2006 HSR&D National Meeting Abstract
1010 — Assessing Facility Performance in Achieving Intensive Glycemic Control: Who Should be Profiled?
Tiwari A (Center for Healthcare Knowledge Management)
Pogach LM (Center for Healthcare Knowledge Management)
Maney M (Center for Healthcare Knowledge Management)
Rajan M (Center for Healthcare Knowledge Management)
Miller DR (CHQOER)
The current public reporting threshold for glycemic control is an A1c >9% (poor control). Efforts to assess excellent control have been hampered by the lack of a validated risk adjustment model. Our objective was to identify veterans with medical and mental health conditions that could attenuate benefits or increase harms from intensive glycemic control and then compare VA facility rankings (<7% A1c threshold) using individuals without contraindications to the entire population.
We utilized the Diabetes Epidemiological Cohort to identify veterans with diabetes less than 65 years of age in FY99 and who were alive at the beginning of FY00. We evaluated the proportion, characteristics, and 4-year mortality of individuals with serious medical or psychiatric comorbidities that were considered contraindications to intensive glycemic treatment. League table rankings of facilities achieving an A1c of <7% in FY00 were determined using the entire population and the intensive glycemic control cohort.
220,922 subjects were identified from 144 facilities. Overall, 73.5% ± 7.2% of individuals were excluded as non-ideal candidates for intensive glycemic management. Four-year unadjusted mortality was 13.2% in the entire population, 14.4% in excluded subjects, and 4.8% in subjects without contraindications. Mean and percent less than 7% A1c levels were comparable in the “intensive glycemic control cohort” (8.06%, 33.3%[19.6 to 50.2]) and the total population (7.97%, 35.2%[17.6-66.7]). Comparison of facility league table rankings indicated that 5 of 28 of the 20% best and 7 of 28 of the 20% worse performing facilities changed at least two decile ranks when the cohort of individuals without contraindications to intensive glycemic control cohort was utilized instead of the unadjusted population.
A significant proportion of persons with diabetes have one or more contraindications for intensive glycemic management. Although A1c control was comparable in the entire population and intensive glycemic control cohort there were marked changes in decile ranks among the facilities in the best and worse two deciles.
Administrative data can be used to identify veterans without contraindications to intensive glycemic control. These cohorts can be used to focus quality improvement efforts and to develop an “excellent glycemic control” performance measure.