Improving the Outcomes of Metabolic Conditions: Managing Momentum To Overcome Clinical Inertia
J. Perlin,L. Pogach
DOI: https://doi.org/10.7326/0003-4819-144-7-200604040-00012
IF: 39.2
2006-04-04
Annals of Internal Medicine
Abstract:Hypertension, dyslipidemia, and diabetes are highly prevalent, and often concurrent, conditions: 50 million Americans have hypertension (1), 38 million have high-risk cholesterol levels (2), and 20.8 million have diabetes (3). The direct (medical services) and indirect (disability and premature death) costs of treating diabetes alone exceed $130 billion per year (4). Amputations, renal failure, visual loss, stroke, heart attack, and premature death reduce the length of life of patients and the quality of life for them and their families. In the last decade or so, several remarkable clinical trials have shown that better control of glycemia, blood pressure, and low-density lipoprotein (LDL) cholesterol level leads to better outcomes (1-3). We now know that surveillance for nephropathy, neuropathy, and retinopathy enables early identification and treatment of diabetes-related complications (5). Guidelines have incorporated the new evidence very quickly. National public education programs and professional societies have disseminated these evidence-based recommendations to the public and to professionals. The new evidence likewise influenced national voluntary consensus standards for performance measurement, which are widely used for accreditation (6) and quality improvement (7). Has this unprecedented momentum created by publicizing evidence and measuring outcomes resulted in translation of evidence into practice? And if not, why? Two studies in this issue (8, 9) provide us with a progress report from both the public health and managed health care plan perspectives. Saaddine and colleagues (8) evaluate changes in the quality of diabetes care from the 1990s to the early 2000s. They measured care by comparing patient-level findings on nationally representative, federally sponsored surveys to standardized national consensus measures for diabetes care. The news is mixed. Encouragingly, the number of individuals with hemoglobin A1c levels between 6% and 8% increased by about 13% to 47%, the proportion of individuals with LDL cholesterol levels less than 3.4 mmol/L (<130 mg/dL) increased by 22% to 64%, and the frequency of aspirin use increased by 13% to 45%. However, mean hemoglobin A1c remained unchanged, as did the percentage (about 68%) of individuals whose blood pressure was less than 140/90 mm Hg. Important process measures, such as foot screening and dilated eye examinations, did not increase. Although better-educated individuals with health insurance generally fared better, these socioeconomic factors did not consistently predict better intermediate outcomes, such as blood pressure control or reduced LDL cholesterol level. The study has weaknesses and strengths. We should be cautious about comparing responses on cross-sectional surveys that were performed years apart, especially with an intervening change in the definition of diabetes. On the other hand, the National Health and Nutrition Examination Survey used a sampling strategy that assures a study sample representative of the U.S. population. Extrapolation of the findings to the U.S. population suggests that millions of Americans with diabetes are benefiting from improved glycemic and cholesterol level control. Unfortunately, care remains poor, not merely suboptimal, for a substantial minority of patients: About 20% have hemoglobin A1c greater than 9% (with 14% having hemoglobin A1c > 10%), 33% have uncontrolled blood pressure (>140/90 mm Hg), and 40% have poor LDL cholesterol level control (>3.4 mmol/L [>130mg/dL]). Less than 50% of patients were taking aspirin, an inexpensive and reasonably safe, yet effective, medication to reduce the risk for coronary heart disease. The main message of Saaddine and colleagues' study is that millions of Americans remain at high risk for the complications of diabetes despite the new evidence and campaigns to educate the public and the medical profession. The term clinical inertia refers to the failure of providers to alter therapy in the face of clear indications for changes (10). Rodondi and colleagues (9) used automated data from a large health maintenance organization to evaluate predictors of medication changes for patients with and without diabetes. The authors defined a single poor glycemic control threshold (hemoglobin A1c > 8%) and varying poor LDL cholesterol level and blood pressure control thresholds on the basis of co-existing conditions, such as diabetes or cardiovascular or renal disease. The baseline prevalence of poor control was 28% for hypertension and 42% for dyslipidemia and glycemic control. The authors defined appropriate care as either an indicated change in treatment or a subsequent measurement that was within the range of adequate control. Most patients with poor control of risk factors70% for hemoglobin A1c, 59% for LDL cholesterol level, 71% for systolic blood pressure, and 82% for diastolic blood pressurewere managed appropriately. Indeed, 70% of the treatment changes occurred within 2 months. The results are especially encouraging for glycemic control since the threshold that defined the need for action was a hemoglobin A1c less than 8%. However, the authors excluded insulin-requiring persons with poor hemoglobin A1c control, which limits the generalizability of this finding. These 2 articles show that diabetes care in the United States is getting better but still falls far short of reasonable goals. The performance of the Veterans Health Administration (VHA) in diabetes care has improved substantially in the past decade so that it is similar to and generally exceeds commercial health care plans (11) and fee-for-service Medicare (12). Moreover, ongoing measurements show that the VHA's performance continues to improve (13). Both the public health and managed care sectors can learn from the VHA experience in systematizing care for more than 1 million veterans with diabetes and overcoming clinical inertia that characterized care as late as 19992000 (14). The VHA developed a national infrastructure to support the delivery of evidence-based care. Key elements included assignment of patients to an identified primary care provider, dissemination of guidelines through pocket cards and internal educational broadcasts, implementation of electronic medical records, and quarterly performance assessment. Each VHA network and medical center director had a performance contract that included fulfillment of reasonable target goals for each accountability measure. Facility managers could implement computerized clinical reminders, academic detailing, and feedback to clinicians but could not offer financial incentives. Cross-sectional analyses demonstrated that most of the variation in outcomes was between VHA medical centers rather than between individual physicians in a medical center (15). Indeed, some facilities far exceed the VHA-wide average rate of lowering of hemoglobin A1c levels over a 2-year period (16). Since all VHA medical centers operate under the same policies and use the same electronic health records, why does facility performance still vary? One evolving area of inquiry suggests the importance of organizational factors. Medical centers with better performance had more frequent internal feedback to clinicians, identified frontline clinicians to lead (champions), and were more likely to accept the guidelines as applicable to their practice (17). These observations suggest that while a health care system may initiate momentum for change, clinicians are essential change agents for local quality improvement efforts. Momentum is now building nationally to improve individual physicians' access to high-quality medical information systems, which was a key factor in the success of the VHA. The American Health Information Community has been chartered to support the President's goal of electronic health records for most Americans within 10 years (18). However, negative trials of diabetes quality improvement interventions in large health care systems with electronic health records (19) underscore the need for continued research to better understand the inter-relationshipsamong patient, provider, and health care systemthat contribute to clinical inertia. Although generalizable solutions are not yet within our reach, the studies in this issue clearly document that physicians must hold themselves to a higher standard than simply maintaining the status quo. The success of the VHA demonstrates that continued managerial momentum can lead to sustained improvement even in disadvantaged populations. We, therefore, offer a modest but achievable proposal for every health care professional, every practice, and every private and public health care system. First, review your own data and processes of care for diabetes. Second, use the principles of evidence-based medicine to identify the highest-priority goals and treatment strategies (20). Third, develop actionable plans to increase the proportion of clinical encounters in which clinicians take appropriate action. Fourth, systematically track progress toward achieving treatment goals. Fifth, give feedback to those who are making patient care decisions (9). We must recognize the deadly consequences of clinical inertia and commit ourselves to the task of overcoming it in our own practices. If we succeed, we will improve the prospect of healthier lives for tens of millions of Americans.