WASHINGTON — Integrating artificial intelligence (AI) with wearable technology may improve the care of patients with liver disease by guiding medication adjustments and alerting clinicians to early signs of disease progression, according to Adam Buckholz, MD, gastroenterologist and assistant professor of medicine at NewYork-Presbyterian/Weill Cornell Medical Center in New York City.
In a presentation at The Liver Meeting 2025: American Association for the Study of Liver Diseases (AASLD) and a subsequent interview with Medscape Medical News, Buckholz highlighted the promise of remote monitoring with wearable devices in hepatology.
“By leveraging the passive data collection of wearable technology, we can collect multiple data points on a second-by-second basis, providing nearly immediate biological insight,” Buckholz said.
AI applications can be instrumental in processing this overwhelming volume of information.
“The machine learning and advanced analytical tools of AI allow for processing and understanding of that data in a way that would otherwise be impossible.”
Tapping Into Wearable Insights
Wearable technology may help address long-standing gaps in treatment, Buckholz said in his presentation. Many patients miss discharge follow-up appointments after hospitalization, and in-person clinical encounters are often brief. In contrast, remotely collected wearable data are more accurate and convenient, less biased and sensitive to change, and unaffected by geographic distance between patients and clinicians.
More than one third of Americans (36%) already use wearable devices, according to a 2025 study in the Journal of Medical Internet Research. But most devices were not developed or validated for identifying health outcomes.
“This represents a big missed opportunity,” Buckholz said. “Early changes happening to patients outside the clinical setting are not being used by their healthcare teams to improve outcomes. Wearables can help patients take charge and be proactive with their healthcare and offer both doctors and patients a stronger sense of control.”
What Makes a Good Biomarker?
Wearables provide insight into standard metrics such as steps, heart rate, sleep, and oxygenation. In addition, they can track body activity/position, mobility, geographic location, falls, heart rate variability, heart arrhythmia, voice characteristics, reaction speed, swallowing pattern, gut motility, and sweat, among other things.
To serve as effective biomarkers, however, this information should be actionable, sensitive, specific, reproducible, scalable, biologically meaningful, and acceptable to patients, Buckholz said.
“By observing patients with wearable technology and then interpreting these data with AI facilitation, researchers are hoping to identify such biomarkers in liver disease.”
Potential applications in liver disease include monitoring improvement or progression in metabolic dysfunction-associated steatohepatitis and detecting early signs of hepatic decompensation.
“These technologies could also help patients take a more active role in their disease and improve communication in periods of both illness and wellness,” Buckholz added.
Current research into wearable technology in liver disease supports these possibilities. Studies of wrist-worn activity monitors have shown that reduced activity is associated with increased waitlist mortality among liver transplant candidates, as well as increased hospital admissions and mortality in patients with cirrhosis. Other investigations with wearables have linked sleep disturbances to poorer post-liver transplant outcomes and explored skin patches and transdermal sensors for detecting blood alcohol levels and inflammatory markers predictive of outcomes in cirrhosis, Buckholz said.
Overcoming Barriers to Adoption
A major barrier to widespread implementation in clinical practices is the so-called “wearable paradox,” whereby early adopters of wearable technology tend to be relatively healthy, whereas those at highest risk are less likely to already use such devices, Buckholz noted. Increasing access, understanding, and uptake in vulnerable populations will therefore be critical.
Additional challenges include determining how to distill massive volumes of wearable data into concise formats that can be incorporated into electronic medical records (EMRs) and easily communicated to patients.
From a research standpoint, Buckholz outlined several necessary steps. First, clinicians need a better understanding of which digital biomarkers are most predictive of gastrointestinal (GI) and liver disease outcomes. For example, although heart rate variability may change during sleep in patients with high-risk liver disease, it remains unclear whether these changes reliably precede clinical deterioration.
Second, clinical trials must demonstrate that acting on digital biomarkers translates into improved outcomes. Implementation research will then be needed to support long-term patient adherence and efficient data transfer to clinicians.
“Finally, we need to build practice models, whether through integration into EMR, useful and simple phone applications, or a combination, that allow communication of results back to patients and let doctors bill for the time spent dealing with these data,” Buckholz said. Incentivizing both patients and providers will also be key to ensuring sustained adoption.
Wearable Technology Beyond the Liver
Wearable technology has potential applications across gastroenterology, said Andres Duarte-Rojo, MD, professor of medicine and surgery at the Northwestern University Feinberg School of Medicine, Chicago, and medical director of Liver Transplantation and the Living Liver Donor program.
AI can serve as an invaluable ally for sorting through the volumes of complex data generated by these devices, Duarte-Rojo told Medscape Medical News. He added that any patient can be a candidate for wearables.
“When tracking a generic feature that affects overall health, such as mobility/exercise or heart rate variability, any patient would benefit from it,” he said. “Cardiorespiratory fitness not only is the ultimate predictor of mortality [but also can] be informative to GI health since both low physical activity and low heart rate variability have been associated with poor outcomes in various chronic GI ailments.”
More targeted applications include monitoring of alcohol use via sweat sensors for alcohol use disorder and liver disease or detecting failed swallowing attempts in patients with dysphagia.
Provider Bottlenecks
Although patient adherence can be a challenge, Duarte-Rojo said the larger obstacle may lie on the provider side.
“For example, if you are tracking a change in voice or mobility to identify hepatic encephalopathy, you need the health IT [information technology] infrastructure to confirm the signal from the wearable is reliable, that signal needs to be filtered to identify the endpoint or event of interest, and that event needs to be displayed on a dashboard a provider can interpret,” he said.
That alert must then prompt timely clinician intervention, such as administering more lactulose or referring the patient to urgent care.
This feedback loop requires technical support, continuous monitoring of the provider’s dashboard, EMR integration, and efficient patient communication to deal with responses ranging from false positives to potential life-threatening emergencies. Consequently, the use of wearables in GI care remains largely limited to research at the moment.
Looking ahead, Duarte-Rojo said calibration of devices for specific GI diseases and implementation trials demonstrating feasibility and cost-effectiveness will be crucial.
Ultimately, he said, “the benefits that wearables will bring to daily care should translate into a more efficient use of healthcare resources, such as decreased hospitalization, in order to convince third-party payers to support this nascent technology and novel healthcare paradigm.”
Buckholz disclosed receiving research funding from the Weill Cornell Clinical and Translational Science Center KL2 award and the AASLD Frank J. Tusa Pilot Award in Hepatic Encephalopathy but having no financial conflicts of interest. Duarte-Rojo disclosed having financial relationships with CIMA Sciences, LLC.
