Application progress of artificial intelligence-driven adaptive intervention systems in adherence management of kidney transplant recipients
10.12464/j.issn.1674-7445.2026090
- VernacularTitle:人工智能驱动的自适应干预系统在肾移植受者依从性管理中的应用进展
- Author:
Junjie XIE
1
;
Gang LI
1
;
Lixin YU
1
;
Junsheng YE
1
Author Information
1. Department of Kidney Transplantation, Organ Transplantation Center, Beijing Tsinghua Changgung Hospital affiliated to Tsinghua University, Beijjng 102218, China.
- Publication Type:ReviewArticle
- Keywords:
Kidney transplantation;
Immunosuppressive drug;
Adherence;
Artificial intelligence;
Adaptive intervention;
Human-machine collaboration;
Digital therapeutic;
Machine learning
- From:
Organ Transplantation
2026;17(6):1009-1017
- CountryChina
- Language:Chinese
-
Abstract:
Poor adherence to postoperative immunosuppressive medication among kidney transplant recipients is a major modifiable risk factor for graft failure. Traditional digital health interventions mostly rely on standardized reminders, which easily lead to reminder fatigue and fail to respond to dynamic changes in patients’ behavioral motivation and clinical risks. Artificial intelligence (AI) can integrate objective clinical indicators, patient-reported outcomes, and digital behavioral phenotypes, shifting adherence management from passive reminders to proactive prediction, risk stratification and adaptive intervention. This article reviews the application mechanism of AI in kidney transplant adherence management, with a focus on data interconnection between in-hospital hospital information system (HIS)/electronic medical record (EMR)/laboratory information system (LIS) and out-of-hospital mobile applications (Apps), the automatic collection of tacrolimus blood concentration via direct connection based on LIS and test data management (TDM) system, integration of the Capability-Opportunity-Motivation-Behavior model with behavioral change techniques, low-disturbance design of real-time adaptive intervention, human-machine collaboration and the boundary of physician-led dose adjustment. Meanwhile, this article differentiates the applicable scenarios of rule engines, machine learning, deep learning and large language models and proposes clinical transformation pathways from the perspectives of model validation, implementation science and ethical supervision, aiming to provide references for constructing a safe, interpretable and sustainable precision adherence management system.
