Abstract
BACKGROUND: Intradialytic hypotension (IDH) is a serious complication of chronic hemodialysis (HD). BestShape is a novel artificial intelligence-driven system developed recently for real-time prediction of IDH. This study aimed to report the clinical results and health-economic benefits following the implementation of BestShape. METHODS: Medical records from two institutions in Taiwan that incorporated BestShape into all HD practices were retrospectively reviewed. Data from HD sessions from January 2020 to December 2023, following BestShape implementation, constituted the "BestShape group." Data from January 2016 through the end of 2019 served as the historical control group. The primary outcome was IDH frequency, and secondary outcomes were rates of cardiopulmonary resuscitation (CPR), falls, mortality, medical personnel satisfaction, and estimated cost reduction. RESULTS: In total, 213 071 HD sessions of 18 141 patients were included (BestShape 152 792 sessions; control 60 279 sessions). The mean monthly IDH rate significantly reduced from 27% in 2019 to 21% in 2023 (P < .001). The need for CPR during dialysis decreased from eight to three times per month, and post-dialysis falls decreased from 21 to seven times. The estimated annual cost savings by implementing BestShape were USD 115 658. The mortality rate during HD was not significantly different before and after BestShape implementation (8% vs 9.3%, P = .260). Medical personnel expressed satisfaction with BestShape, with a mean satisfaction score of 86. CONCLUSION: Preliminary clinical data show BestShape is associated with reductions in IDH, CPR, and falls in patients undergoing HD, and a reduction in healthcare costs and high medical personnel satisfaction.