Kidney Failure Prediction via Creatinine Trend Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting chronic kidney disease (CKD) progression are inadequate, particularly in early stages, leading to unnecessary specialist referrals and increased healthcare burdens, as they often rely on creatinine measurements that may not accurately predict kidney failure timelines.
Innovation Solution
A computer-implemented method using a machine learning model that incorporates recent creatinine or eGFR levels, initial measurements, and statistical parameters derived from linear regression of past data to predict kidney failure within a specified time frame, employing algorithms like gradient-boosted decision trees or neural networks to provide reliable likelihood assessments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If creatinine measurement is used during routine health status tests to detect CKD, then detection simplicity is improved, but detection reliability deteriorates because CKD is often not formally diagnosed in patients with creatinine results indicating CKD
Solution Approach 1:
The patent transforms the single creatinine measurement approach into a multi-parameter prediction system that incorporates creatinine levels, eGFR values, and temporal trends. By changing from a static parameter (single creatinine value) to dynamic parameters (multiple measurements over time with trend analysis), the system maintains ease of operation while significantly improving detection reliability through more comprehensive data evaluation
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between raw creatinine/eGFR measurements and clinical diagnosis. This intermediary processes the measurements, accounts for temporal trends, and generates a predicted likelihood of kidney failure, thereby bridging the gap between simple routine measurements and reliable diagnostic decisions
2Reliability
If specialist referral is made for all CKD patients, then detection completeness is improved, but healthcare system burden worsens due to unnecessary referrals and higher burden on specialists
Solution Approach 1:
The patent applies local quality by differentiating between patient subgroups based on their predicted likelihood of kidney failure. Instead of uniform specialist referral for all CKD patients, the system identifies fast progressors who require specialist intervention while allowing slow progressors to be managed by general practitioners, thereby optimizing resource allocation and reducing unnecessary specialist burden while maintaining complete detection
3Loss of time
If early CKD detection is performed using traditional methods, then early intervention opportunity is improved, but prediction accuracy deteriorates because traditional methods cannot reliably predict kidney failure timelines
Solution Approach 1:
The patent performs preliminary action by predicting the likelihood of kidney failure within specific timeframes (e.g., 1 year, 5 years) before actual kidney failure occurs. This allows clinicians to identify fast progressors early and implement appropriate interventions or monitoring strategies in advance, while the machine learning model provides the necessary prediction accuracy through analysis of temporal trends in creatinine and eGFR measurements
Data Source
AI summary
A computer-implemented method is provided, which determines at a prediction time tp, a likelihood of kidney failure of a patient within an amount of time Δt. The method comprises receiving input data, the input data comprising a recent creatinine level cR or recent eGFR eGFRR, and one or more of the following: (a) an initial creatinine level c0 and either: a time t0 at which the initial creatinine level c0 was measured, or a time interval ΔT0=tp−t0; (b) an initial estimated glomerular filtration rate (eGFR) eGFR0 and either: a time t0 at which the initial eGFR was determined, or a time interval ΔT0=tp−t0; (c) for a plurality of past creatinine level measurements ci measured at a respective times ti, a statistical parameter derived from a linear regression of the plurality of past creatinine level measurements; and (d) for a plurality of past eGFR values eGFRi determined at respective times ti, a statistical parameter derived from a linear regression of the plurality of past eGFR values; and applying a machine-learning model to the input data to generate an output indicating the likelihood of kidney failure within the given amount of time Δt. Corresponding training methods and systems are also provided.