Early cost-effectiveness analysis of the U-Care Renal Platform

Abstract Background Acute kidney injury (AKI) is a common and serious complication in intensive care unit (ICU) patients, with reported incidence ranging from 16% in general hospitalized cohorts to over 20% in some ICU populations, and it increases mortality, prolongs hospitalization, and worsens long-term renal outcomes. The U-Care Renal Platform (UCRP) is a CE-marked AI-powered software as a medical device that, through its Prevention Module via the AKIRA predictive model, predicts moderate-to-severe AKI 24 hours before onset supporting the early application of KDIGO-guided preventive care bundles. We conducted an early cost-effectiveness analysis of UCRP versus standard of care (SoC) and estimated its economic headroom in Italian intensive care. Methods A decision tree (acute phase, 0-90 days) combined with a Markov cohort model (chronic phase, years 1-20) was built from the Italian healthcare payer perspective. The modelled cohort included 2,000 annual ICU admissions with a 40% AKI incidence. UCRP diagnostic accuracy was set at 90% sensitivity and 70% specificity based on internal manufacturer data, with the published Italian external validation (82.5%/83.2%) examined as a scenario. In true-positive patients, early KDIGO-guided treatment was modelled as an approximately 22% relative shift from higher to lower AKI severity stages. Costs were Italian-specific; long-term transition probabilities came from UK NICE HTA data. Costs and QALYs were discounted at 3% annually over a 20-year horizon. Deterministic (±20%), scenario, threshold, and probabilistic (10,000-iteration Monte Carlo) sensitivity analyses were performed. Results In the primary analysis, UCRP was projected to be dominant: an incremental cost saving of €363.42 per patient (total discounted cost €38,059.95 with SoC versus €37,696.53 with UCRP) and an incremental gain of 0.049708 discounted QALYs (9.890 versus 9.940), giving a net monetary benefit (NMB) of €1,854.66 per patient at €30,000/QALY. Ninety-day mortality was projected at 13.72% (SoC) versus 13.29% (UCRP), corresponding to 8.6 additional survivors per 2,000 patients annually. In the probabilistic analysis, mean saving was €175.47 per patient (95% simulation interval -€1,500.72 to €2,016.66) and mean QALY gain was 0.048830 (95% simulation interval: -0.138081 to 0.172619); UCRP was cost-effective at €30,000/QALY in 78.3% of simulations. UCRP remained cost-saving and QALY-gaining across incidence, effect-size, and diagnostic-accuracy scenarios, with savings ranging from €135.93 to €601.21 per patient; the model projected cost saving above an AKI incidence of 4.84% and positive NMB above 1.05%. Results were most sensitive to ICU daily cost, AKI incidence, the 22% severity-shift assumption, and UCRP sensitivity. Conclusions In the base case, the model projected UCRP to be dominant relative to SoC in Italian ICUs, with lower costs and higher QALYs driven by AI-based early AKI detection and guideline-directed treatment. These are conditional model estimates, not observed effects.