As machine learning proliferates, researchers warn that models based on training datasets don’t necessarily translate real-world inputs accurately.

One lab is looking for a new GFR calculation based on a single creatinine value that does not use race and/or gender and/or age. Suggestions include removing the race-based modifier but keeping gender and age, or using other formulas such as MDRD. Another respondent says he is waiting for the final report from the NKF-ASN Task Force on Reassessing the Inclusion of Race in Diagnosing Kidney Diseases. The task force has been working on a solution to GFR estimation that will be safe, equitable, and will not disadvantage any patient by race, sex, or age.

How do separate clinical laboratory sites within a single health system take the first step towards standardization and what tools are available to help?