Back

Diagnostic Accuracy of a Smartphone-Enabled Urinary Albumin-to-Creatinine Ratio Test for Detecting Chronic Kidney Disease Among At-Risk Populations

Jeddah, D.; Bevins, N. B.; Ronen, M.; Tangri, N.

2025-01-21 health informatics
10.1101/2025.01.20.25320142 medRxiv
Show abstract

ObjectiveTo evaluate the diagnostic accuracy of the FDA-cleared Minuteful Kidney Test (MKT), a home-based semi-quantitative urine albumin-to-creatinine ratio (uACR) test, compared to the Beckman Coulter AU 480 analyzer, a standard laboratory quantitative method, for detecting albuminuria in individuals with or at risk for chronic kidney disease (CKD). MethodsUrine samples were obtained from individuals with risk factors for kidney damage. Each sample was analyzed for uACR using both the MKT and the laboratory standard method. Albuminuria was classified according to KDIGO guidelines: A1 (uACR [&le;]30 mg/g, "Normal to mildly increased"), A2 (uACR 30-300 mg/g, "Moderately increased"), and A3 (uACR [&ge;]300 mg/g, "Severely increased"). Sensitivity, specificity, PPV, and NPV of MKT were calculated against the laboratory standard. ResultsOut of 615 collected urine samples, 555 met inclusion criteria. Albuminuria (uACR [&ge;]30 mg/g) was detected in 24.9% of samples using the laboratory method (95% CI, 21.3-28.5%). The MKT demonstrated a sensitivity of 96.4%, specificity of 84.2%, NPV of 98.6%, and PPV of 66.8%. The test accurately categorized 85.6% of samples into the KDIGO albuminuria categories A1, A2, and A3, accurately identifying 100% of laboratory-confirmed A3 samples as abnormal. ConclusionsThe Minuteful Kidney Test shows high sensitivity, exceeding the guideline-required threshold (>85%, P<0.001), and robust specificity for detecting albuminuria in high-risk populations. This device offers potential to improve CKD screening and management and facilitate early intervention in primary care and community settings, where accurate rule-out capabilities are essential.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.