Stories

The CHR-developed tool is 30% more effective than standard-of-care

Executive Update—Of the 10,000 members of Kaiser Permanente Northwest who have chronic kidney disease, 300 will require dialysis or a transplant within five years.  The key is knowing who they are—and now a new tool can identify those patients most in need.

A team of Center for Health Research scientists and Northwest Permanente physicians has developed a risk assessment tool to predict patients’ risk for renal replacement therapy (RRT) within a 1% margin of error.  The tool is nearly three times more effective in identifying patients who will require dialysis than the current guidelines recommended by the National Kidney Foundation.  What’s more, it’s easy to use: All the variables are already in the patient’s electronic medical record (EMR).

The renal replacement therapy risk score was recently published in the American Journal of Kidney Diseases, coauthored by CHR investigators Eric Johnson and David Smith, Northwest Permanente physician Micah Thorp, and McGill University colleague Robert Platt.  For this retrospective study, the team pulled six variables from patients’ EMR: age, sex, estimated glomerular filtration rate (eGFR, a measure of kidney function), and presence/absence of anemia, diabetes, and hypertension.  Each variable was assigned a point score; added together, the variables translated into a predicted risk, which allows Kaiser Permanente to identify the highest-risk patients in greatest need of attention from a nephrologist.  The tool’s predictions were closely aligned with what actually happened to patients over the course of five years, within a 1% margin of error.

Current guidelines from the National Kidney Foundation advise a referral to a nephrologist based on just one patient characteristic, the eGFR (specifically, when eGFR is <30 mL/min).  The new tool uses that measurement as one of six variables to generate a risk score, significantly improving the accuracy of the predictions. As a result, this new tool is 30% more effective than the standard one-variable prediction tool in identifying which patients will progress to dialysis or transplant.

Always in high demand, nephrologists may prioritize workloads more efficiently by using this accurate risk tool to identify which patients need the most care.