News: Machine learning reveals kidney disease through EHR analysis

Machine learning has the ability to analyze electronic health records (EHRs) and identify undetected chronic kidney disease, according to a study published by npj Digital Medicine.

According to the study, patient EHRs can be automatically scanned for blood and urine test results, then the machine learning tool can use a set of established equations and machine learning to process the data and detect chronic kidney disease. The study notes that the algorithm may be able to detect chronic kidney disease in its earliest stages, alerting clinicians to the underlying patient issue.

Early identification of chronic kidney disease is of paramount importance, as it can allow clinicians to use treatments for slowing disease progression before the damage becomes irreversible, study authors say. According to the researchers, roughly one in eight American adults is believed to have chronic kidney disease, yet only one in 10 of those with early stages of the disease are aware of their condition.

The research team noted several possible reasons for underdiagnosis, such as lack of symptoms for patients in the early stages of chronic kidney disease and primary care providers prioritizing immediate patient complaints. Currently, there are two tests needed to detect asymptomatic kidney disease, and the interpretation of these tests is not always straightforward, study authors say.

The new machine learning tool may be able to overcome these obstacles, as it automatically scans EHRs for patient test results, performs calculations that may indicate kidney function and damage, and alerts physicians of the patient’s condition.

Editor’s note: The npj Digital Medicine published study can be found here.