External validation of the Madrid Acute Kidney Injury Prediction Score.

Area of Science:

• Nephrology
• Clinical Prediction Models
• Health Informatics

Background:

• The Madrid Acute Kidney Injury Prediction Score (MAKIPS) is a novel tool for predicting hospital-acquired acute kidney injury (HA-AKI).
• External validation of MAKIPS in diverse clinical settings is crucial for its widespread adoption.
• This study aimed to validate MAKIPS in a tertiary hospital with a distinct patient demographic and case mix.

Purpose of the Study:

• To externally validate the Madrid Acute Kidney Injury Prediction Score (MAKIPS).
• To assess the performance of MAKIPS in predicting hospital-acquired acute kidney injury (HA-AKI) in a new clinical environment.
• To evaluate the generalizability of the MAKIPS tool across different hospital case mixes.

Main Methods:

• An external validation cohort study was conducted.
• Patients admitted to a single tertiary hospital between April 2018 and September 2019 were included.
• Performance was evaluated using discrimination (area under the receiver operating characteristics curve) and calibration plots.

Main Results:

• The incidence of HA-AKI in the validation cohort was 5.3%.
• The MAKIPS demonstrated an area under the curve of 0.798 (95% CI 0.788-0.809), indicating good predictive discrimination.
• Calibration plots revealed a tendency for MAKIPS to overestimate risk at low probabilities (<0.19) and underestimate at moderate probabilities (0.22-0.67).

Conclusions:

• The MAKIPS is a potentially useful tool for predicting HA-AKI risk.
• Its performance suggests it can be applied in hospitals with varying case mix characteristics.
• The score utilizes readily available electronic health record data, facilitating clinical implementation.