Precision Hydration: Causal Machine Learning Pinpoints Sepsis-AKI Patients Who Thrive on Less IV Fluid
medrxiv.orgThe results of this study demonstrate that a causal machine learning (ML) framework can successfully identify patients with sepsis-AKI who benefit from a restrictive fluid strategy. This technology offers a promising, data-driven approach for personalized fluid management that could improve kidney outcomes.
The authors conclude that these findings warrant further investigation through prospective clinical trials.
The causal forest model significantly outperformed the standard random forest model in identifying the heterogeneous treatment effects (HTE) of restrictive IV fluids (AUTOC 0.15 vs. -0.02 in external validation).
When the model recommended restrictive fluids for a subgroup of patients (68.9% of the external validation cohort), those who actually received the restrictive fluids had:
- Significantly higher rate of early AKI reversal (53.9% vs 33.2%)
- Higher rate of sustained AKI reversal (34.2% vs 18.0%)
- Lower rates of MAKE30 (17.1% vs 34.6%)
