Kidney Week 2025 Annual Meeting-Translating Evidence into Real-World Clinical Scenarios in AKI and Critical Care Nephrology

Video Transcription

Video Summary

This symposium on translating evidence into real-world AKI and critical care practice focused on implementation science and artificial intelligence (AI) as complementary tools to close the “know–do gap.”<br /><br />Jeremiah Brown introduced implementation science as the study of methods that speed adoption and sustainment of evidence-based practices, addressing the long delay between research, guidelines, and routine care. He distinguished interventions (“the thing”) from implementation strategies (“how to help people and places do the thing”), and highlighted frameworks such as RE-AIM/PRISM and CFIR, plus hybrid effectiveness–implementation trial designs. He reviewed kidney-related examples (e.g., Dialyzing Wisely, Contrast Risk) and described his team’s VA cath-lab trial using collaboratives, technical assistance, and audit/feedback dashboards to reduce contrast-associated AKI with sustained benefit and favorable cost-effectiveness. He summarized outputs from ADQI 35, which mapped a research agenda for applying implementation methods to AKI, including multidisciplinary engagement, standardized reporting (e.g., STARI), strategies for low-resource settings, and stronger identification of evidence–care gaps.<br /><br />Lily Chan discussed “AI for implementation” (trial screening from EHRs, extracting social determinants from notes using local large language models, faster qualitative theme detection, remote monitoring, decision support, and chatbots for multilingual consent/education) and “implementation for AI,” emphasizing stakeholder involvement, trust, interpretability, workflow fit, bias/inequity, privacy, and post-deployment monitoring for drift.<br /><br />Rupesh Raina reviewed pediatric AKI AI challenges: small/fragmented datasets, heterogeneous phenotypes, high ethical/consent complexity, and limited evidence. He presented emerging pediatric databases, prediction efforts in neonates, transplant-related AKI risk stratification, and the need for precision, globally usable tools.<br /><br />Se-Joon Kim shared bedside lessons from Korea: risk scoring (SPARC), EHR-based AKI alerts that work only when embedded in actionable workflows (including easy nephrology consult triggers), outpatient contrast-AKI prevention automation, and RNN-based AKI prediction with explainable factor outputs. He highlighted real-world barriers (low event rates, staffing disruptions) and the central theme: AI succeeds when integrated with human response, explainability, and workflow.

Asset Subtitle

Moderator(s): Pavan Bhatraju, Jay Koyner

Presentation(s):

Key Concepts of Implementation Science Applied to AKI and Critical Care Nephrology - Jeremiah Brown
Role of Artificial Intelligence and Digital Health in Implementation Science Applied to AKI and Critical Care Nephrology - Lili Chan
Building Artificial Intelligence Solutions for Pediatric Nephrology Challenges - Rupesh Raina
Artificial Intelligence and AKI Detection: Tales from the Bedside - Sejoong Kim

ASN thanks the Korean Society of Nephrology (KSN) and KSN Ambassador, Dr. Sejoong Kim.

Note: Continuing education credits are not being offered for this session.

Meta Tag

Date 11/6/2025

Pathway 1 AKI and Critical Care

Pathway 2 Dialysis

Session ID 506907

Keywords

acute kidney injury (AKI)

critical care practice

implementation science

know–do gap

evidence-based practice adoption

implementation strategies

RE-AIM

CFIR

hybrid effectiveness-implementation trials

contrast-associated AKI prevention

audit and feedback dashboards

electronic health records (EHR)

AI clinical decision support

pediatric AKI prediction

AKI alerts and workflow integration