In recent years, thanks to the implementation of machine learning algorithms, risk prediction models have become indispensable for solving many complex clinical problems that arise in CKD and ESRD. One such model that has been established is the Tangri model, which is a traditional regression model for the prediction of the progression of kidney failure in CKD patients. The data used by Tangri includes routinely available clinical, demographic, and laboratory information.
Clinical Decision Support Systems (CDSS) have an important role in the delivery of evidence-based medicine; their algorithms and models are built on real-world patient data. A recent study showed that physician adherence to guidelines regarding the monitoring of CKD could be improved with automated laboratory-based CDSS.
Panoramic Health's ESKD risk predictor can capture more at-risk patients, outperforming the industry standard Tangri model described above.
Current applications of big data in decision making, innovation, and clinical practice:
•The management of acute kidney injury: classification, detection, and automated alerting of acute kidney injury.
•Clinical decision and risk prediction support: using natural language processing and machine learning for risk prediction of CKD and end-stage renal disease (ESRD).
Potential directions of big data in decision making, innovation, and clinical practice:
•To prompt the activeness of medical imaging: the analysis of urine-formed elements and of renal biopsy images.
•The strengthening of big data-based CDSS and artificial intelligence (AI): the improvement of usability and interpretability of CDSS, and improved/increased collaboration with specialist data scientists.