Development and validation of AI-based triage support algorithms for prevention of intradialytic hypotension.

BACKGROUND: Intradialytic hypotension remains one of the most recurrent complications of dialysis sessions. Inadequate management can lead to adverse outcomes, highlighting the need to develop personalized approaches for the prevention of intradialytic hypotension. Here, we sought to develop and validate two AI-based risk models predicting the occurrence of symptomatic intradialytic hypotension at different time points. METHODS: The models were built using the XGBoost algorithm and they predict the occurrence of intradialytic hypotension in the next dialysis session and in the next month. The initial dataset, obtained from routinely collected data in the EuCliD® Database, was split to perform model derivation, training and validation. Model performance was evaluated by concordance statistic and calibration charts; the importance of features was assessed with the Shapley Additive Explanation (SHAP) methodology. RESULTS: The final dataset included 1,249,813 dialysis sessions, and the incidence rate of intradialytic hypotension was 10.07% (95% CI 10.02-10.13). Our models retained good discrimination (AUC around 0.8) and a suitable calibration yielding to the selection of three classification thresholds identifying four distinct risk groups. Variables providing the most significant impact on risk estimates were blood pressure dynamics and other metrics mirroring hemodynamic instability over time. CONCLUSIONS: Recurrent symptomatic intradialytic hypotension could be reliably and accurately predicted using routinely collected data during dialysis treatment and standard clinical care. Clinical application of these prediction models would allow for personalized risk-based interventions for preventing and managing intradialytic hypotension.

Prolonged patient survival after implementation of a continuous quality improvement programme empowered by digital transformation in a large dialysis network.

BACKGROUND: Treatment of end-stage kidney disease patients is extremely challenging given the interconnected functional derangements and comorbidities characterizing the disease. Continuous quality improvement (CQI) in healthcare is a structured clinical governance process helping physicians adhere to best clinical practices. The digitization of patient medical records and data warehousing technologies has standardized and enhanced the efficiency of the CQI's evidence generation process. There is limited evidence that ameliorating intermediate outcomes would translate into better patient-centred outcomes. We sought to evaluate the relationship between Fresenius Medical Care medical patient review CQI (MPR-CQI) implementation and patients' survival in a large historical cohort study. METHODS: We included all incident adult patients with 6-months survival on chronic dialysis registered in the Europe, Middle East and Africa region between 2011 and 2018. We compared medical key performance indicator (KPI) target achievements and 2-year mortality for patients enrolled prior to and after MPR-CQI policy onset (Cohorts A and B). We adopted a structural equation model where MPR-CQI policy was the exogenous explanatory variable, KPI target achievements was the mediator variable and survival was the outcome of interest. RESULTS: About 4270 patients (Cohort A: 2397; Cohort B: 1873) met the inclusion criteria. We observed an increase in KPI target achievements after MPR-CQI policy implementation. Mediation analysis demonstrated a significant reduction in mortality due to an indirect effect of MPR-CQI implementation through improvement in KPI target achievement occurring in the post-implementation era [odds ratio 0.70 (95% confidence interval 0.65-0.76); P < 0.0001]. CONCLUSIONS: Our study suggests that MPR-CQI achieved by standardized clinical practice and periodic structured MPR may improve patients' survival through improvement in medical KPIs.

Validation of a Novel Predictive Algorithm for Kidney Failure in Patients Suffering from Chronic Kidney Disease: The Prognostic Reasoning System for Chronic Kidney Disease (PROGRES-CKD).

Current equation-based risk stratification algorithms for kidney failure (KF) may have limited applicability in real world settings, where missing information may impede their computation for a large share of patients, hampering one from taking full advantage of the wealth of information collected in electronic health records. To overcome such limitations, we trained and validated the Prognostic Reasoning System for Chronic Kidney Disease (PROGRES-CKD), a novel algorithm predicting end-stage kidney disease (ESKD). PROGRES-CKD is a naïve Bayes classifier predicting ESKD onset within 6 and 24 months in adult, stage 3-to-5 CKD patients. PROGRES-CKD trained on 17,775 CKD patients treated in the Fresenius Medical Care (FMC) NephroCare network. The algorithm was validated in a second independent FMC cohort (n = 6760) and in the German Chronic Kidney Disease (GCKD) study cohort (n = 4058). We contrasted PROGRES-CKD accuracy against the performance of the Kidney Failure Risk Equation (KFRE). Discrimination accuracy in the validation cohorts was excellent for both short-term (stage 4-5 CKD, FMC: AUC = 0.90, 95%CI 0.88-0.91; GCKD: AUC = 0.91, 95% CI 0.86-0.97) and long-term (stage 3-5 CKD, FMC: AUC = 0.85, 95%CI 0.83-0.88; GCKD: AUC = 0.85, 95%CI 0.83-0.88) forecasting horizons. The performance of PROGRES-CKD was non-inferior to KFRE for the 24-month horizon and proved more accurate for the 6-month horizon forecast in both validation cohorts. In the real world setting captured in the FMC validation cohort, PROGRES-CKD was computable for all patients, whereas KFRE could be computed for complete cases only (i.e., 30% and 16% of the cohort in 6- and 24-month horizons). PROGRES-CKD accurately predicts KF onset among CKD patients. Contrary to equation-based scores, PROGRES-CKD extends to patients with incomplete data and allows explicit assessment of prediction robustness in case of missing values. PROGRES-CKD may efficiently assist physicians' prognostic reasoning in real-life applications.

The anaemia control model: Does it help nephrologists in therapeutic decision-making in the management of anaemia? / El modelo de control de anemia: ¿ayuda al nefrólogo en la decisión terapéutica para el manejo de la anemia?

INTRODUCTION: Anaemia is common in haemodialysis patients and treating it with erythropoiesis-stimulating agents (ESAs) is complex due to many factors. OBJECTIVES: To assess the usefulness of the Anaemia Control Model (ACM) in the treatment of anaemia in haemodialysis. METHODS: ACM is a software that predicts the optimal dose of darbepoetin and iron sucrose to achieve target haemoglobin (Hb) and ferritin levels, and makes prescription suggestions. Study conducted in dialysis clinics lasting 18months with two intervention phases (IPs) with ACM (IP1, n:213; IP2, n:218) separated by a control phase (CP, n:219). The primary outcome was the percentage of Hb in range and the median dose of ESAs, and the secondary outcomes were transfusion, hospitalisation and cardiovascular events. Clinical and patient analyses were performed. Hb variability was assessed by the standard deviation (SD) of the Hb. We also analysed the patients with most of the suggestions confirmed (ACM compliant group). RESULTS: ACM increased the percentage of Hb in range: 80.9% in IP2, compared with 72.7% in the CP and reduced the intake of darbepoetin (IP1: 20 [70]; CP 30 [80] µg P=0.032) with less Hb fluctuation (0.91±0.49 in the CP to 0.82±0.37g/dl in IP2, P<0.05), improving in the ACM compliant group. The secondary outcomes decreased with the use of ACM. CONCLUSIONS: ACM helps to obtain better anaemia results in haemodialysis patients, minimising the risks of treatment with ESAs and reducing costs.

An international observational study suggests that artificial intelligence for clinical decision support optimizes anemia management in hemodialysis patients.

Managing anemia in hemodialysis patients can be challenging because of competing therapeutic targets and individual variability. Because therapy recommendations provided by a decision support system can benefit both patients and doctors, we evaluated the impact of an artificial intelligence decision support system, the Anemia Control Model (ACM), on anemia outcomes. Based on patient profiles, the ACM was built to recommend suitable erythropoietic-stimulating agent doses. Our retrospective study consisted of a 12-month control phase (standard anemia care), followed by a 12-month observation phase (ACM-guided care) encompassing 752 patients undergoing hemodialysis therapy in 3 NephroCare clinics located in separate countries. The percentage of hemoglobin values on target, the median darbepoetin dose, and individual hemoglobin fluctuation (estimated from the intrapatient hemoglobin standard deviation) were deemed primary outcomes. In the observation phase, median darbepoetin consumption significantly decreased from 0.63 to 0.46 µg/kg/month, whereas on-target hemoglobin values significantly increased from 70.6% to 76.6%, reaching 83.2% when the ACM suggestions were implemented. Moreover, ACM introduction led to a significant decrease in hemoglobin fluctuation (intrapatient standard deviation decreased from 0.95 g/dl to 0.83 g/dl). Thus, ACM support helped improve anemia outcomes of hemodialysis patients, minimizing erythropoietic-stimulating agent use with the potential to reduce the cost of treatment.

Relative blood volume monitoring during hemodialysis in end stage renal disease patients.

A crucial point in the haemodialysis (HD) treatment is the reliable assessment of hydration status. An inadequate removed volume may lead to chronic fluid overload which can lead to hypertension, left ventricular hypertrophy and heart failure. Therefore, the estimation of the hydration state and the management of a well-tolerated water removal is an important challenge. This exploratory study aims at identifying new parameters obtained from continuous Blood Volume Monitoring (BVM) allowing a qualitative evaluation of hydration status for verifying the adequacy of HD setting parameters (e.g UFR, target dry weight). The percentage of blood volume reduction (BVR%) during HD was compared against a gold standard method for hydration status assessment. The slope of the first 30 minute of blood volume reduction (BVR) was proposed as a useful parameter to identify overhydrated patients.