UNASSIGNED: We developed an algorithm to detect AMR using the 2006-2011 Centers for Medicare & Medicaid Services (CMS) using ICD-10 and billing codes as there is no specific ICD-10 or procedure code for AMR. We then compared healthcare utilization, cost, and risk of graft failure or death in AMR. patients versus matched controls.
UNASSIGNED: The algorithm had a 39.4% true-positive rate (69/175) and a 4.1% false-positive rate (110/2,655). We identified 5,679/101,554 (5.6%) with AMR, who had a nearly 3-fold higher risk of graft failure (hazard ratio [HR], 2.75, 95% confidence interval [CI], 2.50 to 3.03; p < .0001) and death (HR, 2.59; 95% CI, 2.35 to 2.86; p < .0001) at 2 years, nearly 5 times the hospitalizations in the 60 d before AMR diagnosis, and increased nephrology and emergency department visits. Mean AMR attributable healthcare costs were 4 times higher than matched controls, at $13,066 more per patient in the 60 d before AMR diagnosis and $35,740 per patient per year higher in the 2 years after AMR diagnosis.
UNASSIGNED: US kidney transplant recipients with AMR have substantially greater healthcare utilization and higher costs and risk of graft loss and mortality.