Abstract: FR-PO0018
Dynamic Effects of Online High-Volume Hemodiafiltration on Health Outcomes: Longitudinal Study Using Difference-in-Differences and Callaway and Sant'Anna Estimators
Session Information
- Artificial Intelligence and Digital Health at the Bedside
November 07, 2025 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Artificial Intelligence, Digital Health, and Data Science
- 300 Artificial Intelligence, Digital Health, and Data Science
Author
- Silveira Junior, Sergio Antonio Dias Da, Prevent Senior Private Operadora de Saude Ltda, São Paulo, SP, Brazil
Background
Evaluating the causal impact of medical technologies in real-world settings presents methodological challenges, particularly when treatment effects vary over time. We applied a longitudinal difference-in-differences (DiD) framework using Callaway & Sant’Anna estimators to assess the dynamic effects of online high-volume hemodiafiltration (HDF) on key health outcomes among elderly ESRD patients.
Methods
This retrospective observational study included 2,167 elderly ESRD patients (mean age: 72.7 years) treated within a Brazilian private health plan from 2018 to 2023. Patients were categorized into three groups: (1) HD-only, (2) exclusive HDF, and (3) HD→HDF converters. Using DiD models with Callaway & Sant’Anna estimators, we estimated average treatment effects (ATE) over 1, 3, 6, 9, and 12 months for emergency visits, outpatient consultations, and high-complexity procedures. Cox regression models assessed all-cause mortality.
Results
HDF treatment showed increasing benefits over time. ATE estimates at 12 months revealed significant reductions in emergency visits (–0.40, p < 0.001), outpatient consultations (–0.25, p < 0.001), and high-complexity procedures (–0.31, p < 0.01). No significant effects were observed for hospitalizations. Cox models confirmed reduced mortality risk in the HDF-only (HR: 0.574; 95% CI:0.48–0.68; p < 0.001) and HD→HDF groups (HR: 0.607; p < 0.001).
Conclusion
This is the first study in nephrology to apply Callaway & Sant’Anna estimators to assess the dynamic impact of HDF. The findings demonstrate sustained clinical benefits over time, supporting both the broader adoption of HDF and the incorporation of advanced causal inference techniques in kidney health outcomes research.