The quality of diabetes care has made small gains over the last 10 years1 pertaining to the four major risk factors related to diabetes outcomes: A1C, hypertension, lipid control, and tobacco use. While there has been some progress in lipid control, there has been no improvement in curbing tobacco use and very modest progress in glucose control and hypertension management.2 Patients under the age of 40 with diabetes have made the least progress. With the increase in the absolute number of patients with diabetes, there are more patients today in poor control of their condition than 10 years ago.1
The metrics related to diabetes care are disturbing, especially during a period of time when there has been a significant increase in the number of medications focused on glucose control, hypertension, and LDL lowering. In 1997, the first standardized set of diabetes care metrics were developed as part of the Diabetes Quality Improvement Project (DQIP).3 These metrics were partially based on the feasibility of gathering the data, rather than on the factors that had the greatest effect on intermediate outcomes. Efforts to improve quality also saw the proliferation of “guidelines,” “recommendations,” and “standards of care” by international organizations; professional organizations, such as the American Diabetes Association and the American Association of Clinical Endocrinologists; and government-funded committees and programs, such as the United States Preventive Services Task Force (USPSTF), the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure, and the Adult Treatment Panel III. The Agency for Healthcare Research and Quality (AHRQ) maintains a library of evidence-based clinical practice guidelines. The NDEP provides a list of diabetes care standards as well as the Guiding Principles for the Care of People With or at Risk for Diabetes.4
Additional key developments in quality management over the past 10 years include:
National Quality Strategy
Health care innovation and insurance reform are major components of the Patient Protection and Affordable Care Act (ACA). As mandated by the ACA, AHRQ developed the National Quality Strategy5 based on the triple aim “better care, healthy people/healthy communities, and affordable care.” AHRQ solicited the opinions of more than 300groups, organizations, and individuals in developing the strategies that make up the general efforts of the federal government and other stakeholder organizations in improving quality. AHRQ continues to develop specific methods to improve quality and tools to move the process forward.
National Quality Forum
The National Quality Forum (NQF), a 400-member organization initially funded by the federal government in 1999, provides leadership in the development and implementation of guidelines in a broad range of areas of health care topics. NQF has a systematic method of endorsing specific metrics based on a standard process from organizations such as CMS and NCQA. At a federal level, there is integration with the National Quality Strategy.6
Translating Research into Action for Diabetes Study
The TRIAD (Translating Research into Action for Diabetes) study7 was initiated in 1998 and funded by the Centers for Disease Control and Prevention to assess the impact of health plan–sponsored disease management on the quality of diabetes care. Common metrics for diabetes care were being established, and the six center study sites combined data to look at both health plan program design to improve diabetes care and patient factors that potentially would impact outcomes. The study involved 10 health plans and 180,000 adult patients with a number of distinct and varied interventions. The TRIAD study found that health plan interventions directed both at provider groups and at patients significantly improved process measures but failed to improve intermediate outcomes. Accurate data collection was a major challenge and likely decreased the ability to have credibility with providers or provide accurate data to modify interventions going forward. The Veterans Health Administration (VHA),a late addition to the study group, had a relatively mature information system and made significant system changes that were data driven. Consequently, the VHA showed significant improvement in intermediate outcomes in addition to process changes.8
Diabetes Quality Metrics
Based on the modest changes in diabetes care over the last 15 years and the rrenewed dialogue related to aligning reimbursement incentives, risk-sharing, and value-based payment, how should both individual providers and provider groups proceed? Quality metrics are changing to focus on outcome measures primarily by using disease modeling. For example, the CMS Million Hearts® primary care research study is using the American College of Cardiology/American Heart Association (ACC/AHA) 10-year atherosclerotic cardiovascular disease (ASCVD) pooled cohort risk calculator.9 The metrics that have the greatest impact on the risk for stroke and heart disease are the primary serial metrics collected to assess the impact of the program. The National Quality Forum includes a composite measure for diabetes care with five elements that determine outcome, initially developed by MN Community Measurement. One of the best-tested and mature disease models capable of risk assessment for multiple diseases (including diabetes) is Archimedes, developed using the Kaiser database. The risk engine uses a standardized method to generate a “Global Outcome Score,” allowing a systemic approach to measure 10-year risk for a population and comparisons with other health organizations.10 The Global Cardiovascular Risk Score is now incorporated into the NCQA HEDIS measures as a composite metric.11,12
All diabetes metrics are not created equal when measured against the impact on a patient’s health and wellbeing. Pay-for-performance frequently pays a set amount for a variety of metrics, initially developed by health plans and based primarily on the ease of measuring. A limited number of modifiable variables determine 10-year outcome scores. Understanding the impact of A1C, blood pressure control, LDL levels, and smoking status on health assists the provider and patient in prioritizing intervention. The framework for selecting metrics that have the greatest impact on health potentially would greatly reduce the large number of quality metrics to those with the greatest outcome on health.13 Future trends will likely stratify incentives related to desired outcomes, often using composite measures that are intermediate outcomes when considered alone but predict long-term outcomes when entered into a disease risk model.
Focusing Efforts on Upstream Solutions
Moving the risk model upstream improves health outcomes. Patients with diabetes under the age of 40 may have not benefitted from previous efforts to improve self-management skills, as measured by process and outcome metrics. This population will have the greatest risk reduction with quality diabetes care given their time horizon. Risk factor modification has progressively less impact on outcomes as patients enter their last decade of life. This population, by definition, has a limited life expectancy and few metrics that will add months or years of productive life. Quality care for elderly patients with diabetes has yet to be fully defined. What is clear is that quality care in this age group must be sensitive to cost and overall burden of treatment with decreasing rewards. Current cost reduction is frequently driven by Medicare expenditure and complex high-cost patients. Quality of care in this population may reduce cost, but not quality as measured by outcomes.
The evolving quality metrics will re-focus resources to the management of prediabetes and childhood obesity. The team approach required for the management of complex patients needs to also be developed for effective intervention for children and young adults. The USPSTF recommends that health care providers of children aged 6 years and older who are obese should “offer them or refer them to comprehensive, intensive behavioral intervention to promote improvement in weight status.”14 The AHRQ-funded meta-analysis of existing programs recognized moderate evidence that school and community-based intense programs are effective in modifying outcomes related to childhood obesity.15 Office or health organization-led programs have not improved outcomes. Future quality improvement will require a different approach to affect health outcomes as we move toward restructuring the delivery system.
As these developments over the past 10 years suggest, we have come a long way in the journey of defining quality in diabetes care. The DQIP led to the development of the first standardized set of diabetes care metrics in 1997.2 Many of the metrics (glycemic control, lipid control, blood pressure control, etc.) that were used in the DQIP are still used today, although the definitions of “good control” have evolved with the influence of newer data. Efforts to improve quality also led to the proliferation of “guidelines,” “recommendations,” and “standards of care” by professional organizations, international organizations, and government-funded committees and programs.
Today, ACA has placed the measurement of quality and its improvement at the heart of payment reform. With the ACA, the National Quality Strategy was refined and ensured that the framework for quality measures would be well rounded and holistic in its approach. As the literature continues to evolve, the umbrella organization that ensures the measures evolve with it is the National Quality Forum. This “organization of organizations” ensures that all stakeholder voices are heard and can participate in measure development. In that regard, there is increasing interest to find measures that most accurately predict beneficial clinical outcomes that are in the best interest of patients—and are not simply most convenient to measure. Given the diabetes disease burden and costs borne by its patients and the health care system, quality measurement and its improvement and refinement will continue well into the future.