Abstract: FR-PO419
A Novel Individualized Patient Performance Scoring System to Predict Hard Outcomes in Patients on Hemodialysis
Session Information
- Hemodialysis Epidemiology and Outcomes
October 25, 2024 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Dialysis
- 801 Dialysis: Hemodialysis and Frequent Dialysis
Authors
- Santos, Carla, Diaverum AB, Malmo, Sweden
- Lucas, Carlos, Diaverum AB, Malmo, Skåne, Sweden
- Silva, Eliana, Diaverum AB, Malmo, Skåne, Sweden
- Haarhaus, Mathias, Diaverum AB, Malmo, Skåne, Sweden
- Macario, Fernando Jose Gordinho Rocha M, Diaverum AB, Malmo, Skåne, Sweden
Background
Patients in hemodialysis (HD) face high mortality and hospitalization. This study introduces an individual patient performance score (IPPS) as a novel tool for predicting outcomes in HD.
Methods
Multicenter, retrospective study in 19 countries. Between January and June 2023, patients were categorized into 2 groups based on: 1) survival status; 2) presence of a hospitalization event. IPPS were recorded from 24 months until of the index date. Prediction models were developed incorporating IPPS values, demographic data and quality of life SF-36 survey results of 2021 and 2022. Model discrimination was evaluated by the area under the curve (AUC) and 95% confidence intervals (CI). IPPS were developed as a combined score of 8 major key areas: Vascular Access (20 points): type of VA, episodes of thrombosis and infection; HD adequacy (20): eKt/V, blood flow and treatment time; Anemia (20): hemoglobin, transferrin saturation, ferritin; Arterial hypertension (10); Mineral Bone disease, MBD (15): iPTH, phosphorus and calcium; Fluid status (5): percentage of fluid gain; Nutrition (5): albumin and phosphorus and Others (5): influenzae vaccination, transplantation status and previous hospitalization. The final score is obtained by the weighted sum of positive and negative factors, ranging from 0 to 100 and with a higher score representing a better medical performance.
Results
41,125 patients were included, with significant differences in mean IPPS observed between the surviving and deceased groups up to 24 months prior to death and the hospitalized and non-hospitalized groups up to 12 months of the event. This difference was observed in all IPPS areas, with the exception of MBD for death. Multiple logistic regression showed a 4% independent reduction in the odds of death and a 3% independent reduction in the odds of hospitalization with the increase of 1 point in global IPPS (p<0.01). Predictive model demonstrated a very good discrimination for mortality (AUC of 0.76 [0.75; 0.78]) and a good discrimination for hospitalization (AUC of 0.70 [0.69; 0.71]).
Conclusion
IPPS showed a strong correlation with patient survival and hospitalization. The incorporation of IPPS into HD patient care evaluation emerges as an effective strategy for improving critical outcomes, such as mortality and hospitalization.