Go backSearch Session number: 1060

Abstract title: Racial Differences in Diabetic Costs and Control

Author(s):
SL Fultz - Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System University of Pittsburgh School of Medicine
CB Good - Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System University of Pittsburgh School of Medicine
ME Kelley - University of Pittsburgh
MJ Fine - Center for Health Equity Research and Promotion, VA Pittsburgh Healthcare System University of Pittsburgh School of Medicine

Objectives: To determine if differences in diabetic treatment and glycemic control were associated with race.

Methods: Computerized records were extracted for all outpatients receiving diabetic medications from four VA Hospitals in VISN 4 for calendar year 1997. Demographic data were extracted, along with pharmacy data and most recent hemoglobin A1C. Monthly medication and total acquisition costs (including diabetic supplies) for diabetes care were calculated based on pharmacy data. Chart reviews were done for a subgroup of patients at two sites to assess diabetic severity (based on presence of nephropathy, neuropathy and retinopathy) and comorbidity based on the Charlson Index. Logistic regression and ANOVA were used to determine if race was independently associated with cost and glycemic control, while controlling for age and marital status.

Results: Of the 4579 patients, 3258 (71%) were white and 1321 (29%) were non-white. Non-white patients were less likely to be prescribed oral hypoglycemic agents (54.7% vs 68.7%, p<0.005). Although non-white patients were more likely to be prescribed insulin (54.2% vs 41.2%, p<0.005), they were less likely to get glucose self-monitoring supplies (51.3% vs 60.6%, p<0.005). Non-white patients had higher monthly medication costs ($7.93 vs $7.20, p<0.005), lower total monthly costs ($14.73 vs $15.49, p<0.005), and higher hemoglobin A1C (7.8mg/dl vs 7.6mg/dl, p<0.005). In the models for diabetic management costs and glycemic control, non-white race was independently associated with higher total monthly costs and approached significance for greater percentage of patients with glycosylated hemoglobin >9.5mg/dl (p=0.055). In the chart review for diabetic severity and comorbid conditions, no racial differences were observed.

Conclusions: Racial differences exist in the pattern and total cost of care for veterans with diabetes, with trends of worse glycemic control in non-whites. Despite greater use of insulin, total monthly costs were lower for non-white patients due to less frequent prescribing of glucose self-monitoring supplies.

Impact statement: Similar to other conditions, this study demonstrates that racial variation in the patterns of care and glycemic control exists for VA patients with diabetes. Further research is needed to understand why different patterns exist and to eliminate inappropriate variations in care.