Skip to content

References

Every result produced from the workshop cohort is generated from code rather than typed into the site.

Those values come from findings.json and the notebook that produced them.

External claims work differently.

Equations, clinical standards, national estimates, and published evidence come from outside the cohort, so they carry citations that can be followed back to their source.

Levey AS, Stevens LA, Schmid CH, et al. A New Equation to Estimate Glomerular Filtration Rate. Annals of Internal Medicine. 2009;150(9):604–612.

DOI: 10.7326/0003-4819-150-9-200905050-00006

This is the CKD-EPI equation containing the Black race coefficient.

Steps 2 through 6 follow that coefficient downstream: from the eGFR value, to CKD classification, to the model, and finally to a prescribing threshold.

Inker 2021 | CKD-EPI 2021 race-free equation

Section titled “Inker 2021 | CKD-EPI 2021 race-free equation”

Inker LA, Eneanya ND, Coresh J, et al. New Creatinine- and Cystatin C–Based Equations to Estimate GFR without Race. New England Journal of Medicine. 2021;385(19):1737–1749.

DOI: 10.1056/NEJMoa2102953

This is the replacement equation used throughout the workshop.

The implementation here uses the creatinine-only eGFRcr(AS) form.

Delgado 2022 | NKF-ASN Task Force final report

Section titled “Delgado 2022 | NKF-ASN Task Force final report”

Delgado C, Baweja M, Crews DC, Eneanya ND, Gadegbeku CA, Inker LA, Mendu ML, Miller WG, Moxey-Mims MM, Roberts GV, St. Peter WL, Warfield C, Powe NR. A Unifying Approach for GFR Estimation: Recommendations of the NKF-ASN Task Force on Reassessing the Inclusion of Race in Diagnosing Kidney Disease.

The same Task Force report was co-published in two journals:

This report provides the national context used in step 6 for changes in medication eligibility and in step 7 for comparison with the Task Force estimate of roughly 640,000 Black adults crossing eGFR < 60.

Throughout the workshop, Delgado 2022 refers to the AJKD publication. The JASN publication is the same report with a 2021 publication year.

Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney International Supplements. 2013;3(1):1–150.

Official guideline: PDF

This is the source for the G1–G5 eGFR categories used in step 2 and in egfr.ckd_stage.

The categories appear in Chapter 1, Table 5:

GFR category eGFR (mL/min/1.73 m²) Description
G1 ≥90 Normal or high
G2 60–89 Mildly decreased
G3a 45–59 Mildly to moderately decreased
G3b 30–44 Moderately to severely decreased
G4 15–29 Severely decreased
G5 <15 Kidney failure

For the specific chapter containing the classification:

Chapter 1: Definition and classification of CKD. Kidney International Supplements. 2013;3(1):19–62.
DOI: 10.1038/kisup.2012.64

The guideline is called KDIGO 2012 because that is the guideline edition; its journal publication appeared in January 2013.

NCHS | Data presentation standards for proportions

Section titled “NCHS | Data presentation standards for proportions”

Parker JD, Talih M, Malec DJ, Beresovsky V, Carroll M, Gonzalez JF, et al. National Center for Health Statistics Data Presentation Standards for Proportions. Vital and Health Statistics. Series 2, No. 175. 2017 Aug:1–22.

PMID: 30248016

This is the basis for the reliability rule applied in step 7: an estimate must be supported by a sufficient unweighted case count and an acceptable relative standard error.

That is why the CKD projection is reported nationally while the metformin projection remains an in-sample finding.

This report is distributed through NCHS and carries a PMID rather than a DOI.

Vyas 2020 | Race correction across clinical algorithms

Section titled “Vyas 2020 | Race correction across clinical algorithms”

Vyas DA, Eisenstein LG, Jones DS. Hidden in Plain Sight — Reconsidering the Use of Race Correction in Clinical Algorithms. New England Journal of Medicine. 2020;383(9):874–882.

DOI: 10.1056/NEJMms2004740

This paper provides the broader clinical context for the workshop: race can be embedded inside an algorithm or derived measurement and then propagate into downstream decisions even when race is not explicitly present in a later model.

Cited on the landing page.

Cusick 2024 | Race-free eGFR and nephrology referrals

Section titled “Cusick 2024 | Race-free eGFR and nephrology referrals”

Cusick MM, Chertow GM, Owens DK, Williams MY, Rose S. Algorithmic Changes Are Not Enough: Evaluating the Removal of Race Adjustment from the eGFR Equation. Proceedings of the Fifth Conference on Health, Inference, and Learning (CHIL). PMLR 2024;248:619–643.

URL: proceedings.mlr.press/v248/cusick24a.html
Preprint: arXiv:2404.12812

The study follows 547,194 adults through one health system’s transition to the 2021 race-free equation.

eGFR values fell and reclassification into more severe CKD stages increased, while nephrology referral and visit rates did not move.

It provides an important downstream example for the workshop’s central argument: changing an upstream calculation does not guarantee that the downstream system changes with it.

Khazanchi 2026 | Race-free eGFR and transplant waitlist time

Section titled “Khazanchi 2026 | Race-free eGFR and transplant waitlist time”

Khazanchi R, Fleishman A, Eneanya ND, Michelson KA, Diao JA, Morse M, Pavlakis M. Wait Time Modifications for Black Transplant Candidates Affected by Race-Based Kidney Function Estimation. JAMA Internal Medicine. 2026;186(5):517.

DOI: 10.1001/jamainternmed.2026.0001

This study follows the same equation change into transplant waitlist allocation, where affected Black candidates received modifications to credited wait time.

Together with Cusick 2024, it illustrates why this workshop follows the entire supply chain rather than stopping at the equation:

the same upstream change can produce different downstream effects because the decision systems it feeds are different.

NHANES 1999–2004 with linked mortality follow-up

Section titled “NHANES 1999–2004 with linked mortality follow-up”

Centers for Disease Control and Prevention, National Center for Health Statistics. National Health and Nutrition Examination Survey, cycles 1999–2000, 2001–2002, and 2003–2004, with the NCHS Linked Mortality File.

Public-use data.

The workshop analytic cohort is the complete-case, mortality-linked subset used throughout Act II.

The exact cohort is also fingerprinted: its SHA-256 hash is stored in findings.json under _meta.cohort_sha256.

That allows a published workshop result to be traced not only to a notebook and a dataset, but to the exact cohort rows that produced it.

Research, development, production, and presentation by Waverly Rose Brim.