Policy Evaluation Of The Affordable Care Act Graduate Nurse Education Demonstration

University of Pennsylvania, Philadelphia, Pennsylvania.

Linda H. Aiken

University of Pennsylvania.

Barbara Todd

Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania.

Regina Cunningham

Hospital of the University of Pennsylvania.

Heather Brom

Villanova University, Villanova, Pennsylvania.

Morgan E. Peele

University of Pennsylvania.

Matthew D. McHugh

University of Pennsylvania.

Abstract

The US is experiencing a shortage of primary care providers, which could be reduced by the addition of nurse practitioners. However, the ability to increase the supply of nurse practitioners is limited by a shortage of clinical preceptors. The Affordable Care Act’s Graduate Nurse Education Demonstration provided federal funding to offset the clinical training costs of advanced practice nurses. We used data from the American Association of Colleges of Nursing from academic years 2005–06 through 2017–18 to determine whether the demonstration was associated with increased nurse practitioner enrollment and graduation growth. The demonstration was associated with a significant increase in nurse practitioner enrollments and graduations. A promising model of organizing and implementing funding for graduate nursing education nationally was identified. Findings suggest that modernizing Medicare payments for nursing education to support nurse practitioner clinical training costs is a promising option for increasing primary care providers.

One in four Americans, or more than eighty million people, lacked a primary care provider in 2015. 1 The Health Resources and Services Administration projects a 17 percent increase in primary care physician demand yet only 11 percent net growth between 2013 and 2025. 2 Only 8 percent of medical residents matched with primary care residencies in 2018. 3 Evidence suggests that the primary care shortage can be ameliorated by increased numbers of nurse practitioners (NPs). 4-6 NPs have been found in many studies to provide safe and effective care and to achieve high patient satisfaction. 7,8 Although the national supply of NPs has been growing, with more than 325,000 NPs actively licensed in the US, 9 the need for primary care providers continues to exceed their availability.

The Affordable Care Act included $200 million for the Graduate Nurse Education (GNE) Demonstration. Its purpose was to determine whether payments to hospitals for preceptors for NP and other advanced practice nursing students (nurse midwives, nurse anesthetists, and clinical nurse specialists) would enable schools of nursing to expand enrollments and increase graduations, especially of primary care providers. We report here research results on the outcomes of the GNE Demonstration and the feasibility of a new model of geographic consortia tested in the demonstration to operationalize and distribute NP training funds in the future.

Graduate Nurse Education Demonstration

The GNE Demonstration aimed to increase the supply of advanced practice nurses, particularly in primary care, by increasing the participation of clinical preceptors in graduate degree programs. To operationalize this goal, the Centers for Medicare and Medicaid Services (CMS), from July 1, 2012, through June 30, 2018, administered funds to five competitively selected hospitals partnering with a total of nineteen schools of nursing to compensate for reasonable costs incurred while the schools were providing approved clinical training for advanced practice nurses, the vast majority of whom were NPs. 10 The GNE Demonstration placed particular emphasis on growing the primary care provider supply, particularly in underserved areas. CMS thus required that at least 50 percent of GNE-funded precepting hours occur in community-based primary care settings, including federally qualified health centers. Approved clinical preceptors were practicing physicians and advanced practice nurses.

Unlike Medicare funding for medical residents, which is at the postgraduate level, the GNE Demonstration funded clinical training for licensed registered nurses in graduate degree-granting programs, including master’s and doctor of nursing practice (DNP) programs, that produced new-to-practice advanced practice nurses. As part of a degree-granting program, advanced practice nursing students must complete clinical hours in tandem with didactic coursework. The required number of hours varies by program and school of nursing, although it is consistently at least 500 hours for NP programs. CMS funding was limited to providing students with greater access to clinical preceptors and did not affect students’ out-of-pocket educational expenses, including tuition and fees. In addition to paying for clinical training, CMS approved participating schools of nursing to receive modest funds proportional to the number of students in the demonstration for their costs of rapidly expanding enrollments, including funds to update clinical databases and to hire staff to oversee clinical placements and manage site and preceptor recruitment.

The five hospitals selected for funding represented two models: a geographic consortium, in which multiple schools of nursing, hospitals and health systems, and office-based providers within a region were administered by a single hospital funded by CMS (the Hospital of the University of Pennsylvania, in Philadelphia; Memorial Hermann-Texas Medical Center, in Houston; and Scottsdale Healthcare Medical Center, in Arizona), and a single hospital model, in which a teaching hospital partnered with its primary affiliated school of nursing (Duke University Hospital, in Durham, North Carolina, and Rush University Medical Center, in Chicago, Illinois). The Philadelphia and Arizona sites included all schools of nursing with advanced practice nursing programs in their defined geographic target areas; the Houston site included multiple (four) nursing schools, but not all of those with advanced practice nursing programs in the Texas Gulf Coast region.

The Greater Philadelphia Graduate Nurse Education Network site administered by the Hospital of the University of Pennsylvania was the largest funded GNE Demonstration site with preceptor partnerships, with all health systems, many private physician practices, and all nine schools of nursing in the region offering graduate clinical education to advanced practice nurses. This site accounted for nearly 50 percent of the total number of GNE Demonstration schools of nursing. The Scottsdale Healthcare Medical Center site comprised four schools of nursing and spanned the entire state of Arizona. The Memorial Hermann site comprised four schools of nursing in the Texas Gulf Coast region. There were 5,248 total enrolled advanced practice nursing students in participating schools of nursing in the five sites at the outset of the demonstration, which grew to 6,661 students by the fourth demonstration year. 11,12

Challenges Of Increasing Primary Care NP Supply

Nursing educators report that their inability to recruit enough clinical preceptors is the major factor constraining enrollment growth. 11,13-15 Some 9,000 applications from qualified applicants to master’s level nursing programs in 2020 were declined, 14 primarily because of preceptor and faculty shortages. 13,15 The number of applications declined is an overestimate of the actual number of students declined, as students can submit multiple applications.

Recently, the number of health professional students who require clinical preceptors—including medical, nursing, physician assistant, and others—has increased 2,3,9 at the same time that primary care providers are facing growing productivity pressures to see more patients faster, which may be a disincentive to serve as a preceptor for students. The same markets that have numerous schools of nursing with many students needing clinical placements are often the most competitive for health care provider organizations. This creates productivity pressures on potential clinical preceptors that could adversely affect the willingness of medical practices to accept learners.

It is increasingly common for health care organizations to require payment for their professionals to serve as preceptors, creating a problem for enrollment growth, as financing of NP training programs is primarily done by students themselves with limited support from student aid and employer benefits. There are no permanent federal funds to support education costs for advanced practice nurses. 16 Medicare funds for nursing education go mostly to pre-RN licensure diploma nursing programs and account for only about 15 percent of direct graduate medical education funding. 10,16,17 The federal government funds more than $15 billion annually, 80 percent of which comes from Medicare, to hospitals for graduate medical education to offset direct and indirect costs incurred for resident physician training. 18 Yet the national supply of primary care physicians increased only by 18,051 between 2010 and 2016, compared with an increase of 63,874 in the primary care NP supply. 19

These factors were the motivation for the GNE Demonstration. Would the availability of CMS payments through hospitals to clinical preceptors for advanced practice nursing students be associated with an increase in enrollments and graduations, especially in primary care? In this study we examined these GNE Demonstration outcomes while correcting for an important weakness in the CMS-commissioned study on the demonstration 11 by accounting for the impact of market competition on student enrollments and graduations. Moreover, our study focused on the NP workforce, which is critical for increasing the primary care provider supply. In this policy research article we show findings indicating that payments to hospitals to expand preceptors for advanced practice nurse clinical training are effective in increasing enrollments and graduations of primary care providers. We discuss these results in the context of the feasibility, applicability, and benefit of a potential nationwide Graduate Nurse Education Medicare benefit to expand the primary care NP workforce to improve access to primary care and the identification of a geographic consortia model that is successful in emphasizing primary care training.

Study Data And Methods

DATA AND SAMPLE

We used data from the American Association of Colleges of Nursing Annual Survey from academic years 2005–06 through 2017–18, which was administered to deans of accredited, four-year schools of nursing. 20 The survey response rate is higher than 92 percent. The data set included annual enrollments, graduations, and school characteristics for most schools of nursing with baccalaureate and graduate programs. After we eliminated schools of nursing without data during the study period and five outlier schools of nursing with 500 or more NP graduates annually—largely online programs—our analytic sample was 334 schools of nursing, each of which had data across thirteen time points—one point per academic year. Of these schools, nineteen were GNE Demonstration schools of nursing and 315 were non–GNE Demonstration schools of nursing. To be considered an NP graduate or enrollee, a student must have been in an initial pre-NP licensure program, as stipulated by CMS for GNE Demonstration benefit eligibility. This eliminated those who were already advanced practice nurses seeking additional qualifications (for example, doctor of nursing practice).

The 2011 Dartmouth Atlas Hospital Referral Regions data set provided information on regional health care markets for tertiary medical care, which are known as hospital referral regions (HRRs). Each of the 306 HRRs in the country contains at least one hospital that performs major cardiovascular procedures and neurosurgery. 21 HRRs are commonly used to define boundaries of health services markets. We used data from the 2011 Medicare Healthcare Provider Cost Reporting Information System, a data set with information on total hospital discharges, to define market competition. 22

OUTCOMES

The annual number of NP enrollments and graduations per school of nursing was derived from the American Association of Colleges of Nursing Annual Survey.

INDEPENDENT VARIABLE

We divided schools of nursing into a GNE Demonstration treatment group and a comparison group of non–GNE Demonstration schools of nursing, each of which had annual data represented across thirteen time points. The pre–GNE Demonstration period consisted of the academic years from 2005–06 through 2011–12. The GNE Demonstration years were the academic years from 2012–13 through 2017–18. The parameter of primary importance for us was the interaction of these variables (treatment/comparison schools of nursing × predemonstration/demonstration period).

COVARIATES

We adjusted for four covariates: size, academic health center affiliation, and market competition among the schools of nursing and competition in the health services market. Two of these covariates (size and academic health center affiliation) were derived from the American Association of Colleges of Nursing Annual Survey. To account for school of nursing size, we created a categorical variable dividing enrollment into annually varying, equally proportioned terciles: small, medium, and large. Academic health center affiliation was categorized as “affiliated” and “nonaffiliated.”

We used the Dartmouth Atlas and American Association of Colleges of Nursing data sets to create a measure of school of nursing market competition, which is meant to reflect the relative difficulty of recruiting NP preceptors in a region and of attracting NP students to schools of nursing, both of which can affect increases in NP graduates over time. Accordingly, we calculated a Herfindahl-Hirschman Index—a measure of market concentration based on the share of NP students in each school of nursing within an HRR. 23 We also used the Healthcare Provider Cost Reporting Information System and Dartmouth Atlas data sets to calculate a health services Herfindahl-Hirschman Index representing competition between health care systems for patients within a given HRR, which we believe is an indicator of productivity pressures on potential preceptors. To prevent productivity loss, preceptors and health systems may be less likely to accept trainees in competitive health service markets.

STATISTICAL ANALYSIS

In our analysis we first examined changes in enrollments and graduations over time between GNE Demonstration and non–GNE Demonstration schools of nursing and analyzed their characteristic differences, using summary statistics, t-tests, and chi-square tests. We used a difference-in-differences approach to estimate the average change in NP enrollments and graduations at GNE Demonstration versus non–GNE Demonstration schools of nursing during the GNE Demonstration period (academic years 2012–13 through 2017–18) versus the pre–GNE Demonstration period (academic years 2005–06 through 2011–12). Because a difference-in-differences approach compares averages of all GNE Demonstration and non–GNE Demonstration schools of nursing with their respective trends in the pre–GNE Demonstration and GNE Demonstration periods, it helped reduce potential biases from unmeasured confounders. Because post–GNE Demonstration implementation effects are compared with both treatment and control groups’ own baselines before GNE Demonstration implementation, changes over time that affect all schools of nursing (for example, an aggregate increasing trend in NP students nationwide) are differentiated from GNE Demonstration effects. A difference-in-differences approach assumes that growth size is unaffected by preintervention graduate and enrollment counts, so we adjusted for variables that may affect the magnitude of growth. 24 Our unadjusted and adjusted regression models both controlled for a year indicator variable. The adjusted model also controlled for school of nursing characteristics (size and academic health center affiliation) and school of nursing and health services market competition.

LIMITATIONS

Our study was hampered by limitations of the demonstration design, including the short funding period of four years, which was stretched to six years on a no-cost extension; small number of funded sites (N = 5) and participating schools of nursing (N = 19); and preceptor payments only for the number of students that exceeded baseline enrollments rather than all students. Because of the small number of hospitals funded and participating schools of nursing, caution is warranted in generalizing the results.

Study Results

We display average NP graduate and enrollment trends in GNE Demonstration and non–GNE Demonstration schools of nursing from the 2005–06 through 2018–19 academic years ( exhibit 1 ). NP enrollment and graduation averages in the two groups of schools of nursing increased at a relatively similar rate in the predemonstration period. When we compared the demonstration period with the predemonstration period, growth in NP enrollments and graduations was greater at GNE Demonstration schools of nursing than at non–GNE Demonstration schools of nursing.

Average nurse practitioner enrollments and graduations at Graduate Nurse Education (GNE) Demonstration versus non–GNE Demonstration schools of nursing, by academic year, 2006–19

SOURCE Authors’ analysis of American Association of Colleges of Nursing Annual Survey data. NOTE Years represent the end of an academic year (that is, 2006 signifies the 2005–06 academic year).

Exhibit 2 shows that GNE Demonstration schools of nursing tended to be larger and less often affiliated with an academic health center, and they were in more competitive school of nursing and health services markets. Exhibit 3 displays changes in average enrollments and graduations at the two groups of schools of nursing. There was larger relative growth in enrollments and graduations at GNE Demonstration versus non–GNE Demonstration schools of nursing during the demonstration period.

EXHIBIT 2

Descriptive statistics of Graduate Nurse Education (GNE) Demonstration versus non–GNE Demonstration schools of nursing, 2011

SOURCE Authors’ analysis of a 2011 cross-section of American Association of Colleges of Nursing Annual Survey, Healthcare Provider Cost Reporting Information System, and Dartmouth Atlas data sets. NOTES We describe these variables using a 2011 cross-section so as not to inaccurately inflate the statistical significance of time-fixed covariates. p value is

a Measured using Herfindahl-Hirschman Indexes within hospital referral regions. Herfindahl-Hirschman Indexes range from 0 to 10,000. A Herfindahl-Hirschman Index less than 1,500 is considered highly competitive, 1,500–2,500 is considered moderately competitive, and higher than 2,500 is considered less competitive. See Department of Justice. Herfindahl–Hirschman Index (note 24 in text).

EXHIBIT 3

Descriptive statistics of enrollments and graduations at Graduate Nurse Education (GNE) Demonstration versus non-GNE Demonstration schools of nursing, by period, 2006–18

SOURCE Authors’ analysis of American Association of Colleges of Nursing Annual Survey data. NOTE This exhibit displays an average of graduates and enrollments at GNE Demonstration and non–GNE Demonstration schools of nursing in the predemonstration period (academic years 2005–06 through 2011–12) versus the demonstration period (academic years 2012–13 through 2017–18).

Results from the difference-in-differences models revealed that when health services and school of nursing market competition and other school characteristics were adjusted for, the GNE Demonstration was associated with a significant average increase of twenty-eight NP graduates and eighty-nine NP enrollees per school of nursing throughout the demonstration period (p < 0.001) (exhibit 4 ). In the adjusted models, these coefficients translated to 67 percent and 76 percent relative increases in enrollments and graduations, respectively, at GNE Demonstration versus non–GNE Demonstration schools of nursing.

EXHIBIT 4

Average number of nurse practitioner enrollments and graduations at Graduate Nurse Education (GNE) Demonstration schools of nursing versus non-GNE Demonstration schools of nursing, by period, 2006–18

SOURCE Authors’ analysis of American Association of Colleges of Nursing Annual Survey, Healthcare Provider Cost Reporting Information System and Dartmouth Atlas data sets. NOTES N = 334 schools of nursing at 13 points in time. This exhibit shows results from difference-in-differences models. “GNE Demonstration school of nursing × demonstration period” is an interaction term between the number of enrollees and graduates at schools of nursing participating in the GNE Demonstration and the time period of the demonstration. More specifically, the two variables represented in the interaction are GNE Demonstration and non–GNE Demonstration school of nursing and GNE Demonstration versus pre–GNE Demonstration period. Both adjusted and unadjusted models account for a year indicator variable. The adjusted model controls for school size, academic health center affiliation status, school of nursing market competition, and health care market competition.

Discussion

The Graduate Nurse Education Demonstration was associated with a significant increase in NP enrollments and graduations, including in primary care, which was the stated goal of the demonstration. An additional important outcome of the demonstration was the establishment of the feasibility of CMS payments to hospitals to affect NP student enrollments in organizationally independent institutions of higher education and increased clinical training of NPs in community-based settings that had no organizational affiliation with the funded hospital. In the year after the demonstration close, when compared with the last year of GNE Demonstration funding, NP graduates and enrollments decreased by up to 5 percent at GNE Demonstration schools of nursing. Non–GNE Demonstration schools of nursing continued to increase as in prior years. Although we cannot make too much of the findings of a postdemonstration decline based on one year of data, the decline may be further evidence of the impact of preceptor funding on enrollments and graduations, especially in view of growing external factors that diminish the likelihood of health care settings accepting students, including the consolidation of health services into larger systems with potentially weaker ties to local schools of nursing and competitive pressures on large health systems to increase their operating margins to fund acquisitions and mergers.

The GNE Demonstration schools of nursing attributed their growth in enrollments and graduations to increased preceptor availability associated with CMS preceptor payments. 11 Our analysis of the 2018–19 American Association of Colleges of Nursing Annual Survey data (data not shown) shows that 62 percent of schools of nursing that turned away qualified NP applications did so because of insufficient preceptors; these schools of nursing declined forty-eight qualified NP applications, on average, that year. From the beginning of the demonstration, the participating schools of nursing were concerned that the introduction of preceptor payments for such a short period, even though needed, would set a precedent that would adversely affect their ability to recruit preceptors without preceptor funding after the demonstration ended. Thus, the schools of nursing were conservative about increasing enrollments, which in turn likely affected enrollments and graduations toward the end of the demonstration, as shown in exhibit 1 .

A survey of individual preceptors at the Greater Philadelphia Graduate Nurse Education Network site revealed that payment was appreciated but not essential for their participation. 25 It is important to note, however, that the preceptors’ organizations received the demonstration payments and not necessarily the individual preceptors. Indeed, some administrators of large ambulatory practices that were a major source of preceptors demanded direct payments from the schools of nursing to continue serving as preceptors after the demonstration ended. The results of the preceptor survey suggest that the pressure for payments for clinical preceptors is more likely to come from practice administrators rather than from preceptors. 25

The demonstration prioritized developing opportunities for NP students to be placed in federally qualified health centers, although these health centers are mostly small, with only a few exam rooms, and cannot accommodate many students. The objective of rapidly increasing primary care NP enrollments thus had to be met primarily though large, private ambulatory care settings.

Our results that show significant enrollment and graduation increases are consistent with the findings of the CMS-commissioned GNE Demonstration evaluation conducted as a requirement of federal demonstrations. 11 However, our design corrected for an important weakness in the CMS-commissioned study design: a lack of consideration of how market competition affected NP student enrollments and graduations. We accounted for both health services and school of nursing market competition, which may be critical factors affecting preceptor participation, although there is no previous research documenting this. Our results show that GNE Demonstration schools of nursing were, on average, located in more competitive health services and school of nursing markets. Once health services competition, school of nursing market competition, and other school characteristics were adjusted for, GNE Demonstration schools of nursing had significantly higher enrollment and graduation increases than non–GNE Demonstration schools of nursing.

The GNE Demonstration also fostered training and postgraduation employment in underserved areas and community-based primary care settings. Seventy-six percent of GNE Demonstration clinical hours were completed in community-based primary care settings in 2016—even more than required by CMS. 11 At the Greater Philadelphia Graduate Nurse Education Network site, we collected employment and training data from 1,432 NP graduates between the 2013–14 and 2017–18 academic years (data not shown). These data showed that 75 percent of Greater Philadelphia Graduate Nurse Education Network NP graduates trained in primary care, and of those with jobs at the time of survey completion, 66 percent practiced in ambulatory primary care settings. These data collected by the GNE Demonstration sites and the significant increase in enrollments and graduations associated with the demonstration suggest that a financial incentive for NP preceptors, if implemented nationwide, may be a promising strategy for increasing the supply of primary care providers and improving their distribution in community-based sites.

The demonstration allowed preceptor payments only for the number of students above baseline enrollment, as its objective was to increase advanced practice nurse supply. The total cost for training each additional advanced practice nursing student over baseline enrollment in the GNE Demonstration was $47,172, 11 which was calculated based on the amounts paid by CMS for professionals serving as preceptors, administrative costs, and funds to schools of nursing for the costs of rapidly expanding enrollments. The cost to train each additional advanced practice nursing student in the GNE Demonstration ($47,172) was about 30 percent of the median annual cost of training each primary care physician resident ($157,602) trained under the GME Teaching Health Center program. 10,11,26

The Graduate Nurse Education Consortia Model

There have been multiple publications and commission reports on graduate medical education focusing on the appropriateness of the supply of physicians, the shortage of primary care, the geographic maldistribution of providers, and the very high cost of Medicare graduate medical education payments. 27 An Institute of Medicine study recommended the establishment of geographic consortia that would receive and distribute Medicare graduate medical education funding to a range of organizations training physicians, including community-based primary care settings in addition to individual teaching hospitals—hospitals that partner with medical schools, schools of nursing, or other educational programs. 28 In the GNE Demonstration, the Philadelphia and Arizona sites met the definition of a “consortia model” as envisioned for graduate medical education. One of the objectives of the original geographic consortia idea was to produce more primary care physicians. This idea was never implemented for graduate medical education but was very successfully implemented in the GNE Demonstration, resulting in most trainees being in primary care—a result that graduate medical education has never achieved. The consortia model was also attractive to schools of nursing that had NP training programs but were not formally affiliated with a teaching hospital. In our analytic sample of 334 schools of nursing that train NPs and other advanced practice nurses, fewer than 30 percent of them were affiliated with an academic health center ( exhibit 2 ). The consortia model created a central planning mechanism not currently present in communities to allocate existing NP preceptor capacity across multiple schools of nursing and to develop greater capacity if needed, including the development of additional primary care preceptor sites—an important feature in geographic areas with high market competition, such as Philadelphia.

The purpose of the ACA demonstrations was to test the effectiveness and feasibility of innovations to improve various aspects of health care delivery with the possibility that the successful ones would be considered for national implementation. Because it was demonstrated that preceptor payment was an intervention that facilitated higher NP student enrollments and graduations, as reported here, there was also interest in what could be learned about the design and operationalization of a national rollout of preceptor payment for advanced practice nursing clinical training, possibly through a modernization of Medicare payments for nursing education. In previous work, the case has been made for shifting Medicare nursing education funding to graduate nursing education, as is the case for graduate medical education. 10,16,17,29 Advanced practice nurses and graduate education for clinical nurses had not yet developed when Medicare was passed in 1965, which is a reason that graduate nursing education has not been part of Medicare funding along with graduate medical education. Most current Medicare funding of nurse education goes to hospital diploma nursing programs training new RNs, and little funding goes to NP and other advanced practice nursing graduate clinical training.

The success of the Graduate Nurse Education Consortia Model suggests a possible mechanism for distributing Medicare funds nationally for graduate nursing education using either existing or new funding. In keeping with Medicare policy, funds must go to an organization providing care to Medicare beneficiaries rather than to schools of nursing. The Graduate Nurse Education Consortia Model meets that requirement, as it is operated by an eligible hospital that currently receives graduate medical education funding but with a scope that includes preceptor payment in all settings and for eligible trainees in all qualifying NP programs in schools of nursing in the designated region. To operationalize the Graduate Nurse Education Consortia Model for a national program, HRRs could be used to designate geographic areas that would cover the nation, with a single eligible hospital acting as a hub for the consortium.

Our geographic analyses (data not shown) of the American Association of Colleges of Nursing and Dartmouth Atlas data sets revealed that of 304 HRRs in the contiguous US, 61 percent contain both a teaching hospital and a school of nursing with an NP program. The 334 schools of nursing in our sample offering an NP program are located in 193 HRRs. Only 5 percent of HRRs (n = 9) containing a school of nursing with an NP program do not also contain a teaching hospital. A consortia model would fit both urban and geographically larger, rural HRRs. The urban Greater Philadelphia Graduate Nurse Education Network was encompassed in a single HRR; the Scottsdale Healthcare Medical Center was the most geographically far-reaching GNE Demonstration site, involving partnerships with clinical training sites and schools of nursing throughout the entire state of Arizona. Health systems with potential clinical training sites spanning multiple HRRs exist in most rural HRRs.

Conclusion

The ACA Graduate Nurse Education Demonstration was successful in its objective of increasing the supply of NPs in primary care. The demonstration established the potential feasibility of Medicare funding to hospitals to support clinical training in graduate nurse education in degree-granting programs in educational institutions. In addition, it confirmed the feasibility and value of operationalizing Medicare funding by establishing regional consortia, an option proposed for graduate medical education decades ago by the Institute of Medicine but never previously implemented. Modernizing Medicare funding of nurse education is long overdue, and lessons from the GNE Demonstration should be used to inform policy changes.

Acknowledgments

Support for this research was received from the Centers for Medicare and Medicaid Services (CMS); the National Institute of Nursing Research, National Institutes of Health (Grant No. T32NR007104 to Linda Aiken); and the US-Israel Fulbright Postdoctoral Scholars Program (Joshua Porat-Dahlerbruch). The authors are solely responsible for the content of the article. The authors thank their colleagues at the five Graduate Nurse Education (GNE) Demonstration sites, especially Michael Rossi at the Hospital of the University of Pennsylvania; Patricia Hercules and Lori Hull-Grommesh at Memorial Hermann-Texas Medical Center and Patricia Stark at University of Texas at Houston School of Nursing; Melanie Brewer at Scottsdale Healthcare Medical Center and Joan Shaver at the University of Arizona; Kathleen Delaney at Rush University Medical Center and College of Nursing; and Mary Ann Fuchs at Duke University Hospital and Beth Merwin at Duke University School of Nursing in Durham. The authors acknowledge the important contributions of stakeholder organizations including Suzanne Miyamoto and the American Association of Colleges of Nursing and Susan Reinhard, Winifred Quinn, and Peter Reinecke of AARP The authors are also grateful to the many people at CMS who helped them.

Contributor Information

Joshua Porat-Dahlerbruch, University of Pennsylvania, Philadelphia, Pennsylvania.

Linda H. Aiken, University of Pennsylvania.

Barbara Todd, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania.

Regina Cunningham, Hospital of the University of Pennsylvania.

Heather Brom, Villanova University, Villanova, Pennsylvania.

Morgan E. Peele, University of Pennsylvania.

Matthew D. McHugh, University of Pennsylvania.

NOTES

1. Levine DM, Linder JA, Landon BE. Characteristics of Americans with primary care and changes over time, 2002–2015 . JAMA Intern Med . 2020; 180 ( 3 ):463–6. [PMC free article] [PubMed] [Google Scholar]

2. Health Resources and Services Administration. National and regional projections of supply and demand for primary care practitioners: 2013–2025 [Internet] . Rockville (MD): HRSA; 2016. Nov [cited 2021 Nov 18]. Available from: https://bhw.hrsa.gov/sites/default/files/bureau-health-workforce/data-research/primary-care-national-projections-2013-2025.pdf [Google Scholar]

3. Pohl JM, Thomas A, Bigley MB, Kopanos T. Primary care workforce data and the need for nurse practitioner full practice authority . Health Affairs Blog [blog on the Internet] . 2018. Dec 13 [cited 2021 Nov 18]. Available from: https://www.healthaffairs.org/do/10.1377/hblog20181211.872778/full/ [Google Scholar]

4. Rohrer JE, Angstman KB, Garrison GM, Pecina JL, Maxson JA. Nurse practitioners and physician assistants are complements to family medicine physicians . Popul Health Manag . 2013; 16 ( 4 ):242–5. [PubMed] [Google Scholar]

5. Green LV, Savin S, Lu Y. Primary care physician shortages could be eliminated through use of teams, non-physicians, and electronic communication . Health Aff (Millwood) . 2013; 32 ( 1 ):11–9. [PubMed] [Google Scholar]

6. Pohl JM, Thomas A, Barksdale D, Werner K. Primary care workforce: the need to remove barriers for nurse practitioners and physicians . Health Affairs Blog [blog on the Internet] . 2016. Oct 26 [cited 2021 Nov 18]. Available from: https://www.healthaffairs.org/do/10.1377/hblog20161026.057250/full/ [Google Scholar]

7. Newhouse RP, Stanik-Hutt J, White KM, Johantgen M, Bass EB, Zangaro G, et al. Advanced practice nurse outcomes 1990–2008: a systematic review . Nurs Econ . 2011; 29 ( 5 ):230–50, quiz 251. [PubMed] [Google Scholar]

8. Maier CB, Aiken LH, Busse R. Nurses in advanced roles in primary care: policy levers for implementation [Internet] . Paris: Organization for Economic Cooperation and Development; 2017. [cited 2021 Nov 18]. Available from: https://www.oecdilibrary.org/social-issues-migration-health/nurses-in-advanced-roles-in-primary-care_a8756593-en [Google Scholar]

9. American Association of Nurse Practitioners. NP fact sheet [Internet] . Austin (TX): AANP; 2021. [last updated 2021 May; cited 2021 Nov 18]. Available from: https://www.aanp.org/about/all-about-nps/np-fact-sheet [Google Scholar]

10. Aiken LH, Dahlerbruch J, Todd B, Bai G. The Graduate Nurse Education Demonstration—implications for Medicare policy . N Engl J Med . 2018; 378 ( 25 ):2360–3. [PMC free article] [PubMed] [Google Scholar]

11. Centers for Medicare and Medicaid Services. Graduate Nurse Education Demonstration project: final evaluation report [Internet] . Baltimore (MD): CMS; 2019. Aug [cited 2021 Nov 18]. Available from: https://innovation.cms.gov/files/reports/gne-final-eval-rpt.pdf [Google Scholar]

12. Centers for Medicare and Medicaid Services. Evaluation of the GNE Demonstration project, volume I: implementation and impact [Internet] . Baltimore (MD): CMS; 2017. Oct [cited 2021 Dec 6]. Available from: https://innovation.cms.gov/files/reports/gne-rtc-vol1.pdf [Google Scholar]

13. American Association of Colleges of Nursing. Fact sheet: nursing faculty shortage [Internet] . Washington (DC): AACN; 2019. [last updated 2020 Sep; cited 2021 Nov 18]. Available from: https://www.aacnnursing.org/Portals/42/News/Factsheets/Faculty-Shortage-Factsheet.pdf [Google Scholar]

14. American Association of Colleges of Nursing [Internet]. Washington (DC): AACN. Press release, Student enrollment surged in U.S. schools of nursing in 2020 despite challenges presented by the pandemic ; 2021. Apr 1 [cited 2021 Nov 18]. Available from: https://www.aacnnursing.org/News-Information/Press-Releases/View/ArticleId/24802/2020-survey-data-student-enrollment [Google Scholar]

16. Aiken LH, Cheung RB, Olds DM. Education policy initiatives to address the nurse shortage in the United States . Health Aff (Millwood) . 2009; 28 ( 4 ):w646–56. [PMC free article] [PubMed] [Google Scholar]

17. Aiken LH, Gwyther ME. Medicare funding of nurse education. The case for policy change . JAMA . 1995; 273 ( 19 ):1528–32. [PubMed] [Google Scholar]

18. Congressional Budget Office. Consolidate and reduce federal payments for graduate medical education at teaching hospitals [Internet] . Washington (DC): CBO; 2018. Dec 13 [cited 2021 Nov 18]. Available from: https://www.cbo.gov/budget-options/54738 [Google Scholar]

19. Xue Y, Smith JA, Spetz J. Primary care nurse practitioners and physicians in low-income and rural areas, 2010–2016 . JAMA . 2019; 321 ( 1 ):102–5. [PMC free article] [PubMed] [Google Scholar]

20. American Association of Colleges of Nursing. Annual Surveys [Internet] . Washington (DC): AACN; 2021. [cited 2021 Nov 18]. Available from: https://www.aacnnursing.org/News-Information/Research-Data-Center/Annual-Surveys [Google Scholar]

21. Atlas Dartmouth. General FAQ [Internet] . Lebanon (NH): Dartmouth Atlas Project; 2019. [cited 2021 Nov 18]. Available from: https://www.dartmouthatlas.org/faq/ [Google Scholar]

22. Centers for Medicare and Medicaid Services. Healthcare Cost Report Information System [Internet] . Baltimore (MD): CMS; 2011. [last updated 2021 Oct 22; cited 2021 Nov 18]. Available from: https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Cost-Reports [Google Scholar]

23. Department of Justice. Herfindahl–Hirschman Index [Internet] . Washington (DC): DOJ; 2018. Jul 31 [cited 2021 Nov 18]. Available from: https://www.justice.gov/atr/herfindahl-hirschman-index [Google Scholar]

24. Daw JR, Hatfield LA. Matching and regression to the mean in difference-in-differences analysis . Health Serv Res . 2018; 53 ( 6 ):4138–56. [PMC free article] [PubMed] [Google Scholar]

25. Todd BA, Brom H, Blunt E, Dillon P, Doherty C, Drayton-Brooks S, et al. Precepting nurse practitioner students in the Graduate Nurse Education Demonstration: a crosssectional analysis of the preceptor experience . J Am Assoc Nurse Pract . 2019; 31 ( 11 ):648–56. [PMC free article] [PubMed] [Google Scholar]

26. Regenstein M, Nocella K, Jewers MM, Mullan F. The cost of residency training in teaching health centers . N Engl J Med . 2016; 375 ( 7 ):612–4. [PubMed] [Google Scholar]

27. Iglehart JK. Institute of Medicine report on GME—a call for reform . N Engl J Med . 2015; 372 ( 4 ):376–81. [PubMed] [Google Scholar]

28. Institute of Medicine. On implementing a national graduate medical education trust fund . Washington (DC): National Academies Press; 1997. [PubMed] [Google Scholar]

29. Aiken LH, Gwyther ME, Whelan EM. Federal support of graduate nursing education: rationale and policy options . Nurs Outlook . 1996; 44 ( 1 ):11–7. [PubMed] [Google Scholar]