Integrating Chronic Care and Business Strategies in the Safety Net
Appendix A. The Evidence Base for the Chronic Care Model
History of the Chronic Care Model
The initial evidence upon which the Chronic Care Model was based came from evaluations of interventions to improve care. For example, the MacColl Institute participated in a Cochrane Collaboration review of interventions to improve diabetes care in primary care,72 These reviews demonstrated the need for the integrated set of changes called for by the Chronic Care Model. A more recent meta-analysis by Tsai and colleagues confirms these earlier findings and extends them to other conditions.73
Several healthcare organizations began adopting the Chronic Care Model around the turn of the 21st century either through participation in the Improving Chronic Illness Care (ICIC)- sponsored collaboratives or on their own. The second body of evidence about the effectiveness of the Chronic Care Model comes from observational evaluations of that experience. Several early chronic care collaboratives have been evaluated and generally document improvements. Investigators at the Center for Medicare and Medicaid Services (CMS) studied the quality of diabetes care in 134 managed Medicare organizations participating in a diabetes performance measurement program.74 Fleming and colleagues used an organizational assessment tool based on the Chronic Care Model to compare high- (top quartile) and low- (bottom quartile) performing organizations. They found that high-performing organizations were much more likely to organize care delivery in accordance with the Chronic Care Model. They then identified specific systemic features that characterized high-performing organizations and differentiated performance. These included computerized reminders, practitioner involvement on quality improvement teams, guidelines supported by academic detailing, formal self-management programs, and a registry.
Feifer and colleagues studied the relationship between Chronic Care Model implementation and clinical outcomes in nine community-based practices.75 They found a strong correlation between Chronic Care Model implementation and performance measures for diabetes and cardiovascular disease. Most recently, two randomized trials have tested interventions that explicitly used the Chronic Care Model to change primary care for asthma76 and diabetes.77 The Chronic Care Model-based intervention significantly improved asthma quality of life, and the diabetes intervention significantly improved glycemic and lipid control compared to usual care.
Learning from Experience: The Case for a Toolkit
Our experience and a growing body of evidence suggests that implementation of the Chronic Care Model needs to be part of a explicit program of quality improvement, supported by leadership and designed to facilitate learning between practices. In the past, the structure of these improvement efforts frequently has taken the form of Breakthrough Series Collaboratives, which bring together dozens of teams to learn from each other at periodic learning sessions. They then return to their systems to test incremental improvements using Plan-Do-Study-Act cycles. One of the most massive national collaborative efforts was the landmark Health Disparities Collaboratives (HDC) program sponsored by the Health Resources and Services Administration (HRSA) beginning in 1998. In concert with the Institute for Healthcare Improvement, MacColl/ICIC conducted chronic care Breakthrough Series Collaboratives in diabetes, congestive heart failure, asthma, and depression that were attended by pilot community health centers (CHCs) selected by HRSA.
These early collaboratives demonstrated that the Chronic Care Model was a feasible and useful guide to practice redesign and led to measurable improvements in the quality of care.78 In addition, this experience led to two major observations. First, the collaborative structure, although effective as a learning tool, was expensive in terms of staff time and meeting costs. A search began for other, less lengthy and burdensome improvement methods that were still effective. The idea of a manual, or toolkit, first arose in the context of this work.
Second, many of the changes inspired by HDC participation did not sufficiently consider the efficiency and financial health of the participating CHCs. Many changes were made in ways that clearly were not going to be sustainable. To participate in the HDC, CHCs subsidized staff involvement in the collaborative and made new investments in information technology and staff. In addition, planned care often resulted in longer visits, more extensive counseling, or group visits for which reimbursement was often difficult to obtain. Huang and Chin evaluated an early diabetes HDC collaborative and found that implementation of the Chronic Care Model cost the CHC an additional $6.41 to $23.93 per patient.79 This represents a significant portion of a CHC budget. These observations convinced the leaders of the HDC that "a primary care practice is at risk if they simply add the planned care work to their existing systems without stepping back and reengineering their organization."
The pressing need for change, the early evidence of the promise and the limitations of collaboratives, and the requirement to consider both clinical and financial changes all led to the recognition of the need for a coherent set of tools that practices could use. This toolkit is an attempt to fill that need.