is powerful, but it usually operates as an unpredictable black box. In risk adjustment, you simply can’t afford to guess how an algorithm arrived at a billing code. You need a glass box. You need absolute defensibility. Here is how that is finally becoming a reality.

Healthcare IT Today sat down with Chetan Parikh, Founder and CEO of RAAPID, to explore the evolution of risk adjustment technology. We discussed the challenges of relying on standard NLP and why organizations need technology that balances accurate coding with strict regulatory compliance.

What This Conversation Revealed

  • Neuro-symbolic AI offers a glass box approach. By combining large language models with proprietary knowledge graphs, organizations gain high accuracy and fully defensible evidence without the risk of hallucinations.
  • AI reduces the mental load on medical coders. Highly accurate AI tools allow coding teams to stop sweating the small details and start operating at the top of their license.
  • Technology must balance revenue and compliance. The right AI ensures providers get paid for the services they deliver while preventing the regulatory risks of over-billing.

Neuro-Symbolic AI Provides Defensible Evidence

Standard natural language processing casts a wide net but often struggles with precision in complex clinical documentation. Health IT leaders know that adopting large language models brings risks of hallucination, making pure generative AI difficult to trust for revenue cycle applications. The solution, according to Parikh, lies in neuro-symbolic AI.

Parikh explained how RAAPID addresses this industry hurdle by marrying large language models with proprietary knowledge graphs. He noted that their technology focuses on “taking full advantage of the large language models and at the same time making sure that we are not hallucinating”. Parikh further detailed that this approach is all about “converting from a black box to a glass box, where everything is defensible and evidence based.”

Elevating the Role of Medical Coders

Finding and retaining highly skilled medical coding talent is a persistent challenge for provider organizations. When legacy NLP systems only deliver moderate out-of-the-box accuracy, human coders are forced to spend excessive time verifying outputs.

However, with RAAPID’s neuro-symbolic powered AI systems, organizations can achieve more than ninety percent accuracy. This dramatically improves the entire workflow for coding staff. Parikh highlighted this by  stating that “when you have an AI that is as accurate as 91 – 92% out of the box, the coder’s mental load is significantly reduced, and the coders are now operating at the top of the license rather than they trying to identify everything.”

Hitting the Sweet Spot Between Revenue and Compliance

Risk adjustment requires walking a tightrope. If an organization under-codes, they will not capture the true value of care delivered. Conversely, aggressively capturing codes without sufficient documentation triggers intense scrutiny from federal regulators.

“If your AI is unable to identify codes that are truly billable, then you did the work, you provided the service, but you are not getting paid for it,” Parikh summarized. “But you have to make sure to not be overcoding and overbilling.”

Health systems need a middle ground where they capture accurate reimbursement while remaining securely within regulatory boundaries.

The Bottom Line

Risk adjustment technology needs to move beyond good-enough AI with opaque models. As organizations evaluate new AI tools for their revenue cycle, the focus must be on accuracy, defensibility, and operational efficiency. Implementing AI that provides clear evidence pathways, like what RAAPID offers, protects the organization from compliance risks while ensuring fair reimbursement for care delivered.

What Healthcare IT Leaders Are Asking

What is neuro-symbolic AI in healthcare? Neuro-symbolic AI combines the pattern recognition capabilities of large language models with the structured logic of proprietary knowledge graphs. This hybrid approach provides the broad contextual understanding of generative AI while anchoring the outputs in factual, evidence-based rules to prevent hallucinations.

How does AI impact medical coding compliance? Advanced AI improves coding compliance by linking suggested codes directly to documented clinical evidence. By surfacing only defensible codes, the technology helps organizations avoid over-billing while ensuring they capture all appropriate revenue for services rendered.

Why is a “glass box” approach important for risk adjustment? A glass box approach allows human auditors to see exactly how an AI model arrived at a specific coding conclusion. In highly regulated areas like risk adjustment, being able to trace a suggested code back to the exact clinical documentation is essential for defending claims during audits.

Learn more about RAAPID at https://www.raapidinc.com/

Listen and subscribe to the Healthcare IT Today Interviews Podcast to hear all the latest insights from experts in healthcare IT.

And for an exclusive look at our top stories, subscribe to our newsletter and YouTube.

Tell us what you think. Contact us here or on Twitter at @hcitoday. And if you’re interested in advertising with us, check out our various advertising packages and request our Media Kit.

RAAPID is a sponsor of Healthcare Scene



< + > When Phones Aren’t an Option: How UCHealth Modernized Meal Ordering in a Behavioral Health Unit

Designing a modern behavioral health unit often means intentionally leaving out bedside phones – a standard fixture elsewhere in the hospital, but a safety and security concern in psychiatric care.  But when patient’s can’t call in their own meal orders, what then? For UCHealth, that meant taking advantage of an technology that was not yet being used and changing their workflow to accommodate it. Better experiences and higher productivity were the result.

Healthcare IT Today sat down with Jenna Sampson, Nutrition Systems Coordinator at UCHealth. When their new behavioral health unit came online, Sampson and her team had to rethink their organization’s standard phone-based meal ordering process, ultimately deploying an existing app.

What This Conversation Revealed

  • Intentional constraints drive digital workflows. Behavioral health unites routinely exclude bedside phones for safety reasons and because of that UCHealth moved to meal ordering onto a mobile app which ended up being better for everyone.
  • Safety requires strict EHR boundaries. To handle risky free-text allergies in Epic, the system implements a hard stop when allergies are found to be in free-text in the EHR. Nothing moves forward until those allergies are coded discretely into Epic.
  • Frontline tools boost system metrics. Providing nurses with direct mobile access turned a potential chore into a preferred workflow, driving regional app adoption to record highs.

Building Patient Workflows Around Intentional Constraints

When UCHealth opened a new 55-bed behavioral health unit in Fort Collins, the facility featured a very specific design choice. The rooms intentionally lacked bedside phones. Standard practice across the health system relied on patients calling the kitchen to order their meals.

Without patient phones, the burden would fall entirely on the nursing staff to call in the orders. That alternative would have meant nurses calling orders into the call center which in turn would create longer wait times for other patients and bog down the call center with additional, unnecessary call volumes.

The solution was already in their technology stack – the Illumia (formerly CBORD) Patient App.

“We already had the Patient App,” shared Sampson. “We decided to explore it and made it successful.”

Giving nurses direct mobile access transformed the ordering process and avoided a massive call center bottleneck.

Baking Safety into the Epic Workflow

A major challenge in dietary ordering is handling allergies entered manually in the electronic health record. UCHealth uses Epic, which allows clinical staff to input allergies as free text in the “other allergy” field.

A free text entry like “strawberry”, for example,  will not trigger the automated dietary compliance system, creating a serious patient safety risk. The team solved this by creating a new operational workflow. “Any patient that has [something entered into] the other allergy field are ineligible to order through the app to ensure patient safety,” shared Sampson. “When this happens, our staff go into Epic and codify the allergy properly and remove the data from the ‘other allergy’ field. Now that patient’s meals can be ordered through the app.”

The Bottom Line

Tying the mobile app directly to Epic’s allergy compliance engine ensures patient safety remains the top priority. While nurses initially had some resistance, they quickly came around – Sampson reports they now “fight over who gets to put the orders in.” The unit is inputting 99 percent of patient meal orders through the app. That localized success drove the region’s overall patient meal ordering from less than 1 percent to 19 percent. The unit also avoided adding a full FTE to the call center, translating to significant cost savings. The new unit is serving as a potential model for all other units at UCHealth.

What Healthcare IT Leaders Are Asking

How do you handle free-text food allergies in digital ordering?
Free text fields in an electronic health record fail to map to automated dietary compliance systems. The safest approach is to restrict digital meal ordering for any patient with an “other” allergy listed. Clinicians must manually review the chart and convert the free text into a coded, system-recognized allergy before the patient can use self-service apps.

What is the best way to drive clinical adoption of a mobile app?
Removing friction at the point of care is the fastest path to adoption. Pre-loading the required applications directly onto corporate-issued mobile devices ensures immediate access for nursing staff. When a tool genuinely saves time compared to calling a busy contact center, clinical teams will naturally gravitate toward it.

Can localized digital workflows impact system-wide metrics?
Testing a distinct workflow in a controlled environment provides a blueprint for broader rollouts. A near-perfect adoption rate in a single unit can generate enough volume to significantly move regional utilization metrics. This localized data serves as compelling proof to secure buy-in from other clinical departments.

Learn more about UCHealth at https://www.uchealth.org/

Learn more about Illumia at https://illumiatech.com/

Listen and subscribe to the Healthcare IT Today Interviews Podcast to hear all the latest insights from experts in healthcare IT.

And for an exclusive look at our top stories, subscribe to our newsletter and YouTube.

Tell us what you think. Contact us here or on Twitter at @hcitoday. And if you’re interested in advertising with us, check out our various advertising packages and request our Media Kit.



< + > Jimini Health Raises $17M | Ambient Clinical Analytics Secures $5M Strategic Investment

Check out today’s featured companies who have recently raised a round of funding, and be sure to check out the full list of past healthcare IT fundings.


Jimini Health Raises $17M as Behavioral Health Systems Face Growing Pressure to Manage Patient AI Use with Clinical-Grade Infrastructure

Jimini Addresses a Major Market Need as Patients Turn to General-Purpose AI for Mental Health Support, Providers Face Rising Pressure and Opportunity to Implement Safe, Clinician-Supervised Solutions

Funding will Enable AI Development Across Additional Clinical Settings and Expansion with Large Clinical Partners Nationwide

Jimini Health today announced $17 million in seed funding from M13, Town Hall Ventures, LionBird, Zetta Venture Partners, and OneMind, bringing total funding to more than $25 million. The company is building clinician-supervised patient-facing AI infrastructure for large behavioral health provider organizations, enabling providers to deploy patient-facing AI safely, compliantly, and at scale, with licensed clinicians maintaining oversight of every patient interaction.

The Reality Health Systems Can No Longer Ignore

Patients at large behavioral health systems are already using AI for mental health support between appointments, without clinician visibility or control. This cultural shift is happening regardless of whether providers participate, placing new clinical, operational, and legal pressure on behavioral health organizations to respond.

More than 5.4 million U.S. adolescents and young adults now use AI chatbots for mental health advice. More than 1 million people a week have conversations with ChatGPT that include explicit indicators of suicidal planning or intent. Character.AI and Google have already settled wrongful death lawsuits brought by families of teenagers who died by suicide following unsupervised AI interactions.

“When 1 million people a week are discussing suicide with a product that was never designed to handle it, that’s not an edge case, it’s a systemic gap,” said Morgan Blumberg, Partner at M13…

Full release here, originally announced March 31st, 2026.


Ambient Clinical Analytics Secures $5M Strategic Investment and Appoints Brian Tufts as CEO to Accelerate Growth

Ambient Clinical Analytics, a pioneer in software that combines real-time clinical analytics with clinical decision support and workflow tools, today announced the successful closing of a $5 million strategic funding round with key investments from Mairs & Power Venture Capital as well as a Fortune 500 strategic MedTech firm. The company also announced the appointment of Brian Tufts as Chief Executive Officer, signaling a new phase of growth and market expansion.

Healthcare providers today work in a complex environment in which critical patient data is not always readily available and easily interpreted in real time. In addition, care coordination amongst teams, especially in sepsis care, is challenging; often leading to variability in care.

Ambient Clinical Analytics addresses these challenges with an integrated offering that delivers real-time clinical analytics, clinical decision support, and integrated workflow automation to enable clinicians to see the full picture of a patient’s condition as it evolves.

The new capital will be used to accelerate innovation, expand adoption across health systems, and scale Ambient’s product, including its FDA Class II-cleared AWARE platform system. Built on clinically validated algorithms and Mayo Clinic–licensed technology, the company’s platform transforms complex clinical data into intuitive, actionable insights that support faster, more informed decision-making. Additionally, hospitals have leveraged the integrated workflow tools to improve sepsis protocol adherence, which has been correlated to improve both clinical outcomes and financial measures.

Brian Tufts joins Ambient Clinical Analytics with extensive leadership experience from Vantive and Baxter, where he led growth across complex healthcare environments. As CEO, Tufts will focus on expanding Ambient’s market presence, deepening strategic partnerships, and advancing the company’s mission to redefine how care teams operate in high-acuity environments.

“Hospitals do not lack data; yet doctors and nurses have few resources that deliver clarity in a dynamic clinical environment,” said Brian Tufts, Chief Executive Officer…

Full release here, originally announced March 30th, 2026.



Wednesday, April 15, 2026

< + > CommonWell Expands Data Exchanges in Volume and in Purpose

In this video, Paul L Wilder, Executive Director of the CommonWell Health Alliance, discusses the spread of health data exchange as it involves not just providers but new actors such as payers, public health, and patients themselves.

CommonWell, a nonprofit QHIN that started in 2013 and has an enormous reach today, contains IT vendors ranging from startups to big EHR vendors, and providers now as well. For a long time, Wilder says, EHRs supported only unidirectional data exchange: they would allow it to be extracted but not inserted. Now it’s more bidirectional.

While CommonWell is still investing in and supporting FHIR, Wilder noted that the ability of AI to extract key data from plain text documents, and to convert data between formats, makes the FHIR standard less important. Many sites go from source document to their own storage without an intermediate FHIR step.

However, FHIR is valuable for segmenting data and extracting just a few fields. This is important in public health, because transferring complete records on huge numbers of patients creates the security risk of a “honeypot” that could attract attackers.

In contrast, most providers want complete records. Patients do too, although Wilder suggests that in a few cases (such as private notes created by psychotherapists) the provider might be without some data. In our discussion, Wilder describes patients’ interest in their data, and advises that at the very least, they should examine their data for errors.

Wilder also discusses the benefits and challenges of two recent government policies: TEFCA and CMS-aligned networks. One observation he made is that TEFCA applies to covered entities, and thus excludes some important institutions such as free clinics. He also noted that CMS-aligned networks require bilateral agreements, which is very cumbersome to arrange among large groups of institutions.

However, some institutions have vastly increased data exchanges through TEFCA, and therefore increased the number of documents by orders of magnitude. For instance, more errors are being reported to the government because it’s technically easier to report the data. Some providers say they’re getting too much data, but Wilder has little sympathy for that complaint.  He did suggest that as the volume of documents increases, it’s also easier for malicious breaches to get lost and go unnoticed.  That’s a challenge that the industry is going to have to work on.

Check out our interview with Paul Wilder from CommonWell to learn more about the latest on healthcare interopeability.

Learn more about CommenWell Health Alliance: https://www.commonwellalliance.org/

Listen and subscribe to the Healthcare IT Today Interviews Podcast to hear all the latest insights from experts in healthcare IT.

And for an exclusive look at our top stories, subscribe to our newsletter and YouTube.

Tell us what you think. Contact us here or on Twitter at @hcitoday. And if you’re interested in advertising with us, check out our various advertising packages and request our Media Kit.

CommonWell Health Alliance is a proud sponsor of Healthcare Scene.



< + > How Virtual Behavioral Health is Transforming Quality, Access, and Patient Outcomes

The following is a guest article by Dr. Sara Gotheridge, MD, Chief Medical Officer at Array Behavioral Care

Virtual behavioral health has fundamentally changed access to care. At a time when demand continues to outpace supply, virtual models have helped reduce wait times and connect more patients to services more quickly. That progress matters.

But access is only the beginning of the story.

From a clinical perspective, expanding availability alone does not define success. What ultimately matters is whether patients receive care that matches their level of need, whether treatment adjusts as symptoms change, and whether those decisions lead to meaningful, sustained improvement and an improved experience of life. Achieving that requires moving beyond one-size-fits-all approaches and toward care that is intentionally aligned to acuity and the individual, delivering the right care, at the right time, at the right dose.

Access is More than Speed

Virtual behavioral health has improved access not only by increasing availability but by reducing friction at the front door of care. Whether patients are seeking help urgently or entering ongoing treatment, timely connection to the appropriate clinical resource is a critical first step.

In outpatient contexts, one meaningful indicator of this shift is time to first appointment. When patients are seen sooner, care begins earlier, engagement improves, and the risk of drop-off before treatment starts is reduced. In urgent and emergency settings, access looks different. It means rapid clinical assessment, appropriate triage, and clear direction about next steps rather than delay, default escalation, or prolonged waiting.

Across care settings, faster access alone is not enough. Access must also be appropriate. Connecting patients quickly only improves outcomes when they are routed to the level of clinical support that best fits their needs in that moment. This distinction is essential to translating improved access into better quality and patient outcomes. 

Matching Care to Need when Situations are Urgent

When behavioral health needs are urgent, the clinical priority is immediacy paired with appropriateness and risk mitigation. Patients require timely assessment, support that matches the nature of their concern, and clear next steps.

Effective virtual care in these moments focuses on using the most appropriate clinical resource for the situation. Not every urgent behavioral health concern requires the same level of specialization. Some can be addressed through brief clinical intervention, guidance, or rapid connection to ongoing care. Others require escalation and more specialized involvement.

Structured stratification at the point of entry supports consistent clinical decision making. Patients are routed to the appropriate level of care, psychiatrists are reserved for situations where their expertise is most clinically indicated, and higher-intensity interventions are used when they are truly necessary.

This approach is especially important in emergency department settings, where behavioral health demand is high and psychiatric resources are limited. Virtual behavioral health supports timely assessment and appropriate routing, helping emergency departments respond effectively without defaulting to unnecessary psychiatric admission or prolonged boarding. 

Supporting Improvement Over Time in Ongoing Care

For patients engaged in longer-term behavioral health treatment, the clinical challenge shifts from immediacy to sustained improvement.

Here, effective care depends on ongoing assessment and the ability to adjust treatment as symptoms change. Routine use of validated clinical measures allows clinicians to establish baseline severity, monitor progress, and determine whether treatment is producing the intended results.

Outpatient behavioral health designed around ongoing measurement does not rely on static treatment plans. Treatment intensity can change as patients improve, stabilize, or require additional support. When symptoms improve, care can transition to lower-intensity or maintenance-focused approaches that help sustain gains. When progress stalls or symptoms worsen, treatment can be adjusted earlier rather than waiting for deterioration.

Clinical outcomes from large real-world virtual behavioral health populations indicate that patients tend to achieve more consistent and clinically meaningful improvement when care models are structured to assess patient needs, match them to appropriate treatment, measure progress, and adjust care over time, including individuals entering treatment with more complex conditions.

Epic as the Backbone for Consistent, Integrated Care

Delivering this level of consistent, acuity-informed clinical decision-making at scale requires infrastructure that supports integration and continuity.

Operating within an interoperable electronic health record, such as Epic, provides a shared clinical backbone. Documentation, assessment data, and treatment decisions are visible across care settings, allowing clinicians to operate from a common clinical picture regardless of how or where care is delivered.

This integration reduces fragmentation and reinforces accountability. Virtual behavioral health functions as an extension of the health system rather than a parallel experience. Technology supports clinical decision-making without becoming the focus of care.

Designing Virtual Behavioral Health for Better Outcomes

As health systems continue to expand virtual behavioral health, the focus must shift from availability to effectiveness. Access opens the door, but design determines what happens next.

Virtual behavioral health delivers its greatest value when care is aligned to patient need, informed by ongoing assessment, and supported by an integrated clinical infrastructure. When treatment intensity matches need in urgent moments and adjusts appropriately over time, patients improve, clinical resources are used more effectively, and systems scale more sustainably.

The opportunity for health system leaders is to move beyond one-size-fits-all models and invest in care delivery approaches that consistently deliver the right care, at the right time, and at the appropriate level of intensity. When that happens, virtual behavioral health becomes not just more accessible, but demonstrably better.

About Dr. Sara Gotheridge

Sara Gotheridge, MD, is the Chief Medical Officer at Array Behavioral Care, leading clinical teams and quality programs. With 25 years in behavioral healthcare, she has held leadership roles at LifeStance Health, Trilogy Behavioral Healthcare, and a Chicago community mental health center. A former IT professional, she earned her mathematics degree at Indiana University Bloomington and completed her MD and psychiatry residency at Northwestern University. Dr. Gotheridge is a longtime Clinical Instructor at Northwestern and maintains an active practice, focusing on integrated care and improving access to mental health services. She is also a passionate pianist.



< + > This Week’s Health IT Jobs – April 15, 2026

It can be very overwhelming scrolling through job board after job board in search of a position that fits your wants and needs. Let us take that stress away by finding a mix of great health IT jobs for you! We hope you enjoy this look at some of the health IT jobs we saw healthcare organizations trying to fill this week.

Here’s a quick look at some of the health IT jobs we found:

If none of these jobs fit your needs, be sure to check out our previous health IT job listings.

Do you have an open health IT position that you are looking to fill? Contact us here with a link to the open position and we’ll be happy to feature it in next week’s article at no charge!

*Note: These jobs are listed by Healthcare IT Today as a free service to the community. Healthcare IT Today does not endorse or vouch for the company or the job posting. We encourage anyone applying to these jobs to do their own due diligence.



Tuesday, April 14, 2026

< + > Reclaiming the Exam Room: How ModMed Scribe 2.0 Empowers Specialists

In our latest clinical deep dive, Dr. Caleb Masterson from B&B Sinus and Allergy Relief Centers) and Dr. Joel Salamon, Medical Director of Pain Management at ModMed, explore the impact of removing screens from the exam room.

For Masterson, an ENT specialist, this technology goes beyond saving time—it’s reviving the independent medical practice.

“AI in general really opens up an autonomy perspective for physicians that hasn’t been there in the past,” Masterson explained in a recent interview. Noting that the “business of medicine” has historically deterred doctors from private practice, Masterson believes tools like ModMed Scribe 2.0 give a new generation of physicians the confidence to practice medicine on their own terms.

At the core of this shift is how seamlessly the technology handles specialty-specific workflows. ModMed Scribe 2.0 is the first pillar in the company’s broader strategy to deliver an AI-Powered Practice, and its real-world impact is already staggering. Since its launch in November 2025, ModMed Scribe 2.0 has been adopted by over 1,600 providers and used in more than 360,000 patient visits.

Because it was built natively within ModMed EHR, the clinical AI Assistant inherently understands complex medical nomenclature. Masterson highlighted this advantage, noting he can use lay language with a patient—like saying he needs to “take a camera and look in your nose”—and ModMed Scribe 2.0 translates it into proper clinical terminology, actionable orders, and suggested billing codes within seconds.

Dr. Joel Salamon, a pain management physician with 30 years of experience, echoed this precision. “If I see a patient with leg pain, numbness… it’s going to be transcribed into the note as a lumbar radiculopathy,” Salamon said. Unlike generic ‘bolt-on’ AI that simply dumps a summarized conversation into a free-text field for doctors to copy and paste, ModMed Scribe 2.0 places structured data exactly where it belongs, seamlessly routing physical exams or MRI discussions directly into their respective EHR tabs.

By acting as an intelligent observer, the AI eliminates the need for doctors to stare at a screen. For Masterson, this means he is no longer bogged down by mundane clinical workflows and can spend his time “really impressing the patient.”

For Salamon, it translates to unparalleled office efficiency and happier staff. “My notes are literally done and completed and sent off 15 minutes after I see my last patient,” Salamon shared, noting his team now reliably leaves work right at 4:30 PM.

Check out our interview with Dr. Masterson and Dr. Salamon to learn more about how ModMed’s helping specialty medical practices benefit from AI-powered solutions.

Learn more about B&B Sinus and Allergy Relief Centers: https://www.bnbsinusandallergy.com/

Learn more about ModMed: https://www.modmed.com/

Listen and subscribe to the Healthcare IT Today Interviews Podcast to hear all the latest insights from experts in healthcare IT.

And for an exclusive look at our top stories, subscribe to our newsletter and YouTube.

Tell us what you think. Contact us here or on Twitter at @hcitoday. And if you’re interested in advertising with us, check out our various advertising packages and request our Media Kit.

ModMed is a proud sponsor of Healthcare Scene.



< + > Standard AI is a Black Box. Here is Why RAAPID Built a Glass One for Risk Adjustment.

The problem with AI in the revenue cycle is transparency. It is powerful, but it usually operates as an unpredictable black box. In risk adj...