- Email: [email protected]
Improved Clinical Outcomes for Fee-for-Service Physician Practices Participating in a Diabetes Care Collaborative
Joint Commission
Journal on Quality and Safety
Performance Improvement
Improved Clinical Outcomes for Fee-for-Service Physician Practices Participating in a Diabetes Care Collaborative
Rob Benedetti, M.D. Barb Flock, R.H.I.T., C.P.H.Q. Steve Pedersen Melissa Ahern, M.B.A., Ph.D.
he growing prevalence of diabetes mellitus in the United States indicates a serious problem for both the health and health care costs of U.S. citizens. Diabetes is a chronic disease that affected nearly 17 million Americans in 2002, although only 11 million patients had been diagnosed. Diabetes is more prevalent among seniors than among others, with 7 million cases in this age cohort, comprising 20.1% of all people in this age group. Complications related to diabetes include heart disease, stroke, high blood pressure, blindness, kidney disease, nervous system disease, amputations, and dental disease. The total cost of diagnosed cases of diabetes in the United States is estimated to be $98 billion annually.1 The importance of cost-effective care for diabetes patients cannot be overstated. Unfortunately, recent studies show that patients with diabetes in the United States are likely to have suboptimal outcomes, including poor glycemic control, uncontrolled hypertension, and inadequate control of hyperlipidemia.2 These conditions cause greater long-run health care costs. Perhaps the most important barrier to care for patients with diabetes is the acute care orientation of health care organizations.3,4 Although some interventions addressing care for patients with chronic diseases have been effective in intensively managed settings, translating the characteristics of a good chronic care plan into the daily operations of a health care practice is difficult. One important initiative uses an approach that combines use of the Chronic Care Model5–7 and the
T
April 2004
Article-at-a-Glance Background: The Rockwood Clinic in Spokane, Washington, participated in the Washington State Diabetes Collaborative, which promoted spread of the Chronic Care Model. Eleven participating providers managed care for 698 patients with diabetes, while 19 nonparticipating providers had 1,300 patients. Implementing the Chronic Care Model: Rockwood upgraded its clinical information system to allow for creation of a patient registry to track clinical measures and generate performance reports. Components included a referral mechanism to facilitate more frequent use of diabetes educators, monthly reports, and sharing of results and updated clinical information from consulting specialists. Rockwood created a self-management tool kit and implemented patient goal setting and group visits. Outcome Measures: Seven of the 12 patient outcomes were significantly better for participating providers (p < .05). Two favorable outcomes, eye examinations and blood pressure < 130/85 mm Hg, were significantly associated with greater participation levels at p < .05. Discussion: Implementing the Chronic Care Model to improve care, using quality improvement staff and administrative support, required fundamental changes in the system of care delivery. These changes were designed to refocus diabetes care efforts at Rockwood on prevention rather than acute care episodes.
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
187
Joint Commission
Journal on Quality and Safety
Breakthrough Series (BTS).8–15 The Chronic Care Model provides a guide to higher-quality chronic illness management within primary care, whereas the BTS is a process-change method for helping organizations plan and implement change. As described elsewhere, two Washington State Diabetes Collaborative (WSDC) efforts co-sponsored by The Washington State Department of Health, the MacColl Institute, and Pro West/Qualis Health, were conducted: WSDC I (October 1999–November 2000), which included 17 clinic teams from across the state, and WSDC II (February 2001–March 2002), which included 30 teams and 6 health plans9—among them, the Rockwood Clinic. This initiative has been used to improve clinical care of diabetes in 26 health care organizations8 and more recently has added other organizations and has expanded its original focus from diabetes to preventive health. The Chronic Care Model, developed in the mid-1990s by the MacColl Institute for Healthcare Innovation, a division of Group Health Cooperative in Seattle, focuses on helping organizations improve in six interrelated components of health care delivery: (1) the health care organization, (2) community resources, (3) selfmanagement support, (4) delivery system design, (5) decision support, and (6) clinical information systems. These components represent distinct areas of care delivery where improvements can lead to development of empowered patients who interact more effectively with prepared proactive practice teams. Ultimately, this model integrates care across the spectrum where care is delivered, by the patient at home, in the health care organization, and within the community.16 The BTS brings together groups of health care organizations that are committed to making changes in their delivery of health care. The process consists of learning sessions that include creation of improvement plans for each organization, action periods in which organizations and providers start making changes, and feedback loops so participants can pose questions. This study documented process and outcome measures for one organization, Rockwood Clinic in Spokane, Washington, which adopted the Chronic Care Model to redesign its system of care for patients with diabetes.
188
April 2004
Rockwood Clinic Rockwood is a private-sector, fee-for-service, multispecialty group practice with nine primary care sites: five family practice, one internal medicine, and three pediatric. Forty-seven primary care providers are integrated with consulting specialists in a ratio of 40 to 60. In 1996 Rockwood’s board of directors initiated specific, focused activities in quality improvement (QI) and funded the position of QI coordinator. Initial activities focused on Health Plan Employer Data and Information Set17 measurements for preventive care services, patient satisfaction, areas of concern that were identified by physicians, and clinical outcomes for selected diseases. Diabetes care was tracked early in these efforts and set the stage for concerted improvement effort opportunities. Rockwood participated in the WSDC I in 1999 as a pilot project to test the effectiveness of the Chronic Care Model. This venture was a success, and Rockwood decided to participate in the WSDC II in 2001 to promote spread of the Chronic Care Model throughout the organization. This article reports results for 11 primary care providers (8 physicians, 2 nurse practitioners, and 1 physician assistant) who attended three separate collaborative learning sessions and have implemented the model in managing the care of 698 patients with diabetes. It compares the results with those of the 19 primary care physicians who did not participate in the WSDC and the 1,300 patients with diabetes under their care. Providers who attended the learning sessions were selected in a nonrandom process on the basis of ease of implementation in their work site and on their personal interest levels. In participating in the BTS, Rockwood chose to pursue the goal of redesigning the clinical practice for the diabetes population by implementing the six components of the Chronic Care Model. Specific related objectives are listed in Table 1 (page 189).
Implementing the Chronic Care Model Clinical Information Systems To implement the Chronic Care Model, Rockwood’s greatest need was an upgrade in the clinical information system to allow for creation of a patient registry to track clinical measures and generate performance reports. Once this registry was available, practice teams and Rockwood’s QI coordinator used the registry to track clinical measures for
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
Joint Commission
Journal on Quality and Safety
Table 1. The Rockwood Clinic’s Specific Related Objectives* Glycemic Control ■ 85% of patients have one A1C test within the past 12 months ■ 75% of patients have A1C < 9.5 (subsequently lowered to 8.0) Dyslipidemia Therapy ■ 80% of patients have LDL testing within the past 12 months ■ 90% of patients have LDL levels < 130 End Organ Surveillance/therapy ■ ■ ■ ■
■
60% of patients have annual urine protein testing 65% of patients have annual retinal exams 70% of patients have annual foot exams 65% of patients age > 40 have documentation of ASA usage 60% of patients have documented selfmanagement goals
Hypertension Control ■ 75% of patients have BP readings < 140/90 mm Hg ■ 50% of patients have BP readings < 130/85 mm Hg * A1C, glycohemoglobin; LDL, low-density lipoprotein; ASA, acetylsalicylic acid; BP, blood pressure.
patients. For example, practice teams identify lapses in patient care, such as patients who have not had their annual eye exams or patients who have missed their required quarterly visits. The QI coordinator creates monthly reports to providers to summarize their patients’ outcomes and annual reports to compare a provider’s patients with other providers’ patients. This process enhances participating providers’ commitment to use the Chronic Care Model to the extent that (1) they see objective measures of improvements in their own patients and (2) their patients have measurably improved outcomes, compared with patients being treated by nonparticipating providers.
Decision Support Rockwood focused on three decision-support components. First, it improved its referral mechanism to
April 2004
facilitate more frequent use of diabetes educators. It identified diabetes educator visits as a separate service on the diabetes flowsheet to prompt necessary referrals. Second, clinical teams (the physician and the physician assistant or nurse) are provided monthly reports (run charts) that track the performance of their registry population. Providers can easily see trends, compare these results to desired goals, and implement care delivery changes as needed. Finally, all participating teams meet quarterly to share results and receive updated clinical information from consulting specialists. In addition to providing staff and primary care physicians with the latest clinical guidelines, these sessions have fostered an environment of open communication and enthusiasm for the collaborative efforts.
Delivery System Design Rockwood focused on three components of delivery system design: (1) planned visits, (2) group visits, and (3) revision of team roles. Planned visits exclusively for diabetes care were scheduled for patients every three months. Care measurements (such as lab tests and consultant visits) were updated before each patient visit and displayed conveniently on one page to help both providers and patients quickly zero in on what improvements needed to be made. The use of group visits at Rockwood is an emerging new model for patient visits in which 10–12 patients are seen in a two-hour group setting. This provided patients the opportunity to develop support networks with each other. This new approach also improved staff efficiency because three members of the care team can see 10–12 patients in a single two-hour group visit instead of seeing patients individually for visits that were taking 30 minutes each. Team roles at Rockwood are gradually changing. For example, one physician–nurse team modified the nurse’s role to include proactive involvement in patients’ care. The nurse was now responsible for independently completing foot checks, scheduling follow-up visits and lab tests, and providing teaching on insulin, diet, and medications. Patients who see the concerted efforts of an entire care delivery team develop stronger personal commitment to improvement in their diabetes care.
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
189
Joint Commission
Journal on Quality and Safety
Self-Management Components of Rockwood’s approach to increasing patient self-management included (1) creating a selfmanagement tool kit, (2) implementing patient goal setting, and (3) implementing group visits. The self-management diabetes tool kit is given to all diabetes patients. It is a folder of information for the patient, with a letter of invitation for the patient to participate in this new model of care, basic information about diabetes, and information on self-management goal-setting strategies. Another self-management tool, the diabetes care record, was well received; it provides patients with information about their current test results and the clinical goals that are being pursued. Space is provided at the bottom of the form for patients’ goals to be set at each quarterly planned visit. Teams encourage patients to set their own goals by using behavior change techniques such as motivational interviewing.18 This process empowers the patient to begin moving toward selfmanagement by setting specific goals. Finally, the group visits provide patients with an opportunity to develop support networks with other patients who are facing similar challenges.
Community The Chronic Care Model’s conceptual framework provided Rockwood a new view of community resources and external partnerships, which included (1) new methods of collaboration with pharmaceutical companies and health plans, (2) increased community outreach through hosting a health fair with exclusive focus on diabetes, and (3) providing training sessions on the diabetes Chronic Care model for providers in the community. Leaders at Rockwood, including the medical director [R.B.] and the board of directors, wanted a unified approach to caring for patients with diabetes and other chronic diseases. Pharmaceutical companies recognized that support for QI efforts would improve care to patients with chronic diseases. As a result, Rockwood was able to negotiate a mutually beneficial agreement with pharmaceutical companies in which the companies donated educational and disease management funds to aid in implementing the Chronic Care Model. The model supports improved disease management by improving adherence to clinical
190
April 2004
practice guidelines and increasing patients’ involvement in their care, which both lead to improved adherence to pharmaceutical treatment guidelines. Rockwood received help from health plans for assistance in implementing and maintaining the Chronic Care Model, including financial assistance for education and operational redesign, co-hosting of community outreach efforts, and informational exchange on clinical outcomes and financial roadblocks. The health plans provided the initial investment in the belief that this type of diabetes disease management would result in lower future health care costs. Starting in 2001, Rockwood began hosting an annual health fair, Diabetes Day, for all residents in the Spokane area. This event promotes awareness of good diabetes practice through workshops, presentations, clinics, and booths. The health fair is staffed by volunteers from the Rockwood staff, partnering health plans, vendors, and other organizations in the community. More than 550 persons with diabetes and family members attended in 2003. Rockwood staff did not initially understand the best way to put on a health fair, so they reviewed the literature and used trial and error to improve the fair from year to year. In an event evaluation survey mailed out in 2002 (52% of surveys were returned by mail; a $1 bill was included in each survey mailed), almost all the respondents (99.6%) reported that they would recommend Diabetes Day to others; 86.8% stated that they would attend again the following year. The survey also showed that promotion of the event was far more effective with flyers handed out in the physician’s office, mailings to patients known to have diabetes, and local newspaper advertisements than with more expensive media. Both the patient’s experience of the sense of community during the health fair and increase in his or her knowledge base may be a profound predictor of the patient’s ability to improve his or her self-management skills. Patients reported that hearing about how to stay healthy at the health fair enhanced their ability to understand what they had been told in their physicians’ offices.
Health Care Organization Rockwood made two fundamental organizational changes to support the Chronic Care Model work: (1) The job descriptions of the medical director and the
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
Joint Commission
Journal on Quality and Safety
QI coordinator were altered to include improvement in the care of patients with chronic diseases and (2) the Rockwood Clinic Foundation redirected its fundamental mission. The board of directors refocused the foundation, a separate nonprofit organization available to the community to finance and support health care research and education, toward efforts to support and promote research in new systems of health care delivery. As a result, Rockwood can raise funds from local hospital foundations, health plans, and pharmaceutical companies to support the WSDC efforts and raise the community’s awareness of Rockwood’s diabetes care efforts.
Analysis of Patient Outcomes for Participating Versus Nonparticipating Providers One of the Chronic Care Model’s main tenets is stepwise implementation through small and rapid plan-dostudy-act cycles. Using this methodology, ideas and interventions are tested on a small scale and, when proven to be effective, spread to a larger scale. Rockwood used this concept to introduce the Chronic Care Model to selected providers in sequential steps. In 1999 two Rockwood providers tested the Chronic Care Model concept in their practices, which now include care for 207 patients. In 2001, three additional Rockwood providers with 234 patients adopted the model, and in 2002, six additional Rockwood providers with 282 patients began using it.
Baseline Performance Baseline performance for WSDC I and WSDC II providers was similar. For example, glycohemoglobin (A1C) < 8.0 was found in 63% and 68% of WSDC I and WSDC II patients, respectively; blood pressure < 130/85 mm hg was found in 22% and 38% of WSDC I and WSDC II patients, respectively; and urine protein testing was found in 51% and 49% of WSDC I and WSDC II patients, respectively. These pre-intervention measurements were not different from those for the nonparticipating providers (A1C < 8.0 in 70%, blood pressure < 130/85 mm Hg in 36%, and urine protein testing in 46%). This sequential implementation created a natural experiment in which diabetes outcomes could be compared between each of these groups of providers as well as between participating and nonparticipating providers.
April 2004
Despite small-sample problems, a preliminary analysis of available data can be used to test two hypotheses. First, providers using the Chronic Care Model will have better clinical outcomes than those not using it. Second, providers who began using the Chronic Care Model early on will have better clinical outcomes than more recent adopters of the model.
Outcome Measures Each provider had a score for each outcome measure representing the percentage of patients who met each outcome criterion (for example, percentage of patients with A1C < 9.5). These scores were compared between participating and nonparticipating providers by using two-tailed t-tests; t-tests for equal or unequal variances were used, depending on an equality of variances test. Analyses also tested whether length of time in the collaborative model affected outcomes. Scores were compared using F-tests and, when the F-test was significant, post hoc means comparisons. Figure 1 (page 192) compares patient outcomes for patients seeing 11 participating providers and for patients seeing 19 nonparticipating providers. All the participating providers had used the Chronic Care Model for at least one year: ■ Two providers had used the model for three years (1999–2002; Participation Level I) ■ Three providers had used the model for two years (2001–2002; Participation Level II) ■ Six providers had used the model for one year (2002; Participation Level III) Seven of the 12 patient outcomes were significantly better for participating providers (p < .05), and 2 outcomes (blood pressure < 140/90 mm Hg and urine protein tests) approached significance (p < .06 and .08, respectively). Compared to nonparticipating providers, participating providers experienced significantly better rates of A1C < 9.5, low-density lipoprotein (LDL) < 130, eye examinations, foot examinations, blood pressure < 130/85 mm Hg, and patients older than 40 years who were taking acetylsalicylic acid. These results show that patients of participating providers have significantly better control than others of many aspects of diabetes care. In particular, important outcomes unrelated to blood sugar control, such as
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
191
Joint Commission
Journal on Quality and Safety
Summary of Participation Versus No Participation Effects on Patient Outcomes
Provider Satisfaction
Rockwood also conducted a provider satisfaction survey for 13 participating providers and 15 nurses, 4 physician assistants, and 22 other staff members participating in the program. Twenty-nine staff responded to the survey. Most respondents were highly satisfied with the program. Specifically, 71% of respondents said that the quality of their patient care had “significantly improved” since participating in the collaborative, and 76% said that their patients’ satisfaction had “significantly improved.” Importantly, whereas only 28% of respondents were satisfied or very satisfied with the quality of their collaborative work before the program, 78% expressed satisfaction since beginning the collaborative program. Ninety-one percent were either satisfied or very satisfied with the quality of the training they received. In addition, all participants said they would recommend participation in the colFigure 1. Patient outcomes are compared for participation versus laborative to a colleague, and none of them prenonparticipation in a diabetes care collaborative. A1C, glycohemoferred to revert to the previous system of globin; LDL, low-density lipoprotein; ASA, acetylsalicylic acid. diabetes care. In addition, persons who had been at Rockwood blood pressure, hyperlipidemia control, and end organ longer than 5 years expressed the greatest satisfaction surveillance, are often overlooked in a busy clinical with the collaborative in terms of its impacts on quality practice with limited time. Blood pressure < 130/85 mm of care, patient satisfaction, and training. Persons particHg was achieved in 49% and 35% of the patients of paripating in the diabetes collaborative for longer than one ticipating providers and nonparticipating providers, year were more likely than others to rate quality of care, respectively (p = .005). Similarly, LDL < 130 was patient satisfaction, and organization of patient care to achieved in 75% and 45% of patients of participating and be significantly improved. Comparisons of physicians nonparticipating providers, respectively (p = .0009). with other staff revealed no significant differences in As shown in Figure 2 (page 193), length of time in the perceptions of quality of care, patient satisfaction, and collaborative model also affected outcomes. Two favororganization of care or satisfaction with training. able outcomes, eye examinations and blood pressure <130/85 mm Hg, were significantly associated with greater participation levels at p < .05, and three addiDiscussion tional outcomes—LDL testing, blood pressure < 140/90 Improvement of the delivery of care to patients with mm Hg, and self-management goals—were marginally chronic diseases has become a national priority, and the associated with participation (p < .07). Furthermore, the Chronic Care Model has been previously shown to trend for 10 of the 12 outcomes showed that outcomes improve the care of patients with chronic diseases in a were most favorable for the highest participation level. managed care setting. Rockwood, a physician-owned Figure 3 (page 193) shows the results for blood pressure multispecialty medical practice, specifically placed control, the single most important intervention in reducimprovements in the care of patients with chronic dising long-term complications from diabetes. eases into its strategic plan.
192
April 2004
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
Joint Commission
Journal on Quality and Safety
Summary of Length of Participation Effects on Outcomes
Interestingly, outcome improvement was shown less for measurements of blood sugar control (A1C testing and A1C levels) and more for care elements unrelated to blood sugar control, including lipid control, blood pressure control, and end organ surveillance. These results suggest that in a busy clinical practice where focus remains primarily on acute complaints, many aspects of care that are vital to preventing longterm complications may be underemphasized or overlooked. The Chronic Care Model creates a dynamic, productive interaction between an informed, activated patient and a prepared, proactive practice team. This interaction empowers patients and providers to expand their focus from blood sugar control to other elements of care that are vital for prevention of long-term morbidity and mortality. Importantly, the interactions between Rockwood Clinic and the Spokane community improved. Similarly, working collaboratively with pharmaceutical companies and Figure 2. Two favorable outcomes, eye examinations and blood pressure < 130/85 mm Hg, were significantly associated with insurance companies created a mutually greater participation levels. HbA1c, glycohemoglobin; LDL, lowsatisfying dialogue that has created further density lipoprotein; ASA, acetylsalicylic acid. win–win opportunities for those involved. Participation in this model appeared to be meaningUsing existing QI staff and administrative support, ful and satisfying to the physicians and staff. Although Rockwood implemented the Chronic Care Model to staff must spend time and effort learning to use the improve the care of patients with diabetes mellitus. However, implementing this model required fundamental changes in the system of care delivery. These changes required significant startup costs, including major enhancements of information technology infrastructure, disruptive alterations in job tasks and patient care flow, time-consuming community outreach, and increased time devoted to patient education and self-management. Further, these changes required new financing strategies, as well as new management and staff training. To accelerate this transformation, Rockwood created unique community partnerships and revitalized and enhanced existing partnerships. All these efforts were designed to refocus diabetes care efforts at Rockwood on prevention rather than on acute care episodes. Results of this study indicate that processes of care and Figure 3. Length of provider participation in the diaclinical outcomes significantly improved at Rockwood for betes care collaborative was associated with a stepwise, incremental improvement in blood pressure (BP). those providers involved in the Chronic Care Model.
Blood Pressure Outcomes
April 2004
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
193
Joint Commission
Journal on Quality and Safety
Chronic Care Model, the improvements in patient care demonstrated through careful data collection and analysis appeal to the goal-oriented focus of many physicians and engender new enthusiasm. Disseminating the results of the study internally allowed Rockwood administration to promote systemwide change, motivating new physicians and other staff to become participants. J
Rob Benedetti, M.D., is Medical Director and Barb Flock, R.H.I.T., C.P.H.Q., is Director of Quality Improvement, Rockwood Clinic, Spokane, Washington. Steve Pedersen is Research Assistant and Graduate Student and Melissa Ahern, M.B.A., Ph.D., is Associate Professor, Department of Health Policy and Administration, Washington State University, Spokane. Please address reprint requests to Rob Benedetti, M.D., [email protected].
References 1. Centers for Disease Control and Prevention: National Diabetes Fact Sheet: National Estimates on Diabetes. http://www.cdc.gov/ diabetes/pubs/estimates.htm#prev (last accessed Jan. 26, 2004). 2. Harris M.I.: Health care and health status and outcomes for patients with type 2 diabetes. Diabetes Care 23:754–758, Jun. 2000. 3. Etzwiler D.D.: Chronic care: A need in search of a system. Diabetes Education 23:569–573, Sep.–Oct. 1997. 4. Glasgow R.E., et al.: If diabetes is a public health problem, why not treat it as one? A population-based approach to chronic illness. Ann Behav Med 21:159–170, spring 1999. 5. Wagner E.H., et al.: Improving outcomes in chronic illness. Manag Care Q 4:12–25, spring 1996. 6. Wagner E.H., et al.: Organizing care for patients with chronic illness. Milbank Q 74(4):511–544, 1996. 7. Wagner E.H., et al.: A survey of leading chronic disease management programs: Are they consistent with the literature? Manag Care Q 7:56–66, summer 1999. 8. Wagner E.H., et al.: Quality improvement in chronic illness care: A collaborative approach. Jt Comm J Qual Improv 27:63–80, Feb. 2001. 9. Daniel D.M., et al.: A state-level application of the Chronic Illness Breakthrough Series: Results from two collaboratives on diabetes in Washington State. Jt Comm J Qual Saf 30:69–79, Feb. 2004. 10. Berwick D.M.: Continuous improvement as an ideal in health care. N Engl J Med 320:53–56, Jan. 5, 1989.
194
April 2004
11. Berwick D.M.: Developing and testing change in delivery of care. Ann Intern Med 128:651–656, Apr. 15, 1998. 12. Kilo C.M.: A framework for collaborative improvement: Lessons from the Institute for Healthcare Improvement’s Breakthrough Series. Qual Manag Health Care 6:1–13, Sep. 1998. 13. Leape L.L., et al.: Reducing adverse drug events: Lessons from a breakthrough series collaborative. Jt Comm J Qual Improv 26:321–331, Jun. 2000. 14. Measel S., Sershon L., White D.: Reducing adverse drug events and medication errors using rapid cycle improvement. Qual Manag Health Care 6:15–28, Sep. 1998. 15. Flamm B.L., Berwick D.M., Kabcenell A.: Reducing cesarean section rates safely: Lessons from a “breakthrough series” collaborative. Birth 25:117–124, Jun. 1998. 16. Wagner E.H.: Chronic disease management: what will it take to improve care for chronic illness? Eff Clin Pract 1:2–4, Aug.–Sep, 1998. 17. National Committee for Quality Assurance (NCQA): Health Plan Employer Data and Information Set (HEDIS)®. Washington, DC: NCQA, 1999. 18. Baer J.S., et al.: An evaluation of workshop training in motivational interviewing for addiction and mental health clinicians. Drug Alcohol Depend 73:99–106, Jan. 7, 2004.
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
Journal on Quality and Safety
Performance Improvement
Improved Clinical Outcomes for Fee-for-Service Physician Practices Participating in a Diabetes Care Collaborative
Rob Benedetti, M.D. Barb Flock, R.H.I.T., C.P.H.Q. Steve Pedersen Melissa Ahern, M.B.A., Ph.D.
he growing prevalence of diabetes mellitus in the United States indicates a serious problem for both the health and health care costs of U.S. citizens. Diabetes is a chronic disease that affected nearly 17 million Americans in 2002, although only 11 million patients had been diagnosed. Diabetes is more prevalent among seniors than among others, with 7 million cases in this age cohort, comprising 20.1% of all people in this age group. Complications related to diabetes include heart disease, stroke, high blood pressure, blindness, kidney disease, nervous system disease, amputations, and dental disease. The total cost of diagnosed cases of diabetes in the United States is estimated to be $98 billion annually.1 The importance of cost-effective care for diabetes patients cannot be overstated. Unfortunately, recent studies show that patients with diabetes in the United States are likely to have suboptimal outcomes, including poor glycemic control, uncontrolled hypertension, and inadequate control of hyperlipidemia.2 These conditions cause greater long-run health care costs. Perhaps the most important barrier to care for patients with diabetes is the acute care orientation of health care organizations.3,4 Although some interventions addressing care for patients with chronic diseases have been effective in intensively managed settings, translating the characteristics of a good chronic care plan into the daily operations of a health care practice is difficult. One important initiative uses an approach that combines use of the Chronic Care Model5–7 and the
T
April 2004
Article-at-a-Glance Background: The Rockwood Clinic in Spokane, Washington, participated in the Washington State Diabetes Collaborative, which promoted spread of the Chronic Care Model. Eleven participating providers managed care for 698 patients with diabetes, while 19 nonparticipating providers had 1,300 patients. Implementing the Chronic Care Model: Rockwood upgraded its clinical information system to allow for creation of a patient registry to track clinical measures and generate performance reports. Components included a referral mechanism to facilitate more frequent use of diabetes educators, monthly reports, and sharing of results and updated clinical information from consulting specialists. Rockwood created a self-management tool kit and implemented patient goal setting and group visits. Outcome Measures: Seven of the 12 patient outcomes were significantly better for participating providers (p < .05). Two favorable outcomes, eye examinations and blood pressure < 130/85 mm Hg, were significantly associated with greater participation levels at p < .05. Discussion: Implementing the Chronic Care Model to improve care, using quality improvement staff and administrative support, required fundamental changes in the system of care delivery. These changes were designed to refocus diabetes care efforts at Rockwood on prevention rather than acute care episodes.
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
187
Joint Commission
Journal on Quality and Safety
Breakthrough Series (BTS).8–15 The Chronic Care Model provides a guide to higher-quality chronic illness management within primary care, whereas the BTS is a process-change method for helping organizations plan and implement change. As described elsewhere, two Washington State Diabetes Collaborative (WSDC) efforts co-sponsored by The Washington State Department of Health, the MacColl Institute, and Pro West/Qualis Health, were conducted: WSDC I (October 1999–November 2000), which included 17 clinic teams from across the state, and WSDC II (February 2001–March 2002), which included 30 teams and 6 health plans9—among them, the Rockwood Clinic. This initiative has been used to improve clinical care of diabetes in 26 health care organizations8 and more recently has added other organizations and has expanded its original focus from diabetes to preventive health. The Chronic Care Model, developed in the mid-1990s by the MacColl Institute for Healthcare Innovation, a division of Group Health Cooperative in Seattle, focuses on helping organizations improve in six interrelated components of health care delivery: (1) the health care organization, (2) community resources, (3) selfmanagement support, (4) delivery system design, (5) decision support, and (6) clinical information systems. These components represent distinct areas of care delivery where improvements can lead to development of empowered patients who interact more effectively with prepared proactive practice teams. Ultimately, this model integrates care across the spectrum where care is delivered, by the patient at home, in the health care organization, and within the community.16 The BTS brings together groups of health care organizations that are committed to making changes in their delivery of health care. The process consists of learning sessions that include creation of improvement plans for each organization, action periods in which organizations and providers start making changes, and feedback loops so participants can pose questions. This study documented process and outcome measures for one organization, Rockwood Clinic in Spokane, Washington, which adopted the Chronic Care Model to redesign its system of care for patients with diabetes.
188
April 2004
Rockwood Clinic Rockwood is a private-sector, fee-for-service, multispecialty group practice with nine primary care sites: five family practice, one internal medicine, and three pediatric. Forty-seven primary care providers are integrated with consulting specialists in a ratio of 40 to 60. In 1996 Rockwood’s board of directors initiated specific, focused activities in quality improvement (QI) and funded the position of QI coordinator. Initial activities focused on Health Plan Employer Data and Information Set17 measurements for preventive care services, patient satisfaction, areas of concern that were identified by physicians, and clinical outcomes for selected diseases. Diabetes care was tracked early in these efforts and set the stage for concerted improvement effort opportunities. Rockwood participated in the WSDC I in 1999 as a pilot project to test the effectiveness of the Chronic Care Model. This venture was a success, and Rockwood decided to participate in the WSDC II in 2001 to promote spread of the Chronic Care Model throughout the organization. This article reports results for 11 primary care providers (8 physicians, 2 nurse practitioners, and 1 physician assistant) who attended three separate collaborative learning sessions and have implemented the model in managing the care of 698 patients with diabetes. It compares the results with those of the 19 primary care physicians who did not participate in the WSDC and the 1,300 patients with diabetes under their care. Providers who attended the learning sessions were selected in a nonrandom process on the basis of ease of implementation in their work site and on their personal interest levels. In participating in the BTS, Rockwood chose to pursue the goal of redesigning the clinical practice for the diabetes population by implementing the six components of the Chronic Care Model. Specific related objectives are listed in Table 1 (page 189).
Implementing the Chronic Care Model Clinical Information Systems To implement the Chronic Care Model, Rockwood’s greatest need was an upgrade in the clinical information system to allow for creation of a patient registry to track clinical measures and generate performance reports. Once this registry was available, practice teams and Rockwood’s QI coordinator used the registry to track clinical measures for
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
Joint Commission
Journal on Quality and Safety
Table 1. The Rockwood Clinic’s Specific Related Objectives* Glycemic Control ■ 85% of patients have one A1C test within the past 12 months ■ 75% of patients have A1C < 9.5 (subsequently lowered to 8.0) Dyslipidemia Therapy ■ 80% of patients have LDL testing within the past 12 months ■ 90% of patients have LDL levels < 130 End Organ Surveillance/therapy ■ ■ ■ ■
■
60% of patients have annual urine protein testing 65% of patients have annual retinal exams 70% of patients have annual foot exams 65% of patients age > 40 have documentation of ASA usage 60% of patients have documented selfmanagement goals
Hypertension Control ■ 75% of patients have BP readings < 140/90 mm Hg ■ 50% of patients have BP readings < 130/85 mm Hg * A1C, glycohemoglobin; LDL, low-density lipoprotein; ASA, acetylsalicylic acid; BP, blood pressure.
patients. For example, practice teams identify lapses in patient care, such as patients who have not had their annual eye exams or patients who have missed their required quarterly visits. The QI coordinator creates monthly reports to providers to summarize their patients’ outcomes and annual reports to compare a provider’s patients with other providers’ patients. This process enhances participating providers’ commitment to use the Chronic Care Model to the extent that (1) they see objective measures of improvements in their own patients and (2) their patients have measurably improved outcomes, compared with patients being treated by nonparticipating providers.
Decision Support Rockwood focused on three decision-support components. First, it improved its referral mechanism to
April 2004
facilitate more frequent use of diabetes educators. It identified diabetes educator visits as a separate service on the diabetes flowsheet to prompt necessary referrals. Second, clinical teams (the physician and the physician assistant or nurse) are provided monthly reports (run charts) that track the performance of their registry population. Providers can easily see trends, compare these results to desired goals, and implement care delivery changes as needed. Finally, all participating teams meet quarterly to share results and receive updated clinical information from consulting specialists. In addition to providing staff and primary care physicians with the latest clinical guidelines, these sessions have fostered an environment of open communication and enthusiasm for the collaborative efforts.
Delivery System Design Rockwood focused on three components of delivery system design: (1) planned visits, (2) group visits, and (3) revision of team roles. Planned visits exclusively for diabetes care were scheduled for patients every three months. Care measurements (such as lab tests and consultant visits) were updated before each patient visit and displayed conveniently on one page to help both providers and patients quickly zero in on what improvements needed to be made. The use of group visits at Rockwood is an emerging new model for patient visits in which 10–12 patients are seen in a two-hour group setting. This provided patients the opportunity to develop support networks with each other. This new approach also improved staff efficiency because three members of the care team can see 10–12 patients in a single two-hour group visit instead of seeing patients individually for visits that were taking 30 minutes each. Team roles at Rockwood are gradually changing. For example, one physician–nurse team modified the nurse’s role to include proactive involvement in patients’ care. The nurse was now responsible for independently completing foot checks, scheduling follow-up visits and lab tests, and providing teaching on insulin, diet, and medications. Patients who see the concerted efforts of an entire care delivery team develop stronger personal commitment to improvement in their diabetes care.
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
189
Joint Commission
Journal on Quality and Safety
Self-Management Components of Rockwood’s approach to increasing patient self-management included (1) creating a selfmanagement tool kit, (2) implementing patient goal setting, and (3) implementing group visits. The self-management diabetes tool kit is given to all diabetes patients. It is a folder of information for the patient, with a letter of invitation for the patient to participate in this new model of care, basic information about diabetes, and information on self-management goal-setting strategies. Another self-management tool, the diabetes care record, was well received; it provides patients with information about their current test results and the clinical goals that are being pursued. Space is provided at the bottom of the form for patients’ goals to be set at each quarterly planned visit. Teams encourage patients to set their own goals by using behavior change techniques such as motivational interviewing.18 This process empowers the patient to begin moving toward selfmanagement by setting specific goals. Finally, the group visits provide patients with an opportunity to develop support networks with other patients who are facing similar challenges.
Community The Chronic Care Model’s conceptual framework provided Rockwood a new view of community resources and external partnerships, which included (1) new methods of collaboration with pharmaceutical companies and health plans, (2) increased community outreach through hosting a health fair with exclusive focus on diabetes, and (3) providing training sessions on the diabetes Chronic Care model for providers in the community. Leaders at Rockwood, including the medical director [R.B.] and the board of directors, wanted a unified approach to caring for patients with diabetes and other chronic diseases. Pharmaceutical companies recognized that support for QI efforts would improve care to patients with chronic diseases. As a result, Rockwood was able to negotiate a mutually beneficial agreement with pharmaceutical companies in which the companies donated educational and disease management funds to aid in implementing the Chronic Care Model. The model supports improved disease management by improving adherence to clinical
190
April 2004
practice guidelines and increasing patients’ involvement in their care, which both lead to improved adherence to pharmaceutical treatment guidelines. Rockwood received help from health plans for assistance in implementing and maintaining the Chronic Care Model, including financial assistance for education and operational redesign, co-hosting of community outreach efforts, and informational exchange on clinical outcomes and financial roadblocks. The health plans provided the initial investment in the belief that this type of diabetes disease management would result in lower future health care costs. Starting in 2001, Rockwood began hosting an annual health fair, Diabetes Day, for all residents in the Spokane area. This event promotes awareness of good diabetes practice through workshops, presentations, clinics, and booths. The health fair is staffed by volunteers from the Rockwood staff, partnering health plans, vendors, and other organizations in the community. More than 550 persons with diabetes and family members attended in 2003. Rockwood staff did not initially understand the best way to put on a health fair, so they reviewed the literature and used trial and error to improve the fair from year to year. In an event evaluation survey mailed out in 2002 (52% of surveys were returned by mail; a $1 bill was included in each survey mailed), almost all the respondents (99.6%) reported that they would recommend Diabetes Day to others; 86.8% stated that they would attend again the following year. The survey also showed that promotion of the event was far more effective with flyers handed out in the physician’s office, mailings to patients known to have diabetes, and local newspaper advertisements than with more expensive media. Both the patient’s experience of the sense of community during the health fair and increase in his or her knowledge base may be a profound predictor of the patient’s ability to improve his or her self-management skills. Patients reported that hearing about how to stay healthy at the health fair enhanced their ability to understand what they had been told in their physicians’ offices.
Health Care Organization Rockwood made two fundamental organizational changes to support the Chronic Care Model work: (1) The job descriptions of the medical director and the
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
Joint Commission
Journal on Quality and Safety
QI coordinator were altered to include improvement in the care of patients with chronic diseases and (2) the Rockwood Clinic Foundation redirected its fundamental mission. The board of directors refocused the foundation, a separate nonprofit organization available to the community to finance and support health care research and education, toward efforts to support and promote research in new systems of health care delivery. As a result, Rockwood can raise funds from local hospital foundations, health plans, and pharmaceutical companies to support the WSDC efforts and raise the community’s awareness of Rockwood’s diabetes care efforts.
Analysis of Patient Outcomes for Participating Versus Nonparticipating Providers One of the Chronic Care Model’s main tenets is stepwise implementation through small and rapid plan-dostudy-act cycles. Using this methodology, ideas and interventions are tested on a small scale and, when proven to be effective, spread to a larger scale. Rockwood used this concept to introduce the Chronic Care Model to selected providers in sequential steps. In 1999 two Rockwood providers tested the Chronic Care Model concept in their practices, which now include care for 207 patients. In 2001, three additional Rockwood providers with 234 patients adopted the model, and in 2002, six additional Rockwood providers with 282 patients began using it.
Baseline Performance Baseline performance for WSDC I and WSDC II providers was similar. For example, glycohemoglobin (A1C) < 8.0 was found in 63% and 68% of WSDC I and WSDC II patients, respectively; blood pressure < 130/85 mm hg was found in 22% and 38% of WSDC I and WSDC II patients, respectively; and urine protein testing was found in 51% and 49% of WSDC I and WSDC II patients, respectively. These pre-intervention measurements were not different from those for the nonparticipating providers (A1C < 8.0 in 70%, blood pressure < 130/85 mm Hg in 36%, and urine protein testing in 46%). This sequential implementation created a natural experiment in which diabetes outcomes could be compared between each of these groups of providers as well as between participating and nonparticipating providers.
April 2004
Despite small-sample problems, a preliminary analysis of available data can be used to test two hypotheses. First, providers using the Chronic Care Model will have better clinical outcomes than those not using it. Second, providers who began using the Chronic Care Model early on will have better clinical outcomes than more recent adopters of the model.
Outcome Measures Each provider had a score for each outcome measure representing the percentage of patients who met each outcome criterion (for example, percentage of patients with A1C < 9.5). These scores were compared between participating and nonparticipating providers by using two-tailed t-tests; t-tests for equal or unequal variances were used, depending on an equality of variances test. Analyses also tested whether length of time in the collaborative model affected outcomes. Scores were compared using F-tests and, when the F-test was significant, post hoc means comparisons. Figure 1 (page 192) compares patient outcomes for patients seeing 11 participating providers and for patients seeing 19 nonparticipating providers. All the participating providers had used the Chronic Care Model for at least one year: ■ Two providers had used the model for three years (1999–2002; Participation Level I) ■ Three providers had used the model for two years (2001–2002; Participation Level II) ■ Six providers had used the model for one year (2002; Participation Level III) Seven of the 12 patient outcomes were significantly better for participating providers (p < .05), and 2 outcomes (blood pressure < 140/90 mm Hg and urine protein tests) approached significance (p < .06 and .08, respectively). Compared to nonparticipating providers, participating providers experienced significantly better rates of A1C < 9.5, low-density lipoprotein (LDL) < 130, eye examinations, foot examinations, blood pressure < 130/85 mm Hg, and patients older than 40 years who were taking acetylsalicylic acid. These results show that patients of participating providers have significantly better control than others of many aspects of diabetes care. In particular, important outcomes unrelated to blood sugar control, such as
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
191
Joint Commission
Journal on Quality and Safety
Summary of Participation Versus No Participation Effects on Patient Outcomes
Provider Satisfaction
Rockwood also conducted a provider satisfaction survey for 13 participating providers and 15 nurses, 4 physician assistants, and 22 other staff members participating in the program. Twenty-nine staff responded to the survey. Most respondents were highly satisfied with the program. Specifically, 71% of respondents said that the quality of their patient care had “significantly improved” since participating in the collaborative, and 76% said that their patients’ satisfaction had “significantly improved.” Importantly, whereas only 28% of respondents were satisfied or very satisfied with the quality of their collaborative work before the program, 78% expressed satisfaction since beginning the collaborative program. Ninety-one percent were either satisfied or very satisfied with the quality of the training they received. In addition, all participants said they would recommend participation in the colFigure 1. Patient outcomes are compared for participation versus laborative to a colleague, and none of them prenonparticipation in a diabetes care collaborative. A1C, glycohemoferred to revert to the previous system of globin; LDL, low-density lipoprotein; ASA, acetylsalicylic acid. diabetes care. In addition, persons who had been at Rockwood blood pressure, hyperlipidemia control, and end organ longer than 5 years expressed the greatest satisfaction surveillance, are often overlooked in a busy clinical with the collaborative in terms of its impacts on quality practice with limited time. Blood pressure < 130/85 mm of care, patient satisfaction, and training. Persons particHg was achieved in 49% and 35% of the patients of paripating in the diabetes collaborative for longer than one ticipating providers and nonparticipating providers, year were more likely than others to rate quality of care, respectively (p = .005). Similarly, LDL < 130 was patient satisfaction, and organization of patient care to achieved in 75% and 45% of patients of participating and be significantly improved. Comparisons of physicians nonparticipating providers, respectively (p = .0009). with other staff revealed no significant differences in As shown in Figure 2 (page 193), length of time in the perceptions of quality of care, patient satisfaction, and collaborative model also affected outcomes. Two favororganization of care or satisfaction with training. able outcomes, eye examinations and blood pressure <130/85 mm Hg, were significantly associated with greater participation levels at p < .05, and three addiDiscussion tional outcomes—LDL testing, blood pressure < 140/90 Improvement of the delivery of care to patients with mm Hg, and self-management goals—were marginally chronic diseases has become a national priority, and the associated with participation (p < .07). Furthermore, the Chronic Care Model has been previously shown to trend for 10 of the 12 outcomes showed that outcomes improve the care of patients with chronic diseases in a were most favorable for the highest participation level. managed care setting. Rockwood, a physician-owned Figure 3 (page 193) shows the results for blood pressure multispecialty medical practice, specifically placed control, the single most important intervention in reducimprovements in the care of patients with chronic dising long-term complications from diabetes. eases into its strategic plan.
192
April 2004
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
Joint Commission
Journal on Quality and Safety
Summary of Length of Participation Effects on Outcomes
Interestingly, outcome improvement was shown less for measurements of blood sugar control (A1C testing and A1C levels) and more for care elements unrelated to blood sugar control, including lipid control, blood pressure control, and end organ surveillance. These results suggest that in a busy clinical practice where focus remains primarily on acute complaints, many aspects of care that are vital to preventing longterm complications may be underemphasized or overlooked. The Chronic Care Model creates a dynamic, productive interaction between an informed, activated patient and a prepared, proactive practice team. This interaction empowers patients and providers to expand their focus from blood sugar control to other elements of care that are vital for prevention of long-term morbidity and mortality. Importantly, the interactions between Rockwood Clinic and the Spokane community improved. Similarly, working collaboratively with pharmaceutical companies and Figure 2. Two favorable outcomes, eye examinations and blood pressure < 130/85 mm Hg, were significantly associated with insurance companies created a mutually greater participation levels. HbA1c, glycohemoglobin; LDL, lowsatisfying dialogue that has created further density lipoprotein; ASA, acetylsalicylic acid. win–win opportunities for those involved. Participation in this model appeared to be meaningUsing existing QI staff and administrative support, ful and satisfying to the physicians and staff. Although Rockwood implemented the Chronic Care Model to staff must spend time and effort learning to use the improve the care of patients with diabetes mellitus. However, implementing this model required fundamental changes in the system of care delivery. These changes required significant startup costs, including major enhancements of information technology infrastructure, disruptive alterations in job tasks and patient care flow, time-consuming community outreach, and increased time devoted to patient education and self-management. Further, these changes required new financing strategies, as well as new management and staff training. To accelerate this transformation, Rockwood created unique community partnerships and revitalized and enhanced existing partnerships. All these efforts were designed to refocus diabetes care efforts at Rockwood on prevention rather than on acute care episodes. Results of this study indicate that processes of care and Figure 3. Length of provider participation in the diaclinical outcomes significantly improved at Rockwood for betes care collaborative was associated with a stepwise, incremental improvement in blood pressure (BP). those providers involved in the Chronic Care Model.
Blood Pressure Outcomes
April 2004
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations
193
Joint Commission
Journal on Quality and Safety
Chronic Care Model, the improvements in patient care demonstrated through careful data collection and analysis appeal to the goal-oriented focus of many physicians and engender new enthusiasm. Disseminating the results of the study internally allowed Rockwood administration to promote systemwide change, motivating new physicians and other staff to become participants. J
Rob Benedetti, M.D., is Medical Director and Barb Flock, R.H.I.T., C.P.H.Q., is Director of Quality Improvement, Rockwood Clinic, Spokane, Washington. Steve Pedersen is Research Assistant and Graduate Student and Melissa Ahern, M.B.A., Ph.D., is Associate Professor, Department of Health Policy and Administration, Washington State University, Spokane. Please address reprint requests to Rob Benedetti, M.D., [email protected].
References 1. Centers for Disease Control and Prevention: National Diabetes Fact Sheet: National Estimates on Diabetes. http://www.cdc.gov/ diabetes/pubs/estimates.htm#prev (last accessed Jan. 26, 2004). 2. Harris M.I.: Health care and health status and outcomes for patients with type 2 diabetes. Diabetes Care 23:754–758, Jun. 2000. 3. Etzwiler D.D.: Chronic care: A need in search of a system. Diabetes Education 23:569–573, Sep.–Oct. 1997. 4. Glasgow R.E., et al.: If diabetes is a public health problem, why not treat it as one? A population-based approach to chronic illness. Ann Behav Med 21:159–170, spring 1999. 5. Wagner E.H., et al.: Improving outcomes in chronic illness. Manag Care Q 4:12–25, spring 1996. 6. Wagner E.H., et al.: Organizing care for patients with chronic illness. Milbank Q 74(4):511–544, 1996. 7. Wagner E.H., et al.: A survey of leading chronic disease management programs: Are they consistent with the literature? Manag Care Q 7:56–66, summer 1999. 8. Wagner E.H., et al.: Quality improvement in chronic illness care: A collaborative approach. Jt Comm J Qual Improv 27:63–80, Feb. 2001. 9. Daniel D.M., et al.: A state-level application of the Chronic Illness Breakthrough Series: Results from two collaboratives on diabetes in Washington State. Jt Comm J Qual Saf 30:69–79, Feb. 2004. 10. Berwick D.M.: Continuous improvement as an ideal in health care. N Engl J Med 320:53–56, Jan. 5, 1989.
194
April 2004
11. Berwick D.M.: Developing and testing change in delivery of care. Ann Intern Med 128:651–656, Apr. 15, 1998. 12. Kilo C.M.: A framework for collaborative improvement: Lessons from the Institute for Healthcare Improvement’s Breakthrough Series. Qual Manag Health Care 6:1–13, Sep. 1998. 13. Leape L.L., et al.: Reducing adverse drug events: Lessons from a breakthrough series collaborative. Jt Comm J Qual Improv 26:321–331, Jun. 2000. 14. Measel S., Sershon L., White D.: Reducing adverse drug events and medication errors using rapid cycle improvement. Qual Manag Health Care 6:15–28, Sep. 1998. 15. Flamm B.L., Berwick D.M., Kabcenell A.: Reducing cesarean section rates safely: Lessons from a “breakthrough series” collaborative. Birth 25:117–124, Jun. 1998. 16. Wagner E.H.: Chronic disease management: what will it take to improve care for chronic illness? Eff Clin Pract 1:2–4, Aug.–Sep, 1998. 17. National Committee for Quality Assurance (NCQA): Health Plan Employer Data and Information Set (HEDIS)®. Washington, DC: NCQA, 1999. 18. Baer J.S., et al.: An evaluation of workshop training in motivational interviewing for addiction and mental health clinicians. Drug Alcohol Depend 73:99–106, Jan. 7, 2004.
Volume 30 Number 4
Copyright 2004 Joint Commission on Accreditation of Healthcare Organizations