Artificial Intelligence-Assisted Care After Hospital Discharge for People With Decompensated Cirrhosis

Trial details

Who runs it

Beijing Tsinghua Chang Gung Hospital

Trial code

NCT07931404

Record updated

12 November 2026

Original description

This study will compare two ways of supporting adults with cirrhosis after they leave hospital. Cirrhosis is severe scarring of the liver. Eligible participants will have been hospitalized for liver-related complications and will be ready for discharge. The main question is whether adding an artificial intelligence (AI)-assisted management service to usual care reduces unplanned hospital care caused by worsening liver disease during the next 90 days.Read full original descriptionShow less

Development and Evaluation of an AI Agent-Assisted Proactive Postdischarge Management Model for Patients With Decompensated Cirrhosis

This study will compare two ways of supporting adults with cirrhosis after they leave hospital. Cirrhosis is severe scarring of the liver. Eligible participants will have been hospitalized for liver-related complications and will be ready for discharge. The main question is whether adding an artificial intelligence (AI)-assisted management service to usual care reduces unplanned hospital care caused by worsening liver disease during the next 90 days.

The study plans to enroll 180 adults at three centers. Participants will be assigned by chance, with an equal chance of receiving usual care alone or usual care plus AI-assisted proactive management. Participants receiving the additional service, or their authorized caregivers, will complete a brief daily health report taking about 3-5 minutes and may report symptoms when they occur. The system helps organize reports and identify information requiring attention. Nurses check the information, and physicians make clinical decisions. The system does not independently diagnose illness, prescribe treatment, or change medication.

Both groups will continue their usual medical care and have study follow-up contacts on days 7, 30, 60, and 90. The study will compare unplanned emergency visits, hospital observation, acute same-day treatment, and readmissions related to worsening liver disease. Safety and the burden of using and delivering the service will also be assessed. Participants with urgent symptoms should seek emergency care without waiting for a response from the study platform.

Inclusion: Aged 18 years or older on the date of signing the informed consent form, with no upper age limit; cirrhosis confirmed by a hepatologist based on clinical findings, laboratory tests, imaging and/or endoscopy, and histopathology when necessary; The index hospitalization is related to decompensated cirrhosis or associated complications, with evidence of decompensation such as ascites, overt hepatic encephalopathy, or variceal bleeding, or hospitalization for infection, worsening renal function, or other related complications in a patient with previous decompensation; the patient currently meets the criteria for discharge; Plans to return home or to an out-of-hospital setting that supports follow-up, with randomization completed on the day of discharge before initiation of the intervention and the 90-day observation period beginning at discharge; Is alert and has the capacity to understand the study and independently provide informed consent; the participant or an authorized reliable caregiver can complete mobile reporting or telephone-assisted data entry as permitted by the study; Voluntarily provides written informed consent and separate consent for necessary data processing, and agrees to 90-day outcome ascertainment conducted at the same frequency in both study groups.

Exclusion: Receiving end-of-life care, or participation is considered to impose a clearly disproportionate burden by the treating clinical team; Concurrent participation in another interventional study that clearly conflicts with this study; Persistent severe impairment of consciousness or cognition that prevents the participant from independently providing informed consent; No feasible means of contact, or inability to obtain the assistance necessary for study participation; Near-term hospitalizations are primarily driven by tumor progression or planned oncological treatment, such that cirrhosis-related outcomes cannot be reliably distinguished from tumor-related care.

ClinicalTrials.gov NCT07931404