Introduction: KidneyIntelX is a multiplex immunoassay of 3 plasma biomarkers with 7 clinical variables combined using machine-learning to generate a risk score for progressive decline in kidney function over 5 years in individuals with early-stage diabetic kidney disease (DKD) .

Methods: A Population Health defined Mount Sinai approved care pathway for DKD patients informed by the KidneyIntelX test was introduced into the Mount Sinai Health System in New York, NY as part of a Real World Evidence (RWE) study (NCT04802395) . Decision impact of medication management (anti-hypertensives, SGLT2 inhibitors/GLP1 agonists) and specialist referral was tracked. An interim analysis was performed for 1) Assessing comparability between RWE and a published clinical validation cohort based on KidneyIntelX risk score distribution; 2) Determining if necessary EHR clinical fields were captured; 3) Identifying early evidence of KidneyIntelX impact on provider decision-making.

Results: Between Mar-Nov 2021, 1,112 patients had KidneyIntelX test results, with post-test follow up to 36 weeks. The risk breakdown of RWE population was similar to the clinical validation cohort [High 14% vs. 17%, intermediate 40% vs. 37% and low risk 46% vs. 46%]. EHR record review for care changes confirmed ability to meet study objectives. Compared to patients scored low risk, there were changes in anti-hypertensives (OR 3.0; 95% CI 1.6-5.6) , initiation of SGLT2i/GLP-1a (OR 6.4; 95% CI 2.9-14) and increased referrals to nephrologists, endocrinologists, or dieticians (OR 2.8;95% CI 1.7-4.7) in patients scored as high-risk (Figure) .

Conclusions: KidneyIntelX was successfully deployed in a health care system in a comparable population to the validation cohort with high data capture fidelity. Application of guideline-based

therapies and specialist referral increased in the proportion to reported risk level by 3-6 and >2-fold, respectively.

Disclosure

J. Tokita: None. M. J. Donovan: Employee; Renalytix. R. Fields: None.

Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.