The Accountable Primary Care Model: Beyond Medical Home 2.0

December 12, 2014
Thomas D. Doerr, MD

,
Herbert B. Olson, FSA

,
Deborah C. Zimmerman, MD

The American Journal of Accountable Care, December 2014, Volume 2, Issue 4

The new Accountable Primary Care Model delivers Triple Aim-like outcomes. This model is empirical, measurable, programmatic, scalable, and transplantable.

ABSTRACT

Objective: To test a novel model of primary care designed to improve healthcare outcomes.

Methods: We used literature searches, data analytics, and interviews with high-performing physicians to create the Accountable Primary Care Model. We enhanced the “4 Cs” Primary Care Model (first contact; comprehensive care; continuous care; coordinated care) with the addition of 5 more Cs: physician credibility, collaborative learning, cost-effectiveness, capacity expansion, and career satisfaction. Then, using a retrospective quasi-experimental design, we compared 3 years of outcomes data from members in a Medicare Advantage plan in a virtually integrated delivery system with relevant benchmarks, including matched patients in fee-for-service Medicare.

Results: The Medicare Advantage health plan averaged 29,556 members with costs that were 58.93% of those in fee-for-service Medicare. Spending was shifted when appropriate from inpatient to outpatient settings and from specialists to primary care physicians. After allocating $80 per member per month for primary care costs, the health plan had medical cost ratios from 2010 to 2012 averaging 72.7% before surplus distribution. Member benefits were the best in the market, evidenced by the average maximum out-of-pocket expense limit of $1875. The next best plans averaged $2800. The health plan was rated 4.5 stars each year by CMS for quality, patient experience, access to care, and care process metrics. Of the primary care physicians surveyed, 90.2% would probably or definitely recommend this plan to another physician. About 91.4% of specialists surveyed were satisfied or very satisfied with this plan.

Conclusions: The Accountable Primary Care Model offers improved clinical outcomes, cost, and patient satisfaction—the Triple Aim formulated by the Institute for Healthcare Improvement—along with the additional aim of physician satisfaction. It is the Triple Aim Plus One.Introduction

Volume-based reimbursement has limited the deployment of the Institute of Medicine/Starfield’s Primary Care Model (including the 4 Cs: first contact; comprehensive care; continuous care; coordinated care)1,2 despite its better healthcare outcomes at lower costs and validation of an assessment tool.3 Emerging evidence from controlled trials of the Patient Centered Medical Home (PCMH) model 3 to 5 years post implementation shows that it has produced little, if any, savings in the total cost of care. For example, Friedberg et al published a controlled trial 3 years after implementation of the PCMH for 39,559 patients in a multi-payer environment and found no total cost savings.4 Analysis of the retrospective study by Flottemesch et al reveals that a 10% increase in the Patient Practice Connections Readiness Survey (a proxy for medical home-ness) correlates with a 0.35% reduction in total annual costs of care for 58,391 patients with diabetes during years 3 through 5 post implementation.5 Wang et al found no significant changes in total medical costs during the third year post implementation of the PCMH in 1891 patients with diabetes.6 Furthermore, none of these studies includes the cost of supporting the PCMH, estimated to be $148,347 to $163,347 annually per primary care physician (PCP).7

Development of the APC Model

To enable physician transformation to accountability, Lumeris developed the Accountable Primary Care (APC) Model. Lumeris is a company that designs, builds, operates, measures, and optimizes value-based care models. This model embraces the 4 columns (Cs), or pillars, that support the Institute of Medicine/Starfield Primary Care Model,2 and contributes 5 additional Cs. The goal of this model is to foster the Institute for Healthcare Improvement’s Triple Aim of improved clinical outcomes, cost, and patient satisfaction, along with the additional aim of physician satisfaction—it is the Triple Aim Plus One.

The APC Model is built on the Lumeris data analytics department’s multivariate regression analysis of relationships between 23 hypothesized variables and the medical cost ratio to discover critical success factors. We also contracted with the Olin School of Business at Washington University in St. Louis for an independent discovery of critical success factors. They utilized multivariate regression analysis to identify the 13 most important independent variables that impacted cost/revenue ratios at the PCP level. They then used forecast error variance decomposition of 2010 data to validate the model they generated from 2009 data. In addition, in 2012, the primary author of this manuscript interviewed 10 high-performing PCPs and hospitalists with more than 200 years of collective experience in diverse settings such as medical groups, independent practice associations, and a physician-hospital organization. The high-performing physicians were identified based on Lumeris internal data analytics focusing on population health quality metrics, patient satisfaction, and costs of care. During the same time period, this author performed Google Scholar searches of the grey and medical literature about topics and synonyms related to the 9Cs. He then searched the references in, and the articles citing, each study, as appropriate. After obtaining full text copies of the 1377 articles that passed initial screening, he reviewed each, analyzed the key findings, and integrated them into coherent themes. All of this information informed the development of the APC model as we utilized qualitative interpretative analyses to discover larger meaning and theory.

The APC Model (The 9 Cs)

The APC Model has 4 levels. Level 0 is no PCP engagement, or the absence of accountability; PCPs are paid one-off for piecework. Levels 1, 2, and 3 represent low, medium, and high PCP engagement, respectively. Organizations may progress through these levels sequentially. Some organizations will have aspects of more than 1 level of the model implemented at the same time. Each level is implemented through a playbook that includes granular and specific actions for 8 different actors—the types being: administrator; analyst; care manager; chief medical officer; office staff (registered nurse, medical assistant); primary care physician/ nurse practitioner/physician assistant; population health manager; and regional medical director—and roles in the healthcare system by day, week, month, and quarter.

C1: First contact refers to the patient’s initial entry point into the healthcare system. When new health or medical needs arise, it is usually optimal for the first contact to be with the PCP. In a nationally representative sample of more than 20,000 episodes of care, episodes that began with PCP visits, as opposed to some other source of care in the system, were ultimately 53% less costly. Controlling for emergency department (ED) visits, health status, sociodemographics, and other relevant variables did not materially change this cost differential.8

C2: Comprehensive care means PCPs offer a large range of services across the spectrum of patient care needs, for all but the most uncommon problems. With the APC Model, office visits for older or more complex patients are usually scheduled for 30 minutes to proactively address both manifest and latent issues. Longer visits also enable compliant documentation and accurate diagnosis coding for comprehensive health risk status assessments. These assessments enable claims-enhanced, data-driven collaborative population management, as well as appropriate risk-adjusted revenue assignment in Medicare Advantage (MA) and many accountable care organizations, Medicaid, and state insurance exchange programs.

C3: Continuous, longitudinal, person-focused care means that PCPs and patients work together in the context of longstanding relationships to reach mutual decisions that transcend episodic care. Accountable PCPs tend to follow their patients across care settings as they attend to their care in hospitals, nursing homes, and hospices. Person-focused care includes discussing expectations, discerning preferences, setting priorities, and engaging with family, as well as the usual personalized prevention, screening, and advance care planning.

C4: Coordinated care is widely admired but narrowly understood and practiced in the United States. Transitions to and from EDs, inpatient facilities, and skilled nursing facilities are obvious high-stakes care coordination opportunities. Half of Medicare beneficiaries have 5 or more chronic conditions.9 And every referral to a specialist—with accompanying requirements for information transfers, medication reconciliations, and adjustments to the patient’s overall care plan— may be viewed as a type of care transition. Accountable PCPs, those practicing at the top of their professional training and providing comprehensive primary care, decrease the number of care coordination events as they enhance patient convenience.

C5: Credible primary care earns patients’ trust in the PCP’s competence, beneficence, and professionalism. The strategic significance of this high level of trust is manifest as patients are willing to place confidence in PCP recommendations that may run contrary to their expectations or apparent wishes. Credibility is needed to counter the prevailing currents of Internet-driven consumerism and supply-sensitive care, which tend to undermine PCPs’ authority as they promote patient-specific and individually designed care agendas. PCPs earn credibility with patients primarily by providing the first 4 Cs, but also by honoring patients’ perspectives with active listening and respectful explanations, and by their professionalism in providing excellent customer service.10

One accountable PCP said that without credibility, physicians are nothing but referral clerks. Patients who do not trust their PCPs often want their care from specialists, even for health problems within the realm of primary care.

C6: Collaborative learning requires cooperation across institutional roles, organizational boundaries, and care settings. Both PCPs and the integrators11 of the healthcare system (often the payer or hospital) make important contributions. Payers contribute clinical and claims data transparency aggregated from across the continuum of care. Clinicians perform the work of augmenting this shared, actionable longitudinal information with their own patient data. Clinicians' knowledge of their patients is complementary to payers'. It is more nuanced (as it includes patients' current preferences) but incomplete in its breadth. The system integrator typically funds the analytics, while clinicians engage with the resulting rules-based clinical decision support messages. These 2 parties also take accountability for various tasks and responsibilities in collaborative population management, built on a foundation of collaborative learning and trust.12

C7: Cost-effective care naturally results from the other Cs, but it is enhanced by proactive, intentional PCP-led efforts. These are empowered by important infrastructure items that help dedicated care teams succeed: educational programs, care management, clinical quality and cost data sharing, opportunity analyses, and clinical decision support from the integrator. Cost-effective care is the embodiment of the Triple Aim Plus One, as it optimizes the trade offs among population health, patient experience, and cost for each healthcare decision. Most of the time, cost-effective care is based on the 5 Rs derived from Lean manufacturing: right diagnosis, right care, right place, right provider, and right time.13

While payers have an important role in cost-effective care, individualized decisions weighing trade offs and patient preferences are best made by physician-led care teams, in conjunction with patients and their families. Collaborative payers have the responsibility to provide vital integrated clinical and claims data with cost transparency. They also have the incentives and resources available to sift through the medical literature to identify evidence-based, fiscally responsible and, whenever possible, individualized clinical decision support messages and deliver them as close to the point of care as possible. These activities support the practices and processes by which PCPs can bend the cost curve.14

C8: Capacity expansion is a requirement for the APC Model, because it causes dramatic shifts of care from specialists to PCPs and from hospitals to the outpatient setting. The requisite primary care can be estimated as the number of patients in a PCP panel multiplied by the expansive responsibilities articulated in C1 through C7 that must be managed by PCP-led teams in order to progress toward the Triple Aim Plus One.15 The collapsing interest in primary care among recent medical school graduates demands reinvention of the fee-for-service model.16 PCP productivity expansion is the best solution for this impending crisis. A number of approaches—including e-mail visits, team- or pod-based care, and the use of nurse practitioners and other clinicians in appropriate situations—promise to narrow or even eliminate the PCP-to-future-needs gap.17 C9: Career satisfaction is markedly low among PCPs in the United States, with 36% saying they are dissatisfied with their careers—a rate 2 to 3 times higher than in Western European countries. 18 Career satisfaction is dependent on nonmonetary as well as monetary considerations. For example, pediatrics is among the lowest paid specialties, but pediatricians have one of the highest work-life balance self-ratings of all specialties.19 Career satisfaction is directly related to issues such as meaning, control, order, work-life balance, and remuneration. Lumeris internally articulated the physician satisfaction (fourth aim) as “Triple Aim Plus One” in 2009. Other thought leaders, including Wallace,20 Bodenheimer,21 Wagner, 22 and the Institute for Healthcare Improvement have proposed this fourth aim.

Table 1

Suggested metrics for the 9 Cs are presented in .

Methods

After creating the APC Model, we looked back at its performance in an MA plan with a PCP gatekeeper model. All 3 levels of the APC model are represented by various PCP organizations in the outcomes data. The health plan had 25,423 members 65 years and older in 2010, 30,091 in 2011, and 33,155 in 2012. These members were matched to Medicare beneficiaries of the same county, age, and gender from a 5% Medicare fee-for-service (FFS) population limited data set each year.

The populations were intentionally not risk-adjusted by diagnosis codes, to eliminate any possible impact of more thorough diagnosis coding in the MA plan. In this analysis the MA plan’s contracted hospital rates were normalized to 100% of Medicare allowable. In a typical contract, the payer was allocated 15% of the total revenue to cover its costs. PCPs received 80% of the shareback (distribution of surplus) in global risk contracts. The percentage of shareback that was directly linked to performance on quality and process metrics increased over the 3-year period. By 2012, this percentage was typically 25% of the PCP’s 80% shareback. Approximately 1400 specialists contracted with the health plan. They were paid 110% of Medicare rates to compensate for their added work collaborating with PCPs. They were assured that they would receive more referrals with better documentation from PCPs if they provided high-value care. The plan contracted with 22 hospitals, of which 7 were independent and 15 were in 3 hospital systems. The provider profiles for 2010 to 2012 were similar.

Two cohorts were defined for comparison. The APC Model cohort consisted of all members of the Essence Healthcare plan who were 65 years and older. The 5% FFS sample consisted of a sample of Medicare FFS data matched for age, gender, and county. Claims data for the 2 cohorts were analyzed for each of the 3 years.

Results

Table 2

During calendar years 2010 through 2012, an MA plan (Essence Healthcare, #H2610) in Missouri and southern Illinois had costs that averaged 58.93% of a matched FFS Medicare cohort adjusted for age, gender, and county ().

Figure 1

contrasts this APC Model health plan with the matched FFS population for calendar years 2010, 2011, and 2012. Each cohort was divided into 10 deciles based on total costs of care for each member of the cohorts. Each year was considered to be independent of the other 2.

Figure 2

is an analysis based on classifying claims into 1 of 10 groups. The number of claims in each group (not the costs of care) were compared. Note that the reductions in costs in Figure 1 were greater in the higher-cost deciles than in the lower-cost deciles, reflecting some influence of expenditures on preventive services.

While the number of accountable PCP visits is 197% of the fee-for-service unmanaged control group, this underestimates the increase in primary care services because it does not account for the longer duration of many of these PCP visits and many APC model services (eg, virtual visits) that are not reimbursable in the FFS remuneration system and therefore not billed. This is offset by the 67% decrease in the total number of specialist claims. Acute hospital bed days were 50% lower and skilled nursing facility days 64% lower. Occupational, physical, and speech therapy were 89% lower, and outpatient radiology 71% lower than FFS. The APC model simultaneously increased the number of preventive service claims by 1400%.

This health plan was rated at 4.5 Stars28 by CMS in 2012, 2013, and 2014 based partially on 2011, 2012, and 2013 Health Care Effectiveness Data and Information Set (HEDIS)/Star quality metrics for patient care delivered in 2010, 2011, and 2012, along with patient experience, access to care, and care process metrics. The top 19% to 30% of MA plans nationally are 4.5 or 5 Stars each year.29 Of note, the plan was rated at 5 Stars in the Consumer Assessment of Health Plans member survey questions C24 (getting needed care) and C28 (health plan quality) for each of these years.

In 2011 DSS Research performed a survey of PCPs contracted with this MA plan; the response rate was 44%. Of the 102 PCP respondents, 90.2% would definitely or probably recommend this MA Health plan to another physician. At the 95% confidence level, this is significantly higher than the national average (83.8%, n = 4529).30 In 2010, DSS Research performed a survey of specialists contracted with this MA plan; the response rate was 19.4%. About 91.4% of the 169 specialist respondents were satisfied or very satisfied with this plan. At the 95% confidence level, this is significantly higher than the national average (88.9%, n = 5810).

If the PCPs were paid $80 per member per month (PMPM)31—twice the FFS Medicare remuneration—and hospital contracts were normalized to 100% of Medicare, the MA plan would have medical cost ratios of 73.4% in 2010, 72.2% in 2011, and 72.5% in 2012, before surplus distribution. These low medical cost ratios occurred despite having the most generous member benefits, as measured by maximum out-of-pocket expenses (which were $1875, while the next lowest were $2800, looking at plans from 2010 to 2012), of any zero-premium-dollar MA health maintenance organization in its market.

Discussion

As seen in Figure 2, the profound redistributions of medical services from hospitals and skilled nursing facilities to the outpatient setting, and from specialists to PCPs, are at the heart of the APC Model. Specialists are empowered to practice at the top of their training, as the APC model frees them from many routine follow-up visits and basic care that is provided by PCPs. The overall result is improvement in outcomes from both providers. These changes are consistent with Landon and Roberts’ vision for specialty care under global payment systems.32 The 5-Star member survey ratings for access to care and quality of care suggest that the APC Model reins in supply-sensitive care33 and increases preventive care services while maintaining quality and patient satisfaction. C1 through C9, but especially C5, produce these changes.

Other models of care can be analyzed through the lens of the 9 Cs. The Direct Primary Care Medical Home concept developed by Qliance Medical Management Inc can include all of the 9 Cs except for C6.34,35 The APC Model has greater emphasis on C5 and C7, but both care models include longer visits, a greater range of primary care office services, substantially more revenue for primary care, and virtual visits. The savings generated by the APC model provide the enhanced funding for primary care that is required from patients through the age-related monthly fees of the Direct Primary Care model. This enables the APC model to serve patients of all socioeconomic strata, since it provides the most generous benefits in the market in a zero-dollar-premium MA plan. Financial outcomes of the Direct Primary Care model are unknown.

Table 3

The Primary Care Model defined C1-4, but these expansions of primary care are less substantially emphasized in the Medical Home and Wagner's Chronic Care models (). The Medical Home as deployed today is an important building block for the transition from volume to value, but it does not go far enough. Care coordination fees typically ranging from $3 to $15 PMPM don’t cover the costs of practice reengineering and primary care expansion. PCP expenses in the APC model are estimated to be $80 PMPM—twice the typical Medicare FFS compensation—for a Medicare population, consistent with Goroll’s model of comprehensive payment for comprehensive adult primary care.31

Several additional models of primary care include some of the 9 Cs. The Collaborative Care Model developed by Aetna and NovaHealth is focused mostly on C6, utilizing payer-sponsored care managers.41 Its encouraging outcomes were limited to 750 members of 1 Independent Physician Association. ThedaCare’s Collaborative Care Units are a redesign of inpatient care that has lowered costs of inpatient care by 25% and improved quality metrics.42 Thedacare also reported improvements in outpatient quality.43 Their model incorporates C6 and C7. The CareMore model includes retraining hospitalists into extensivists, who provide hospitalist care and follow these patients in the office for a visit or two before releasing them back to the care of their PCPs.44 The extensivists effectively coordinate care transitions (C4) and expand PCP capacity (C8). This model lowered inpatient utilization, length of stay, and readmissions, but overall financial outcomes are not available. HealthCare Partners Medical Group developed a care model that targets highneed patients at risk of hospitalization using multidisciplinary care teams and Homecare Teams. They bring high-risk conditions under control and return the patients to care with their PCPs. The organization’s Medicare patients use 800 acute hospital days per 1000 people per year. While their model is focused on team care rather than PCPs, they nonetheless utilize parts of C1, C2, C4, C6, C7, and C8.45 Limited information is available about Boeing/Regence’s pilot study of the Intensive Outpatient Care Program, which included C4, C6, C7, and C8 to reduce the costs of care by 20% while improving quality in the sickest quintile of the population.46 The New Model of Family Medicine was developed by the American Academy of Family Physicians and TransforMED. The model promotes all the Cs except C5 and C7, but it is not very granular and lacks published outcomes.47,48 The Institute of Medicine’s 10 New Rules for 21st Century Health Care include C1, C3, C4, C6, and C7, but the plan lacks detailed recommendations and outcomes.49 Dartmouth Clinical Microsystems offers a well-established, dynamic, customizable approach to systems improvement, rather than a specific model of care.50 Implementations in primary care may utilize Wagner’s Chronic Care Model and include C1 through C4, and C8.51 We could not identify any published systematic reviews or meta-analyses of Clinical Microsystems’ cost and quality outcomes in primary care.

Limitations

This study has several limitations. The APC model was partially derived from some physicians participating in the testing of the same model. In addition, the model was tested retrospectively in a single health plan in 1 market. It is also unclear to what extent the health plan’s gatekeeper model and PCP compensation through shared global risk contracts with quality incentives contributed to the outcomes, above and beyond the APC Model itself. The APC Model is designed for scalable implementations and various levels of the model are being deployed in multiple markets, but these independent outcomes data are not yet available.

The demonstrated strength of the APC Model is its extension of the proven Primary Care Model and the strong support for C5 through C9 in the medical literature. Note that this performance was not obtained within an Integrated Delivery System. The contracted PCPs came from medical groups, a Physician Hospital Organization and several Independent Physician Associations of varying cohesiveness. This improves the likelihood that the model may be replicable outside an Integrated Delivery System. The APC Model fosters the transition from volume- to value-based care by parsing it into 3 levels or steps. This model is completely consistent with the vision for healthcare reform articulated by Mostashari, Sanghavi, and McClellan.52

Conclusion

The 2011 Medical Home provides building blocks for better healthcare. The 9 Cs of the APC Model extend the Medical Home model as they provide an empirical, programmatic road map for physician-led teams that invigorates the practice of high-value primary care medicine.

The APC Model enables a win-win-win system for patients, providers, and payers. All 3 benefit from improved patient satisfaction, higher-quality care, and lower costs of care. Collectively, the framework of the 9 Cs of the APC Model offers new hope for our beleaguered healthcare system, and for primary care in particular. Based on these results in a 4.5-Star MA plan, it is possible to utilize the 9 Cs to progress toward the Triple Aim Plus One.Acknowledgments: We thank Bart Hamilton, PhD, Tat Chan, PhD, and their colleagues at the Olin School of Business, Washington University in St. Louis, for their contribution to the analytical foundation of the APC model. We thank DSS Research for performing and providing the physician satisfaction survey research.

Author Disclosures: All authors are employees and shareholders of Essence Group Holdings Company. This includes Lumeris, which provides services to Essence Healthcare, the health plan in this study. Essence Group Holdings Company funded and had a role in the collection of data. This information was previously presented at the Patient Centered Primary Care Collaborative Conference on October 14, 2013, Bethesda, MD; and the 4th Annual Accountable Care Organization Congress on November 5, 2013, Los Angeles, CA.

Address Correspondence to: Thomas D. Doerr, MD, c/o Lumeris, 13900 Riverport Dr, St. Louis, MO 63043. E-mail: APCmodel@lumeris.com.REFERENCES

1. Institute of Medicine. A Manpower Policy for Primary Health Care: Report of a Study. Washington, DC: National Academies Press; 1978.

2. Starfield, Barbara. Primary Care: Concept, Evaluation and Policy. New York: Oxford University Press; 1992.

3. Shi L, Starfield B, Xu J. Validating the adult primary care assessment tool. J Fam Pract. 2001;50(2):161-175.

4. Friedberg MW, Schneider EC, Rosenthal MB, Volpp KG, Werner RM. Association between participation in a multipayer medical home intervention and changes in quality, utilization and costs of care. JAMA. 2014;311(8):815-825.

5. Flottemesch TJ, Anderson LH, Solberg LI, Fontaine P, Asche SE. Patient-centered medical home cost reductions limited to complex patients. Am J Manag Care. 2012;18(11):677-686. [Notes: Analysis consisted of recalculations of Table 5 to combine the three cohorts, with weighting for the differing numbers of patients in each.]

6. Wang QC, Chawla R, Colombo CM, Snyder RL, Nigam S. Patient-centered medical home impact on health plan members with diabetes. J Public Health Manag Pract. 2014;20(5):E12-E20.

7. Deloitte Center for Health Solutions. The Medical Home: disruptive innovation for a new primary care model. http://www.deloitte.com/assets/Dcom-UnitedStates/Local%20Assets/Documents/us_chs_MedicalHome_w.pdf Published 2008. Accessed August 13, 2014.

8. Forrest CB, Starfield B. The effect of first-contact care with primary care clinicians on ambulatory health care expenditures. J Fam Pract. 1996;43(1):40-48.

9. Thorpe KE, Howard DH. The rise in spending among Medicare beneficiaries: the rise of chronic disease prevalence and changes in treatment intensity. Health Aff (Millwood). 2006;25(5):378-388.

10. Thom DH, Hall MA, Pawlson G. Measuring patients’ trust in physicians when assessing quality of care. Health Aff (Millwood). 2004;23(4):124-132.

11. Berwick DM, Nolan TW, Whittington J. The triple aim: care, health and cost. Health Aff (Millwood). 2008;27(3):759-769.

12. Doerr TD, Bak R, Ingari F, Gribble D. The collaborative payer model: new hope for Medicare and primary care. http://lumeris.com/thought-leadership/point-of-view/collaborative-payer-pov. Published November 24, 2008. Accessed April 4, 2014.

13. Mitsuishi M, Udea K, Kimura F, eds. Manufacturing Systems and Technologies for the New Frontier. London: Springer-Verlag; 2008.

14. Sirovich B, Gallagher PM, Wennberg DE, Fisher ES. Discretionary decision making by primary care physicians and the cost of U.S. health care. Health Aff (Millwood). 2008;27(3):813-823.

15. Margolius D, Bodenheimer T. Transforming primary care: from past practice to the practice of the future. Health Aff (Millwood). 2010;29(5):779-784.

16. Schwartz MD, Durning S, Linzer M, Hauer KE. Changes in medical students’ views of internal medicine careers from 1990 to 2007. Arch Intern Med. 2011;171(8):744-749.

17. Green LV, Savin S, Lu Y. Primary care physician shortages could be eliminated through use of teams, nonphysicians and electronic communication. Health Aff (Millwood). 2013;32(1):11-19.

18. Schoen C, Osborn R, Doty MM, Squires D, Peugh J, Applebaum S. A survey of primary care physicians in eleven countries, 2009: perspectives on care, costs, and experiences. Health Aff (Millwood). 2009; 28(6):w1171-w1183.

19. Shanafelt TD, Boone S, Tan L, et al. Burnout and satisfaction with work-life balance among US physicians relative to the general US population. Arch Intern Med. 2012;172(18):1377-1385.

20. Wallace JE, Lemaire JB, Ghali WA. Physician wellness: a missing quality indicator. Lancet. 2009; 374(9702):1714-1721.

21. Margolius D, Bodenheimer T. Transforming primary care: from past practice to the practice of the future. Health Aff (Millwood). 2010;29(5):779-784.

22. Ed Wagner. The fourth aim: primary care and the future of American medical care. California: California Association of Public Hospitals and Health Systems. http://www.caph.org/content/upload/AssetMgmt/PDFs/EdWagner.pdf Published December 6, 2011. Accessed April 4, 2014.

23. Consumer Assessment of Healthcare Providers and Systems Consortium. CAHPS clinician & group surveys - version: 12-month survey with patient centered medical home (PCMH) items. Agency for Healthcare Policy and Research. http://gmcboard.vermont.gov/sites/gmcboard/files/Quality%20Blueprint%20cahps%20pcmh%20adult.pdf. Updated September 1, 2011. Accessed April 4, 2014.

24. NCQA. Standards and guidelines for NCQA's Patient-Centered Medical Home. http://store.ncqa.org/index.php/recognition/patient-centered-medical-home-pcmh/2011-pcmh-standards-and-guidelines-epub-single-user.html. Published 2011. Accessed November 25, 2014.

25. Thom DH, Ribisl KM, Stewart AL, Luke DA. Further validation and reliability testing of the Trust in Physician scale. the Stanford Trust Study Physicians. Med Care. 1999;37(5):510-517.

26. Konrad TR, Williams ES, Linzer M, et al; SGIM Career Satisfaction Study Group; Society of General Internal Medicine. Measuring physician job satisfaction in a changing workplace and a challenging environment. Med Care. 1999;37(11):1174-1182.

27. Moore LG, Wasson JH. The ideal medical practice model: improving efficiency, quality and the doctor-patient relationship. Fam Pract Manag. 2007;14(8):20-24.

28. Gretchen Jacobson, Tricia Neuman, Anthony Damico, Jennifer Huang; Kaiser Family Foundation. Medicare Advantage Plan star ratings and bonus payments in 2012. http://kaiserfamilyfoundation.files.wordpress.com/2013/01/8257.pdf Published November 2011. Accessed April 4, 2014.

29. Myers N. Reaching for the stars: medical directors feel pressure to score Medicare Advantage points. Manag Care. 2014;19(6):15-18.

30. DSS Research website. https://www.dssresearch.com. Accessed November 26, 2014.

31. Goroll AH, Berenson RA, Schoenbaum SC, Gardner LB. Fundamental reform of payment for adult primary care: comprehensive payment for comprehensive care. J Gen Intern Med. 2007; 22(3):410-415.

32. Landon BE, Roberts DH. Reenvisioning specialty care and payment under global payment systems. JAMA. 2013;310(4):371-372.

33. John E. Wennberg, Elliott S. Fischer, David C. Goodman, Jonathan S. Skinner; The Dartmouth Institute for Health Policy and Clinical Practice. Tracking the care of patients with severe chronic illness: the Dartmouth Atlas of Health Care 2008. www.dartmouthatlas.org/downloads/atlases/2008_Chronic_Care_Atlas.pdf. Published 2008. Accessed April 4, 2014.

34. Wu WN, Bliss G, Bliss EB, Green LA. Practice profile. a direct primary care medical home: the Qliance experience. Health Aff (Millwood). 2010;29(5):959-962.

35. American Academy of Family Physicians. Direct primary care: an alternative practice model to the fee-for-service framework. http://www.aafp.org/dam/AAFP/documents/practice_management/payment/DirectPrimaryCare.pdf. Published April 2014. Accessed August 7, 2014.

36. Mayes R. Moving (realistically) from volume-based to value-based health care payment in the USA: starting with Medicare payment policy. J Health Serv Res Policy. 2011;16(4):249-251.

37. Miller HD. From volume to value: better ways to pay for health care. Health Aff (Millwood). 2009;28(5):1418-1428.

38. Institute of Medicine; Donaldson MS, Yordy KD, Yohr KN, Vanselow NA, eds. Primary Care: America’s Health in a New Era. Washington, DC: National Academy of Sciences; 1996.

39. Bodenheimer T, Wagner EH, Grumbach H. Improving primary care for patients with chronic illness. JAMA. 2002;288(14):1775-1779.

40. Coleman K, Austin BT, Brach C, Wagner EH. Evidence on the Chronic Care Model in the new millennium. Health Aff (Millwood). 2009;28(1):75-85.

41. Claffrey TF, Auostini JV, Collet EN, Reisman L, Krakauer R. Payer-provider collaboration in accountable care reduced use and improved quality in Maine Medicare Advantage plan. Health Aff (Millwood). 2012;31(9):2074-2083.

42. Touissaint J. Writing the new playbook for U.S. health care: lessons from Wisconsin. Health Aff (Millwood). 2009;28(5):1343-1350.

43. Toussaint J, Milstein A, Shortell S. How the Pioneer ACO Model needs to change: lessons from its best-performing ACO. JAMA. 2013;310(13):1341-1342.

44. Do H. Medical “extensivists” care for high-acuity patients leading to reduced hospital use. Agency for Healthcare Research and Quality Innovations Exchange website. https://innovations.ahrq.gov/profiles/medical-extensivists-care-high-acuity-patients-across-settings-leading-reduced-hospital-use. Published October 13, 2010. Accessed November 25, 2014.

45. Feder JL. Predictive modeling and team care for high-need patients at HealthCare Partners. Health Aff (Millwood). 2011;30(3):416-418.

46. Milstein A, Kothari PP, Fernandopulle R, Helle T. Are higher-value models replicable? Health Affairs Blog website. http://healthaffairs.org/blog/2009/10/20/are-higher-value-care-models-replicable/ Published October 20, 2009. Accessed April 4, 2014.

47. Task Force 1 Writing Group; Green LA, Graham R, Bagley B, et al. Task Force 1. report of the Task Force on patient expectations, core values, reintegration, and the new model of family medicine. Ann Fam Med. 2004;2(suppl 1):S33-S50.

48. Martin JC, Avant RF, Bowman MA, et al; Future of Family Medicine Project Leadership Committee. The Future of Family Medicine: a collaborative project of the family medicine community. Ann Fam Med. 2004;2(suppl 1):S3-S32.

49. Committee on Quality of Health Care in America; Institute of Medicine. Crossing the Quality Chasm: A New Health System for the 21st Century. Washington, DC: National Academies Press; 2001.

50. Nelson EC, Batalden PB, Godfrey MM, Lazar JS. Value by Design: Developing Microsystems to Achieve Organizational Excellence. San Francisco: Jossey-Bass; 2011.

51. Godfrey MM, Nelson EC, Wasson JH, Mohr JJ, Batalden PB. Microsystems in health care: part 3. planning patient-centered services. Jt Comm J Qual Saf. 2003;29(4):159-170.

52. Mostashari F, Sanghavi D, McClellan M. Health reform and physician-led accountable care: the paradox of primary care leadership. JAMA. 2014;311(18):1855-1856.