Comparisons of HbA1c among healthcare plan administrative units are currently being made without case mix adjustment (CMA) because validated CMA models are not available. Furthermore, use of last available A1c value does not permit a longitudinal assessment of change in persons at risk. The availability of linked datasets over several years for persons with diabetes allows us to study and monitor the progress of individual patients and to develop the trajectories of HbA1c levels or other key diabetes quality standards for each individual patient. The average progresses of individual patients in the same VHA facility (after risk adjustment) can be viewed as a measure of improvement for the facility or center.
1) Determine directions and magnitudes of institution-level changes in HbA1c measured continuously as well as dichotomously from FY98 to FY99 to ascertain the amount of improvement in performance.
2) Evaluate respective contributions of individual and institutional factors and their joint influences to performance improvement
3) Explore empirical connections between cross-sectional and longitudinal ratings of institutional performance and develop a conceptual understanding of similarities and differences between the two modes of performance evaluation and risk adjustment approaches.
In order to faithfully represent institutional change in performance, we will utilize a three-level hierarchical growth curve approach to model change patterns over time. This model allows for the representation of hierarchical structure for repeated measures that are nested within individuals, and individuals in turn nested within institutions.
Cross Sectional: We determined thresholds for predicted A1c corresponding to unadjusted thresholds (9.5%, 9.0%, 8.5% and 8.0%), counted individuals at each station above the adjusted threshold, then generated station-level observed-to-expected ratios for profiling. We compared stations placed in the 5th and 95th percentiles using raw A1c values with those using station-level observed-to-expected ratios. A1c assay variation caused 81% of station level clustering The final case-mix adjustment model had R2=10.4%. Thirty random 50% samples demonstrated stability in parameter estimates and R2. Profiling resulted in 50-75% disagreement of stations placed in the 5th and 95th percentile with and without case-mix adjustment at all thresholds. Even with lab instrument variability removed, case-mix adjustment using just administrative data results in large differences in which stations are placed in the 5th or 95th percentiles at all tested thresholds.
Longitudinal: We identified a cohort of 284,895 patients from 125 facilities with 816,721 A1c tests from October 1, 1998 to September 30, 2000. We found that the preponderance of facilities showed monthly declines in within-patient A1c over the study period (mean change of -0.0148 A1c units per month, range -0.074 to 0.042.) Individual facilities varied in their monthly change, with 105 facilities showing monthly declines (70 significant at 0.05 level) and 20 showing monthly increases (5 significant at 0.05 level.) Case-mix adjustment resulted in modest changes (mean change of -0.0131 case-mix adjusted A1c units per month, range -0.079 to 0.043.)
Our findings indicate that case mix adjustement of HbA1c using cross sectional data is feasible, and may be necessary for fair comparisons. There is also substantial variation in facility-level longitudinal changes in A1c levels. We propose that evaluation of change in A1c levels over time can be used as a measure to reflect quality of care provided to populations of individuals with chronic disease.
- Pogach LM, Tiwari A, Maney M, Rajan M, Miller DR, Aron D. Should mitigating comorbidities be considered in assessing healthcare plan performance in achieving optimal glycemic control? The American journal of managed care. 2007 Mar 1; 13(3):133-40.
- Aron D, Rajan M, Pogach LM. Summary measures of quality of diabetes care: comparison of continuous weighted performance measurement and dichotomous thresholds. International journal for quality in health care : journal of the International Society for Quality in Health Care / ISQua. 2007 Feb 1; 19(1):29-36.
- Maney M, Tseng CL, Safford MM, Miller DR, Pogach LM. Impact of self-reported patient characteristics upon assessment of glycemic control in the Veterans Health Administration. Diabetes Care. 2007 Feb 1; 30(2):245-51.
- Pogach LM, Rajan M, Aron DC. Comparison of weighted performance measurement and dichotomous thresholds for glycemic control in the Veterans Health Administration. Diabetes Care. 2006 Feb 1; 29(2):241-6.
- Pogach L, Xie M, Shentue Y, Tseng CL, Maney M, Rajan M, Tiwari A, Kolassa J, Helmer D, Crystal S, Safford M. Diabetes healthcare quality report cards: how accurate are the grades? The American journal of managed care. 2005 Dec 1; 11(12):797-804.
- Thompson W, Wang H, Xie M, Kolassa J, Rajan M, Tseng CL, Crystal S, Zhang Q, Vardi Y, Pogach L, Safford MM. Assessing quality of diabetes care by measuring longitudinal changes in hemoglobin A1c in the Veterans Health Administration. Health services research. 2005 Dec 1; 40(6 Pt 1):1818-35.
- Miller DR, Christiansen CL, Chen X, Fincke BG, Tseng C, Pogach L. Improvements in VA Patient Glycemic Control Over Time with Differences by Race and Co-Morbid Disease. Poster session presented at: VA HSR&D National Meeting; 2008 Feb 14; Baltimore, MD.
- Miller DR, Caspard H. Prevalence of metabolic syndrome in U.S. Veterans Health Administration patients. Paper presented at: Rocky Mountain Metabolic Syndrome Annual Symposium; 2006 May 3; San Francisco, CA.
- Maney M, Tseng CL, Safford M, Miller D, Pogach L. Impact of Self-Reported Social/Health Data on Facility Level Ranking of Improved Glycemic Control. Paper presented at: VA HSR&D National Meeting; 2006 Feb 1; Arlington, VA.
- Tiwari A, Pogach L, Maney M, Rajan M, Miller D. Assessing Facility Performance in Achieving Intensive Glycemic Control: Who Should be Profiled? Paper presented at: VA HSR&D National Meeting; 2006 Feb 1; Arlington, VA.
- Pogach LM. Diabetes Healthcare Quality Report Cards: How Accurate Are The Grades? Paper presented at: AcademyHealth Annual Research Meeting; 2005 Jun 1; Boston, MA.
Technology Development and Assessment
Diabetes, Quality assessment, Risk adjustment