Abstract: FR-PO102
Can Augmented Intelligence Assist in Delivering Continuous Renal Replacement Therapy? A Scoping Review
Session Information
- AKI: Epidemiology, Risk Factors, Prevention
November 04, 2022 | Location: Exhibit Hall, Orange County Convention Center‚ West Building
Abstract Time: 10:00 AM - 12:00 PM
Category: Acute Kidney Injury
- 101 AKI: Epidemiology‚ Risk Factors‚ and Prevention
Authors
- Hammouda, Nada, The University of Texas Southwestern Medical Center, Dallas, Texas, United States
- Neyra, Javier A., University of Alabama at Birmingham, Birmingham, Alabama, United States
Background
Despite advances in renal replacement therapy (RRT) technologies, the overall mortality rate for ICU patients on RRT is over 50%. Yet, CRRT delivery is not standardized, and there are no validated quality indicators to assess structure, processes and/or outcomes of programs. Digital health and artificial intelligence (AI) technologies have transformed most aspects of health service delivery, including patient diagnostics, risk classification, clinical decision support, and even workflow optimization such as patient scheduling. The current state of literature on the applications of AI in CRRT delivery remains unknown. We aimed to characterize the state of existing literature on the use of AI in CRRT delivery, and identify current gaps and future research priorities.
Methods
We searched PubMed, OVID Embase, Web of Science, Cochrane, Scopus and ProQuest, from inception onwards, for original papers published or translated in English. Study summaries were tabulated and analyzed for insight on the current state of research and potential future directions.
Results
12 papers were selected, 6/6 (50%) in 2021, and 7/12 (60%) focused on machine learning to augment CRRT delivery. All innovations were in the design/validation phase of development. Primary research interests focused on early indicators of CRRT initiation, clinical prognostication, and identifying risk factors for mortality. Secondary research priorities included dynamic CRRT monitoring, predicting complications, and automated data pooling for point-of-care analysis. Identified literature gaps included implementation barriers, quality indicators, bias ascertainment, and quantifying social or machine-generated healthcare disparities.
Conclusion
Research on AI applications in CRRT delivery grows exponentially but the field remains premature. Future studies are needed on validation, structural implementation, quality assurance, bias and equity ascertainment. The next breakthrough in the field should be smart task delegation: its utility, hazards and overall benefit for patients, clinicians, and health systems.