What is the clinical evidence from this study?

Study design: Observational. Population: Acute kidney injury (AKI) requiring renal replacement therapy (n=3369). Intervention: Hierarchical Reinforcement Learning for Renal Replacement Therapy (HRRT) vs. Clinician-led care. Primary outcome: Estimated mortality rate (ARR 6.1).

What question did this study set out to answer?

The study aims to create a clinical decision support system using hierarchical reinforcement learning to improve decision-making in renal replacement therapy.

June 22, 2026npj Digital Medicine0 citationsOpen Access

HRRT: hierarchical reinforcement learning for renal replacement therapy decision support

QXQianyi XuQingdao University FWFeng WuNational University of Singapore ZTZi Yi Christopher ThongNational University Hospital

Key Result

The HRRT AI-driven clinical decision support system reduced the estimated ICU mortality rate by 6.1 percentage points (from 47.7% to 41.6%) compared to clinician-led outcomes.

Key Points

The study aims to create a clinical decision support system using hierarchical reinforcement learning to improve decision-making in renal replacement therapy.
Conducted a retrospective multi-center study involving 2467 ICU stays of 1439 patients from a US hospital for model training.
Evaluated model performance on two external validation sets with ICU stays from the Netherlands and China.
Implemented hierarchical reinforcement learning to develop personalized recommendations for renal replacement therapy.
Estimated mortality rate decreased by 6.1 percentage points from 47.7% to 41.6% compared to clinician-led outcomes, 95% CI: 45.2–50.0 to 35.6–47.2.

Study Design

Type

Observational (n=3,369)

Multicenter

Yes

Structured PICO

Does an AI-driven HRRT clinical decision support system reduce mortality in ICU patients with AKI compared to clinician-led care?

Population

3,369 ICU patients with acute kidney injury requiring renal replacement therapy from hospitals in the US, Netherlands, and China.

Exposure

Hierarchical Reinforcement Learning for Renal Replacement Therapy (HRRT), an AI-driven clinical decision support system

Comparator

Clinician-led outcomes (empirical experience)

Outcome

Estimated mortality ratehard clinical

An AI-driven hierarchical reinforcement learning model for renal replacement therapy decision support may reduce ICU mortality in patients with acute kidney injury compared to standard clinician-led care.

Main Result

Absolute Risk Reduction: 6.1

Absolute Event Rate: 41.6% vs 47.7%

Absolute Risk Reduction: 6.1%

Abstract

Renal replacement therapy (RRT) is a critical intervention for patients with acute kidney injury (AKI). However, clinical decision-making regarding the timing of initiation, modality selection, optimal ultrafiltration rate, and weaning criteria remains highly complex and exhibits significant practice variation. Current RCTs fail to guide adaptive strategies, leaving dynamic decision-making heavily reliant on empirical experience. Consequently, a substantial gap persists in dynamically adjusting treatment in response to patient-specific clinical progression. We conducted a retrospective multi-center study to develop H ierarchical R einforcement Learning for R enal R eplacement T herapy (HRRT), a holistic clinical decision support system (CDSS) that covers the full decision-making process in RRT. 2467 Intensive Care Unit (ICU) stays of 1439 patients from a US hospital were used for training and internal testing. The model’s performance was evaluated on two external validation sets, where we selected 1085 ICU stays of 1085 patients from Netherlands and 1230 stays of 845 patients from China. The estimated mortality rate decreased by 6.1 percentage points from 47.7% (95% CI: 45.2–50.0) to 41.6% (95% CI: 35.6–47.2) compared to clinician-led outcomes. These findings imply that an AI-driven, holistic approach to RRT can reduce inappropriate practice variation and lower ICU mortality by recommending personalized treatment adjustments.

问 AI

Bookmark

View Full Paper