Each year, ambitious minds in technology, finance, and healthcare attend HLTH. The 2025 edition of HLTH welcomed more than 12,000 people who navigated a labyrinth of escalators and exhibition halls across the Venetian – a journey that demanded both strategy and stamina.
AI Everywhere, Proof Nowhere
Walking the showroom floor, you’d notice that every booth, every panel was touting a ‘next-gen AI-powered platform,’ often just steps away from a puppy park or pop-up pickleball court. While AI dominated every conversation and headline with buzzwords flying so fast, they could make even a data scientist flinch, there was a clear sense of fatigue and realism. The conference even had a dedicated “AI pavilion,” which made me smirk, given that it was impossible to distinguish a non-AI company from an AI company. Almost everyone is selling AI now.
Beyond the excitement, the gap between promise and proof is wide. But here is what I noticed – Many of the AI technologies on display worked beautifully in controlled settings, yet few showed the resilience needed for the realities of a hospital room. Beyond the slick demos and marketing gloss, few could answer the real questions: Where is this actually deployed? Can clinicians touch and feel the technology working at scale in the real hospital room? Very few AI vendors seemed to be realistic about the challenges of privacy, operational burden, and deployment feasibility.
We’re watching AI promises being sold by the truckload, and investors’ exuberance is funding this frenzy – often without demanding proof of real operational savings. Money and contracts are flowing on the strength of narratives and PowerPoints without clear understanding of measurable outcomes and demonstrated results. This isn’t sustainable.
Big Tech’s Bet
OpenAI, Anthropic, and Nvidia showed up in force this year, signaling their ambition and conviction that healthcare is ripe for disruption. It’s natural for them to think “we move fast, we have the best models, we have scale, and we can crack this.” But history is instructive. Microsoft, Google, Amazon and others once announced bold moves, only to eventually readjust their strategies. Healthcare is fundamentally different and winning requires more than scale – it requires trust, clinical empathy, and the ability to operate seamlessly within regulated workflows.
Trust and Integrations
The truth is, this industry does not reward speed; it rewards trust, integration, and workflow compatibility. AI may be non-deterministic, and false positives are inevitable, yet frontline clinicians must not just use the technology but trust it. This is a problem that cannot be solved by the world’s best AI engines. It’s a human, incentive, and workflow problem.
Even so, the bright spots are real. There’s genuine optimism around administrative automation and ambient speech technology. In pharma and clinical research, AI is already accelerating drug discovery and improving clinical trial recruitment.
Providers are now adopting AI at pace, but with discernment. Leading health systems evaluate solutions based on technical maturity, production readiness, scalability, risk-level, and rapid ROI that builds organizational confidence.
Ambient documentation and revenue cycle automation continue to lead adoption; they deliver measurable ROI, tangible efficiency gains, and minimal workflow disruption. We see automation adoption increasing in people-intensive areas like prior authorization, patient engagement, and front-office operations.
The Next Buzz
After last year’s Ambient documentation frenzy, Agentic AI is getting all the attention. Vendors are now showcasing ‘AI Agents’ that promise to handle everything from charting to patient navigation. Yet, many of these offerings still feel nascent, generic, and devoid of true clinical depth. One could possess the world’s most advanced large language model, but without understanding the nuance of clinicians’ work or the importance of a handoff, even the shiniest technology will fail.
A Reality Check
Beyond the AI noise, HLTH still offered moments of grounded discussions. Mark Cuban’s remarks on pharmacy benefit managers stood out. Through his Cost Plus Drugs, he is challenging opaque pricing and pushing for transparency in drug costs, where market dynamics set prices instead of middlemen. There were also conversations about value-based care, nurses finally having a seat at the table, and the need for Interoperability and Platform thinking.
This sentiment reflects a long-overdue truth: Innovation that ignores the frontline is innovation destined to stall.
Healthcare transformation isn’t tidy – it’s messy. Building something transformative requires grit, persistence and a willingness to go deep, not just walk through the neon.
Epic UGM Recap: AI, Virtual Care, and Healthcare’s Next Frontier
Every August, healthcare leaders converge on Epic’s Verona, Wisconsin, campus for the annual User Group Meeting (UGM). This year’s event drew nearly 10,000 health executives, clinicians, and innovators eager to see where healthcare technology is heading.
Artificial Intelligence dominated conversations as Epic revealed a new AI solution suite that had been speculated about extensively in the weeks leading up to the event. The offerings unveiled by Epic CEO Judy Faulkner included:
Art: a clinician-facing AI colleague and scribe that can pull vitals, update family history, place orders, and draft notes in real time.
Emmie: a patient-facing AI assistant embedded in MyChart, helping patients interpret labs, schedule care, and navigate health needs.
Penny: a revenue cycle assistant that streamlines coding and insurance appeals.
Cosmos AI: Epic’s foundation model trained on more than 300 million patient records and designed to predict readmissions and major health events.
With over 150 AI projects in play, Epic isn’t just layering AI on top of its EHR — it’s embedding intelligence across workflows, shaping what STAT dubbed as Epic’s “Doctor Strange moment” in envisioning possible patient futures.
Industry reaction was swift. Some CIOs celebrated the potential to ease documentation burdens and reduce administrative waste. Others raised eyebrows about how native Epic AI will reshape the competitive landscape. What’s undeniable is that the AI race in healthcare is accelerating, and hospitals are taking notice.
Market Reaction: Hype to Hope
In hallways and after-hours conversations, UGM attendees grappled with the implications of Epic’s moves. While many saw the announcements as proof that AI in healthcare is “no longer just hype,” others noted that health systems face budget pressure and remain cautious about new spending. Trade-in programs and real-world ROI will influence adoption.
The buzz underscored a critical reality: as hospitals explore AI, digital health, and virtual care, they want solutions that work today and are well positioned to keep up with what’s ahead. That’s where Caregility’s story resonated.
Embedded AI for Inpatient Virtual Care
Among the approximately 40 Toolbox exhibitors at UGM, Caregility’s solutions stood out in the Inpatient Virtual Care category by delivering what competitors could not:
Live Epic integration, not native solution demos. Attendees saw firsthand how Caregility embeds AI-driven workflows directly inside Epic applications.
Bedside TV integration without Epic intervention. Caregility is the first and only platform to implement Epic Bedside TV integration independently, recently deployed at Mobile Infirmary’s new Smart Unit.
GPU-powered edge computing. Every multi-sensor Caregility Duo device includes a built-in GPU, enabling advanced computer vision at the bedside for real-time insights without reliance on cloud-only processing.
“Being one of the first live with Epic Bedside TV integration shows that we can deliver innovation without creating complexity,” said Caregility Product Manager Ben Cassidy.
Caregility Chief Innovation Officer, Bin Guan, elaborated, “Our integration with Epic is about giving hospitals practical innovation they can deploy today and build on tomorrow. By embedding AI directly in virtual care workflows within Epic, we’re reducing friction and creating value where clinicians work every day.”
Extending Epic’s Vision
Epic’s AI announcements highlight the power of predictive modeling. Actionable bedside data remains essential to success. Sensor-based technologies like those integrated into the Caregility Connected Care™ platform allow care teams to see, hear, and monitor at the point of care in real time. This unlocks tremendous potential for capturing data in the patient room, redefining what’s possible in care delivery. API-first design supports rapid integration with emerging technologies to enable next-generation care models.
“At Caregility, we’re committed to building on this momentum,” said Cassidy. “Our work with Epic isn’t just about integrating technology. It’s about helping hospitals deliver safer, smarter, more sustainable care — one room, one patient, one workflow at a time.”
During UGM, Epic released its annual Honor Roll list, spotlighting health systems pioneering digital health to improve patient care. 64 organizations were awarded for excellence in leveraging Epic’s EHR. We congratulate everyone recognized, including the many Caregility customers on the list! See the full list here.
Epic’s focus on AI and introduction of an Inpatient Virtual Care Toolbox category at the 2025 UGM confirmed that AI and virtual care are no longer side projects. They are strategic imperatives shaping the next era of healthcare.
When Virtual Nursing and AI Collide: Q&A with Mary Washington Healthcare
In our recent webinar When Virtual Nursing + AI Collide: Lessons from the Trenches, Mary Washinton Healthcare’s AVP of Hospital Operations Debra Marinari and Information Systems Analyst Trudy Osborne sat down with Caregility CNO Wendy Deibert to discuss their journey integrating remote nurses and artificial intelligence into inpatient care. Here are some key takeaways and practical strategies shared by our expert panelists.
1. How did you start your virtual nursing journey, and what were the initial challenges?
Mary Washington Healthcare began its virtual nursing journey a little over two years ago to modernize the health system’s approach to patient care. Reflecting on her 30-year nursing career, Marinari noted that many of the processes they used in the past would be considered outdated by today’s standards. She sees Virtual Nursing and AI as natural next steps in the health system’s journey to safer, more efficient patient care. One early challenge was getting experienced nurses to embrace the technology, but once they saw the benefits of improved support during staffing shortages, adoption quickly followed.
2. How is telehealth helping make care teams more agile?
Marinari and Osborne emphasized that telehealth has enhanced their ability to iterate quickly and uncover new opportunities for efficiency gains in nursing workflows. With a centralized Virtual Nursing hub on-site, remote nurses can triage and respond to inbound calls from the bedside, leaning on built-in backup coverage when multiple requests come in simultaneously. Mary Washington has seen improvements in nurse turnover rates and staff retention by implementing Virtual Nursing.
3. What workflows do Virtual Nursing and AI support?
In addition to supporting virtual admissions, discharges, and second signature verifications, virtual nurses also play a crucial role in supporting specialized workflows. For instance, virtual nurses can easily support MRI checklists and help determine the model of medical device (i.e. pacemakers) a patient has before procedures. The integration of Social Determinants of Health (SDOH) questionnaires into the virtual admissions process has increased completion rates to 98%, a dramatic improvement from before Virtual Nursing implementation. These workflows help ensure high-quality care and compliance.
4. How do you manage staffing and nurse-to-patient ratios for your Virtual Nursing program?
Mary Washington’s Virtual Nursing team is staffed with three virtual nurses per 12-hour shift, operating 24/7. The full-time roles were hired specifically for Virtual Nursing, with a minimum of two years of direct nursing experience required. Initially, the team was aligned by unit but later pivoted to a triage model that allows remote nurses to handle calls across multiple units as they come in. This flexibility has been key to managing staffing efficiently. The health system opted to add extra staffing support during peak hours when patient activity picks up between 11 am and 11 pm.
5. What platforms and AI tools are integrated with your Virtual Nursing program?
Osborne noted that the organization uses Epic for electronic health records, with virtual nurses leveraging secure chat within Epic to keep care teams connected. Mary Washington is in the process of integrating Critical Alert as their nurse call platform. Additionally, the organization is exploring health AI solutions like computer vision and ambient listening to assist with fall prevention and other early patient interventions. They are also lab-testing devices for contactless patient vital sign monitoring to improve early detection of patient deterioration and health issues.
6. How have Virtual Nursing and AI impacted patient care and staff efficiency?
One benefit of adding a remote nurse component to bedside care teams has been reduced documentation time, which has been a pain point for many nurses. By reassigning routine tasks, bedside nurses can focus more on direct patient care while remote nurses streamline workflows due to fewer interruptions. Although pilot programs are still in the early stages, the healthcare organization is excited about the promise of AI applications in areas like fall prevention and vital sign trending, which can further improve patient safety and nurse satisfaction.
7. Have patients or staff expressed concerns about virtual nursing or AI?
Marinari noted that patient resistance to Virtual Nursing has been minimal, with only one case involving a mental health patient who preferred in-person care. The staff’s attitude toward Virtual Nursing and AI has become increasingly positive, particularly as Mary Washington has fine-tuned its remote support processes over the last year. Marinari and Osborne actively work with nursing staff to continue to iterate and identify new workflows.
8. Are there specific metrics you’re tracking to measure success?
Since implementing Virtual Nursing, nurse satisfaction has improved significantly, and documentation time has decreased, both of which have been major wins for Mary Washington. The organization has also expanded its Virtual Nursing program into the emergency department and is actively tracking metrics related to nurse retention, patient outcomes, and fall rates to assess the program’s ongoing success.
9. What hardware supports your Virtual Nursing and AI implementation?
Dual-camera APS200 telehealth edge devices from Caregility, which include 40x zoom, far-end pan/tilt/zoom video capabilities, and night vision, are deployed in patient rooms to facilitate virtual nurse engagement. Bedside staff can press a vLert button to request remote nurse support. A minimalist, ceiling-mounted radar puck device supports contactless vitals capture. Osborne appreciates that deployment was straightforward, with centralized device management and strong vendor support throughout the process. The organization has also installed anti-ligature devices in the emergency department and is exploring facial scanning technology as a potential tool to support ED triage. The use of advanced hardware continues to evolve as the VN program expands.
10. What’s next for your Virtual Nursing and AI journey?
Looking ahead, Marinari expressed excitement about expanding AI capabilities, particularly in the area of computer vision for fall prevention. While the organization has made strides in reducing fall rates, this remains an area of focus, and they believe AI can further enhance early intervention. Additionally, the team is testing contactless devices that can provide early indicators of patient health trends, which they hope will lead to improved patient outcomes. They are also exploring new use cases in the emergency department and continue to work on integrating AI tools with existing platforms like Epic.
Do you have additional questions about Virtual Nursing and AI? Set up a discovery call to connect with our team of telehealth specialists and customer referral sites.
As in almost every aspect of life, Artificial Intelligence (AI) has entered the healthcare space, driving healthcare leaders to simultaneously be optimistic and concerned. The potential for AI to modernize care delivery and deliver on the promise of improving patient outcomes, increasing caregiver efficiency, and allowing practitioners to work at the top of their certifications is garnering understandable attention. However, like an iceberg, there is much more beneath the surface to consider before diving into AI adoption.
The Hidden Layers of Health AI
As we approach the AI iceberg, it’s imperative to consider the unseen, equally important, and often challenging aspects of effective implementation. The healthcare AI market is nascent and in flux, as evidenced by the spate of recent acquisitions, company failures, and startups entering the market. With regulatory frameworks still pending, healthcare AI remains a moving target. A measured approach to adoption is crucial, particularly in the high-stakes world of patient care, where poor execution will cost money and can put programs and patient lives at risk.
On the surface is the shiny object called Artificial Intelligence, but right out of the gate, a great deal of technology that is touted as AI is nothing more than automated data gathering with some logic around what to do when certain data is captured. While this can still prove to be valuable and in some cases, even more impactful than true AI, it should be considered as a different potential tool.
True AI will certainly capture data, but the difference is what is being done with that data and what inferences are deduced from comparing that data against a proven historical model to provide predictive analysis. That predictive analysis can be in the form of a diagnosis, a recommended care plan, or an alert indicating patient decline. The permutations are endless but what all AI engines have in common is the need to compare current data against known models. AI designed to learn will add new data to the model to continue to learn and enhance the accuracy.
Responsible health AI implementation warrants a look below the surface to expose the hidden challenges and considerations when evaluating technology for use in healthcare.
Data Model Considerations
As mentioned above, the success of AI in healthcare depends largely on the quality of the training data used to develop models. Data quality, representation across diverse patient populations, and model accuracy are vital to ensuring that AI systems can be trusted to make sound clinical decisions without bias. Health systems should require transparency from AI vendors and rigorous testing to ensure that models are accurate and support reliable outcomes. It is important to ask critical questions including: How was the model built? Where did the data come from? Will you be using my data? Is the data anonymized? Where is the data stored? Can you opt out of having your data used? The bottom line is, the data model is what drives the accuracy of an AI engine, and the answers to these questions will provide valuable insight on the viability of the engine.
For AI to be effective, it must be embraced by staff. There is often apprehension that AI will replace human workers. It’s important to implement AI in a way that is clinically impactful, enhances workflows, and minimizes disruption. A key question to ask clinical leadership when evaluating AI technology is, “Does it add value?” This value can come in the form of decision support, productivity gains, or actionable information. The question to ask the business is whether this is worth paying for, and if so, how much.
As an example, there is AI technology that can count the number of patient coughs per hour and provide an alert on the frequency increasing and the probable cause of the cough. Is this valuable? Who will get this information? Will it create false alerts from visitors coughing? This is one example, and one you may find value in, but ask the challenging questions and think through the impact on the clinical team, the volume of potential data and alerts, and how this will fit into the operational flow. And also, is it valuable enough to pay for the service?
AI solutions can generate large volumes of protected health information (PHI), adding to patient data vulnerability. Safeguarding the PHI generated and processed by health AI solutions is paramount. Solutions that support local edge processing can enhance security by keeping PHI within the confines of the healthcare facility, minimizing the transmission of sensitive information over the internet.
Additionally, the data captured by AI should be used responsibly. The data required for models to support machine learning AI means the solution provider will want to leverage data from your patients. You should have the option to opt out of this data participation and if you do opt in, you must be confident in how this data is being handled and protected.
This plays into another critical component of the use of AI in healthcare: patient acceptance. Implementing AI raises questions about patient rights, awareness, and consent. Health systems should consider provisions for offering patients insight into the AI solutions being used by the care facility and offer clear patient opt-out options when applicable.
I have had numerous conversations with healthcare executives and innovation teams across the country, and there is an often glaring disconnect between them and the clinical team on the floor about what is feasible. There is an overwhelming desire to expand nurse-to-patient ratios and AI is seen as the panacea for the staffing shortages and high cost of care. However, what many fail to realize is that there are a number of potential logistical challenges that can derail the intended benefits of AI. The potential for a massive amount of information and alerts that need to go somewhere can lead to caregivers becoming desensitized and potentially missing critical warnings. Evaluation of any AI solutions must include understanding the impact on the clinical team and the logistics of fitting it into the workflows as an augmentation.
AI implementation in healthcare is not a one-size-fits-all proposition. Scalable, multi-solution setups require flexible foundational infrastructure that can support a variety of technologies, both native and third-party, on-premises and cloud-based. As the market evolves, health systems will want to avoid getting locked into siloed solutions that may become obsolete as technology advances. Agile infrastructure that allows for adaptability and growth is key, enabling organizations to integrate new AI tools as they emerge and extend use cases where it makes sense.
Due to the processing power required and the massive amount of data required by AI engines, many providers utilize cloud-based solutions. Understanding the impact on your network and the amount of data being transmitted can be an important aspect of determining if the solution can scale.
AI is undeniably the shiny new object in healthcare, but it’s really not about technology. It’s about our ability to create solutions that solve problems for caregivers. We’re standing at a technological pivot point in healthcare, and leaders must approach AI implementation with open eyes, looking beyond the hype to understand the full scope of challenges and opportunities that come with implementation.
Field testing and clinical feedback are essential to ensure that AI tools meet the real-world needs of healthcare professionals. This is not a race to adopt the latest technology. It is an intentional move toward more modern, future-proof care delivery models that better serve patients and healthcare organizations. By taking a measured, thoughtful approach to AI implementation, health systems can navigate the hidden challenges of the AI iceberg and chart a course toward intelligent, truly transformative care.
Artificial Intelligence (AI) is steadily becoming a part of our everyday lives. From self-driving cars to targeted social media ads, AI is revolutionizing nearly every industry. Healthcare is no exception. At Caregility, we value AI as a tool that is capable of reducing errors, supporting clinicians, and keeping patients safe. We also recognize that AI is not without its limitations and cannot replace the experience, compassion, and empathy of professional healthcare workers.
The Utilization of AI in Healthcare
The healthcare industry is no stranger to incorporating technology to deliver groundbreaking treatments while reducing costs and improving patient outcomes. Robotic surgeries, approved in 2000, are just one example of machine learning. In recent years, AI technology has become commonplace in our smartphones, watches, and other wearable devices to monitor our vital signs, ovulation cycles, cardiac rhythms, glucose levels, and sleep patterns.
The applications in development appear limitless. The following are additional examples of AI in healthcare that are already underway, or we can expect to encounter in the near future:
Enhanced genome sequencing to generate targeted treatments for genetic conditions
Reduced time and money to discover, develop, and test new drugs
Interpretation of diagnostic images to support clinical decision-making
Virtual reality technology that allows medical students to practice surgical procedures
Brain-machine interfaces to help patients with neurological difficulties communicate or function
Assistive technologies that promote safety and autonomy so aging or disabled populations can remain in their homes
These cutting-edge concepts require deep learning, a type of advanced machine learning in which AI learns and adapts from its data without human intervention. There are many layers of deep learning, where algorithms use their neural networks, which function like the human brain, to recognize patterns and predict outcomes.
AI may have the advantage of processing information much faster than the human brain, but it has yet to master the qualities of human nature and emotion. The goal is not for AI to replace the expertise of a skilled clinician but to optimize the way care teams capture and use clinical information. Still, it would be naive to underestimate the potential power of AI, given its relatively quick integration, growth, and impact. This is why we must begin the conversation about how to responsibly utilize AI without overstepping boundaries.
What is Responsible Health AI?
As advancements in AI continue, we must consider the impacts on transparency, privacy, cost, efficiency, and person-centered care. Since there is hardly anything more personal or private than our health information, we must navigate a delicate balance that keeps ethics and patient care at the forefront. According to the editorial Responsible AI in Healthcare: Opportunities, Challenges, and Best Practices, responsible AI “seeks to ensure that AI systems are developed and deployed in a manner that is ethical, fair, transparent, accountable, and beneficial to all users.”
How Caregility Uses AI to Empower Nursing Care
Nurses comprise the largest component of the healthcare workforce, with over 5.2 million active nurses nationwide. Not surprisingly, nurses spend the majority of their time in direct patient care. They perform ongoing assessments, administer medications, and complete tasks at the direction of physicians. While nurses may constantly interact with patients, they report a lack of “presence,” meaning their ability to share in the human experience with patients. Nurses frequently cite increasing documentation requirements, high-acuity workloads, and staffing shortages as factors that limit quality time with patients.
This is where Caregility’s AI capabilities make a difference. By automating certain tasks, we can afford nurses the time necessary to provide compassionate, individualized care. Here are the ways we accomplish this:
Augmented Observation. One nurse can’t be everywhere. Computer vision technology acts as an extra layer of patient safety protection to detect behaviors that could result in patient injury or adverse events, such as falls, elopement, or violence. Our iObserver application incorporates remote patient sitters to observe up to 12 patients simultaneously. With two-way audio and visual and Augmented Observation capabilities, the virtual observer can intervene rapidly by redirecting the patient or sending an alert for assistance to the bedside team.
Vitals Trending. Monitoring resting heart rate, respiratory rate, and movement is crucial in evaluating a patient’s health status. In most traditional healthcare settings, vital signs are obtained at prescribed intervals or as needed, which can delay care if the patient is deteriorating. Furthermore, frequent vital sign checks can be disruptive to the patient’s rest and recovery. With Caregility’s continuous and contactless vitals monitoring, nurses can capture patient vitals continuously, review trends over time, and receive alerts when a significant change in patient condition is identified.
Vitals Scanning. Make the most of telehealth visits with virtual vital sign scanning. Facial scanning software obtains the patient’s blood pressure, heart, and respiratory rate in less than a minute to support remote clinical evaluations. This information is useful for patient follow-up visits, as well as remote monitoring in the home setting.
AI will never replace or replicate the warm touch, listening ear, or clinical judgment of a nurse. However, Responsible Health AI can transform the healthcare landscape through enhanced patient access to care and provider collaboration. The key is to develop processes that enable clinicians to work more efficiently without creating disconnection or distrust. Caregility remains committed to this goal by integrating progressive technology and innovative workflows without sacrificing excellent human-based care.
HIMSS24 Takeaways: Healthcare’s New Table Stakes
In the wake of several challenging years for healthcare, health IT conference-goers showed up at the 2024 HIMSS exhibit with renewed enthusiasm and a discerning eye on ROI.
Attendees from healthcare institutions across the globe convened in Orlando, March 3-6, to explore solutions for “creating tomorrow’s health,” as the conference theme promised. The forum offered providers an opportunity to consider future-state possibilities and fresh approaches to persistent problems like process inefficiency and staffing challenges. AI and hybrid care solutions were on tap to answer both needs.
Responsible AI and the Hospital Room of the Future
In healthcare’s post-pandemic era, augmented intelligence and hybrid care models are bringing additional dimensions to the patient care team. The embrace of virtual care and AI at the patient bedside is redefining what we consider to be table stakes in care delivery.
Like the ViVE conference that drew healthcare stakeholders to the West Coast in the weeks prior, HIMSS24 showcased an overwhelming array of health AI innovations. As Caregility COO Mike Brandofino observed, you couldn’t walk the conference floor without bumping into an AI company.
“The hospital room of the future is really about how you can use sensors in the room to augment information for the caregivers, thereby stretching what they can do,” said Brandofino.
“Our approach is what we call responsible AI. What we’ve done is focused on how we can enhance the information that’s getting to the caregiver. Things like vitals scanning – can we, just through an image of the patient’s face, get the blood pressure of the patient and give that to the clinician who is looking at the patient right now?”
Those solutions are closing care gaps, improving care quality for patients, and saving time for care teams. “We notify caregivers of changes that they need to pay attention to,” said Brandofino. “We don’t want to replace the caregivers. What we want to do is augment the information they have so they can spend more time maximizing their certification.”
Caregility’s “Intelligent Hospital Room of the Future” conference exhibit demonstrated ways health systems are using virtual collaboration, AI, and digital health integrations to build next-gen hybrid care models like radar-supported Virtual Nursing and computer vision-assisted Virtual Patient Observation.
A Hippocratic Oath for Health AI
Still, some providers are wary of adopting nascent technologies without proof of value and some reassurance.
As momentum builds for AI regulation on a national level, healthcare leaders are rising to meet the call for accountability within the medical field. During HIMSS24, Microsoft announced the creation of the Trustworthy & Responsible AI Network (TRAIN), “which aims to operationalize responsible AI principles to improve the quality, safety, and trustworthiness of AI in health.”
Fellow members of the TRAIN consortium include over a dozen leading health systems, OCHIN, and TruBridge. The group will share AI best practices, provide tools to enable the measurement of AI-related outcomes, and facilitate the development of “a federated national AI outcomes registry that will capture real-world outcomes related to efficacy, safety, and optimization of AI algorithms.”
As AI transparency initiatives unfold, the 2025 conference circuit could see a new system of checks and balances in place to help providers vet vendor offerings. For its part, HIMSS24 presented promising paths forward for providers keen on balancing innovation and ethics.
Empowering Compassionate Care: Caregility’s Journey with AI in Telehealth
Author: Paul Oliver, CRO, Caregility
Some years ago, while in a discussion about technology innovation, the CIO of a world-famous healthcare institution spoke about his focus on compassionate care. That conversation struck a chord with me. At its very core, healthcare delivery is a human undertaking that is skillfully tailored to meet the needs of the individual patient. We should never lose sight of that fact. Digital health solutions must be a means to that end: delivering high-quality, compassionate, efficient, and safe patient care across the continuum of prevention, diagnosis, treatment, recovery, and follow-up. Technology must blend into the background and silently help human caregivers do their best work.
With this mission in mind, our team at Caregility is now embedding AI technologies into our proven virtual care platform to bring individualized patient care to a new level. The compassionate care belief system is a driving force as we set about that work.
We recognize the immense potential for AI in healthcare, but the consequences of getting it wrong, even in a tightly defined domain such as virtual care, could be harmful to care teams and patients. We’ve seen AI applied in broad strokes – delivering a standard set of capabilities to every patient without consideration of their individual needs. We believe that approach is a mistake, and we are committed to applying AI in a responsible way that respects the needs of individual patients and adapts accordingly.
Steering Principles for AI in Telehealth
To ensure our introduction of AI creates value for care teams and contributes to individualized patient care, we’ve embraced a few key steering principles. We believe these principles will help us stay true to our compassionate care beliefs, focus our development work, and allow us to adapt quickly as the underlying technology matures and we receive real-world feedback from our customers. Those steering principles include:
Partnering our clinical and development teams closely on product direction. Our clinical solutions team, comprised of licensed healthcare practitioners with decades of experience in bedside nursing and virtual care program rollouts across leading healthcare institutions, plays a significant role in ensuring that the new AI enhancements we bring to market will positively impact the day-to-day experience of bedside clinicians and their patients.
Responding to every patient as an individual. The more an AI-driven service is customized to the individual patient, the more meaningful the clinical insights that AI helps reveal are. Striving to achieve this level of personalized care delivery helps ensure actionability without introducing nuisance alarms.
A focus on Return-on-Investment. We are committed to being clear-eyed about the financial and clinical outcomes new AI enhancements must deliver to justify the resource investment in time, money, and opportunity.
Caregility’s AI Roadmap
At Caregility, we are providing an optional set of AI-driven services that enhance specific virtual clinical programs, such as Virtual Nursing and Continuous Observation. As we build out these and other AI solutions, in order to adhere to our guiding principles, Caregility is employing an agile software development approach to release AI enhancements for virtual care in phases.
The initial phase is a minimum viable product (MVP) that embeds selected AI capability into our platform, delivering a targeted subset of our overall AI vision to Caregility customers and their patients. We will deploy the MVP to early field trial customers willing and able to partner closely with our clinical solutions team. Together they will provide our development team with a feedback loop that answers essential questions: Are we heading in the right direction? Is our vision for AI enhancement something that solves problems that exist in this institution, or should we adjust? Do our customers trust that this technology shows promise?
We’ll then make modifications and improvements to deliver subsequent phases, using customer feedback to drive value at every stage: Does it increase efficiency? Does it improve care quality? Does it positively impact patient and staff safety?
Our goal is to create additional value for our customers through the Caregility point-of-care telehealth systems that they have already deployed and are planning to deploy. With powerful edge computing and high-resolution cameras and microphone arrays that simultaneously support live two-way audio/video sessions and multimedia streaming, the Caregility Cloud™ platform was built with medical-grade AI integration in mind.
Intelligent Telehealth Endpoints
Each Caregility point-of-care system we deploy today is equipped with sophisticated, purpose-built microphone arrays and HD cameras that can introduce both remote clinician support and AI-enhanced monitoring to care teams. Here are a few examples that illustrate what purpose-built means:
Patient rooms are often noisy environments. Caregility audio processing is purpose-built to perform well in this environment so that remote clinicians can pick up on alarms and hear patients and bedside teams clearly. That same capability can be dual-purposed to feed AI language processing components that enable ambient listening and documentation.
To support patient observation at night, Caregility cameras are purpose-built with infrared night vision. That capability also ensures that computer vision-based AI performs well under low-light patient room conditions.
Our point-of-care devices can send audio and video streams to multiple locations simultaneously to support simultaneous workflows. That means a remote doctor can consult with a patient at the same time the video stream is feeding our AI engine to monitor for patient safety issues such as the bed rail positioning.
The question we seek to answer through AI enhancement is: What tasks can AI help us automate to augment the work of care teams?
AI-Enhanced Audio, Video, and Radar for Virtual Care
As we set out to introduce clinical AI to our virtual care platform, we’re focused on three key areas.
Computer vision enhancements will analyze patient room video streams to look for safety risks, best practice adherence, and workflow optimization opportunities. If the engine detects something that requires human intervention, our intent is to flag it for the right member of the care team. We will leverage our existing iConsult and iObserver applications as the main way to surface useful AI-driven insights to care team members, with incremental updates expected. We want insights to be actionable, not disruptive. Customer feedback from early field trials will inform our roadmap.
In the outpatient context of telehealth, we plan to extend video stream processing to virtual visits to gather patient vital signs (i.e., respiratory rate, heart rate, blood pressure, and body temperature) from facial video analysis. We’ll present this live data to the remote clinician so they can see it as part of their remote consultation.
Acoustic-based AI will listen to audio streams for patterns that can alert staff to patient stress or behavioral issues. In the inpatient context, we are researching embedded AI to identify medically relevant parts of conversations between clinical staff and patients to relieve some of the clinical note-taking burden for care teams. Ambient listening will inform structured clinical data capture for nurses to review before being documented in the EHR.
We are also working to integrate an optional radar sensor with our point-of-care devices to support contactless vitals streaming in inpatient care. Trending heart and respiratory rates over time could signal deterioration or changes in the patient’s condition. Our goal is to support personalized compassionate care and alert the appropriate clinician if vital signs diverge away from that patient’s baseline. Sepsis is a major patient safety factor in our hospitals, and we believe widespread adoption of this type of technology will help attack that problem and others.
The AI journey ramps up for Caregility in 2024, when we release our first commercial offerings addressing two of our key focus areas: Augmented Observation and Vitals Trending. We will partner closely with our early field test customers to measure the impact those solutions have on key performance indicators. Customer feedback will fuel our next wave of intelligent telehealth enhancements aimed at advancing compassionate, personalized care.
AI’s Potential in Inpatient Clinical Care
Traditionally, healthcare has been regarded as a laggard in embracing IT innovations, primarily due to the inherent complexity of care delivery and the stringent regulatory environment. However, this perception is undergoing a significant transformation. The widespread adoption of acute telehealth has played a pivotal role in this shift, providing care teams with valuable hands-on experience that has helped build trust in health technology. As a result, healthcare professionals have become more comfortable with and open to leveraging digital health tools in their daily practices.
This newfound receptivity to digital enablement has set the stage for a remarkable leap forward in the industry. This is evidenced by the recent surge of interest in artificial intelligence (AI) in healthcare. The proliferation of generative AI and the urgent need to find new solutions to the ongoing staffing crisis are further fueling interest in exploring AI’s place in clinical care.
AI-Enhanced Telehealth
Although headway has been made regarding operational use cases for machine learning-based AI in backend process improvements, healthcare teams are eager to identify and implement AI solutions that can enhance clinical workflows, produce more precise diagnoses, and improve patient outcomes. For the many hospitals equipped with synchronous, audio/video-based telehealth services at the bedside, infrastructure exists to introduce AI services at the point of care as well.
In the realm of virtual care, the evolution of intelligent telehealth endpoints has been remarkable. As telehealth has evolved from audio-only interactions to feature-rich video encounters, camera and microphone quality have improved to keep pace. One noteworthy development is the integration of edge computing capabilities into telehealth endpoints, enabling the support of AI applications. New multiplexing technology supports multiple cameras, allowing care teams to simultaneously support virtual patient engagement and video-based AI solutions.
In addition to high-fidelity camera and microphone arrays, new sensors such as radar technology are making their way into telehealth platforms. These sensors are capable of continuously capturing valuable clinical data while minimizing disruption to the care process. These advancements in telehealth are allowing care teams to infuse AI services such as Augmented Observation and Vitals Trending into bedside care processes, arming clinical teams with patient safety reinforcement tools that reduce pressure on staff and lead to better outcomes.
Promise and Precautions in Clinical AI
While the adoption of AI in healthcare holds immense promise, it also comes with its set of challenges and precautions.
In a recent interview with Healthcare IT Today editor John Lynn at the 2023 HLTH conference, Caregility Chief Product and Engineering Officer Kedar Ganta lauded AI’s ability to bring in ambient intelligence from audio, video, and sensor feeds, “whether it’s collecting vitals or documenting notes for the clinician.” This technology operates in the background, collecting vital patient information and supporting care processes without causing disruption.
Ganta does note, however, that amidst the excitement surrounding AI, the topics of trust and accountability often receive insufficient attention. Establishing patient and provider trust in AI systems is crucial. “This is where regulation comes into the picture,” says Ganta. “It’s a balance between over-regulating something versus promoting innovation.” Reliable data and transparency in AI solutions are essential for building trust, and the inner workings of AI algorithms should be shared with providers to enhance visibility into these factors.
Additionally, “providers should have the ability to override the AI decision” Ganta advises. Creating a regulatory framework and an independent body to oversee AI in healthcare is vital to address these concerns. As AI continues to reshape the healthcare landscape, these discussions about trust, regulation, and accountability are essential for harnessing the full potential of AI while safeguarding patient wellbeing.
As the healthcare industry continues to embrace these advancements, the future holds exciting possibilities for improved patient outcomes and experiences.
AI-Enhanced Telehealth: Hope or Hype?
ChatGPT and a plethora of other AI-powered applications are rapidly gaining popularity in today’s tech-driven world. In healthcare, AI and machine learning algorithms are being adopted to drive efficiency in patient-facing and back-office settings alike.
One of the clinical frontiers gaining attention is the augmentation of virtual care programs with AI tools such as computer vision, ambient clinical intelligence, and contactless monitoring. By bringing these AI enhancements into virtual workflows in the inpatient setting, healthcare organizations hope to positively impact patient safety, clinical outcomes, care team experience, and operational performance.
During a recent fireside chat, Caregility President and COO Mike Brandofino sat down with Healthcare Innovation editor Mark Hagland to explore the practicality, best practices, and perils associated with selecting and adopting AI technology to advance telehealth.
AI’s Potential in Acute Virtual Encounters
AI is showing promise in clinical use cases in acute care settings where staffing shortages and burnout are prominent. As Brandofino sees it, one of AI’s biggest benefits is in “augmenting the information that a clinician or caregiver has access to with more clinical insight than they’d be able to gather on their own.” When evaluating tools, he encourages stakeholders to consider the impact: “Is it taking tasks away that can potentially save staff time? Is it a tool that adds to productivity?”
One of the AI functions Brandofino sees potential in is radar-based contactless monitoring. These tools continuously capture patient vitals such as heart rate and breathing rate, as well as track motion in the room. This allows caregivers to see trends over time.
“The AI part of that is the algorithms can detect changes in that pattern that mean something,” Brandofino explains. He offers a practical use case example. “That radar device can tell you, based on telemetry, that a patient is starting to wake up. Now think of a post-op situation where the nurses have to be there when the patient wakes up disoriented. Can you just have a contactless device notify them when the patient is starting to wake up so they can get in there then instead of sitting there for 30 minutes waiting?”
Automated, contactless vitals monitoring also accelerates the frequency and timeliness of clinical documentation.
“If you think about what happens with nurses as they do their rounds and take vital signs, many times they don’t get that information into the EHR until the end of their shift or hours later,” Brandofino notes. AI tools can gather vital signs many times throughout the day and put it through an algorithm to evaluate if the patient is getting better or worse. This allows care teams to intervene earlier and potentially improve outcomes.
Ambient clinical intelligence uses AI tools like natural language processing to draft clinical notes and reports, posing similar efficiency benefits. In care environments where resources are thin and burnout is high, those incremental time savings can add up.
Caveats to Consider When Adopting Health AI
As you evaluate AI solutions to bring into patient care delivery, it’s important not to become enamored with the technology before understanding where it fits into the patient care workflow. Brandofino recommends including all stakeholders—clinical, IT, and operations—in evaluations. “How are you going to support your device fleet? Consider the clinical workflow as well as the experience on the patient side.”
“Think about the operational logistics of supporting what you’re doing,” Brandofino advises. “What we suggest to our customers is to understand the impact that you’re going to have on the staff on the floor and think about what that is going to be like at scale.” Nurses are some of the most interrupted people in healthcare. The last thing you want to do is introduce new tools that add to their stress level, whether that be an overabundance of false alarms or device overload.
Given the newness of many AI tools entering the market, it’s also important to consider who you’re partnering with. Has the tool been implemented in one or two patient rooms or thousands of rooms? Are there examples of in-market success that can offer a roadmap?
Combining AI and Telehealth to Empower Caregivers
By integrating AI with virtual care, healthcare organizations can modernize care delivery with innovative new tools and keep the human element of care intact. AI can drive intelligent clinical alerting, while virtual engagement channels serve as a bridge for immediate staff intervention. When combined, these resources amplify what virtual teams are capable of supporting remotely, doubling down on reducing the burdens on bedside staff.
“We believe that combining that remote nurse with smart technology to help gather telemetry in the room will be really impactful in improving care for patients in the long term,” says Brandofino. “Don’t feel like you have to put in a siloed solution just to get access to AI technology. Look for players that are capable of integrating with what you already have. If you already have high-end cameras and mics in the room with edge processing, what else can you gather in the patient room to give to caregivers?”
Ultimately, healthcare organizations that focus on applying technologies that solve real problems that exist today around the shortage of nurses, productivity, and quality of life for staff will have the most impact.
Interested in learning more about AI-enhanced hybrid care solutions? Contact us today!
