The Primary Care–Specialty Income Gap: Why It Matters
Thomas Bodenheimer, MD; Robert A. Berenson, MD; and Paul Rudolf, MD, JD
A large, widening gap exists between the incomes of primary care
physicians and those of many specialists. This disparity is important
because noncompetitive primary care incomes discourage medical
school graduates from choosing primary care careers.
The Resource-Based Relative Value Scale, designed to reduce the
inequality between fees for office visits and payment for proce-
dures, failed to prevent the widening primary care–specialty income
gap for 4 reasons: 1) The volume of diagnostic and imaging pro-
cedures has increased far more rapidly than the volume of office
visits, which benefits specialists who perform those procedures; 2)
the process of updating fees every 5 years is heavily influenced by
the Relative Value Scale Update Committee, which is composed
mainly of specialists; 3) Medicare’s formula for controlling physician
payments penalizes primary care physicians; and 4) private insurers
tend to pay for procedures, but not for office visits, at higher levels
than those paid by Medicare. Payment reform is essential to guar-
antee a healthy primary care base to the U.S. health care system.
Ann Intern Med. 2007;146:301-306.
For author affiliations, see end of text.
come gap is widening.
According to surveys done by the Medical Group
Management Association, median physician income for all
primary care increased by 9.9% from 2000 to 2004, com-
pared with a 15.8% increase for all non–primary care spe-
cialties. During those years, median income for family
practice physicians increased 7.5% to $156 000, median
income for invasive cardiologists increased 16.9% to
$428 000, median income for hematologists and oncolo-
gists increased 35.6% to $350 000, and median income for
diagnostic radiologists increased 36.2% to $407 000. Table 1
shows trends in physician compensation.
Two other sources, Medical Economics and the Center
for Studying Health System Change, confirm these trends
(1–4). The income disparity is not explained by a differ-
ence in hours worked per week (5). Fifteen percent of
full-time family practice physicians earned less than
$100 000 in 2004, whereas 20% of invasive cardiologists,
25% of neurosurgeons, and 14% of orthopedists had in-
comes of $600 000 or more (1).
Does this income gap matter? Yes. Although incomes
of primary care physicians are far higher than the earnings
of most persons in the United States, and the public has
little sympathy for physicians who cry poor, the lower in-
come of primary care physicians is a major factor leading
U.S. medical students to reject primary care careers (6, 7).
The percentage of U.S. medical graduates choosing family
medicine decreased from 14% in 2000 to 8% in 2005 (8).
Seventy-five percent of internal medicine residents eventu-
ally become subspecialists or hospitalists rather than gen-
eral internists (9). Because office visit fees are relatively low,
primary care physicians schedule many short, rushed visits
to keep afloat financially, which potentially compromises
patient outcomes (10) and fosters the unsustainable physi-
cian work life that contributes to students’ avoidance of
primary care careers (11). With a median debt of $120 000
for public and $160 000 for private medical schools, med-
ical students are further discouraged from choosing careers
ncomes of primary care physicians are well below those
of many specialists, and the primary care–specialty in-
in primary care because of the noncompetitive income
WHY IS THE PRIMARY CARE–SPECIALTY INCOME GAP
The Resource-Based Relative Value Scale (RBRVS)
system (13), adopted by Medicare in 1992 and copied in
part by private insurers, was designed to lessen the fee
disparity between office visits—the bread and butter of
primary care—and procedures provided by specialists.
What happened? Why has the primary care–specialty in-
come gap widened rather than narrowed? (Appendix, avail-
able at www.annals.org).
Under the RBRVS system, the 2005 Medicare fee for
a typical 25- to 30-minute office visit to a primary care
physician in Chicago was $89.64 for a patient with a com-
plex medical condition (Current Procedural Terminology
[CPT] code 99214). The fee is calculated by multiplying
the relative value unit (RVU) for the 99214 CPT code
(2.18) by the 2005 Medicare conversion factor (37.8975)
and adding a geographic adjustment. The 2005 Medicare
fee was $226.63 for a gastroenterologist in the outpatient
department of a Chicago hospital performing a colonos-
copy (CPT code 45378), which is of similar duration to
the office visit. Colonoscopy performed in a private office
in Chicago, which differs from the hospital setting because
the gastroenterologist pays for equipment and nursing
time, would cost $422.90. Office visits are considered eval-
uation and management services (history, physical exami-
nation, and medical decision making), whereas colonosco-
pies are an example of a procedural service.
Conversion of tables into slides
Annals of Internal Medicine
© 2007 American College of Physicians 301
WHY IS THE GASTROENTEROLOGIST PAID MUCH
MORE THAN THE PRIMARY CARE PHYSICIAN FOR
WORK TAKING ABOUT THE SAME AMOUNT OF TIME?
Three factors determine RVUs: work, practice ex-
pense, and malpractice costs. Work and practice expenses
each account for about half of the RVU value; malpractice
is a small fraction. The main difference between the office
visit and the colonoscopy lies in the work portion of the
RVU. The colonoscopy work RVU is higher because the
“intensity” (skill, effort, judgment, and stress) is deemed to
be greater for procedures than for cognitive visits. In the
research on which the RBRVS is based, the intensity com-
ponent was determined by surveys and interviews with
physicians in many specialties (13); procedural specialists
tended to view procedural work as being more intense than
To explain the widening income gap between proce-
dural specialists and primary care physicians, it is helpful to
examine 4 factors.
1. The Key Role of Volume
Physician income is the product of the number of
services performed multiplied by the fee for each service.
Volume growth may increase income even if the fee per
service decreases. For example, a Medicare colonoscopy
performed in 2004 paid 15% less than a colonoscopy per-
formed in 2001. Yet, the number of Medicare colonosco-
pies in those years increased by 30%.
From 1999 to 2003, the Medicare volume of evalua-
tion and management services (mainly office visits) in-
creased by 15%. Major procedures (mostly surgeries) in-
creased at a similar pace of 14%, but “other procedures”
(endoscopies and minor surgeries) increased by 26%, diag-
nostic tests increased by 36%, and imaging increased by
45% (14) (Table 2 ). This trajectory continued in
2004 (16), and patterns for non-Medicare populations are
similar (17). The reasons for volume growth are discussed
The volume trend—evaluation and management ser-
vices growing more slowly than imaging and “other proce-
dures”—hurts primary care because evaluation and man-
agement services account for more than 80% of Medicare
payments to primary care physicians (19, 20). “Other pro-
cedures” account for 73% of such payments to dermatol-
ogists and 53% to gastroenterologists. Radiologists and car-
diologists receive the bulk of Medicare payments for
imaging (16). Because of technologic improvement and
increasing efficiency over time, specialists can perform
many procedures more quickly (21), whereas office visits
cannot be shortened without reducing the quality of care
or patient and physician satisfaction (10, 22–24).
In summary, the volume of many procedures per-
formed by specialists has increased more rapidly than office
visits, at times in dramatic fashion, contributing to faster
income growth of some specialists compared with that of
primary care physicians.
2. The Relative Value Scale Update Process
Medicare agreed to allow the American Medical Asso-
ciation (AMA) to create a consensus process for recom-
Table 1. Median Pretax Compensation of Physicians, 1995–2004*
Specialty Median Compensation, U.S.$10-Year Change, %5-Year Change, %
19952000 2004 1995–2004 2000–2004
All primary care
Family practice (without obstetrics)
* Data are from the Medical Group Management Association, Englewood, Colorado, December 2005. Compensation refers to net income.
The Primary Care–Specialty Income Gap
302 20 February 2007 Annals of Internal Medicine Volume 146 • Number 4
mending updated RVU values. The AMA, along with spe-
cialty societies, created the Relative Value Scale Update
Committee (RUC) to recommend RVU changes to the
Centers for Medicare & Medicaid Services (21, 25).
The RUC is composed of 29 members, of whom 23
are named by specialty societies (including 3 from primary
care specialties); 26 are voting members. Although primary
care physicians provide about half of Medicare physician
visits, primary care makes up only 15% of RUC’s voting
members. Specialty societies request changes in RVUs for
procedures performed by their members.
Each year, the RUC meets to recommend RVUs for
new CPT codes. Through the creation of new codes, many
specialties receive increased RVUs for procedures they can
bill under more highly reimbursed codes. For example, in
2004, 20 new codes were created for placement, revision,
and repair of central venous catheters, replacing 4 previous
In addition, every 5 years the RUC considers potential
revisions to RVUs for all existing CPT codes. In this up-
date process, the RUC requires that specialty societies sur-
vey at least 30 of their members to provide data justifying
a proposed RVU change. The Medicare Payment Advisory
Commission (MedPAC) has criticized this survey method
as biased toward reviewing procedures felt to be underval-
ued because specialty societies have little incentive to re-
quest evaluation of potentially overvalued procedures that
might result in a reduced RVU. Potentially overvalued pro-
cedures are rarely discussed at the RUC (21).
At the 5-year update meetings, a two-thirds vote is
required to recommend an RVU change. Observers report
that many RUC members from procedural specialties tend
to vote in favor of requested increases. If a proposed RVU
change does not receive sufficient votes, it does not die but
is referred to a subcommittee that generally proposes an
The 1995 RUC 5-year review resulted in gains for
primary care and some specialties. In 2000, evaluation and
management codes were not discussed, which hurt primary
care. The RUC recommended 469 increases and only 27
reductions in non–evaluation and management codes. The
Centers for Medicare & Medicaid Services accepted more
than 95% of RUC recommendations (25). Some examples
of surgical procedures whose work RVUs have increased
since 2000 are gastrectomy, colectomy, simple mastec-
tomy, and lower leg amputation (26). In contrast, office
visit RVUs did not increase from 1995 to 2005.
In the RUC, discussion of evaluation and manage-
ment code valuation is contentious because of the sustain-
able growth rate (SGR) formula for Medicare physician
payment. Under the SGR, the total amount that Medicare
pays physicians each year is limited by a formula based on
growth of the Medicare population, increase in physician
practice expenses, and change in gross domestic product. If
total expenditures exceed an SGR-defined target because of
service volume increases, Medicare reduces the conversion
factor in the subsequent year, thereby cutting fees equally
for all physicians. Physician payment under SGR is a pie: If
1 specialty receives a larger slice, others must accept smaller
Evaluation and management services make up more
than 50% of total Medicare physician payments. Even
small increases in office visit RVUs would create a dramatic
increase in total Medicare physician spending, thus trigger-
ing a conversion factor reduction. Specialists making little
income from evaluation and management services are
therefore wary of such RVU increases. Procedure increases
are less contentious because no single procedure has suffi-
cient volume to perceptibly increase total Medicare spend-
In summary, the RUC process favors increases in pro-
cedural and imaging reimbursement for 3 reasons: specialty
society influence in proposing RVU increases, the special-
ist-heavy RUC membership, and the desire of RUC spe-
cialists to avoid increases in evaluation and management
RVUs. With their ability to create new codes and influence
RVU updates, many procedural specialists can influence
fees in a way that observers find to substantially overvalue
procedural and imaging services. Moreover, high fees may
encourage physicians to increase the volume of profitable
services, leading to even higher income gains and greater
spending growth (27).
For a service to qualify for an RVU increase, a re-
quester must prove to the RUC that the work of perform-
ing the service has increased over the past 5 years. In the
2005 to 2006 negotiation over evaluation and manage-
ment increases, the cognitive specialists who requested
work RVU increases for evaluation and management ser-
vices presented evidence that the intensity of these services
was greater because 1) the patient population is aging and
is being seen more often for multiple chronic illnesses as
opposed to being seen for single, acute symptoms, 2) pa-
tients are taking more medications, and 3) decision making
is more complex because of the availability of more diag-
nostic information and treatment options (28).
In the RVC deliberations, some members questioned
Table 2. Change in Volume of Services per Medicare
Beneficiary, 2000 to 2005*
Office visit, established patient
Cardiovascular stress test
Magnetic resonance imaging
* Data are from reference 15.
The Primary Care–Specialty Income Gap
20 February 2007 Annals of Internal Medicine Volume 146 • Number 4 303
the evidence showing that evaluation and management ser-
vices were increasing in intensity. Furthermore, they be-
lieved that higher work RVUs were inappropriate because
primary care physicians could bill a higher level of service
for patients with more complicated conditions. Some ob-
servers thought that money played a role in the delibera-
tions because evaluation and management services account
for such a large portion of Medicare spending; an increase
in evaluation and management RVU values would trigger a
substantial reduction in the Medicare conversion factor.
After 6 months of discussion, including 2 face-to-face
meetings and several conference calls, the RUC recom-
mended increases in RVUs for some evaluation and man-
agement codes. Some observers thought that MedPAC’s
recent criticism of the RUC (20), questioning the domina-
tion of specialists over primary care, heightened the pres-
sure on the RUC to increase evaluation and management
On 21 June 2006, the Centers for Medicare & Med-
icaid Services proposed a 37% increase in the work RVU
for intermediate office visits (99213) and substantial in-
creases for some other evaluation and management codes
commonly used by primary care physicians (29) in accor-
dance with RUC recommendations. This announcement
left the impression that primary care physicians would en-
joy major increases in Medicare payments. In fact, the net
increase is far smaller: 5% for 2007. As 1 payment expert
commented, “The large print giveth, the small print taketh
away.” A careful reading of the CMS proposal reveals that
family physicians and general internists will receive 5%
more in allowed Medicare charges in 2007 than in 2006
(30). This increase is far lower than might be expected
because work RVUs represent only half of the total RVU
and because SGR-related budget neutrality requirements
reduce the payment increase.
3. Sustainable Growth Rate Hurts Primary Care
From 1999 to 2004, total Medicare physician pay-
ments increased rapidly because of volume growth from
imaging, minor procedures, and diagnostic tests—not
from evaluation and management services (14). This vol-
ume growth is largely responsible for Medicare physician
payments that exceed the SGR target, thereby triggering a
reduction in the Medicare conversion factor. The decrease
in the conversion factor affects not only those responsible
(specialists benefiting from imaging, minor procedures,
and diagnostic tests) but all physicians. Volume-based in-
come gains of procedural and imaging specialists can create
fee-based income losses for primary care physicians.
The 2005 Medicare conversion factor was 1% below
the 2001 level. The 2006 conversion factor was scheduled
to decrease by 4.4%, but Congress acted to hold it at the
2005 level. Under the SGR formula, however, the conver-
sion factor is scheduled to decrease by 5% per year from
2007 through 2015 (31). If such a reduction actually oc-
curs, primary care gains from increased evaluation and
management payment resulting from the 2005 five-year
review would be canceled out.
4. Private Insurance Aggravates the Income Gap
Compared with the inequities of Medicare’s physician
payment, the situation for private insurance is even more
problematic. For most specialties, the majority of payments
derive from privately insured patients. A 2002 study of 34
private insurers found that most use a fee schedule mod-
eled on RBRVS, but that many offer specialty-specific con-
version factors. On average, private insurers paid primary
care physician office visits at 104% of Medicare’s fee,
whereas surgical, diagnostic procedure, and imaging codes
were paid at 119% to 120% of Medicare fees. Markets
with large single-specialty groups were associated with even
higher specialty fees. Surgical codes have been paid as high
as 330% of Medicare fees while radiology and diagnostic
procedures may attain 250% of Medicare fees. The pri-
mary care–specialty fee gap is greater for private plans than
for Medicare (32).
Another study independently corroborated that private
insurers paid an average of 104% of Medicare fees for
office visits and 133% for procedures and imaging studies
in 2001 (33).
These studies confirm that private insurer payment
favors specialty care over primary care to a greater degree
than does Medicare. The long-standing ability of specialists
to negotiate higher rates than those of primary care physi-
cians from private plans has been enhanced by the recent
advent of market-dominant, single-specialty groups with
market power to command higher fees (34). Moreover, the
trend toward performance of procedures in ambulatory
surgery centers and other facilities owned by specialist phy-
sicians enables these physicians to earn both professional
and facility fees.
ADDRESSING THE PRIMARY CARE–SPECIALTY
Patients, specialists, and the entire health system need
a healthy primary care base. In a 1997 patient survey, 94%
valued having a primary care physician who knew their
medical problems. Eighty-nine percent wanted their pri-
mary care physician to participate in the specialty referral
process (35). Patients with a regular generalist physician
have lower overall costs than those without a generalist
physician (36–38). Increased ratios of primary care physi-
cians to population are associated with reduced hospitaliza-
tion rates for 6 ambulatory care–sensitive conditions (39).
Health care costs are higher in regions with greater special-
ist-to-population ratios (40–43).
Primary care practice is not viable without a substan-
tial increase in the resources available to primary care phy-
sicians. The American College of Physicians (ACP), the
American Academy of Family Physicians (AAFP), and
MedPAC have recommended changes to rescue primary
care from what the ACP has called an “impending col-
The Primary Care–Specialty Income Gap
304 20 February 2007 Annals of Internal Medicine Volume 146 • Number 4
lapse” (12). The MedPAC, whose 17 members are ap-
pointed for 3-year terms by the U.S. Comtroller General,
has been concerned with primary care because, as a watch-
dog of Medicare costs, it views a high ratio of specialists to
population as a cost driver while a greater number of pri-
mary care physicians may help contain costs (21).
In the short term, Medicare and private payers need to
review and modify their reimbursement approaches to shift
payments from procedural and imaging services to evalua-
tion and management services. The 2006 increase in some
evaluation and management codes recommended by the
RUC is a small and inadequate step in that direction. The
MedPAC has suggested that RUC membership include
more primary care physicians, has recommended that
CMS review overvalued services, and has discussed altering
the SGR system to protect primary care physicians who are
the victims of, but are not responsible for, most expendi-
ture growth (21). Furthermore, MedPAC has recom-
mended payments for care coordination services targeted to
chronically ill Medicare patients, many of whom are man-
aged by primary care physicians (44). The ACP recom-
mends that Medicare substantially increase evaluation and
management RVUs, pay for time spent on telephone and
e-mail consultations, and reimburse care coordination for
patients with chronic conditions (12). The ACP and AAFP
call for a revision of the SGR process.
For the long term, it would be desirable to develop
new payment models that blend the best of fee-for-service,
capitation, and salary, while mitigating each approach’s de-
ficiencies (45, 46). For example, primary care physicians
who are caring for patients with multiple chronic condi-
tions could be paid on the basis of capitation-like princi-
ples, though avoiding the problems created by previous
managed care capitation systems. (47, 48). Surgeons and
other specialists responsible for episodes of care over a de-
limited time period might be paid case rates on the basis of
diagnoses, and specialists providing one-time professional
services might continue to be paid fee-for-service.
It is unclear whether the medical profession—with dif-
ferent specialties having distinct monetary interests and dif-
ferent estimations of the professional value of their work—
can agree on substantial changes in payment policy on its
own. Public and private payers, working with physicians,
have a common interest in promoting a vibrant primary
care sector as a medical home for patients and families, a
home that, if properly supported, can contribute to sub-
stantial reductions in health care costs (49).
From University of California, San Francisco, San Francisco, California,
and the Urban Institute and Arent Fox, PLLC, Washington, DC.
Potential Financial Conflicts of Interest: Consultancies: P. Rudolf
(American College of Physicians, American Geriatrics Society)
Requests for Single Reprints: Thomas Bodenheimer, MD, Building
80-83, San Francisco General Hospital, 1001 Potrero Avenue, San Fran-
cisco, CA 94110; e-mail, TBodenheimer@fcm.ucsf.edu.
Current author addresses are available at www.annals.org.
1. Lowes R. The earnings freeze. Now it’s everybody’s problem. Med Econ.
2005;82:58-62, 64, 66-8. [PMID: 16300220]
2. Crane M. Earnings. Time to call a code? Med Econ. 2001;78:74-6, 79, 83-6.
3. Tu HT, Ginsburg PB. Losing ground: Physician income, 1995-2003. Track-
ing Report No. 15. Washington, DC; Center for Studying Health System
Change; June 2006.
4. Center for Studying Health System Change. Physician survey, 1996-2001.
Washington, DC; Center for Studying Health System Change. Accessed at http:
//CTSonline.s-3.com/psurvey.asp on 24 August 2006.
5. Weiss GG. Exclusive survey: productivity takes a dip. Med Econ. 2005;82:
86-7, 89, 91-3. [PMID: 16400970]
6. Rosenblatt RA, Andrilla CH. The impact of U.S. medical students’ debt on
their choice of primary care careers: an analysis of data from the 2002 medical
school graduation questionnaire. Acad Med. 2005;80:815-9. [PMID: 16123459]
7. Newton DA, Grayson MS, Thompson LF. The variable influence of lifestyle
and income on medical students’ career specialty choices: data from two U.S.
medical schools, 1998-2004. Acad Med. 2005;80:809-14. [PMID: 16123458]
8. Pugno PA, Schmittling GT, Fetter GT Jr, Kahn NB Jr. Results of the 2005
national resident matching program: family medicine. Fam Med. 2005;37:555-
64. [PMID: 16145633]
9. West CP, Popkave C, Schultz HJ, Weinberger SE, Kolars JC. Changes in
career decisions of internal medicine residents during training. Ann Intern Med.
2006;145:774-9. [PMID: 17116922]
10. Zyzanski SJ, Stange KC, Langa D, Flocke SA. Trade-offs in high-volume
primary care practice. J Fam Pract. 1998;46:397-402. [PMID: 9597997]
11. Dorsey ER, Jarjoura D, Rutecki GW. Influence of controllable lifestyle on
2003;290:1173-8. [PMID: 12952999]
12. American College of Physicians. The impending collapse of primary care
medicine and its implications for the state of the nation’s health care: a report
from the American College of Physicians. Philadelphia: American College of
Physicians; 30 January 2006.
13. Hsiao WC, Braun P, Yntema D, Becker ER. Estimating physicians’ work
for a resource-based relative-value scale. N Engl J Med. 1988;319:835-41.
14. Hackbarth G. Medicare payment to physicians. Washington, DC; Medicare
Payment Advisory Commission; 17 November 2005.
15. U.S. Government Accountability Office. Medicare Physician Services.
Washington, DC: U.S. Government Accountability Office; July 2006.
16. Medicare Payment Advisory Commission. Healthcare Spending and the
Medicare Program. A Data Book. Washington, DC: Medicare Payment Advisory
Commission; June 2005.
17. Woodwell DA, Schappert SM. National Ambulatory Medical Care Survey:
1993 Summary. U.S. Department of Health and Human Services. Publication
6:0029. Hyattsville, MD: National Center for Health Statistics; December 1995.
18. Bodenheimer T. High and rising health care costs. Part 3: the role of health
care providers. Ann Intern Med. 2005;142:996-1002. [PMID: 15968014]
19. Medicare Payment Advisory Commission. Report to Congress. Growth in
the Volume of Physician Services. Washington, DC: Medicare Payment Advisory
Commission; December 2004.
20. Medicare Payment Advisory Commission. Report to Congress. Impact of
Resource-Based Practice Expense Payments for Physician Services. Washington,
DC: Medicare Payment Advisory Commission; December 2004.
21. Medicare Payment Advisory Commission. Report to Congress. Washington,
DC: Medicare Payment Advisory Commission; March 2006.
22. Dugdale DC, Epstein R, Pantilat SZ. Time and the patient-physician rela-
tionship. J Gen Intern Med. 1999;14 Suppl 1:S34-40. [PMID: 9933493]
23. Mechanic D. Physician discontent: challenges and opportunities. JAMA.
2003;290:941-6. [PMID: 12928472]
24. Wetterneck TB, Linzer M, McMurray, Douglas J, Schwartz MD, Bigby J,
et al. Worklife and satisfaction of general internists. Arch Intern Med. 2002;162:
649-56. [PMID: 11911718]
25. American Medical Association. RBRVS: Resource-based relative value scale.
Chicago: American Medical Association; 2005. Accessed at www.ama-assn.org/go/
byUS medical students.JAMA.
The Primary Care–Specialty Income Gap
20 February 2007 Annals of Internal Medicine Volume 146 • Number 4 305