Incident 79: Kidney Testing Method Allegedly Underestimated Risk of Black Patients

Description: Decades-long use of the estimated glomerular filtration rate (eGFR) method to test kidney function which considers race has been criticized by physicians and medical students for its racist history and inaccuracy against Black patients.

Tools

New ReportNew ReportNew ResponseNew ResponseDiscoverDiscoverView HistoryView History
Alleged: Chronic Kidney Disease Epidemiology Collaboration developed and deployed an AI system, which harmed Black patients and African-American patients.

Incident Stats

Incident ID
79
Report Count
3
Incident Date
1999-03-16
Editors
Sean McGregor, Khoa Lam

CSETv0 Taxonomy Classifications

Taxonomy Details

Full Description

A 2020 study conducted in the Mass General Brigham health system demonstrated that a popular algorithm for estimating kidney function underestimated the risk to African-American patients. This bias could lead to inequitable outcomes, such as not being placed on a kidney transplant waiting list. The equation, known as the Chronic Kidney Disease Epidemiology Collaboration estimated Glomerular Filtration Rate (CKD-EPI eGFR) equation, includes a race multiplier for African-Americans. When researchers removed the race multiplier, 33.4% of African-American patients in their study were reclassified into more severe risk categories.

Short Description

A 2020 study conducted in the Mass General Brigham health system demonstrated that a popular algorithm for estimating kidney function included a race multiplier, which underestimated the risk to African-American patients.

Severity

Unclear/unknown

Harm Distribution Basis

Race

Harm Type

Harm to physical health/safety

AI System Description

An equation created by the Chronic Kidney Disease Epidemiology Collaboration to calculate Glomerular Filtration Rate (GFR)

System Developer

Chronic Kidney Disease Epidemiology Collaboration

Sector of Deployment

Human health and social work activities

Relevant AI functions

Cognition

AI Techniques

machine learning

AI Applications

statistical projection

Location

Boston, MA

Named Entities

Chronic Kidney Disease Epidemiology Collaboration, Mass General Brigham health system, Mass General Hospital, Brigham and Women's Hospital

Technology Purveyor

Chronic Kidney Disease Epidemiology Collaboration

Beginning Date

6/2019

Ending Date

6/2019

Near Miss

Unclear/unknown

Intent

Unclear

Lives Lost

No

Infrastructure Sectors

Healthcare and public health

Data Inputs

creatinine levels, age, sex, race

A yearslong push to remove racist bias from kidney testing gains new ground
statnews.com · 2020

For years, physicians and medical students, many of them Black, have warned that the most widely used kidney test — the results of which are based on race — is racist and dangerously inaccurate. Their appeals are gaining new traction, with …

Examining the Potential Impact of Race Multiplier Utilization in Estimated Glomerular Filtration Rate Calculation on African-American Care Outcomes
link.springer.com · 2020

BACKGROUND: Advancing health equity entails reducing disparities in care. African-American patients with chronic kidney disease (CKD) have poorer outcomes, including dialysis access placement and transplantation. Estimated glomerular filtra…

How an Algorithm Blocked Kidney Transplants to Black Patients
wired.com · 2020

BLACK PEOPLE IN the US suffer more from chronic diseases and receive inferior health care relative to white people. Racially skewed math can make the problem worse.

Doctors often make life-changing decisions about patient care based on algo…

Variants

A "variant" is an incident that shares the same causative factors, produces similar harms, and involves the same intelligent systems as a known AI incident. Rather than index variants as entirely separate incidents, we list variations of incidents under the first similar incident submitted to the database. Unlike other submission types to the incident database, variants are not required to have reporting in evidence external to the Incident Database. Learn more from the research paper.