A trio of companies has introduced new advances in artificial intelligence designed to improve patient outcomes and reduce clinician burnout.

RapidAI announced a new tool that helps front-line clinicians make decisions when minutes count in acute stroke care, where "time is brain," without waiting for additional imaging.

Meanwhile, Clew Medical is unveiling new FDA-cleared machine learning algorithms that not only reduce patient condition alerts, but have also been shown to increase accuracy in detecting clinical deterioration.

Finally, a new dataset on obesity from Dandelion based on longitudinal patient records could unlock insights on efficacy and side effects faster than could trial-and-error prescribing in a class of popular diabetes and weight loss drugs, the company said.

Stroke AI imaging

The U.S. Food and Drug Administration has cleared San Mateo, California-based RapidAI's AngioFlow module, the company said.

The module delivers perfusion imaging analysis directly within its interventional platform to enhance clinical confidence, streamline imaging workflows and potentially improve patient outcomes.

Time is a stroke-patient variable at both small or rural hospitals and large health systems, where transfers can run down the clock, threatening life or future quality of life. Real-time AI analysis at the point of care can provide caregivers with immediate information that can make a huge difference in what happens to patients after a stroke.

For care teams working in the Rapid AI interventional suite, they can now generate qualitative perfusion maps within minutes to help assess ischemic change in brain regions with reduced cerebral blood flow, thus saving time and reducing redundant imaging, the company said.

"By avoiding unnecessary scans, stroke patients can receive the timely care that can be the difference between being able to walk out of a hospital to their homes versus being discharged to a skilled nursing facility," said Dr. Abhishek Singh of the Creighton University School of Medicine in Omaha, Nebraska in a statement from RapidAI.

"We can now support stroke AI imaging along the entire patient pathway, from the initial non-contrast CT scan all the way to the interventional suite," added Karim Karti, RapidAI CEO.

Patient deterioration AI

Boston-based Clew Medical, a clinical surveillance platform, announced that its second-generation AI/ML models predicting patient deterioration have received 510(k) clearance from the FDA.

"AI and machine-learning technology in this space must undergo the same level of scrutiny and diligence in design, development, testing and validation as other medical devices used by clinicians," Paul Roscoe, Clew's CEO, said in a statement Monday.

Clearance included the FDA's approval of Clew's pre-authorized change control plan, which will allow certain future changes to the system-input data set without needing to file for a new clearance, the company noted.

The Clew platform, which received initial clearance as a medical device in 2021, offers health systems an early identification of a patient’s risk of deterioration that is five times more accurate than alerts, according to a study by UMass Memorial Medical Center and WakeMed Health & Hospitals published in November in CHEST.

Where false alerts are notorious for overburdening hospital staff, the researchers wanted to know, "Do ML alerts, telemedicine system-generated alerts or biomedical monitors have superior performance for predicting episodes of intubation or administration of vasopressors?"

The algorithms were trained to predict intubation and vasopressor initiation events among critically ill adults, and the study comparing them to traditional alerts found that on average, 98% of bedside monitoring alarms were false positives, according to Clew Medical. 

"ML-derived notifications for clinically actioned hemodynamic instability and respiratory failure events represent an advance because the magnitude of the differences of accuracy, precision, misclassification rate and pre-event lead time is large enough to allow more proactive care and has markedly lower frequency and interruption of bedside physician workflows," the researchers said in the published abstract.

Obesity data trove

Leveraging its consortium of nonacademic medical center health system partners, Dandelion Health has launched a new library of all clinical data – structured and unstructured data that is refreshed quarterly – for a class of medications used to treat type 2 diabetes and obesity.

According to the New York-based startup, the dataset contains the full longitudinal patient records for millions of patients – thousands of whom have taken Glucagon-like peptide-1 agonists, and 200,000 of whom are on prescribed GLP-1 agonists – and not just from electronic health records. 

The library includes image and waveform data island content in clinical notes, offering unprecedented insight into patient journeys and the impact of GLP-1s, the company said in its announcement Tuesday.

"By making the rich, multi-dimensional data offered by unstructured modalities available and accessible – and connected to real-world treatment patterns and outcomes – people who use the library can answer key questions about the critical role that GLP-1 based treatments will play in clinical care," said Shivaani Prakash, Dandelion’s head of data. 

The data could enable medical researchers to understand how GLPs affect everything from the physical structure of the heart to specific side-effect profiles, Dandelion said, noting that the company is working on proofs of concept. One academic medical center research partner developed an algorithm that segments abdominal CTs to quantify fat loss and muscle and bone preservation. 

Using the dataset researchers can evaluate the quality of weight loss through biomarkers found in body scans, compare the efficacy of treatments, demonstrate secondary benefits, quantify side effects associated with taking GLP-1s and develop precision medicine tools that match patients to more effective treatment plans. 

"Our GLP-1 dataset will help cardiometabolic disease enter its precision medicine era," Elliott Green, cofounder and CEO of Dandelion Health, said in a statement.

Obesity care lags far behind that of immunology, oncology and other specialties, he noted. 

"What got those markets to where they are today was data – data that revealed underlying mechanisms of disease, how individuals’ diseases look different and consequently, how they might respond to therapy differently."

The data could be used to personalize patient care by helping providers understand which patients would be at higher risk for bone thinning or loss of muscle mass, and then placing high-risk patients on adjuvant medications or therapies, increasing screenings or prescribing a different drug, a spokesperson for the company said.  

The article was updated with comments from a Dandelion Health spokesperson.

Andrea Fox is senior editor of Healthcare IT News.
Email: afox@himss.org
Healthcare IT News is a HIMSS Media publication.