pyAKI: An Open Source Tool for Acute Kidney Injury Classification
Author Information
Author(s): Porschen Christian, Ernsting Jan, Brauckmann Paul, Weiss Raphael, Würdemann Till, Booke Hendrik, Amini Wida, Maidowski Ludwig, Risse Benjamin, Hahn Tim, von Groote Thilo
Primary Institution: University Hospital Münster, Germany
Hypothesis
The pyAKI pipeline provides a standardized solution for implementing KDIGO criteria in time series data for acute kidney injury classification.
Conclusion
The pyAKI pipeline demonstrates high accuracy in classifying acute kidney injury using standardized KDIGO criteria.
Supporting Evidence
- pyAKI achieved an accuracy of 1.0 in all categories compared to physician annotations.
- The pipeline provides a comprehensive solution for consistent AKI classification.
- Validation against expert annotations demonstrated pyAKI’s robust performance.
- pyAKI is the first open-source solution for implementing KDIGO criteria in time series data.
Takeaway
The pyAKI tool helps doctors quickly and accurately identify kidney problems in patients using computer data, making it easier to help them.
Methodology
The pyAKI pipeline was developed and validated using the MIMIC-IV database, implementing KDIGO guidelines for AKI diagnosis based on serum creatinine and urine output data.
Potential Biases
Potential biases may arise from the reliance on the quality of input data and the limitations of the algorithms used.
Limitations
The study's sample size was small, and the accuracy of the Cockcroft-Gault formula for estimating GFR may vary in critically ill patients.
Participant Demographics
Patients included in the study were critically ill individuals from the MIMIC-IV database, with a mean ICU stay of 176.73 hours.
Statistical Information
P-Value
1.0
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website