Using Predictive Analytics to Move Kidney Care Upstream
Clinical team reviews information on a tablet in a clinical office

Of the 37 million Americans who have kidney disease, an estimated 90% don’t know they have it. Often, it can go undetected because patients can present few or no symptoms until they approach or transition into kidney failure necessitating dialysis or transplantation to sustain life, often in an urgent, unplanned manner called “crashes.”

Previously, DaVita Clinical Research (DCR) developed and validated models that can help care teams identify individuals with undiagnosed chronic kidney disease (CKD) with 72% accuracy[1] and identify 75% of patients at risk of experiencing kidney failure in 6–18 months.[2]

“Predictive models are not a replacement for clinical decision-making,” Dr. Steven Brunelli, MCSE, vice president for DCR, says. “They can, however, provide crucial insights that bring patients to medical attention in order to initiate important conversations and help inform treatment choices. When physicians have data indicating an increased risk of negative outcomes, they can intervene with the goal of prevention and to help patients achieve better health.”

DCR recently presented a new validated machine learning algorithm to expand predictive analytics for CKD care. This new model, first presented at the American Society for Nephrology (ASN) Kidney Week event in 2021, enables clinicians to identify patients with CKD (stages 3–5) at risk of near-term hospitalization.[3]

Patients with CKD are at higher risk of hospitalization—often associated with higher medical costs, increased morbidity and increased risk of kidney failure. This new hospitalization-risk model can help care teams deliver the right care and interventions at the right time—for the right patients.

The goal? By intervening at the right time, care teams aim to help delay—or even halt, when possible—disease progression. When kidney failure cannot be prevented, predictive analytics can help inform conversations between the patient and their nephrologist about transplant and dialysis modalities sooner, with the goal of achieving smoother, planned transitions to treatment when needed.

Advancing Kidney Care with the Right Tools

In July 2019, the federal government announced an initiative to transform kidney care through better diagnosis, treatment and preventive care. This Executive Order also outlined goals to refocus primary treatment for end stage kidney disease (ESKD) on transplantation and home dialysis modalities.

As DaVita, along with others in the kidney care community, aligned with these goals, DCR helped in this transformation through the development of proprietary predictive modeling.

More than 30 DaVita data scientists take part in developing and validating operational predictive models. DCR is uniquely positioned to build out these new capabilities because of DaVita’s more than 20 years of expertise in kidney care—data scientists are able to leverage more than 1 billion CKD and ESKD data points, relationships with health technology leaders and experience with deploying capabilities and new models into care management workflows.

DCR is a wholly-owned subsidiary of DaVita Inc., and conducts focused, retrospective studies that help innovate to drive improved clinical outcomes in kidney care.

To learn more about DCR, please visit DaVitaClinicalResearch.com.

 

[1] 2020 DaVita Internal Analysis; using claims data

[2] 2019 DaVita Integrated Kidney Care (IKC) Internal Analysis

[3] 2021 DaVita Internal Analysis