Predicting which patients with serious liver disease are at greatest risk of dying in the hospital is a major challenge for physicians. Cirrhosis, caused by factors such as alcohol, hepatitis, or excess fat, severely damages the liver, leading to repeated hospitalizations and life-threatening complications like kidney failure, severe infections, and hepatic encephalopathy. Once hospitalized, patients with cirrhosis face a high risk of death, but identifying those most in danger remains difficult. Now, researchers have developed an artificial intelligence (AI) model that can help save lives by providing a better prediction of which hospitalized liver patients are at greatest risk of dying.

The AI model was developed by a global research team, led by Virginia Commonwealth University (Richmond, VA, USA), using a prospectively collected database of over 7,000 cirrhosis patients treated at 121 hospitals across six continents to see if AI could help physicians assess their hospitalized patients with cirrhosis. Detailed information on patient admissions, complications, treatments, and outcomes was gathered to test how well four different models could predict which patients would die in the hospital, including traditional statistical and machine learning techniques. The team found that the Random Forest machine learning model was the most effective, detecting hidden warning signs with higher accuracy than the traditional logistic regression model. Even when the model was simplified to include only the 15 most important risk factors—such as kidney failure, brain complications, and infections—it remained highly effective and easy to use globally.

By providing reliable risk estimates from just these inputs, the tool helps health providers decide whether to intensify care, initiate hospice discussions, or consider early transfer or transplant. The model was further validated using data from 29,000 U.S. military veterans treated in VA hospitals, where it continued to outperform older scoring systems despite differences in patient demographics. The results, published in Gastroenterology, highlight the model’s clinical relevance. The tool is now being made available to hospitals worldwide for use in caring for patients with advanced liver disease. Its ease of use and accuracy mean better planning, targeted care, and timely communication with families. Looking ahead, the team is promoting broader adoption of the tool to ensure that cirrhosis patients receive the most appropriate care at the right time.

“This means a doctor can have more confidence about which patients need the most urgent care, which ones might need hospice discussions with family members, who could need transfer to better-equipped hospitals and which patients are likely to recover,” said Jasmohan Bajaj, M.D., corresponding author of the study. “Medically and nonmedically, we can better approach the patient if we have a better handle on the patient’s condition.”

http://www.gzjiayumed.com/en/index.asp