93IMPROVING CHRONIC KIDNEY DISEASE CARE IN PRIMARY CARE PRACTICES Chester Fox, Renee Cadzow, Brittany Cooper, Samantha Sessamen, Meredith Snyder, Primary Care Research Institute, Buffalo, New York. The purpose of this study is to improve treatment of CKD in primary care practices by increasing diagnosis of CKD, diagnosing and treating anemia, improving proteinuria with use of ACE/ARB medications, checking for bone disease, treating vitamin D deficiency, and avoiding medicines such as NSAIDS which are harmful to the kidney. Twelve practices have been recruited – 6 were randomly assigned to a basic intervention; 6 to receive an enhanced intervention. All practices received a quick reference guide summarizing the Kidney Disease Outcomes Quality Initiative (KDOQI) chronic kidney disease care guidelines. The 6 enhanced intervention sites were assigned a Practice Enhancement Assistant (PEA) to promote implementation of the guidelines. Six practices also received CINA, a computer point of care decision support protocol engine, which integrates with the practices’ existing electronic health record. Data is collected through manual chart review by a research assistant or PEA at the 6 sites without computer decision support. Data is extracted by CINA at the remaining 6 sites. At sites with support of a PEA, data is reported regularly to the practices throughout the 2 year study as part of the quality improvement cycle. This paper will present baseline, 3 month, and 6 month data for each arm of the study with results of chi-square and t-test analyses demonstrating whether differences exist between participating practices.

94TRANSFUSION BURDEN AMONG CHRONIC KIDNEY DISEASE (CKD) PATIENTS NOT ON DIALYSIS. Kathleen M. Fox1, Jerry Yee2, Ze Cong3, John Brooks4, Lois Lamerato2, Jeffrey Petersen3, Shravanthi Gandra3

1Strategic Healthcare Solutions, Monkton, MD; 2HenryFord Health System, Detroit, MI, USA;3Amgen, Inc., Thousand Oaks, CA; 4University of Iowa, Iowa City, IA. Transfusion patterns are not well characterized in non-dialysis (ND)-CKD patients. This study describes rates and units of blood transfused and hemoglobin (Hb) level at the time of transfusion in chronic, moderately-severe anemic, ND-CKD patients in routine practice. A retrospective observational cohort study of administrative data from Henry Ford Health System identified 374 adult, ND-CKD patients with chronic anemia (Hb <10 g/dL and subsequent medical care as either erythropoiesis-stimulating agent (ESA) therapy, blood transfusion, or second Hb <10 g/dL) between 01/2004-06/2008. Exclusions included those with prior diagnoses of cancer, renal or liver transplant, end-stage renal disease, acute bleeding, trauma, sickle cell disease, or aplastic anemia. At least 1 transfusion (mean of 2 units per episode; range, 1-4) was administered in 20% (75/374) at mean follow-up of 459 days. The mean Hb level closest and prior to a transfusion was 8.8 g/dL (SD ± 1.5); 20% had Hb 7-7.9, 36% had Hb 8-8.9, and 13% had Hb 9-9.9 g/dL. In conclusion, transfusions were prevalent and the mean hemoglobin level at the time of transfusion was 8.8 g/dL among ND-CKD patients with anemia. To reduce the transfusion burden, clinicians should consider other anemia treatments including ESA therapy.

95DOPPS PRACTICE MONITOR FACILITY SAMPLE REPRESENTS OVERALL US HEMODIALYSIS POPULATION DS Fuller, RL Pisoni, DM Dykstra, JM Albert, BW Gillespie, F Tentori, M Turenne, FK Port, and BM Robinson. Arbor Research Collaborative for Health, Ann Arbor, MI USA. In Nov 2010, the Dialysis Outcomes and Practice Patterns Study (DOPPS) launched the DOPPS Practice Monitor (DPM) as a public website (www.dopps.org/DPM) for reporting up-to-date statistics and trends for US dialysis practice, prior to and following implementation of the new CMS ESRD Prospective Payment System (PPS) in Jan 2011. Data are reported within 3-4 mo of collection and updated 3x yearly. DPM study sites were randomly selected to represent a wide breadth of US facility types (dialysis organization size; rural/non-rural; hospital/free-standing). Here, we present data evaluating the representativeness of the recruited DPM sample. CMS facility-level data from 2007 were used to compare the mean and distribution of 8 measures in 137 DPM facilities vs. remaining CMS facilities (N=4206) treating >20 adult chronic HD patients (Table). The DPM sample did not significantly differ from CMS overall (p>0.10) for 7 of 8 tested measures (except lower % black patients). Additional analyses illustrate the capacity of the DPM to report trendsin care not only for average facilities, but also among facilities at the tails of the distribution for the patient characteristics and practices in the table (e.g., nearly 25% of DPM facilities have >50% black

ialysis care with implementation of the PPS and other policy hanges.

patients). In sum, the DPM sample is generally representative of the US HD population, and can serve as an early warning system for monitoring trends in dc

96CLINICAL AND ACE GENETIC PREDICTORS OF INCIDENT CHRONIC KIDNEY DISEASE IN NON-DIABETIC HYPERTENSIVE VETERANS Maple M. Fung, Rany M. Salem, Dalal Naqshbandi, Daniel T. O’Connor. VA San Diego Healthcare System and University of California, San Diego. La Jolla, CA. USA.

Chronic kidney disease (CKD; an MDRD estimated GFR <60 ml/min/1.73 m2) is common and is associated with increased morbidity and mortality. The ACE gene has been associated with kidney diseases. We performed a prospective study in non-diabetic hypertensive veterans without CKD recruited from San Diego VA primary care clinics. We used an electronic medical record (EMR) and evaluated functional ACE genetic variants to determine clinical and genetic predictors of incident CKD. Adjusted hazard ratios and concordance statistics for progression to incident CKD (by the MDRD equation) using Cox regression were performed, censoring for death and loss to follow-up, over a mean follow-up of 3.2 years in the entire cohort, and in white and African American subgroups. 768 non-diabetic hypertensive veterans without CKD were included in this study, predominantly male (96.6%) and Caucasian (67.2%). After a mean of 3.2 years, 2.7% died and 35.2% developed incident CKD. We validated that well accepted CKD risk factors, eGFR, hemoglobin, history of cancer and coronary artery disease at the start of the study, and increased number of blood pressure lowering agents were predictors of incident CKD. Additionally, a polymorphism in the ACE gene (G2350A; rs4343) was strongly predictive of incident CKD in univariate and multivariate models in the entire sample and in the ethnic specific analyses. In hypertensive non-diabetic veterans, incident CKD was predicted by age, baseline renal function and anemia, and comorbid conditions such as coronary artery disease. Genetic variation within the renin-angiotensin system may also play an influential role.

NKF 2011 Spring Clinical Meetings Abstracts

Am J Kidney Dis. 2011;57(4):A1-A108A40