Study Suggests Artificial Intelligence Improves Donor Kidney Allocation Process

New York, NY - May 9, 2019 – A new study that analyzed 10 years’ worth of data and hundreds of variables about deceased donor kidneys and the recipients who received them has found that with artificial intelligence (AI)—specifically, training machines to learn from past experience—predictive outcomes are more accurate and will continue to improve over time. Kidneys can then be allocated more efficiently, and with less waste than ever before, according to the author.

 

The study titled “Predicting Deceased Donor Kidney Transplant Outcomes: Comparing KDRI/KDPI with Machine Learning” was presented at the National Kidney Foundation’s 2019 Spring Clinical Meetings in Boston this week.

 

About 21,000 kidney transplants are performed each year and the majority of the kidneys donated are from deceased donors. Thousands retrieved each year are never transplanted for a variety of reasons. Kidneys retrieved from deceased donors are currently allocated using the Kidney Donor Risk Index (KDRI) and administered by the United Network for Organ Sharing (UNOS).

 

“There is insurmountable evidence that we are not taking full advantage of these discarded kidneys,” said lead author and PhD candidate at The University of Iowa Eric Pahl, of OmniLife, Inc. “We are throwing away a lot of kidneys. We need to be better stewards of the gift we have in deceased donor organs.”

 

Pahl said the study suggests that hundreds of kidneys retrieved from deceased donors and then never transplanted may be saved using AI through machine learning. Further studies are planned, including a clinical trial that is scheduled to begin in August using the machine learning method, Pahl said.

 

It is one of NKF’s goals to reduce the number of discarded kidneys after retrieval from deceased donors and has supported initiatives that move the bar closer to eliminating the kidney transplant waitlist.

 

Pahl said that the study looked at hundreds of variables, rather than the 10 variables the current system examines when allocating a kidney. They also looked at characteristics of the recipient as well as the donor kidney and found this added significant predictive value.

 

“It just makes common sense that we should incorporate variables for the recipient,” Pahl said. “It would seem important for better predictable outcomes.”

 

Pahl is a Health Informatics Ph.D. student at the University of Iowa and participates in the Interdisciplinary Graduate Program in Informatics (IGPI). With the help of his IGPI mentors, including primary adviser Hans Johnson, Associate Professor of Electrical and Computer Engineering; Informaticist Nick Street, Tippie College of Business; Management Sciences; and Transplant Surgeon Alan Reed, UI Hospitals and Clinics, Pahl focused on the task of using AI and machine learning to compensate for missing data.

 

In Boston, thousands of professionals in nephrology gathered to share the latest information, treatment and studies of CKD. Teams presented hundreds of studies during the “poster presentations.”

 

For the past 27 years, nephrology healthcare professionals from across the country have come to NKF’s Spring Clinical Meetings to learn about the newest developments related to all aspects of nephrology practice; network with colleagues; and present their research findings. The NKF Spring Clinical Meetings are designed for meaningful change in the multidisciplinary healthcare teams’ skills, performance, and patient health outcomes. It is the only conference of its kind that focuses on translating science into practice for the entire healthcare team.  This year’s Spring Clinical Meetings will be held May 8-12 in Boston, MA.

 

Healthcare professionals can join NKF to receive access to tools and resources for both patients and professionals, discounts on professional education, and access to a network of thousands of individuals who treat patients with kidney disease.

 

In the United States, 30 million adults are estimated to have chronic kidney disease—and most aren’t aware of it.  1 in 3 American adults are at risk for chronic kidney disease.  Risk factors for kidney disease include diabetes, high blood pressure, heart disease, obesity and family history.  People of African American, Hispanic, Native American, Asian or Pacific Islander descent are at increased risk for developing the disease.  African Americans are 3 times more likely than Whites, and Hispanics are nearly 1.5 times more likely than non-Hispanics to develop end stage renal disease (kidney failure).