The structure of healthcare delivery in the United States differs substantially from health systems in other well-resourced nations in several ways. The US system depends on broad use of physician assistants (PAs) and nurse practitioners (NPs) to expand access to medical services and relies on a highly variable range of payment sources. These sources include out-of-pocket payment, charity, private insurance, publicly provided or underwritten insurance, and federal services such as Medicare, Medicaid, public health, Indian Health Service, Bureau of Prisons, military, community health centers, and Veterans Health Administration (VHA).1,2 Now in the second decade of the new millennium, the US healthcare workforce faces complex challenges as the population ages, patients with chronic diseases live longer, and the Affordable Care Act (ACA) expands public and private health insurance.
Various supply and demand models address the resulting increase in need for services by examining the effect of potential changes to the traditional health workforce.3-6 One proposed solution is greater use of PAs and NPs.7,8 Although the number of US medical school graduates is expected to increase, the anticipated growth is proportionally small when compared with forecasted increases in the number of PA and NP graduates.9,10 As the healthcare system evolves and providers other than physicians increasingly deliver medical services, payment for services may become a critical policy issue.
Since the mid-1990s, reports have suggested that PAs and NPs manage a greater proportion of patient visits reimbursed by lower-payment options (such as Medicaid, Medicare, workers' compensation, charity, pro bono, and out-of-pocket). One national study revealed that visits paid by Medicaid and other government payment sources accounted for a proportionally larger share of visits to PAs and NPs compared with physicians.11 Another study of Medicare recipients found significant differences in types of payments for PA and NP visits compared with physician visits. Medicare-Medicaid dual enrollees more frequently reported that a PA or NP was their principal provider of care; those with private insurance in addition to Medicare more often identified a physician as their primary provider.12 Lin and colleagues found that the distribution of “expected source of payment” (private versus Medicare/Medicaid) for patient care was similar for NPs compared with PAs when either served as the sole providers of care.13
Outpatient hospital visits remunerated by private insurance account for 37% of all medical visits in the United States.2 Sixty-one percent of private physician office visits are compensated by private insurance.14 In community health centers (CHCs), where the majority of patients are either uninsured or have Medicaid, only 17% of office visits are paid for by private insurance.2,15 Notably, both CHCs and hospital outpatient departments use PAs and NPs at higher rates than do private family medicine physicians.16 In general, healthcare providers receiving relatively lower payments for medical services relative to private insurance are more likely to employ PAs and NPs than healthcare entities that primarily receive reimbursement from private insurance.17
This study sought to determine if the source of payment for a medical visit correlates with the likelihood that care will be provided by a specific type of provider. A possible correlation is of interest because studies have shown that source of payment influences physicians' decisions about the type of services to offer patients and that Medicaid influences patients' access to physician care.18-22 However, the potential for payment source to influence the likelihood of receiving services from a provider other than physician has not been fully addressed. One perception is that differentiation in the type of healthcare provider for a patient is partly predetermined by the source of payment, consistent with relative wage differences between providers. In other words, NPs and PAs may be more likely to provide care for patients when the payment is less relative to private health insurance.
Any effect of payment source is not strictly binding, as the relationship is likely moderated by the relative availability of physicians versus other types of providers. Rural areas, for example, contain a proportionally higher distribution of PAs and NPs and lower distribution of physicians than urban areas.16 To address the influence of access to a particular provider based on geography, an analysis of the relationship between provider and payment source based on location was undertaken, with regressions stratified by the five National Center for Health Statistics (NCHS) county categories of urbanization using the NCHS Urban-Rural Classification system (Table 1).
In outpatient settings, a number of factors likely determine whether a physician, PA, or NP sees a patient. If the propensity for a patient to be scheduled with a particular type of provider is in part determined by the patient's mode of payment, consideration may be given to the difference in marginal revenue gained from the patient's source of payment and the marginal cost of hourly wage for the provider being seen. The assumption is that physicians are paid a higher wage than NPs and PAs regardless of specialty or employment setting.23
Because Medicaid and Medicare typically reimburse at lower rates than private insurance, physicians may have less incentive to treat a patient whose primary source of payment is Medicaid.24 Taking this a step further, assuming there are no other influences and looking at the bottom of the payment scale, a physician would be more likely to treat a patient paying with Medicaid than a patient without the ability to pay. Similarly, patients with lower-reimbursing insurance or no insurance would be more likely to be treated by PAs and NPs because their marginal cost of labor is less than that of a physician. The premise is that employers of healthcare professionals will maximize their revenue stream when labor costs and reimbursement are optimized. This national study was undertaken to see whether PAs and NPs in hospital outpatient departments are in fact more likely to provide care for patients with lower levels of reimbursement.
Data were drawn from the National Hospital Ambulatory Medical Care Survey (NHAMCS) outpatient department visits for 2006-2010 (with repeated cross-sections). The study sample includes more than 123,000 observations representing more than 74 million outpatient visits. A more detailed description of the survey can be found on the CDC website (http://www.cdc.gov/nchs/ahcd.htm). The NHAMCS data in particular were chosen because PA and NP visits can be readily separated from physician visits. This is an important distinction from many other available data sources because incident-to services provided by PAs or NPs are not attributed to physicians. All analyses were completed using Stata/SE 13.1 for Windows (Stata Statistical Software: Release 12. College Station, TX: StataCorp LP; 2011).
The primary analysis was a multinomial logistic regression with type of provider as the dependent variable. For prevalence estimates, the survey was weighted by visit to reflect the design and sampling scheme of the NHAMCS for each year. Sample weights were used to approximate the prevalence estimates and adjust for the complex sampling design, which is described elsewhere.17 Predicted probabilities based on the final regression models are included in the results section (Table 2).
The dependent variable was a nominal categorical variable indicating whether a patient was treated by a physician, PA, or NP. The independent variable was the patient's source of payment for the visit. The statistical model tests the likelihood that a patient was treated by a provider other than a physician as a function of the patient's primary source of payment. All regression models controlled for patient sex, racial/ethnic group status, total number of chronic conditions (such as hypertension and hyperlipidemia), and age. Additional contextual controls included the median household income of the patient's home postal code, percentage of the population in the patient's postal code with a bachelor's degree or higher; the census-based geographic region of the country (North, South, East, West), and the metropolitan area status (urbanity) of the location of the outpatient department.
To characterize the sample by location, Table 1 shows that 28% of all weighted hospital outpatient department visits occurred in large metropolitan areas at one end of the five-level urbanity scale, and about 31% of outpatient department visits occurred in sparsely populated areas at the other end. The percentage of physician visits in each of the five areas of urbanity followed closely with the overall percentage of visits in each of these areas; however, the distribution of NP and PA visits was notably different. Although more than half of the physician visits occurred in larger metropolitan areas, the majority of PA visits (62%) and NP visits (51%) were in more sparsely populated areas. Table 1 shows that the percent of PA and NP visits increases as the location of services becomes more rural and the availability of physician services decreases.
Overall, physicians provided care for about 83% of the visits during the study period (Table 1). PAs provided care for 5% of visits and NPs for 11%. Among all visits, 47% occurred in large urban areas and 22% in rural areas. Almost a third (31%) of the physician visits occurred in a large/central metropolitan area; 18% took place in rural areas. The bulk of PA and NP visits occurred in nonmetropolitan (rural) areas.
Figure 1 illustrates that all three providers allocated an equal proportion of their visits to providing uncompensated care (3.4% to 3.8%) and care to privately insured patients (33.3% to 34.5%). What differs is that proportionally PAs and NPs provided more care to patients paying out of pocket than physicians did (9.4% versus 6.6%). Additionally, physicians and NPs spent the largest share of their efforts treating Medicaid patients; however, the share of NP (42%) visits paid by Medicaid was larger than the share of either PA (34%) or physician (38%) Medicaid visits.
Figure 2 reveals the association between setting and provider by source of payment. When the same compensation array is stratified by urban-rural across a continuum of the five population densities (metro-central to nonmetro) some of the findings become amplified. Regardless of metropolitan size, the combined percentage of visits paid by private insurance and Medicare dollars (the top of the payment scale) was higher for physicians than PAs or NPs. Across all settings, NPs saw a larger proportion of patients whose primary source of payment was Medicaid than physicians or PAs. Compared with physician visits, a greater share of NP and PA services was paid out of pocket (spanning 2006 to 2010).
In the full sample, with type of provider as the dependent variable, patients whose primary source of payment was Medicaid were proportionally 32% more likely to be treated by NPs than physicians and 16% more likely to be treated by PAs (Table 2). Patients paying out of pocket were 42% more likely to be treated by a PA and 60% more likely to be treated by an NP. Uncompensated visits were 76% more likely to be managed by a PA.
Patients' sources of payment appear to correlate with whether they see a physician, PA, or NP. This differentiation tends to occur when compensation for a patient visit does not involve third-party insurance. In this analysis of hospital outpatient department ambulatory care visits, when private insurance is not the mode of payment, the likelihood that a PA or NP will provide services increases. The hypothesis that source of payment influences type of provider merits further investigation. Identifying when this occurs by policy or system and then analyzing outcomes in terms of cost and quality of patient care is recommended. Considering the fact that PAs and NPs often provide care comparable with physician care, such a strategy may benefit the healthcare system as well as the patient, and may be an optimal means of maximizing division of labor or coordinating care.25,26
This type of analysis is of interest because, generally speaking, payment method tends to affect physician behavior. For example, physicians who are compensated in fee-for-service settings are more likely to perform procedures than those who are in capitated systems.27,28 Various studies have shown that payment by fee-for-service, capitation, and salary influences physician activity levels and productivity, with increases in productivity and procedure compensation proportional to levels of remuneration.29,30 These observations transcend a number of countries and are not unique to the United States.28 When policies are implemented to curb health expenditures by controlling fee levels, changes in productivity tend to occur. For example, physicians increase the volume of service supply, or provide services that attract higher fees.27 Although the effect of using PAs and NPs as a strategy to offset low revenue has not been reported in the public domain, the results of this study suggest that some lower-remuneration patient visits may be diverted to NPs and PAs.
Some evidence indicates that, absent a remuneration system, patients seeking primary care are not differentiated. In the VHA system, NPs and PAs attend about 30% of all ambulatory care encounters. NPs, PAs, and physicians fill similar roles in VHA medical care, although patients of PAs and NPs are slightly less complex than those of physicians.26 PAs appear to be as productive as physicians in the occupational medicine setting, where industrial injuries account for a large portion of the caseload. These visits involving workers' compensation tend to be prepaid and are largely undifferentiated by provider type. PA and NP compensation in family medicine, however, is about half that of a physician, so using PAs and NPs for low-compensation visits has the potential to improve revenue to the organization.31,32
Although these findings do not represent a causal relationship, the correlation between type of provider and source of payment should not be ignored. This dataset also does not address variation in the use of PAs and NPs across outpatient settings other than hospital outpatient departments. The public use version of the NHAMCS data prohibits users from controlling for the relative ratio of outpatient department physicians to NP and PA staff, a factor that could influence the likelihood of receiving treatment from a provider other than a physician.
In addition, patients may have been treated by providers with whom they had prior encounters in the outpatient department—longitudinally or as an episode. This noninclusion is a limitation of the NHAMCS dataset. Finally, if a patient had been treated in the outpatient department before a recorded visit, the data do not reflect whether the subsequent visit is with the same provider.
An ideal study would assign the average reimbursement amounts to each source of third-party payment to further investigate if amount has any predictive effect. Doing so would identify an upward trending supply curve for each source of insurance. The supply curve could summarize how much more likely physicians are to provide care as the level of reimbursement increases within each insurance type, and the slope on the curves could be compared with similar curves plotted for PAs and NPs. Because different factors influence the accessibility to physician care across the myriad of settings in the United States, this study suggests sources of payment should be considered one of them.
In the US multipayer system, it appears that, compared with physician visits, a higher proportion of NP and PA patient encounters are associated with lower levels of reimbursement. This is observed with greater proportionality as the area's population decreases. Furthermore, it appears that physicians may be more likely to attend visits with higher remuneration, such as private insurance and Medicare. Knowing the higher proportion of lower-reimbursement visits attended by PAs and NPs sets the stage for further inquiry into payment-based research, as it may help project the effects of expansions in Medicaid coverage on access to physician care. As the proportion of all types of providers delivering uncompensated and out-of-pocket care to uninsured patients is likely to decrease, research will likely shift toward differences in access by the type of insurance.
1. Shi L, Singh DA. Essentials of the US Health Care System
. 4th ed. Jones & Bartlett Publishers; 2012.
2. Shi L, Lebrun LA, Hung LM, et al. US primary care delivery after the Health Center Growth Initiative: comparison of health centers, hospital outpatient departments, and physicians' offices. J Ambul Care Manage
3. 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
5. Kirch DG, Henderson MK, Dill MJ. Physician workforce projections in an era of health care reform. Annu Rev Med
6. Petterson SM, Liaw WR, Phillips RL Jr, et al. Projecting US primary care physician workforce needs: 2010-2025. Ann Fam Med
7. Dill MJ, Salsberg ES. The complexities of physician supply and demand: projections through 2025. Association of American Medical Colleges; 2008.
8. Sargen M, Hooker RS, Cooper RA. Gaps in the supply of physicians, advance practice nurses, and physician assistants. J Am Coll Surg
9. Auerbach DI. Will the NP workforce grow in the future? New forecasts and implications for healthcare
delivery. Med Care
10. Hooker RS, Muchow AN. Supply of physician assistants: 2013-2026. JAAPA
11. Hing E, Hooker RS, Ashman JJ. Primary health care in community health centers and comparison with office-based practice. J Community Health
12. Hooker RS, Cipher DJ. Physician assistant
and nurse practitioner
prescribing: 1997-2002. J Rural Health
13. Lin SX, Hooker RS, Lenz ER, Hopkins SC. Nurse practitioners and physician assistants in hospital outpatient departments, 1997-1999. Nurs Econ
14. Hing E, Uddin S. Visits to primary care delivery sites: United States
, 2008. NCHS Data Brief
15. Hing E, Uddin S. Physician assistant
and advance practice nurse care in hospital outpatient departments: United States
, 2008-2009. Hyattsville, MD: CDC, National Center for Health Statistics; 2011.
16. Hooker RS, Benitez JA, Coplan BH, Dehn RW. Ambulatory and chronic disease care by physician assistants and nurse practitioners. J Ambul Care Manage
17. Hing E, Hall MJ, Xu J. National Hospital Ambulatory Medical Care Survey: 2006 outpatient department summary. Natl Health Stat Report
18. Shen J, Andersen R, Brook R, et al. The effects of payment
method on clinical decision-making: physician responses to clinical scenarios. Med Care
19. Long SH, Settle RF, Stuart BC. Reimbursement and access to physicians' services under Medicaid. J Health Econ
20. McManus M, Flint S, Kelly R. The adequacy of physician reimbursement for pediatric care under Medicaid. Pediatrics
21. Cohen MD. Access to health care: rhetoric vs. reality. Disch Plann Update
22. Decker SL, Doshi JA, Knaup AE, Polsky D. Health service use among the previously uninsured: is subsidized health insurance enough. Health Econ
24. Zuckerman S. How much will Medicaid physician fees for primary care rise in 2013? Evidence from a 2012 survey of Medicaid physician fees. The Urban Institute; 2012.
25. 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
26. 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
27. Ransom SB, McNeeley SG, Kruger ML, et al. The effect of capitated and fee-for-service remuneration on physician decision making in gynecology. Obstet Gynecol
. 1996;87(5 Pt 1):707–710.
28. Scott A, Hall J. Evaluating the effects of GP remuneration: problems and prospects. Health Policy
29. Jegers M, Kesteloot K, De Graeve D, Gilles W. A typology for provider payment
systems in health care. Health Policy
30. Simoens S, Giuffrida A. The impact of physician payment
methods on raising the efficiency of the healthcare
system: an international comparison. Appl Health Econ Health Policy
31. Cawley JF, Hooker RS. Physician assistants in American medicine: the half-century mark. Am J Manag Care
32. Hooker RS. Physician assistants in occupational medicine: how do they compare to occupational physicians. Occup Med (Lond)