The decision to start dialysis is largely determined by patient-related factors not recorded in USRDS data, according to findings published in the May 2017 issue of the journal Medicine.

Researchers from Arbor Research Collaborative for Health and the University of Michigan analyzed data from the United States renal Data System (USRDS) for nearly one million patients who started dialysis between 1995 and 2012.

The investigators wrote that the factors leading to dialysis initiation may include estimated glomerular filtration rate (eGFR) trajectory over time, worsening nutrition or frailty, hyperkalemia, metabolic acidosis, patient/provider preference, or perceived benefits and burden of dialysis, among other factors,

In their analysis, the investigators calculated the proportion of variation in eGFR at dialysis initiation that is explained by variables measured in the USRDS. Measured variables include patient-level factors (such as age and dialysis modality), facility-level factors (such as total number of patients and nurse-to-patient ratio), and county-level factors (such as unemployment rate and high school graduation rate).

Measured patient-level variables explained 10.7% of the total variation in eGFR at dialysis start, and measured facility- and county-level factors explained 1.2% of the variation. However, 78.9% of patient-level variation and 9.2% of variation between facilities and counties remained unexplained.

“It is striking that so much of the variation in eGFR at dialysis initiation is not explained by variables in the USRDS,” said lead author Yun Li, PhD. “Our study showed that we need to collect different types of data to truly understand why patients start dialysis when they do.”

The fact that most of the variation in eGFR occurred at the patient level indicates that the mean eGFR at dialysis start is relatively similar across different physicians, but the eGFR of patients treated by the same physician tend to vary substantially, according to the researchers.

“These findings suggest the need to collect data prospectively from advanced chronic kidney disease patients approaching the need for dialysis, rather than rely on data from patients starting dialysis therapy as with the USRDS dataset,” the authors wrote.