In the United States, projections suggest a significant increase in the demand for primary care services. Increases in the prevalence of chronic illness due to an aging population will be a large contributor to increased demand.1 More than 90% of Americans age 65 years or older report having at least one chronic illness, and the number of Americans in this age group is expected to double by 2050.2,3 Similarly, estimates suggest that increases in the number of insured patients due to the Affordable Care Act could raise the demand for primary care by an additional 15 to 24 million visits annually.4
This increase in demand for services is leading primary care organizations to explore new models of care delivery. Successful management of chronic illnesses is longitudinal and complex, requiring care from a range of healthcare providers.5,6 Organizing these providers into teams to maximize coordination of services is one goal of the patient-centered medical home (PCMH) model.7,8 One frequently discussed approach to primary care teams involves including multiple clinicians—such as physicians, physician assistants (PAs), and NPs—who can play a variety of roles.9-11
Based on theories of team effectiveness and roles, effective incorporation of multiple providers into teams requires well-defined roles.12-14 Roles are defined by the division of tasks between team members.14,15 Because of overlapping competencies with physicians, differences in state and institutional scope of practice policies, and individual negotiations with collaborating physicians, PAs and NPs can perform a variety of roles.16-18 Provider roles (physician, PA, or NP) can be conceptually defined according to level of involvement—as a usual provider or a supplemental provider of care.17,19,20 A single provider can perform different roles for different panels of patients: for example, a PA can be a usual provider for a panel of patients, but also a supplemental provider on the panel of another primary care provider.17,19,20
The role of usual provider of care, often called primary care provider or PCP, has philosophical as well as operational significance. The philosophical foundation of primary care is based on the preeminence of the relationship between a primary care provider and patient.20 This relationship is believed to develop as a result of the usual provider's consistent provision of a wide range of healthcare services over time. Historically, the role of usual provider is operationalized in a primary care clinic as a primary care provider's panel of patients. The primary care provider has an agreement with this group of patients to provide the full range of primary care services and coordinate all the patients' healthcare services over time, thus ensuring that patients' healthcare needs are met.20 However, as the primary care needs of an aging population with chronic illness have increased in scope and complexity, it has become increasingly difficult for a single primary care provider (physician, NP, or PA) to provide all the required care, necessitating a team of clinicians.10,21-23
The reality of caring for a group or panel of patients makes it difficult to ensure that the usual providers of care perform all of the primary care visits for patients on their panel, resulting in other clinicians in the clinic providing some of the visits.10,22 Supplemental providers of care are clinicians who do not act as a usual provider to a patient but provide a minority of primary care visits to patients who are not on their panel.15,19,20,24,25 A study done in the Veterans Health Administration (VHA) suggests that patients with diabetes see their usual providers for about 75% of their visits.26 The remainder of the visits are handled by providers performing a supplemental role. This supplemental role could be intentional, such as two clinicians having an explicit agreement to comanage a patient. Alternatively, it could be less intentional, as when one clinician sees a same-day visit patient on another clinician's panel for an acute need. One study conducted in more than 20 primary care clinics associated with a large multispecialty physician group reported that an average of about five clinicians (one usual provider and four supplemental providers) served a panel of primary care patients.24
Several studies have compared outcomes for patients seen by physicians and NPs and those seen by physicians, PAs, and NPs, and have demonstrated similar patient outcomes.27-32 However, each study has a variety of limitations. Several of the studies only compare physicians and NPs.27-29 Two of these studies use Medicare data, which has a high risk of misclassifying the usual provider due to incident to billing and inability to identify the specialty of NPs.28,29 Another study looked only at the profession performing a usual provider role and did not take into account the supplemental providers.32 Only one study evaluated all three clinician types in a range of roles and demonstrated that different roles are associated with different patterns of patient outcomes.15,31 However, this previous study was performed in clinics associated with one academic institution, limiting its generalizability.
The VHA has been a leader in innovative primary care models. It is the largest integrated delivery system in the United States and the largest employer of PAs and NPs.33,34 In 2012, the VHA provided primary care to more than 6.33 million patients nationally in 990 outpatient clinics.35,36 About one-third of primary care visits delivered in the VHA are with NPs and PAs.37 The VHA also adopted a team-based PCMH model, the Patient-Aligned Care Team (PACT), that expanded use of PAs and NPs in primary care.38,39 Each PACT is led by an NP, PA, or physician who is responsible for the medical care of a panel of patients, making the VHA an ideal source of information on primary care provider roles.40
Data sources and sample construction
This study used electronic health record (EHR) data from the VHA. Data from the EHR included inpatient and outpatient visit information. Each visit contained International Classification of Diseases Clinical Modification 9th revision (ICD-9-CM) diagnosis codes; CPT-4 procedure codes; clinician identifiers; and the date, place, and type of service. The study was approved by the institutional review board of the Durham Veterans Affairs Medical Center.
The study sample consisted of adults (age 18 years and older) with diabetes who were being treated pharmaceutically at VHA primary care clinics (final sample, N = 609,668). Specifically, patients must have had a diabetes diagnosis (ICD-9-CM codes 250.xx) associated with at least one VHA inpatient admission and/or at least two VHA outpatient visits in fiscal year (FY) 2012 (October 1, 2011 to September 30, 2012) (N = 1,049,638) and a filled prescription for insulin and/or an oral hyperglycemic agent the same year (N = 830,602). These patients had to have one or more VHA primary care visits in FY 2012, identified using VHA administrative codes indicating a primary care clinic. Patients were excluded if they did not also have an outpatient visit with a diabetes diagnosis in FY 2013 (October 1, 2012 to September 30, 2013) (Figure 1).
Each patient was assigned a home VA facility as the clinic most frequently visited for primary care in FY 2012. To be retained in the cohort, patients had to have a home VA facility with at least 100 cohort members in FY 2012. The veteran's PCP was the provider most often visited in the home VA's primary care clinic in FY 2012. We excluded patients who most frequently saw a resident physician because of the dual responsibility of care held by both the resident and attending physician. We also excluded patients whose home VA facility was more than 1,000 miles from their home ZIP code or was not in one of the 50 US states or the District of Columbia. We did not include patients with inconsistent identifiers in the VA Corporate Data Warehouse or no information on body mass index (BMI). We examined only patients from VA facilities with at least two professions of primary care providers practicing in the facility. Finally, we excluded veterans whose PCP had fewer than 10 patients.
Panel and role definition
A panel-level dataset was constructed from the 609,668 including veterans with diabetes. Each panel was defined by the provider and the patients for whom the provider delivered the majority of primary care visits in FY 2012.
Provider role on the panel was defined by level of involvement on the panel. A categorical variable representing the profession of the PCP (that is, usual) and supplemental providers was divided into categories. Only one clinician was assigned as the PCP for each patient, and that clinician's profession was recorded as the PCP profession. However, several providers could serve in a supplementary role to a patient, so the supplemental provider categorization might represent more than one provider. However, the profession of supplemental providers can be represented by more than one individual.
- Physician only was defined as a panel of patients who had physicians as the usual provider of care and had 10% or less of their visits with any PA or NP within the home clinic.
- Physician PCP plus NP supplemental and physician PCP plus PA supplemental were defined as panels of patients for whom a physician was the usual provider of care and had at least 10% of visits with primary care NPs or PAs within the home clinic, but fewer than the usual provider.
- NP PCP plus physician supplemental and PA PCP plus physician supplemental were defined as panels of patients for whom an NP or PA was the usual provider of care and had at least 10% of visits with primary care physicians within the home clinic, but fewer than the usual provider.
- NP PCP no physician and PA PCP no physician were defined as panels of patients that had NPs or PAs as the usual provider of care and had less than 10% of their primary care visits with a physician.
We examined the relationship between the provider team composition and hemoglobin A1C, systolic BP, and low-density lipoprotein cholesterol (LDL-C). Continuous outcomes included the mean of all FY 2013 outpatient measurements of A1C, systolic BP, and LDL-C. Because of the large number of patients represented in the dataset, a priori clinical significance thresholds were set at 0.3% for A1C, 3 mm Hg for systolic BP, and 5 mg/dL for LDL-C, which we have previously used in research with this particular cohort.32
Variables with empirical evidence of association with provider type and patient outcomes were included as control variables in the model.25,32,41,42 Additional patient-level variables measured patient demographics, medical complexity, mental health diagnoses, and healthcare use.25,32,41,42 Categorical variables were used for race, ethnicity, marital status, and copay status (Table 1). Continuous variables were used for age, BMI, and distance of home address from clinic in miles. An indicator variable represented homelessness of the patient (1 = homeless at any point during the year). Patient medical complexity was measured by the prospective Diagnostic Cost Group (DCG) comorbidity measure calculated by the VHA, originally designed to predict cost of care but validated to measure medical complexity in the VHA population.43,44 The algorithm uses demographic and diagnostic information to assign each patient a DCG score, normed so that the average Medicare patient has a score equal to 1.45 Indicator variables were created for mental health diagnoses (post-traumatic stress disorder, mood disorder, substance abuse, dementia, and other mood disorder). Two healthcare use variables were also used. One patient-level use variable, percentage of visits with PCP, was calculated by dividing the total number of visits a patient had in a year with the usual provider by the total number of primary care visits in the home clinic.
Facility-level variables also were constructed and assigned to each patient based on the home clinic. Categorical variables represented the region of the United States (Northeast, Midwest, South, West) and metropolitan status of the clinic location (metropolitan area, micropolitan area, small town/ rural). Facility size was represented by a categorical variable of the number of patients seen at the facility in FY 2012 based on tertiles of facility size (size 1 = 704 to 2,379, size 2 = 2,384 to 5,582, size 3 = 5,602 to 43,738). The availability of an endocrinology specialty visit within the clinic was represented as a dichotomous variable. Our previous work found that PCP assignment was not associated with state scope of practice regulations. Therefore, we did not include variables for PA or NP scope of practice.42
Analyses were conducted using SAS 9.4 and SAS Enterprise Guide 7.1. All analyses were performed at the patient level. Hierarchical linear mixed models with random intercepts to account for clustering both by VHA facility and PCP were used to analyze continuous outcomes using SAS PROC MIXED. All models included each of the control variables described above.
A total of 609,668 patients with diabetes were identified in VHA primary care clinics. Patients were predominantly male (96.7%), non-Hispanic (90.7%), and white (71%). The mean DCG was 1, indicating that, on average, patients were as complex as the average Medicare patient (Table 1). Physicians were the usual providers of care for about 77% of the patients and PAs or NPs were the usual providers of care for about 22% of the patients. About 14% of patients received all of their primary care from PAs or NPs (Table 2).
Statistical differences in intermediate diabetes outcomes by usual and supplemental provider type were observed, but all were below the a priori thresholds for clinical significance (Table 3). When compared with physician-only care, physician PCP plus NP supplemental care had an average of 0.61 mg/dL higher LDL-C (95% CI 0.22, 0.99). Physician PCP plus PA supplemental care demonstrated, on average, 0.04% higher A1C (95% CI 0.02, 0.07) and 0.08 mm Hg lower systolic BP (95% CI -0.30, 0.14). Similar patterns were seen when NPs and PAs were the PCP and physicians were providing supplemental care. When compared with physician-only care, NP PCP plus physician supplemental care had an average of 1.9 mg/dL increase in LDL-C (95% CI 1.51, 2.29). PA PCP plus physician supplemental demonstrated an average of 0.03% increase in A1C (95% CI 0.004, 0.06) and 1.4 mg/dL increase in LDL-C (95% CI 0.82, 1.99). Similar outcomes for patients who only saw NPs or PAs in primary care were observed. When compared with physician-only care, patients with NP PCP with no physician involvement, on average, had 0.06% lower A1C (95% CI -0.07, -0.04), 0.39 mm Hg lower systolic BP (95% CI -0.52, -0.26), and 0.81 mg/dL higher LDL-C (95% CI 0.51, 1.12). Patients with PA PCP with no physician involvement had 0.03% lower A1C (95% CI -0.06, -0.01) and 0.53 mg/dL higher LDL-C (95% CI 0.06, 0.99).
Using diabetes as a model, this study examined whether patients with different types of usual and supplemental primary care providers have different chronic illness outcomes than patients seen only by physicians. No clinically meaningful differences in intermediate diabetes outcomes were observed based on the profession of either the usual or supplemental provider.
Although this study adds additional support to the conclusions of previous studies that patient outcomes are similar for those with physicians, NPs, and PA as PCPs, it also contributes in two other important aspects.30,32 First, the methodological approach used more closely reflects the reality that team-based care often means that supplemental providers frequently provide care to patients, and that these supplemental providers can affect quality of care delivered.24,31 Unlike the other study that compared the effectiveness of usual and supplemental provider types, no clinically significant differences in patient outcomes were observed. These differences in study findings could be due to a variety of factors, including differences in patient populations, institutional policies, team designs, and approaches to coordination between clinicians.8,46-48
The second unique contribution is our finding that among the 14% of patients who received care only from PAs or NPs (that is, patients who had less than 10% physician involvement in their primary care), outcomes were not clinically significantly different from those of patients with physician-only care (that is, patients who had less than 10% of PA or NP involvement in their primary care). Although these findings are similar to the findings of the other studies, this is the first time effectiveness of PA-only and NP-only care have been assessed in a national sample.27,31,32 The capacity to identify a significant number of patients who received primary care only from PAs and NPs likely reflects the VHA's expanded use of NPs and PAs as team leaders in their primary care team model. The finding that patients receiving primary care PA-only and NP-only care have such similar outcomes to patients receiving physician-only primary care indicates possible potential for improved efficiency, by means of transferring primary care of even moderately complex patients to less-expensive providers. However, such a disruptive change might be challenging for healthcare delivery organizations. Although studies indicate that patients generally are satisfied with care delivered by NPs and PAs, some patients prefer to have physicians as their usual providers.25,49-51 These patient preferences, paired with professional dominance of medicine, may make such a significant change to the social and economic organization of primary care difficult.52
This study has several strengths and limitations. The study examined outcomes among more than 600,000 patients in more than 500 primary care clinics across the United States with uniform scope of practice. This is ideal for controlling for variation in state and regional characteristics. Similarly, the patients in the VHA are predominantly male and generally more complex than similar patients outside the VHA. The complexity of the patients makes this sample ideal for understanding the clinical capacities of primary care physicians, NPs, and PAs. Despite these positives, the findings may not be generalizable to patients with chronic illness in other healthcare systems or with other sociodemographic characteristics.
A second limitation is that despite the evidence suggesting that NPs and PAs consult with their collaborating physicians relatively rarely in the VHA primary care setting, we cannot rule out undocumented consultations with physicians by PAs and NPs that potentially affected the quality of care provided.53
Another limitation is that PA, NP, and physician supplemental roles are a composite value of each profession serving on the panel. Multiple PAs, NPs, and physicians provide care to some panels, so we are unable to determine their individual roles, which may vary.
Redesigning primary care to meet the needs of patients with chronic illness is a focus for many healthcare organizations, and incorporating a range of clinicians, including physicians, NPs, and PAs, is a commonly proposed strategy.21,23 With growing evidence that the quality of care delivered by the professions is similar, this wide range in implementation of roles likely reflects organizational goals and cultures rather than the capacities of each profession.28-32 The results of this study suggest that physicians, PAs, and NPs can perform a variety of roles that can reflect the needs of the organization and patient population while maintaining quality of care. Although this adaptability can provide great benefit to healthcare delivery organizations and the system in general, conclusions regarding the “ideal” role should be approached with caution. More research on the effect of each role on access, cost, quality of care, and satisfaction of providers and patients is needed before answers will become clear.
1. Dall TM, Gallo PD, Chakrabarti R, et al. An aging population and growing disease burden will require a large and specialized health care workforce by 2025. Health Aff (Millwood)
2. Hung WW, Ross JS, Boockvar KS, Siu AL. Recent trends in chronic disease, impairment and disability among older adults in the United States. BMC Geriatr
3. Ortman JM, Velkoff VA, Hogan H. An Aging Nation: The Older Population in the United States. Report P25-1140. Washington, DC: US Census Bureau; 2014.
4. Hofer AN, Abraham JM, Moscovice I. Expansion of coverage under the Patient Protection and Affordable Care Act and primary care
utilization. Milbank Q
5. Wagner EH. Chronic disease management: what will it take to improve care for chronic illness. Eff Clin Pract
6. Wagner EH. The role of patient care teams in chronic disease management. BMJ
7. American College of Physicians. Joint Principles of the Patient-Centered Medical Home (PCMH). www.acponline.org/acp_policy/policies/joint_principles_pcmh_2007.pdf
. Accessed February 20, 2019.
8. Jackson GL, Powers BJ, Chatterjee R, et al. The patient centered medical home. A systematic review. Ann Intern Med
9. Bodenheimer T, Pham HH. Primary care
: current problems and proposed solutions. Health Aff (Millwood)
10. Østbye T, Yarnall KS, Krause KM, et al. Is there time for management of patients with chronic diseases in primary care
. Ann Fam Med
11. Green LV, Savin S, Lu Y. Primary care
physician shortages could be eliminated through use of teams, nonphysicians, and electronic communication. Health Aff (Millwood)
12. Biddle BJ, Thomas EJ, eds. Role Theory: Concepts and Research
. New York, NY: John Wiley & Sons, Inc.; 1966.
13. Salas E, Cooke NJ, Rosen MA. On teams, teamwork, and team
performance: discoveries and developments. Hum Factors
14. Katz D, Kahn RL. The Social Psychology of Organizations
. New York, NY: John Wiley & Sons, Inc.; 1966.
15. Everett CM, Thorpe CT, Palta M, et al. The roles of primary care
PAs and NPs caring for older adults with diabetes
16. Mittman DE, Cawley JF, Fenn WH. Physician assistants in the United States. BMJ
17. Sibbald B, Laurant M, Scott A. Changing task profiles. In: Saltman RB, Rico A, Boerma, eds. Primary Care in the Driver's Seat? Organizational Reform in European Primary Care
. Berkshire, England: Open University Press; 2006.
18. Pittman P, Leach B, Everett C, et al. NP
and PA privileging in acute care settings: do scope of practice laws matter. Med Care Res Rev
19. Hooker RS, Everett CM. The contributions of physician assistants in primary care
systems. Health Soc Care Community
20. Starfield B. Primary Care: Balancing Healthy Needs, Services, and Technology
. New York City, NY: Oxford University Press; 1998.
21. Yarnall KS, Østbye T, Krause KM, et al. Family physicians as team
leaders: “time” to share the care. Prev Chronic Dis
22. Yarnall KS, Pollak KI, Østbye T, et al. Primary care
: is there enough time for prevention. Am J Public Health
23. Institute of Medicine. Primary Care: America's Health in a New Era
. Washington, DC, 1996.
24. Everett CM, Thorpe CT, Palta M, et al. Division of primary care
services between physicians, physician assistants, and nurse practitioners for older patients with diabetes
. Med Care Res Rev
25. Everett CM, Morgan P, Jackson GL. Primary care physician assistant
and advance practice nurses roles: patient healthcare utilization, unmet need, and satisfaction. Healthc (Amst)
26. Smith VA, Morgan PA, Edelman D, et al. Utilization and costs by primary care
provider type: are there differences among patients of physicians, NPs, and PAs. Health Aff (Millwood)
. In press.
27. Mundinger MO, Kane RL, Lenz ER, et al. Primary care
outcomes in patients treated by nurse practitioners or physicians: a randomized trial. JAMA
28. Lutfiyya MN, Tomai L, Frogner B, et al. Does primary care diabetes
management provided to Medicare patients differ between primary care
physicians and nurse practitioners. J Adv Nurs
29. Kuo YF, Goodwin JS, Chen NW, et al. Diabetes
mellitus care provided by nurse practitioners vs primary care
physicians. J Am Geriatr Soc
30. Virani SS, Akeroyd JM, Ramsey DJ, et al. Comparative effectiveness of outpatient cardiovascular disease and diabetes
care delivery between advanced practice providers and physician providers in primary care
: implications for care under the Affordable Care Act. Am Heart J
31. Everett C, Thorpe C, Palta M, et al. Physician assistants and nurse practitioners perform effective roles on teams caring for Medicare patients with diabetes
. Health Aff (Millwood)
32. Jackson GL, Smith VA, Edelman D, et al. Intermediate diabetes
outcomes in patients managed by physicians, nurse practitioners, or physician assistants: a cohort study. Ann Intern Med
33. Woodmansee DJ, Hooker RS. Physician assistants working in the Department of Veterans Affairs. JAAPA
34. Robinson KR. Roles of nurse practitioners in the US Department of Veterans Affairs. In: Sullivan-Marx EM, McGivern DO, Fairman JA, Greenberg SA, eds. Nurse Practitioners: The Evolution and Future of Advanced Practice
. 5th ed. New York, NY: Springer Publishing Co.; 2010.
35. Klein S. The Veterans Health Administration: Implementing Patient-Centered Medical Homes in the Nation's Largest Integrated Delivery System. The Commonwealth Fund, Report No. 1537, September 2011.
36. Veterans Health Administration. Office of the Assistant Deputy Under Secretary for Health for Policy and Planning. Selected VHA Statistics: FY 2012-2013 End of Year. December 19, 2013.
37. Morgan PA, Abbott DH, McNeil RB, Fisher DA. Characteristics of primary care
office visits to nurse practitioners, physician assistants and physicians in United States Veterans Health Administration facilities, 2005 to 2010: a retrospective cross-sectional analysis. Hum Resour Health
38. Nelson KM, Helfrich C, Sun H, et al. Implementation of the patient-centered medical home in the veterans health administration: associations with patient satisfaction, quality of care, staff burnout, and hospital and emergency department use. JAMA Intern Med
39. Piette JD, Holtz B, Beard AJ, et al. Improving chronic illness care for veterans within the framework of the patient-centered medical home: experiences from the Ann Arbor Patient-Aligned Care Team
Laboratory. Transl Behav Med
40. Yano B, Stark R, Fihn S, et alPatient aligned care teams (PACT) demonstration lab initiative. Veterans Affairs Health Services Research and Development meeting, Washington, DC, 2011.
41. Everett CM, Schumacher JR, Wright A, Smith MA. Physician assistants and nurse practitioners as a usual source of care. J Rural Health
42. Morgan P, Everett CM, Smith VA, et al. Factors associated with having a physician, nurse practitioner, or physician assistant
as primary care
provider for veterans with diabetes
43. Ellis RP, Ash A. Refinements to the Diagnostic Cost Group (DCG) model. Inquiry
44. Maciejewski ML, Liu CF, Derleth A, et al. The performance of administrative and self-reported measures for risk adjustment of Veterans Affairs expenditures. Health Serv Res
45. Asch SM, McGlynn EA, Hogan MM, et al. Comparison of quality of care for patients in the Veterans Health Administration and patients in a national sample. Ann Intern Med
46. Lemieux-Charles L, McGuire WL. What do we know about health care team
effectiveness? A review of the literature. Med Care Res Rev
47. Shortell SM, Wu FM, Lewis VA, et al. A taxonomy of accountable care organizations for policy and practice. Health Serv Res
48. Press MJ, Rajkumar R, Conway PH. Medicare's new bundled payments: design, strategy, and evolution. JAMA
49. Roblin DW, Becker ER, Adams EK, et al. Patient satisfaction with primary care
: does type of practitioner matter. Med Care
50. Laurant M, van der Biezen M, Wijers N, et al. Nurses as substitutes for doctors in primary care
. Cochrane Database Syst Rev
51. Dill MJ, Pankow S, Erikson C, Shipman S. Survey shows consumers open to a greater role for physician assistants and nurse practitioners. Health Aff (Millwood)
52. Friedson E. Professional Dominance: The Social Structure of Medical Care
. New Brunswick, NJ: Transaction Publishers; 1970.
53. Fletcher CE, Copeland LA, Lowery JC, Reeves PJ. Nurse practitioners as primary care
providers within the VA. Mil Med