Chronic kidney disease often hides in plain sight, surfacing only after the damage is well advanced, and primary care clinicians are usually the first to glimpse it. At the University of California, San Francisco, a team has built a tool that lives inside the electronic health record and does the patient-by-patient sorting that busy clinicians rarely have time for: finding the disease, weighing how dangerous it is, and suggesting a path forward. The work, led by Ivan Villasenor and presented at ASN Kidney Week 2026, shows that such automation can match hand calculations and fit into everyday p
Treat CKD Tool Integrates EHR Data to Flag Kidney Disease and Guide Primary Care Treatment
Two risks that travel together in kidney disease are scored side by side.
Mark
What does Treat CKD actually do that a clinician couldn't do alone?
Mimi
It does the sorting automatically. It reads the chart, identifies CKD, classifies risk by KDIGO criteria, and calculates kidney failure and cardiovascular risk, then suggests next steps.
Mark
Which next steps?
Mimi
Albuminuria testing, blood pressure management, cholesterol-lowering therapy, cardio-kidney therapies, and nephrology referral.
Luke
Before we get excited, what was actually checked?
Mimi
Two things. The risk classification and the scores matched manual calculations, fully concordant. And clinicians at UCSF called it easy to use and well integrated.
Luke
But that's accuracy of calculation, not patient outcomes. The source doesn't say anyone's health improved.
Mimi
Right. It shows the tool computes correctly and is acceptable, not that it changes outcomes.
Mark
And the clinician feedback?
Luke
Reported as descriptions from primary care professionals in one health system. The source gives no numbers on how many, so I'd hold that loosely.
Mimi
The authors themselves only claim it supports further evaluation as a scalable strategy.
Mark
So what to watch?
Luke
Whether it's tested beyond UCSF, and whether results get published in full.
Der Puls
- Kidney disease is frequently missed in primary care, and the cost of that silence is measured in years of progression that could have been slowed.
- A new decision support tool, Treat CKD, now sorts through chart data on its own, flagging disease and scoring kidney failure and heart risk side by side.
- Its risk scores match manual calculations exactly, and UCSF clinicians found it easy to use without it adding burden to their already crowded workflows.
- The tool is being shown to the nephrology world at ASN Kidney Week, but it has only been tested within UCSF's own system and awaits broader trials before it can be called a scalable answer.
Chronic kidney disease often hides in plain sight, surfacing only after the damage is well advanced, and primary care clinicians are usually the first to glimpse it. At the University of California, San Francisco, a team has built a tool that lives inside the electronic health record and does the patient-by-patient sorting that busy clinicians rarely have time for: finding the disease, weighing how dangerous it is, and suggesting a path forward. The work, led by Ivan Villasenor and presented at ASN Kidney Week 2026, shows that such automation can match hand calculations and fit into everyday practice, though whether it can carry that promise beyond one health system remains a question for the years ahead.
A primary care clinician at the University of California, San Francisco, can open a patient's chart and find that the sorting has already been done. The tool waiting there is Treat CKD, a clinical decision support program built into the electronic health record. Its task is to find chronic kidney disease, judge how serious it is, and suggest what comes next.
The team behind it, led by corresponding author Ivan Villasenor, an MD candidate at the UCSF School of Medicine, will present the study at ASN Kidney Week 2026, held October 22 to 25. The tool works from data already in the record. It applies the KDIGO criteria to classify risk, estimates kidney failure with the Kidney Failure Risk Equation, and estimates cardiovascular risk with the American Heart Association's PREVENT equations. Because the two risks often travel together in kidney disease, they are scored side by side.
From these scores, the tool offers guidance on albuminuria testing, blood pressure, cholesterol therapy, cardio-kidney treatments, and when to refer a patient to a kidney specialist. It is aimed at primary care, where most people with early kidney disease are first seen.
The investigators tested two things: whether the numbers held up, and whether clinicians would welcome the tool. The automated risk classifications and scores matched hand calculations completely. Clinicians who used it described it as easy, clinically useful, and light enough to fit their workflows, which is often where such tools fail.
The authors judge the tool acceptable for routine primary care and see grounds for wider evaluation as a scalable way to detect and treat kidney disease. Yet the evidence so far comes from one health system and has been presented, not yet tested in larger trials. That broader testing is the work still ahead.
A primary care clinician at the University of California, San Francisco, can now open a patient's chart and find a tool that has already done the sorting. It is called Treat CKD, a clinical decision support program built into the electronic health record, and its job is to find chronic kidney disease, gauge how dangerous it is, and tell the clinician what to do next.
The investigators who built and validated it will present their work at ASN Kidney Week 2026, running October 22 to 25. The corresponding author is Ivan Villasenor, an MD candidate at the UCSF School of Medicine. The study is titled "Treat CKD: An Electronic Health Record (EHR)-Integrated Clinical Decision Support Tool for Risk-Based CKD Care in Primary Care."
The tool works from data already sitting in the record. It identifies CKD and classifies risk using the criteria from Kidney Disease: Improving Global Outcomes, known as KDIGO. It then calculates the risk of kidney failure with the Kidney Failure Risk Equation, and it estimates cardiovascular risk with the American Heart Association's PREVENT equations. Two risks that travel together in kidney disease are therefore scored side by side.
From those scores, Treat CKD generates recommendations across the main points of care. These cover testing for albuminuria, managing blood pressure, cholesterol-lowering therapy, and what the source calls cardio-kidney therapies. It also advises on when to refer a patient to nephrology. The recommendations are aimed at primary care, where most people with early kidney disease are first seen.
The team tested two things: whether the numbers were right, and whether clinicians would accept the tool. On accuracy, Treat CKD correctly confirmed the KDIGO-based risk classification. Its Kidney Failure Risk Equation and PREVENT scores were fully concordant with those calculated by hand, so the automation reproduced the manual work without discrepancy.
On acceptance, primary care health care professionals used it within UCSF's system and described it as easy to use and clinically relevant. They also said it fit into their workflows without adding substantial burden, a common point of failure for decision support tools.
The investigators conclude that Treat CKD accurately automates identification and risk classification, estimates the risk of cardiovascular events and progression to kidney failure, and offers guideline-concordant treatment advice. Clinicians found it acceptable and appropriate for routine primary care use. The authors say this supports further evaluation of the tool as a scalable approach to detecting and treating CKD.
That last phrase marks the limit of what has been shown. The findings come from UCSF's system and are being presented, not yet tested in wider trials. The next step is that broader evaluation.
Bemerkenswerte Zitate
Treat CKD accurately automates CKD identification and risk classification and gives guideline-concordant recommendations, supporting further evaluation as a scalable detection and treatment strategy.— Ivan Villasenor, corresponding author, UCSF School of Medicine (paraphrased)
Möchten Sie die ganze Geschichte?
Das Original lesen bei News-Medical ↗
Health & Medicine