We outline the experience of Southern California Kaiser Permanente, a large integrated health maintenance organization, in implementing the chronic kidney disease (CKD) definition and staging guidelines of the Kidney Disease Outcomes Quality Initiative (KDOQI) from 2002 to 2008, including estimated glomerular filtration rate (eGFR) implementation, algorithm for GFR range assignment and reassignment, and practical modifications of CKD staging for population management. We departed from the KDOQI CKD definition and staging as follows: for stages 1 to 2, we required "macroproteinuria" rather than "microalbuminuria" as the marker of kidney damage; for stage 3, we included individuals with macroproteinuria, diabetes mellitus based on diabetic registry, or eGFR + 1/2 age less than 85; and for stage 5, we included only individuals not receiving renal replacement therapy. In an adult population of 2.5 million members, we identified 2.9% (72,005) for CKD population management (0.1%, 0.2%, 1.7%, 0.15%, and 0.01% with stages 1, 2, 3, 4, and 5, respectively). Outpatient visits with a nephrologist in the past 12 months for the prevalent CKD population increased modestly from 2003 to 2008 from 20% to 24%. Nephrologists see a higher risk subset, including 77% of patients with stages 4 to 5, 45% of prevalent patients with CKD stages 1 to 5 with the last urine protein level greater than approximately 1 g, and 21% of patients with stage 3 in the past 12 months, but only 4% of patients with eGFR of 30 to 59 mL/min/1.73 m(2) not meeting our criteria for stage 3. Primary care providers see the majority of patients with stages 1 to 5 in the course of a year (85%) and are aware of kidney disease (79% coded for kidney disease). Other quality indicators during the 12-month window include the following: for patients with prevalent CKD stages 1 to 5, a total of 56% with last blood pressure greater than 129/79 mm Hg, 21% missing qualitative proteinuria, 16% missing angiotensin-converting enzyme inhibitor or angiotensin receptor blocker, 11% missing low-density lipoprotein cholesterol, 40% with low-density lipoprotein cholesterol level greater than 100 mg/dL, 50% of patients with diabetes with hemoglobin A(1c) level of 7% or greater; for prevalent patients with CKD stages 3 to 5, a total of 14% missing hemoglobin level and 13% with hemoglobin level less than 11 mg/dL; and for prevalent patients with CKD stages 4 to 5, a total of 2.5 hospital d/patient and 62% not attending instructional classes for modalities of renal replacement therapy. Optimal start of end-stage renal disease therapy, defined as the proportion of patients with stages 4 to 5 who either started peritoneal dialysis therapy directly, started hemodialysis therapy using an arteriovenous fistula, or received a preemptive renal transplant, was 54%. Comprehensive CKD care is possible within a large health maintenance organization, but with substantial opportunity for improvement remaining.