Improvements in Diabetes Processes of Care and Intermediate Outcomes: United States, 19882002

Context As the target of many quality improvement programs, positive change in diabetes care is a good marker for progress toward better health care. Content The authors analyzed measures of diabetes care from national population-based surveys that were conducted between 1988 and 2002. Improvements occurred in the proportion of patients with hemoglobin A1c between 6% and 8%, low-density lipoprotein (LDL) cholesterol levels less than 3.4 mmol/L (<130 mg/dL), annual influenza vaccination, and aspirin use. Blood pressure did not change. Substantial proportions of patients still had poor control of LDL cholesterol levels, glycemia, and blood pressure. Implications Despite some progress, population-based measurements show that care for many Americans with diabetes falls far short of targets. The Editors Diabetes currently affects 20.8 million people in the United States (1), and that number is projected to reach 39 million by the year 2050 (2). If current trends continue, 1 in 3 Americans will develop diabetes sometime in his or her lifetime, and those with diabetes will lose, on average, 10 to 15 life-years (3). In 2002, diabetes cost the nation an estimated $132 billion in direct and indirect costs (4). There is, however, a growing array of effective and cost-effective treatments to help prevent or delay diabetes complications and also diabetes itself (5-17). Diabetes care has been suboptimal and varied in the United States (18-21). The National Diabetes Quality Improvement Project, founded in 1997, developed a comprehensive set of measures of diabetes quality of care (22). These measures have been incorporated into the Health Plan Employer Data and Information Set, the American Diabetes Association Provider Recognition Program, the American Medical Association Diabetes Measures Group, the Veterans Administration performance monitoring program, and other activities. The Diabetes Quality Improvement Project partners now continue their work as a coalition of 13 influential private and public national organizations called the National Diabetes Quality Improvement Alliance. The Alliance develops, maintains, and promotes the use of an updated standardized measurement set (the Alliance measures) for quality of diabetes care (23). We previously established a national benchmark for diabetes quality of care in the United States for the years 1988 to 1995 by using the standard measurements recommended by the Diabetes Quality Improvement Project (18). On the basis of nationally representative data collected in 1999 to 2002, we report the changes in the quality of diabetes care from the 1990s to 2000s by using the standardized Alliance measures for both time periods. Methods Surveys We used data from 2 federally funded, nationally representative surveys: the National Health and Nutrition Examination Survey, 19881994 (NHANES III) and 19992002 (NHANES 19992002), and the Behavioral Risk Factor Surveillance System, 1995 (BRFSS 1995) and 2002 (BRFSS 2002). As previously explained (18), we used both BRFSS and NHANES to obtain data on all the process and intermediate outcome measures needed for the analysis. In our report, we refer to NHANES III and BRFSS 1995 as baseline surveys and NHANES 19992002 and BRFSS 2002 as recent surveys. We analyzed data from each survey separately. Table 1 presents the indicators used and their respective data source. Table 1. National Diabetes Quality Improvement Alliance and Additional Indicators of Diabetes Processes and Outcomes of Care National Health and Nutrition Examination Survey The NHANES consists of nationally representative samples of the U.S. civilian, noninstitutionalized population. Samples were obtained by using a stratified multistage probability design with planned oversampling of older and minority groups. Household interviews were conducted to ascertain sociodemographic characteristics and medical and family history. After the household interview, clinical examinations were conducted at a mobile examination center. Detailed descriptions of the design and data collection of each survey have been published elsewhere (24-27). Data from NHANES were self-reported (demographic characteristics and clinical variables) or were obtained during the clinical examination (hemoglobin A1c, cholesterol level, triglycerides level, and blood pressure level). Hemoglobin A1c measurements were standardized to the Diabetes Control and Complications Trial. Cholesterol levels were standardized by using the criteria established by the Centers for Disease Control and Prevention and the National Heart, Lung, and Blood Institute Lipid Standardization Program II. For persons who fasted for more than 8 hours and had triglyceride levels less than 4.5 mmol/L (<400 mg/dL), the Friedewald equation was applied to calculate low-density lipoprotein (LDL) cholesterol level. We log-transformed triglyceride levels because data were not normally distributed. We used the average of each person's blood pressure readings that were taken in the seated position during the clinical examination. Because we did not have data on annual testing for microalbuminuria, we assessed the absence of microalbuminuria, defined as albumin-to-creatinine ratio greater than 30 g/mg in spot urine collection (28). We analyzed the data for all indicators regardless of respective treatment status. Behavioral Risk Factor Surveillance System The BRFSS is an ongoing random-digit telephone survey of the noninstitutionalized U.S. adult population in each of the 50 states and the District of Columbia. Detailed descriptions of the design and data collection of the BRFSS have been published elsewhere (29). We used the diabetes-specific module that contains questions on clinical and preventive care practices to collect information from the participants with diabetes. Participants We included adults 18 to 75 years of age who reported a previous diagnosis of diabetes by a health care professional. We excluded women with gestational diabetes. We analyzed data from 1024 participants in NHANES III and 750 participants in NHANES 19992002 who selfreported a diagnosis of diabetes and who completed the clinical examination. We analyzed data from 3065 persons in BRFSS 1995 and 13078 persons in BRFSS 2002 who identified themselves as having diabetes. Participants reporting diabetes in all surveys were similar in age, sex, education, smoking, and insurance status at each point of time. Among participants of the recent surveys compared with those of the baseline surveys, the proportion of women and non-Hispanic white persons was lower and the proportion of participants with more than a high school education and an annual household income of $20000 or more was higher (Table 2). The proportion of people with diabetes who use insulin was also lower in the recent surveys but was statistically significant only in BRFSS 2002. Table 2. Characteristics of Participants 18 to 75 Years of Age with Self-Reported Diabetes in the National Health and Nutrition Examination Survey, 19881994 and 19992002, and Behavioral Risk Factors Surveillance System, 1995 and 2002 Performance Measurement Set We assessed the quality of diabetes care by using the Alliance measurement set (22) (Table 1). We used the Alliance measures of diabetes care wherever data were available, and we also examined additional measures that may be indicators of quality care in the future: pneumococcal vaccination, diabetes education, annual dental examination, and self-monitoring of blood glucose level. The BRFSS did not have a question about smoking counseling. We, therefore, used the proportion of smokers who tried to quit smoking. Questions about aspirin use were asked only every other year, so we used data from BRFSS 1996 for this variable. Statistical Analysis We conducted statistical analyses by using SAS for Windows software, version 7.0 (SAS Institute, Inc., Cary, North Carolina), for data management. We used SUDAAN software (Research Triangle Institute, Research Triangle Park, North Carolina) to obtain point estimates and SEs based on sampling weights to produce national estimates accounting for the complex survey design. We used Taylor series linearization for variance estimation. We computed the percentage of respondents who reported receipt of each measure. We examined the diabetes care measures by age, sex, race or ethnicity, education, insulin use, and health insurance status because our previous analysis had variations by these factors (18). However, insulin users were not asked to fast; hence, we did not examine LDL levels by insulin use. We used multiple logistic regression and predictive margins to estimate the probability of receiving or meeting the care measure after controlling for all known potential confounders. Predictive margins are a type of direct standardization, where the predicted values from the logistic regression models are averaged over the covariate distribution in the population (30). This statistic has several advantages over the odds ratio: It is not influenced if the outcome is not rare; a comparison group is not required; and it provides a measure of absolute difference rather than relative difference. We included an interaction term between time and each measure in the models to allow estimation of the probability for each period. To assess the difference in the percentage change between the 2 comparison groups, we tested the interaction term of each demographic characteristic and clinical variable (age, sex, race or ethnicity, education, insulin use, and health insurance status) with time. Role of the Funding Source No funding was received for this study. Results Half of the quality care measures that we analyzed improved between the baseline and recent surveys, and the only measure that worsened was the proportion of participants with hemoglobin A1c < 6%. We observed absolute increases for LDL levels less than 3.4 mmol/L (<130 mg/dL) (22 percentage points), annual lipid profil