Diabetes is one of the most common chronic illnesses in the United States, affecting approximately 7% of the population or 21 million people.1,2 Disease incidence is increasing each year; estimates indicate the number of individuals diagnosed with diabetes will increase 44% by the year 2020.3 Inpatient and emergency services already consume 62% of the health care expenditure attributable to diabetes in the United States.3 Diabetes is one of the most frequently managed chronic diseases in primary care practice.4 However, primary care clinics see a variety of patients with many different problems and are not specifically organized to deliver care for chronic illnesses such as diabetes.5–10
In recent years, the concept of improving diabetes care through principles of the Improving Chronic Illness Care (ICIC) Model, delivered through team management, has led to improved measures of diabetes care11–14 and has received endorsements from the American Diabetes Association and the American Association of Diabetes Educators.15 The ICIC Model guides primary care practices in developing higher-quality care for patients with chronic illnesses—patients whose acute symptoms typically dominate their visits with providers, leaving little time to devote to chronic illness management.16,17 The model identifies six essential elements: community resources and policies, health care organization, self-management support, delivery system design, decision support, and clinical information systems.16 A key component of this model is building interprofessional primary care teams to provide acute, chronic, and preventive care using evidence-based guidelines.18 The complexities of diabetes management and the multiple chronic complications associated with diabetes may be best managed by an interprofessional team of providers trained to deliver high-quality, consistent care.
Intervention studies have demonstrated that chronic illness management is more successful when care is provided by teams that include nurse case managers and clinical pharmacists.18 However, many practices fail to provide the requisite training in chronic illness management and interprofessional team building necessary to realize such benefits.18 In a systematic review of studies of diabetes care programs featuring elements of the ICIC Model, 32 of 39 studies found that interventions based on the ICIC Model improved at least one process or outcome measure for patients.19 A later randomized clinical trial, implementing the ICIC Model throughout linked primary care clinics in an underserved urban community, demonstrated significant decreases in patients' glycosolated hemoglobin (HbA1c) and non-HDL cholesterol and a significant increase in self-monitoring of blood glucose in the intervention groups compared with groups that received provider education only or usual care.11
Others have tested implementation of the ICIC Model but were unable to show that the improvements in quality measures were correlated with specific ICIC Model elements within a large medical group, though certain clinical parameters in the patient population improved.13 In a related study, a qualitative approach was used to study the process of implementation of the ICIC Model and identified many barriers to implementation within the care system, including competing priorities, lack of specificity and agreement about the desired process changes, and lack of engagement of physicians.20 Prior published reports of implementing chronic illness training in medicine residency programs have identified numerous barriers and challenges,21,22 but none reported the impact on the patient populations of implementing this training. Nuovo et al21 implemented the chronic care curriculum for diabetes care using a multidisciplinary team that included family practice residents, physician and behavioral science faculty, a certified diabetes educator, the practice manager, and the clinic director of a family practice center. Key barriers to implementation were the lack of a patient registry, inadequate buy-in from faculty, residents, staff, and health system leaders, and the need to maintain clinical productivity and bill for services. The medicine residents' confidence improved with regard to managing diabetes and motivating patients in self-management activities, but no patient outcomes were reported.
The purpose of our study was to implement an interprofessional model of chronic illness care, delivered by teams of learners who were trained in the ICIC Model, and to describe the outcomes for a sample of patients with type 2 diabetes, as well as the outcomes for interprofessional learners.
This study was designed as a nonrandomized, parallel-group clinical trial of 384 adult patients with type 2 diabetes who were receiving care in two general internal medicine clinics within the University of California, San Francisco (UCSF) ambulatory care clinics. The organization had no specific focus on chronic care before the intervention. The intervention group (n = 221) received diabetes care based on the ICIC Model from teams of interprofessional clinical learners, and the control group (n = 163) received usual care from internal medicine residents only. All patients were followed prospectively for 18 months, from July 2002 through December 2003. The objective was to assess the impact of interprofessional team care compared with usual care on three major categories of outcomes: processes of care, clinical status, and health care utilization.
Patients in the intervention group received care from either the interprofessional diabetes management team composed of primary-care-track medicine residents, nurse practitioner (NP) students, and pharmacy students. Patients in the control group received usual care from an assigned traditional-track internal medicine resident. Patients were not randomly assigned but were allocated to medicine residents before the beginning of the study by medical center staff not affiliated with the research study. Patients were specifically matched with residents only in cases of language needs; the remaining patients were arbitrarily allocated. There were a total of three interprofessional teams, each including four to five medicine residents, two NP students, and two pharmacy students, supervised and cross-precepted by interprofessional faculty from schools of medicine (n = 2), nursing (n = 1), and pharmacy (n = 1). A total of 120 learners (56 second- and third-year medicine residents participating in a three-year program, 29 second-year NP students participating in a two-year program, and 35 fourth-year pharmacy students participating in a four-year program) were involved in interprofessional team care. Medicine residents and pharmacy students were assigned to the rotation, and NP students chose to participate on the basis of expressed interest in chronic illness management. On the basis of existing rotation schedules, NP student rotations were 10–12 weeks, pharmacy students participated for 12 weeks, and medicine residents participated on rotating ambulatory blocks for two months at a time during two years of their three-year primary care residency.
The 28 traditional-track internal medicine residents responsible for the usual care patients were supervised by attending internal medicine faculty and had no formal exposure to interprofessional faculty. The UCSF institutional review board granted an exempt certification for this study, because only aggregate data stripped of all identifiers were analyzed. The interprofessional faculty oriented the clinic staff, developed the curriculum, informed institutional leaders, and planned the interprofessional collaboration in the six months before the study was implemented. The project was supported by a two-year grant from the Robert Wood Johnson Foundation–Partners in Quality Education, which paid for 5% to 10% of four interprofessional faculty members' time (the rest was donated), 10% of a data analyst's time, and a half-time program manager.
Intervention: Team care
Intervention team members participated in the chronic illness curriculum for one half-day each week, which included a 60-minute didactic presentation, a 30-minute clinical discussion session that focused on patient management and quality improvement, and 2.5 hours of clinic visits with patients. The curriculum, developed and delivered by the interprofessional faculty, was based on the ICIC Model and emphasized improving quality of care and diabetes outcomes. Weekly presentation topics covered various aspects of diabetes care, including insulin management, use of oral antidiabetic agents, hypertension and diabetes, lipid management in diabetes, aspirin use, glucose monitor instruction and use, foot care, nutrition, exercise, motivational interviewing, and patient self-management.
The curriculum was designed as an integrated process, which made it easier to rotate students in and out of the program more efficiently without losing continuity of patient care. A supplemental online course focusing on the practical management aspects of diabetes was also made available. The principles of the ICIC Model were introduced during two didactic seminars at program implementation and reinforced throughout the curriculum. Population-based quality improvement was learned by planning and implementing quality improvement projects using the Plan, Do, Study, Act (PDSA) approach recommended by the ICIC Model (Table 1). Learners received quarterly patient panel reports on process of care benchmarks and clinical status markers from the registry. These data were used to identify deficiencies in processes of care and clinical status and to plan quality improvement projects. Over the course of the study, the teams developed and tested more than 15 small quality improvement projects using the PDSA strategy (Table 1) as part of the implementation of the ICIC Model. The projects included mail and phone outreach and group visits with defined subpopulations (e.g., patients with elevated HbA1c, smokers, etc.) of the sample, enhanced self-management tools, and interprofessional collaboration interventions. The teams were supported by interprofessional faculty and by specialists in diabetes and nutrition. Assessment of clinical status and treatment decisions were handled jointly by the teams. Communication was maintained by medical chart entries at each patient visit and through weekly conferences.
Delivery system redesign
Following each seminar, patients were seen in individual 30-minute appointments by one or more of the team members. Patients were scheduled into a new format of planned diabetes visits. Follow-up appointments were scheduled with either the same team member or one from another discipline, depending on identified patient needs, to facilitate continuity of care. Roles of team members overlapped, but in general, medicine residents focused on medical management. NP students took over as functional coordinators of care with primary responsibility for lifestyle assessment and patient goals, active follow-up, and support of patient self-management. Pharmacy students took responsibility for medication therapy management, insulin initiation, smoking cessation counseling, and patient education on medications and glucose monitors. Quality of care was the collective responsibility of the entire team.
Clinical information and decision support systems
An electronic clinical information system supplied clinical data, such as blood pressure (BP), weight, and laboratory data from clinic visits. A clinical registry of diabetic patients was created, and intervention teams were provided with feedback of panel summary results. The decision support system reported parameters of diabetes care: eye exam, foot exam, smoking status, HbA1c, LDL-C, and urine microalbumin (MA) with electronic reminders of due dates and target goals. All laboratory data were entered automatically into the electronic clinical information record when blood or urine tests were done in the clinical laboratory. However, data entry for process of care variables depended on the provider's note for assessment of smoking status, self-management, completion of foot and eye exam, and pneumococcal vaccination.
Patient self-management activities, such as self-monitoring of blood glucose, foot care, diet and exercise modification, diabetes education resources, and participation in planned visits, were addressed through individual and small-group appointments with members of the team and through population-based quality improvement projects. All patients in the intervention group were targeted for individual coaching in self-management activities by the NP or pharmacy students. Group visits to reinforce these skills were scheduled in response to opportunities for skill building identified through the PDSA approach.
Comparison control group
The control patients received usual care from internal medicine residents (n = 28) in the same general medicine clinic but on different days and times from the intervention group. Allotted appointment time lengths were the same for the comparison group as the intervention group. The clinical information and decision support systems also were available to providers giving usual care. The control residents and patients were not exposed to interprofessional team care learners, the ICIC Model, registry data reports, the didactic curriculum, or the case discussion and quality improvement curriculum. The control residents had internal medicine preceptors who were not associated with the diabetes project and had no formal contact with the nursing and pharmacy faculty.
Learners in interprofessional team care (n = 120), and internal medicine residents performing usual care (n = 28), as well as NP students (n = 9) and pharmacy students (n = 7) not participating in the chronic care curriculum or involved with patients in the chronic care clinic were asked to complete the Robert Wood Johnson/Partners in Quality Education “Take Care to Learn” Trainee Survey at baseline and after completing clinical rotations.24 This survey, developed by the Center for Health and Public Service Research at New York University in 2002, included items measuring four principal domains of providing diabetes care (accessing information and support, process of care, preparation for care delivery, and success providing diabetes care) and reflected the elements of the ICIC Model.
Data preparation and analysis
Data from the clinical information system and the appointment and billing system were downloaded in Excel format by the institutional registry coordinator, stripped of all identifiers, aggregated into cohorts, and then uploaded to the SPSS statistical data management program (SPSS, Chicago, Illinois). Of the 384 patients who had baseline data, 59 patients (29 intervention, 30 control) never came in for a follow-up visit after the baseline encounter; these patients were retained in the analyses based on intention to treat.
Process of care measures included assessment of HbA1c, LDL-C, MA, BP, foot exam, pneumococcal vaccination, discussion of self-management goals, smoking status, and ophthalmology referral. Process of care measures were dichotomous variables assessing whether recommended procedures were done or not done at the appropriate intervals. Clinical status measures included HbA1c, LDL-C, MA, and BP values. Health care utilization measures included planned clinic visits, urgent care visits, emergency department visits, and hospitalizations. Baseline process of care and clinical status measures were taken as the most recent result within the 24 months prior to study commencement and compared with the final measure taken nearest to the end of the study. Health care utilization measures included visits and hospitalizations during the 24 months prior to study commencement for baseline and during the 18 months of the study for the final assessment. Utilization data were retrieved from the institution's appointment and billing system database.
Intention-to-treat statistical analyses were performed with independent t tests for normally distributed continuous variables, Mann-Whitney U tests for nonnormally distributed variables, and continuity-adjusted chi-square tests for categorical variables to assess differences between groups at baseline and final measures. The statistical significance level was set at α = .05. The sample “n” for specific clinical status variables varied depending on the availability of data for each patient during the window of comparison. When the data were missing, it was because either the assessment was not done or the assessment was not clinically indicated as determined by the provider. It was not possible to differentiate between these two explanations for missing data.
Table 2 shows the demographic characteristics of the two cohorts; there were no significant differences between them. The total sample had a mean age of 64 years and was 54% female, 15% Hispanic, and 69% non-Caucasian; 33% were non-English speaking. This sample is characteristic of the type 2 diabetes adult population living in this major West Coast city. Although nearly all patients registered in this system were insured, the sample was socioeconomically challenged in that 68% were insured by Medicaid or Medicare.
Analysis of the process of care measures showed there were no statistically significant differences between groups at baseline (Table 3). However, there were statistically significant differences between the groups by the end of the study regarding processes of care. Intervention patients more frequently received assessments of HbA1c (79% versus 67%; P = .01), LDL-C (69% versus 55%; P = .009), BP (86% versus 79%; P = .08), MA (40% versus 30%; P = .05), smoking status assessment (43% versus 31%; P = .02), and foot exams (38% versus 20%; P = .0005).
There were no significant differences between groups at baseline regarding mean values of clinical status measures. Though mean values of clinical status measures for some outcomes (e.g., LDL-C) improved over time for both groups, there were no significant differences between groups at the final assessments for any clinical status measures. Table 4 shows the means and standard deviations of clinical status measures for those with data.
There were no significant differences between groups at baseline regarding health care utilization measures (Table 5). However, there were significant differences between the groups regarding the final health care utilization measurements over the course of the study. Intervention patients had significantly more planned general medicine visits (7.9 ± 6.2 versus 5.7 ± 5.7, P = .006) than did control patients. Although not statistically significant, the intervention group also had fewer emergency room visits (1.1 ± 2.3 versus 1.8 ± 4.9, P = .17) and hospitalizations (0.52 ± 1.2 versus 0.74 ± 2.0, P = .68) than did control patients. The intervention group had a total of 9.9 visits during the 18-month observation period versus 8.8 total visits for the control group. This 1.1 difference in total visits comprised an increase of 1.7 scheduled visits and a decrease of 0.5 acute care (non-ED) visits in non-primary-care settings for the intervention group.
Post hoc analysis revealed significant differences in baseline data between patients who did not return for at least one visit during the study and those who did. Patients who did not return for a visit during the study had significantly higher diastolic BP (77.5 versus 71.8, P = .01) and tended not to have a prescription for ACE/ARB (71% versus 37%, P < .001).
Table 6 compares responses of the interprofessional team learners and the control learners to survey items relevant to the elements of the ICIC Model. Responses were obtained from 81% of the team learners and 64% of the comparison group. Although responses were sought from participants in both groups at the end of participation, completion of the survey was not mandatory. Overall, at the end of the intervention period, and in contrast to the comparison group, learners in the interprofessional diabetes management intervention group improved significantly in all measured components of the ICIC Model. These learners felt significantly more prepared to access and use the decision support system, clinical information system, and community resources at the end of one year. Also, they felt significantly more successful in delivering diabetes care and providing self-management support to patients. The comparison group failed to improve on all but one of the items.
The results of our study showed that interprofessional learners, supported by interprofessional faculty, were able to learn and implement population-based quality improvement initiatives based on the ICIC Model. Remarkably, despite relatively short rotations and constantly changing composition, the teams learned to work together to improve the quality of diabetes care in one half-day, weekly clinic and demonstrated significant improvements in the processes of care. The patients cared for by the interprofessional teams had improved processes of care, more scheduled visits, and a trend toward less use of urgent care than control patients. The intervention group had more total visits during the 18-month observation period than did the control group. However, the difference in total visits comprised an increase in planned visits and a decrease in unscheduled acute care (non-ED) visits in non-primary-care settings. Therefore, although the intervention patients had more total visits, it is likely that the increase was associated with better care—and predictably less cost—because of the more appropriate care setting. It is worth noting that 10% of the intervention patients (29 of 221) never came in for scheduled clinic visits and could not have benefited from the ICIC Model. Nevertheless, the intervention group improved more than the control group. This finding suggests that the potential for the ICIC Model to improve the processes of care is even greater than what our modest results were able to show.
It is not possible to determine the significance of the differences uncovered between the baseline data of those patients who did return for a visit during the study and those who did not return. It is likely that these patients were not on an ACE/ARB because they had fewer encounters with primary care. The nonreturning patients may have been more socioeconomically challenged (e.g., a greater proportion of nonreturning patients were on Medicare/Medicaid), which also may have adversely affected their clinical status. Based on these preliminary findings, more powerful interventions to make contact with these patients should be investigated.
The novelty of these findings lies in teaching the principles of improving chronic illness care to interprofessional learners and having them devise and implement the quality improvement interventions. The results show that this educational approach is both feasible and effective. Interprofessional team care by learners was effective in improving quality of care for adult patients with diabetes who were followed in general medicine clinics. The chronic illness framework for the team care intervention resulted in more appropriate health care utilization. Results of the learner evaluations suggest that learners exposed to this training model significantly improve in their ability to provide care for type 2 diabetes and feel more successful. The significant gains in providing self-management support suggest that having learners participate in interprofessional teams leads to the understanding that all team members and patients themselves share responsibility for self-management of diabetes.
Our study had some important limitations. Study patients were preassigned to the medicine residents in both groups and were not randomized. Still, there were no significant differences between the two groups at baseline. The intervention was implemented in an existing clinical practice system in which clinic staff were used to an established delivery system, and there were no additional resources to dramatically change the system before the intervention. Although the learners themselves adjusted to working together, challenges related to overlapping roles and communication barriers were persistent issues. Interprofessional learners were unclear as to how to negotiate overlap in expertise, especially between the NP students and the primary care medicine residents. In one example of a communication barrier, a team member inadequately documented a selected treatment, which led to confusion among the other team members and delays in communication for treatment decisions made by another team member. Because the learners were enrolled in different training programs that had different rotational schedules, the team members were constantly changing, so the development of a true sense of “team” was challenging. Another potential limitation may have been that the comparison group residents were internal medicine residents in the traditional track, whereas those participating in the intervention were internal medicine residents in the primary care track. It is possible that the primary care internal medicine residents may have had a greater propensity toward ambulatory care than the internal medicine residents in the traditional track. Both groups of UCSF residents, however, spend a substantial amount of training time in the ambulatory setting. Additionally, patient panels were not stable over time as patients enrolled or left the clinic registry. Therefore, our comparisons are based on the sample available, although the majority of the patients were present for all 18 months of the study. New patients who joined the registry were not included in the analysis, because they lacked the initial baseline data required for analysis.
Although we hoped to see improvement in the clinical outcomes of the intervention patients as compared with the control patients, we were limited by the relatively small size of the groups, the surprisingly good clinical status at baseline (creating a ceiling effect), and the relatively short duration of the study. Clinical status measures improved in both groups, which may be attributable to new clinical practice guidelines for diabetes management that were released during the study period. The wide distribution of these guidelines may have influenced the diabetes management of both groups of providers, potentially blunting the effect of the study intervention on clinical status outcomes. Additionally, data available for our analysis came from the institutional clinical information system and the appointment and billing systems. If patients received care outside our institutional setting during the study period, that information would not have been captured in our analysis.
The strengths of the study were implementation in a real clinic setting, delivering ongoing care to a population of socioeconomically challenged urban patients with type 2 diabetes, and using teams of interprofessional learners all enrolled in clinical training programs. A key element of success was the strong, consistent support from a core and constant group of interprofessional faculty who served as clinical preceptors for the interprofessional learners. It is notable that improvement in process measures and health care services utilization is possible and sustainable under these conditions.
This study is of significance to medical educators interested in interprofessional education and curriculum development. Going forward, it is important to touch on some critical lessons learned from this experience. Of significant importance is the establishment of a committed interprofessional core of faculty to guide curriculum development and implementation. The faculty team is the primary program resource for negotiating clinical logistics and curricular requirements. Additionally, it cannot be overlooked that a strong, comprehensive patient registry is vital for long-term program success. Any program interested in chronic illness management must invest heavily in development and maintenance of a patient registry. Program efficiency and continuity of care are heavily reliant on reliable patient panel data.
The ICIC Model calls for team-based care, but using interprofessional teams of learners, who led their own quality improvement interventions, is a novel approach. The Institute of Medicine report, Crossing the Quality Chasm,25 called for clinical education and training, including collaborative team-based care, to improve quality of care for patients. The report emphasized the need to introduce meaningful interprofessional experiences into health education training programs. Although in some academic settings students have implemented interprofessional educational experiences for the health professions,26 structured interprofessional curricula remain rare among US universities with health professions schools. The largely positive interprofessional experience of the diabetes program in our institution has led to an ongoing interprofessional weekly chronic illness clinic where primary care teams of medicine residents, NP and pharmacy students, and residents continue to learn and provide care together to patients with any chronic illness, including diabetes.
This research was supported by a grant from the Robert Wood Johnson Foundation, Partners for Quality Education. The authors thank Josh Adler, MD, for his role providing administrative support for collecting panel data and for championing the program with institutional leaders; Kathy Julian, MD, and Cindy Lai, MD, for primary care medicine residency direction and assisting with program implementation; Michael Potter, MD, and Robert Rushakoff, MD, for help in developing the curriculum, and Andrew Leeds, PharmD, JoAnne Saxe, MS, NP, and Kellie Sheehan, MS, NP, for participating as faculty preceptors.
1 Blonde L. State of diabetes care in the United States. Am J Manag Care. 2007;13(suppl 2):S36–S40.
2 Centers for Disease Control and Precention. National Diabetes Fact Sheet: General Information and National Estimates on Diabetes in the United States, 2005. Atlanta, Ga: Centers for Disease Control and Prevention; 2005.
3 Hogan P, Dall T, Nikolov P. Economic costs of diabetes in the US in 2002. Diabetes Care. 2003;26:917–932.
4 Hing E, Cherry DK, Woodwell DA. National Ambulatory Medical Care Survey: 2004 summary. Adv Data. June 23, 2006:1–33.
5 Clark CM, Fradkin JE, Hiss RG, Lorenz RA, Vinicor F, Warren-Boulton E. Promoting early diagnosis and treatment of type 2 diabetes: The National Diabetes Education Program. JAMA. 2000;284:363–365.
6 Legorreta AP, Liu X, Zaher CA, Jatulis DE. Variation in managing asthma: Experience at the medical group level in California. Am J Manag Care. 2000;6:445–453.
7 McBride P, Schrott HG, Plane MB, Underbakke G, Brown RL. Primary care practice adherence to National Cholesterol Education Program guidelines for patients with coronary heart disease. Arch Intern Med. 1998;158:1238–1244.
8 Ni H, Nauman DJ, Hershberger RE. Managed care and outcomes of hospitalization among elderly patients with congestive heart failure. Arch Intern Med. 1998;158:1231–1236.
9 Perez-Stable EJ, Fuentes-Afflick E. Role of clinicians in cigarette smoking prevention. West J Med. 1998;169:23–29.
10 Young AS, Klap R, Sherbourne CD, Wells KB. The quality of care for depressive and anxiety disorders in the United States. Arch Gen Psychiatry. 2001;58:55–61.
11 Piatt GA, Orchard TJ, Emerson S, et al. Translating the chronic care model into the community: Results from a randomized controlled trial of a multifaceted diabetes care intervention. Diabetes Care. 2006;29:811–817.
12 Nutting PA, Dickinson WP, Dickinson LM, et al. Use of chronic care model elements is associated with higher-quality care for diabetes. Ann Fam Med. 2007;5:14–20.
13 Solberg LI, Crain AL, Sperl-Hillen JM, Hroscikoski MC, Engebretson KI, O'Connor PJ. Care quality and implementation of the chronic care model: A quantitative study. Ann Fam Med. 2006;4:310–316.
14 Siminerio LM, Piatt G, Zgibor JC. Implementing the chronic care model for improvements in diabetes care and education in a rural primary care practice. Diabetes Educ. 2005;31:225–234.
15 Bodenheimer T. Interventions to improve chronic illness care: Evaluating their effectiveness. Dis Manag. 2003;6:63–71.
16 Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness: The chronic care model, Part 2. JAMA. 2002;288:1909–1914.
17 Wagner EH, Glasgow RE, Davis C, et al. Quality improvement in chronic illness care: A collaborative approach. Jt Comm J Qual Improv. 2001;27:63–80.
18 Wagner EH. The role of patient care teams in chronic disease management. BMJ. 2000;320:569–572.
19 Bodenheimer T, Wagner EH, Grumbach K. Improving primary care for patients with chronic illness. JAMA. 2002;288:1775–1779.
20 Hroscikoski MC, Solberg LI, Sperl-Hillen JM, Harper PG, McGrail MP, Crabtree BF. Challenges of change: A qualitative study of chronic care model implementation. Ann Fam Med. 2006;4:317–326.
21 Nuovo J, Balsbaugh T, Barton S, et al. Development of a diabetes care management curriculum in a family practice residency program. Dis Manag. 2004;7:314–324.
22 Landefeld CS, Callahan CM, Woolard N. General internal medicine and geriatrics: Building a foundation to improve the training of general internists in the care of older adults. Ann Intern Med. 2003;139:609–614.
23 Ury WA, Rahn M, Tolentino V, et al. Can a pain management and palliative care curriculum improve the opioid prescribing practices of medical residents? J Gen Intern Med. 2002;17:625–631.
24 Greene J, Yedidia MJ. Provider behaviors contributing to patient self-management of chronic illness among underserved populations. J Health Care Poor Underserved. 2005;16:808–824.
25 Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academy Press; 2001.
26 Johnson AW, Potthoff SJ, Carranza L, Swenson HM, Platt CR, Rathbun JR. CLARION: A novel interprofessional approach to health care education. Acad Med. 2006;81:252–256.