Provider Opt-Out Under Medicare Private Contracting

Abstract

The 1997 Balanced Budget Act (BBA) permits private contracting for care between Medicare beneficiaries and providers who have opted out of Medicare. This article examines the number and characteristics of providers who have opted-out, their role in the provision of Part B services, and their impact on beneficiary access from 1998 to 2002. Opt-out providers differ from providers remaining in Medicare with respect to specialty, practice characteristics, and Medicare Program activity. Very few providers found opting-out attractive and the departure of this small group of providers appears not to have created access problems for beneficiaries.

Private Contracting and Opt-Out

Physicians and non-physician providers have not always been content with the Medicare Program' s Part B payment policies and administrative procedures. The Medicare fee schedule (MFS) enacted in 1992 has reallocated physicians' income resulting in income reductions for some specialties. More recently, the negative overall updates to the MFS due to the link between the sustainable growth rate factor and growth in the GDP has reduced Part B payments to all specialty groups. Policymakers and advocates feared that providers who faced revenue cuts to changes in Part B payment policies would cut back on services to Medicare beneficiaries by simply telling Medicare beneficiaries that their practices were not accepting new patients. If sufficiently numerous, providers who were discontent with Medicare payment policies and cut back on treating Medicare patients could potentially create access problems for Medicare beneficiaries.

In response to congressional concern regarding reduction in access to Medicare physician services and proposed legislation by Senator Kyl in 1997, the eventual compromise between the Clinton Administration and Congress, represented in section 4507 of the 1997 BBA, allows physicians and other selected providers of Part B services to opt-out of the Medicare Program and establish, in writing, private contracts with Medicare beneficiaries for all covered Part B services except those services provided for emergency and urgent care. Not all providers may opt out. Private contracting was only authorized for physicians, osteopaths, and selected non-physician providers (clinical, psychologists, clinical social workers, physician assistants, nurse practitioners, clinical nurse specialists, certified registered nurse anesthetists, certified nurse midwives) in the 1997 BBA. The Medicare Prescription Drug Improvement, and Modernization Act of 2003 (MMA) extended private contracting to podiatrists, dentists, and optometrists effective December 2003.

Under these private contracts, beneficiaries are liable for payment of the costs of care provided and could not bill Medicare. Providers sending opt-out affidavits to their Part B claims carriers would be prohibited from billing Medicare for services provided to program beneficiaries or receiving payment linked to Medicare health maintenance organization (HMO) capitation payments for 2 years after the effective date of opt-out on the affidavit. Providers opting-out of Medicare, however, may order services for Medicare patients to be delivered by providers who have not opted-out.

The congressional debate on private contracting was posed in terms of the following issues: beneficiary access, freedom of beneficiaries to choose their physician and supplement payment with personal funds, and patient privacy. The debate surrounding these concerns is described by Hoff (1998). His presentation of these issues with anecdotal illustrations, however, points to a common antecedent, refusal to treat new Medicare patients. He argues that refusal of Medicare patients occurs because of payment inadequacy and the perception of onerous, overzealously enforced administrative requirements. Hoff presents several examples where, he argues, Medicare Part B payment policies may be flawed and misdirected. However, many of the issues he raises date back to the beginning of the Medicare Program. Results of physician surveys by the Physician Payment Review Commission (1997) and the MedPAC (Schoenman and Chang, 1999; Schoenman and Feldman, 2003) indicate that physicians have been displeased with low fees, excessively stringent documentation, coding requirements, and administrative red tape in fee-for-service (FFS) Medicare. However, most physicians have not refused to treat new Medicare patients or limit the percentage of Medicare patients seen. Indeed, physicians indicated far greater displeasure with capitated managed care providers regarding these issues than Medicare (Schoenman and Chang, 1999; Schoenman and Feldman, 2003).

Some of the concerns raised by Hoff may very well be related to increased efforts to control Medicare fraud and abuse, and increased prosecution of health care fraud by the Department of Justice. The nationwide expansion of Project Restore Trust in 1997 and the review of physician billing for Medicare services as part of the Physicians at Teaching Hospitals has made the possibility of an audit a real risk for providers that are either engaged in fraud or careless in documentation. Prior to these efforts, few Medicare Program resources were devoted to fraud and abuse control, and audits of Medicare provider records were infrequent.

The overall effect that private contracting was expected to have on beneficiary access is not clear. The economically rational expectation would be that Medicare beneficiaries would quickly leave physicians and other providers who have chosen private contracting. Nevertheless, some beneficiaries may feel that they must sign private contracts to continue care with their current provider. A previous study of differences in physician costs and physician choice among Medicare beneficiaries suggests that Medicare beneficiaries are generally reluctant to switch physicians, even when clear economic incentives appear to promote switching (Rice, Nelson, and Colby, 1992). Their study, however, only examined Medicare beneficiaries switching from non-participating to Medicare participating providers that involves a relatively small actual change in out-of-pocket payments.

Although much of the legislative debate concerning Medicare private contracting has been couched in terms of beneficiary issues, the impetus behind it is more plausibly from providers disaffected from Medicare and the direction of recent Medicare policy changes.

Provider Status

Physicians' (and other providers') activity in Medicare has been intertwined with Part B payment policy throughout the history of the program. When the Medicare Program began, Part B services were paid allowed charges. As an incentive to provide care to Medicare beneficiaries, providers were also permitted to charge more than the Medicare allowed amount with beneficiaries being liable for charges above this amount (balance billing). While payment for Medicare Part B services has been made under a resource-based relative value (RBRVS) fee schedule since 1992, this option is still available to Medicare providers under certain conditions.

Providers under Medicare can choose to submit claims for patients and accept the Medicare allowed amount as payment in full (i.e., accept assignment). Alternatively, physicians can refuse assignment on claims and bill the patient the full charge of services. The patient is reimbursed the Medicare allowed amount minus copayment. Providers may choose on a claim-by-claim basis to accept assignment unless they are Medicare participating providers who accept assignment on all Medicare Part B claims. This was done to allow access to top line services priced above Medicare payment rates (Colby et al., 1995), to avoid the appearance of creating fixed fees for physicians, and to ensure that beneficiaries would maintain access to physician services.

Previous studies have indicated that physicians who charged more for office visits were more likely to not accept assignment and balance bill Medicare beneficiaries (Paringer, 1980; Mitchell and Cromwell, 1982; Rice and McCall, 1983; Rodgers and Musacchio, 1983; Mitchell, Rosenbach, and Cromwell, 1988). Balance billing has been more prevalent in rural areas, but less prevalent in areas with large poverty populations (Physician Payment Review Commission, 1993). This reflects the prohibition against balance billing Medicare-Medicaid dually eligible beneficiaries in effect since the beginning of the Medicare Program.

Use of balance billing varies across physician specialties (Centers for Medicare & Medicaid Services, 2003). Balance bills from radiologists and gynecologists are noticeably higher per patient than the average balance bill. While Colby et al. (1995) found a clear difference across specialty groups in balance billing incidence, interspecialty variation in balance billing had decreased substantially in 1999 (Centers for Medicare & Medicaid Services, 2003).

Most providers, however, have chosen to accept assignment and have Medicare pay them directly. As of 1984, physicians and other providers can become a Medicare participating provider and accept assignment on all Part B claims. Physicians who become participating providers receive several benefits. MFS payments are 5 percent higher for participating providers. Participating providers have their names included in the Medicare Participating Physician/Supplier Directory and have expedited processing of electronic claims. As with balance billing, participating provider rates vary by specialty with psychiatrists, anesthesiologists, and general practice physicians having the lowest participation rates among specialty groups (Physician Payment Review Commission, 1993).

In 1990, 80 percent of allowed charges were for assigned claims and 44 percent of physicians were participating providers (Physician Payment Review Commission, 1993). Assignment and participating provider rates have increased to where, in 1996, 96 percent of allowed charges were for assigned claims and 78 percent of physicians were participating providers. As of January 2002, 89.3 percent of Part B providers were Medicare participating providers (Centers for Medicare & Medicaid Services, 2002). These factors have reduced balance billing and successfully placed limits on beneficiary out-of-pocket spending for Medicare Part B services (Physician Payment Review Commission, 1997).

In order to control spending for Part B services, Medicare began limiting the amount providers could balance bill in 1992. The implementation of the MFS not only tightened physician revenue but made assignment more restrictive. As of 1993, balance billing by non-participating providers was limited to 115 percent of their fee schedule amount (95 percent of fee schedule amount for participating providers). Thus, balance billing could not exceed 109.25 percent of the participating provider fee schedule amount. Unassigned Medicare Part B claims (containing balance billing) were 19.1 percent of Medicare Part B claims in 1990, but only 1.9 percent of total claims in 2001 (Centers for Medicare & Medicaid Services, 2002). Limiting charge rules effectively control cost sharing for balance bills since providers can be required to refund overcharges. Although more providers during the last 15 years took assignment on their billing, a small group of providers chose not to take assignment. These providers retained the option to decide their fee amounts and whether to accept assignment, although at the cost of reduced fees.

Why Providers May Opt-Out

The previous discussion suggests several possible reasons why physicians and other providers might opt-out of Medicare. Opting out may be a response to specialty-specific reductions in income due to cuts in the MFS. In this case, opting-out is a direct response to policy changes that should be experienced quickly. Research on MFS impacts, however, indicates that physicians have generally increased patient volume to maintain revenue in response to MFS cuts (Nguyen and Derrick, 1997; Zuckerman, Norton, and Verilli, 1998). Thus, cuts in the MFS would need to be very substantial in any one year to change physicians' responses. At any rate, MFS changes from 1997 to 2000 appear to have increased rather than decreased physicians' Medicare volumes in the aggregate.

Physicians and other providers may also choose to opt-out in response to historical patterns of fee limitations for their specialty groups, problems with administrative procedures, balance billing limits, and claims denials. This group would include providers who wish to be price setters who can charge fees without limits via unrestricted balance billing and providers who work in small solo practice settings. The latter group may prefer to avoid all but the most necessary paperwork and, thus, bill all patients directly and avoid interface with government and private insurance payers.

Provider opt-out may also be a reaction to changes in Medicare. Some providers may feel that the greater regulatory control imposed by the MFS, administrative reforms such as stricter limits on non-assigned claims, and increased efforts toward control of fraud and abuse in Medicare (i.e., more stringent review of claim coding accuracy, review of patient medical records, increased fines and prosecution) represents excessive Federal Government involvement in the practice of medicine. Some providers may wish to provide non-traditional services (e.g. holistic therapies) or extended preventive, case management and on call services (e.g. concierge or boutique practices) that are not covered by Medicare and may require providers to fill out an advance beneficiary notice¹ for each visit. Some providers may simply wish full payment at the time of service to avoid spending time with insurance paperwork and cash-flow uncertainty.

Although data for physician activity under private contracts are not available, with currently available data one can observe whether providers who opt-out are concentrated within specialty groups or States, and any differences in practice characteristics and billing patterns when compared with providers remaining in Medicare. Examination of small-area location patterns and past service use by beneficiaries treated by providers who have opted-out can provide information regarding the effect of private contracting on Medicare beneficiaries' access to Part B services.

Data

While Medicare data cannot be used to directly examine the activities of private contracting providers after opt-out, Medicare data can provide some information concerning demographic and geographic characteristics of these providers and their Part B service provision prior to opt-out. The provider data recorded in the mandated opt-out affidavits are reported to CMS in standardized reports from Part B carriers and are used to create the Opt-Out Provider File. This file contains data for the 2,839 physicians who opted-out of Medicare from January 1, 1998 to December 12, 2002. These data are merged with two CMS provider data files. The first of these is the Medicare Provider Identifier and Eligibility Registry File that contains the unique physician identifier number (UPIN), demographic, specialty, and program information for all U.S. physicians and other providers who bill Medicare for Part B services. The second data source is the Medicare 1997 Provider Summary File. This file contains aggregated total billing data for providers who billed Medicare during calendar year (CY) 1997. These data are used to describe demographic characteristics and billing patterns for opt-out and non-opt-out providers.

This article also examines the amount and type of care received by Medicare beneficiaries from opt-out providers in the year prior to opt-out and the Medicare services received by these beneficiaries from active Medicare providers in the year after opt-out. The providers who opted-out of Medicare during CY 1998 (1998 provider cohort) were identified from the opt-out provider file and a UPIN finder file was created to extract all Part B physician/supplier claims for these physicians for the 12-month period prior to opt-out. Claims for 33,245 beneficiaries were extracted and a health insurance claim finder file was created. All Medicare claims for services performed by non-optout providers for the same 12-month period were then extracted and all Medicare claims for this beneficiary group for the 12 months after opt-out were extracted. Beneficiaries who died, were HMO members, or were not eligible for Part B services during the study period were dropped from the study cohort and 27,203 Medicare beneficiaries remained in the 1998 analysis cohort. A similar file for physicians who had opted-out of Medicare during CY 1999 was also created. (The analysis of these data is not included in this article and is available on request from the author.)

Provider Opt-Out Trends

From 1998 to 2002, 2,839 physicians, clinical psychologists, and other providers chose to opt-out. This represents 0.42 percent of the physicians and other providers eligible to opt out under the 1997 BBA This low percentage may reflect the limited knowledge among physicians concerning opting-out, as MedPAC's 1999 physician survey indicated (Schoenman and Chang, 1999). Figure 1 shows the total number of providers opting-out by year, along with the cumulative total of providers opting out. Growth in the ranks of providers electing private contracting has been uneven over time. These data indicate that the greatest number of opt-outs (721) occurred during 2002 surpassing the 710 providers who opted out during 1998, the first year of the provision.

Figure 1. Total Number of Providers Opting Out, by Year: 1998-2002.

As shown in Table 1, providers who have opted-out of Medicare are slightly older than those remaining in the program and have slightly more practice settings than non-opt-out providers. Nearly 3 percent of opt-out providers have no active Medicare practice settings. These providers are often retired or hospital-based teaching faculty or specialists who normally do not treat Medicare beneficiaries (e.g., pediatricians, child psychologists).

Table 1. Characteristics of Opt-Out and Non-Opt-Out Providers: 1998-2002.

Characteristic	Opt-Out	Non-Opt-Out
Number	2,839	677,988
Percent Active	97.2	100
Age	56	53
Number of Practice Settings	2.4	2.3
	Percent
Participating Providers	30.5	81.3
Solo Practice	71.6	45.7
Group Practice	28.4	54.3

SOURCE: Centers for Medicare & Medicaid Services: Data from the Opt-Out Carrier Reports and the Medicare Provider Identifier and Eligibility Registry, 1998-2002.

The most striking differences between these two groups are that opt-out providers are far more likely to be in solo practice (71.6 percent) than providers remaining in the Medicare Program (45.7 percent) and while 81.3 percent of providers remaining in the Medicare Program are participating providers, only 30.5 percent of opt-out providers are Medicare participating providers. This reflects, in part, the specialty composition of the opt-out provider group. These differences reflect dissimilarities in specialty group composition and practice volume between opt-out providers and providers remaining in the Medicare Program.

Table 2 shows the distribution of opt-out providers across States. New York has the greatest number of opt-out providers (475), followed by California (363), and Texas (166). A comparison of the distributions of opt-out providers to all active providers shows that the distribution of opt-out providers generally reflects the distribution of the U.S. medical care provider population. Also shown in Table 2 is the percent of opt-out providers among all active providers eligible to private contract within the State. Alaska has the highest percent of active providers opt-out (2.5 percent). The percentage of opt-out providers did not exceed 1 percent in any other State. In New York and California, the States with the largest number of opt-outs, opt-out providers represent only 0.84 and 0.52 percent of active providers in these States, respectively.

Table 2. Distribution of Opt-Out and Active Providers, by State: 1998-2002.

State	Opt-Out Providers	Percent	All Active Providers	Percent	Percent Opt-Out in State
Alabama	14	0.5	7,725	1.1	0.18
Alaska	31	1.1	1,258	0.2	2.46
Arizona	70	2.5	9,966	1.5	0.70
Arkansas	17	0.6	6,265	0.9	0.27
California	363	12.8	70,177	10.3	0.52
Colorado	86	3.0	10,986	1.6	0.79
Connecticut	57	2.0	7,563	1.1	0.75
Delaware	5	0.2	1,959	0.3	0.26
District of Columbia	50	1.8	4,237	0.6	1.18
Florida	142	5.0	36,478	5.4	0.39
Georgia	116	4.1	16,291	2.4	0.71
Hawaii	6	0.2	3,424	0.5	0.18
Idaho	14	0.5	2,166	0.3	0.65
Illinois	123	4.3	27,503	4.0	0.45
Indiana	18	0.6	13,213	1.9	0.14
Iowa	10	0.4	7,100	1.0	0.14
Kansas	21	0.7	5,882	0.9	0.36
Kentucky	12	0.4	8,803	1.3	0.14
Louisiana	27	1.0	12,739	1.9	0.21
Maine	13	0.5	4,436	0.7	0.29
Maryland	87	3.1	17,635	2.6	0.49
Maine	54	1.9	29,553	4.3	0.18
Michigan	81	2.9	24,719	3.6	0.33
Minnesota	30	1.1	10,951	1.6	0.27
Mississippi	9	0.3	4,323	0.6	0.21
Missouri	27	1.0	14,431	2.1	0.19
Montana	2	0.1	2,105	0.3	0.10
Nebraska	4	0.1	3,886	0.6	0.10
Nevada	19	0.7	3,379	0.5	0.56
New Hampshire	10	0.4	3,821	0.6	0.26
New Jersey	100	3.5	24,382	3.6	0.41
New Mexico	13	0.5	3,781	0.6	0.34
New York	475	16.7	57,693	8.3	0.84
North Carolina	59	2.1	17,466	2.6	0.34
North Dakota	0	0	1,816	0.3	0.00
Ohio	45	1.6	28,311	4.2	0.16
Oklahoma	20	0.7	6,342	0.9	0.32
Oregon	56	2.0	8,130	1.2	0.69
Pennsylvania	83	2.9	34,118	5.0	0.24
Rhode Island	7	0.2	2,558	0.4	0.27
South Carolina	26	0.9	8,691	1.3	0.30
South Dakota	0	0	1,560	0.2	0.00
Tennessee	33	1.2	14,197	2.1	0.23
Texas	166	5.8	40,211	5.9	0.41
Utah	17	0.6	4,693	0.7	0.36
Vermont	2	0.1	2,078	0.3	0.10
Virgina	65	2.3	13,678	2.0	0.48
Washington	108	3.8	13,904	2.0	0.78
West Virginia	4	0.1	4,431	0.7	0.09
Wisconsin	33	1.2	13,119	1.9	0.25
Wyoming	2	0.1	1,088	0.2	0.18
Puerto Rico, Territories	7	0.2	6,506	1.0	0.11
Total	2,939	100	680,747	100	0.42

SOURCE: Centers for Medicare & Medicaid Services: Data from the Opt-Out Carrier Reports and the Medicare Provider Identifier and Eligibility Registry, 1998-2002.

Some of the differences described here reflect variation in the distribution of physician/provider specialties across opt-out and non-opt-out provider groups. Previous studies have indicated that provider participation in Medicare varies by specialty group (Gillis, Lee, and Wilke, 1992). Because of this, disaggregated data for two specialty groups, psychiatrists and primary care physicians (general and family practice and internal medicine) will be examined.

Specialty Group Variation

Table 3 shows the (1) distribution of opt-out and active providers by specialty, (2) the number and percent of opt-outs by specialty, and (3) the number and percent of active providers in specialties eligible to private contract. Also shown in Table 3 are the percent of opt-outs among eligible active providers. Psychiatry was the largest specialty group to opt-out, representing 31.1 percent of all opt-outs, and had the highest percent of a major specialty group opt out (2.5 percent). Nearly 43 percent of opt-outs providers were in specialties providing psychiatric care. However, the within specialty opt-out rates for clinical psychologists and clinical social workers are far lower than those for psychiatrists.

Table 3. Distribution of Opt-Out and Active Providers, by Specialty: 1999-2002.

Specialty	Opt-Out Providers	All Active Percent	All Active Providers	Percent of All Active	Percent of Specialty
Total	2,839	100.0	682,747	100.0	100.00
Primary Care	717	25.3	205,528	30.1	0.35
General Practice	192	6.8	29,199	4.3	0.66
Family Practice	307	10.8	76,399	11.2	0.40
Internal Medicine	218	7.7	99,930	14.6	0.22
All Psychiatric Providers	1,213	42.7	108,835	15.9	1.11
Psychiatry	882	31.1	34,946	5.1	2.52
Neuropsychiatry	10	0.4	159	0	6.29
Clinical Psychology	195	6.9	34,157	5.0	0.57
Clinical Social Work	126	4.4	39,573	5.8	0.32
Obstetrics/Gynecology	249	8.8	35,886	5.3	0.69
Dermatology	58	2.0	9,478	1.4	0.61
Plastic and Reconstructive Surgery	84	3.0	5,380	0.8	1.56
General Surgery	45	1.6	27,667	4.1	0.16
Anesthesiology	38	1.3	35,129	5.1	0.11
Cardiology	11	0.4	19,488	2.9	0.06
Orthopedic Surgery	49	1.7	22,408	3.3	0.22
Diagnostic Radiology	26	0.9	24,656	3.6	0.11
Emergency Medicine	35	1.2	27,080	4.0	0.13
All Other Specialties	314	11.1	161,212	23.6	0.19

SOURCE: Centers for Medicare & Medicaid Services: Data from the Opt-Out Carrier Reports and the Medicare Provider Identifier and Eligibility Registry, 1998-2002.

Primary care physicians (general practice, family practice, internal medicine) accounted for over 25 percent of all opt-outs. Other specialties with a disproportionate number of opt-outs are obstetrics/gynecology, dermatology, and plastic and reconstructive surgery. Although cardiologists, anesthesiologists, orthopedic surgeons, emergency medicine specialists, radiologists, and general surgeons were specialties negatively impacted by Medicare fee schedule relative value unit reductions from 1997-2000, very few of these physicians chose to opt out.

Table 4 compares characteristics of opt-out and non-opt-out providers for two specialties: psychiatrists (including neuropsychiatry) and primary care physicians (general practice, family practice, and internal medicine). These analyses show the differences between opt-out and non-opt-out providers within specialty.

Table 4. Characteristics of Opt-Out and Non-Opt-Out Providers, by Specialty: 1999-2002.

Characteristic	Psychiatrists1		Primary Care2
	Opt-Out	Non-Opt-Out	Opt-Out	Non-Opt-Out
Number	892	34,236	717	204,834
Percent Active	97.4	100.0	96.8	100.0
Age	63	55	54	53
Number of Practice Settings	1.9	2.1	3	2.4
	Percent
Participating Providers	13.5	70.8	35.3	77.5
Solo Practice	88.5	55.6	59.3	42.5
Group Practice	11.5	44.4	40.7	57.5

¹Includes neuropsychiatry.

²Includes general and family practice and internal medicine.

SOURCE: Centers for Medicare & Medicaid Services: Data from the Opt-Out Carrier Reports and the Medicare Provider Identifier and Eligibility Registry, 1998-2002.

For psychiatrists, 2.6 percent of opt-out providers were inactive providers. Many of the inactive psychiatrists may have been either in a hospital-based practice or in a specialized practice (e.g., child psychology, substance abuse) where the likelihood of treating Medicare patients was low.

Opt-out psychiatrists had fewer practice sites than psychiatrists remaining in the Medicare Program. While psychiatrists are more likely to be in solo practice and less likely to be participating providers than other providers, only 13.5 percent of opt-out psychiatrists were participating providers and 88.5 percent of opt-out psychiatrists were in solo practice. In contrast, 70.8 percent of non-opt-out psychiatrists were participating providers and only 57.5 percent are in solo practice.

For primary care physicians, 3.2 percent of opt-outs were inactive providers. Opt-out primary care physicians also are more likely to be in solo practice than primary care physicians remaining in the program. Only 35.3 percent of primary care physicians opting out are participating providers compared to 77.5 percent of primary care physicians who did not opt out.

Billing Patterns

Table 5 shows data on billing patterns of opt-out and non-opt-out providers for those providers billing Medicare in 1997. These data show that opt-out providers, as a group, were responsible for a minuscule percentage of total submitted Medicare claims (0.17 percent) and total allowed charges (0.15 percent) during 1997. Individually, opt-out providers submitted fewer claims for a smaller average annual amount than providers remaining in the Medicare Program. Although providers who did not opt out submitted nearly 93 percent of claims and 94 percent of charges as participating providers, opt-out providers submitted only 66.3 percent of their Medicare claims and 72.0 percent of their Medicare charges as participating providers. They were far more likely to submit non-assigned claims than other providers. On average, opt-out providers submitted over one-third as many claims (643) as other providers (1,732). Average annual allowed charges per provider for opt-out providers were $27,932. This was only about 32 percent as great as those of non-opt-out providers ($86,495). However, these patterns vary by specialty group.

Table 5. Billing Patterns of Opt-Out and Non-Opt-Out Providers Who Billed Medicare, by Speciality Group: 1997.

Variable	All Specialties		Psychiatry		Primary Care
	Opt-Out	Non-Opt-Out	Opt-Out	Non-Opt-Out	Opt-Out	Non-Opt-Out
Number of Providers	2,188	474,492	704	25,046	574	146,055
Number of Claims Submitted	1,406,483	821,767,950	124,138	16,269,286	613,220	308,059,206
Total Allowed Charges (Millions)	$61.1	$41,041.4	$8.3	$940.8	$19.7	$10,596.3
Percent Participating of Total Claims	66.3	92.9	40.9	86.9	68.3	90.0
Percent Non-Assigned of Total Claims	23.1	2.6	43.3	3.4	21.2	3.6
Average Number of Claims Submitted	643	1,732	176	650	1,068	2,109
Average Annual Allowed Charges per Provider	$27,932	$86,495	$11,738	$37,564	$34,320	$72,550

SOURCE: Centers for Medicare & Medicaid Services: Data from the Opt-Out Carrier Reports and the Medicare Provider Identifier and Eligibility Registry, 1998-2002 and Medicare Part B Claims, 1997.

Both psychiatrists and primary care physicians show differences in patterns of billing from those for all specialties. Psychiatrists who opted out were far more likely to submit non-assigned claims than non-opt-out psychiatrists. They also submitted far fewer claims for fewer dollars than other psychiatrists. Primary care physicians submit more claims for lower average amounts than the average Medicare provider. Those who opted out of Medicare were far more likely to submit non-assigned claims than other providers. Opt-out primary care physicians submitted far fewer claims for fewer dollars than other providers.

Geographic Variation in Supply

In order to determine if any small areas were adversely impacted by providers opting out of Medicare, the percentage of opt-out providers was examined by three-digit ZIP Code areas for all providers. Table 6 shows the 15 ZIP Code areas with the highest percentages of opt-out to total active providers. Relatively few small areas were disproportionately affected by providers opting out of Medicare. Only five areas had more than 3 percent of local physicians/providers opt out. Some of the areas have small populations and additional losses of Medicare providers either through opting out, moving, retirement, or death could potentially create a health care provider shortage. In these 15 areas, rates of opting out would have to be significantly higher to affect beneficiary access.

Table 6. Highest Percentages of Opt-Out to Total Active Providers, by ZIP Code Areas: 1997.

Area1	Active	Opt-Out	Percent
Virgin Islands (008)	112	4	3.57
Anchorage, Alaska (995)	692	24	3.47
Boulder, Colorado (803)	434	14	3.23
Annapolis, Maryland (214)	281	9	3.20
Greenwich-Norwalk, Connecticut (068)	971	30	3.09
Berwick, Maine (039)	68	2	2.94
Spencer, Iowa (513)	104	3	2.88
Guymon, Oklahoma (739)	35	1	2.86
Rock Springs - Jackson, Wyoming (830-831)	74	2	2.70
New York, New York (100-101)	12,015	321	2.67
Santa Barbara, California (931)	561	14	2.50
Palo Alto, California (943)	953	23	2.41
Capon Bridge, West Virginia (267)	42	1	2.38
Santa Fe, New Mexico (875)	389	9	2.31
Prescott - Sedona, Arizona (863)	229	5	2.18

¹Three-digit ZIP Codes are shown in parentheses.

SOURCE: Centers for Medicare & Medicaid Services: Data from the Opt-Out Carrier Reports and the Medicare Provider Identifier and Eligibility Registry, 1998-2002.

Since the vast majority of local areas had negligible numbers of providers opt out, access problems that may occur in these areas may be due to factors other than opt-out-induced changes in provider supply. It does not appear likely, even in the 15 areas identified in Table 6, that any Medicare beneficiaries were limited to seeking treatment from providers who had opted out because there were no other available providers in the area. Although the Seattle, Washington area has been identified as having the lowest percent of physicians (55 percent) willing to accept all new Medicare patients in 12 study areas (Trude and Ginsburg, 2002), providers electing private contracting were less than 2 percent of active physicians.

There is a possibility that provider opt-out may lower use of Part B services in other ways. For example, the process of changing providers may not be smooth for all Medicare beneficiaries, which may result in beneficiaries not seeking medical care (even from other providers) because their regular provider has opted out.

Beneficiary Demographics

In the 12 months prior to opt-out, 27,203 beneficiaries were seen by providers in the 1998 opt-out cohort (Table 7). Only 502 beneficiaries (1.8 percent) had no claims for services from non-opt-out providers during the 12-month period prior to opt-out. Thus, this small group received all Medicare services from providers who chose to private contract in 1998. In the 12 months after the 1998 cohort opted out, only 619 beneficiaries (2.3 percent) received no covered services from active Medicare (non-opt-out) providers. These beneficiaries may have remained with their past providers and were treated under private contract. Alternatively, this group may not have required Medicare services during 1998.

Table 7. Covered Services, by Beneficiary Demographics and Specialty Group: 1997-1998.

Variable	1997 Opt-Out (n=27,203)	1997 Non-Opt-Out (n=26,701)	1998 Non-Opt-Out (n=26,584)	Change 1998–1997	Percent Change
Total	193,706	1,909,286	1,908,873	-194,119	-9.2
Age
0-64 Years	26,379	251,237	230,507	-47,109	-17.0
65-74 Years	99,160	912,832	918,659	-93,333	-9.2
75-84 Years	56,076	616,724	615,705	-57,095	-8.5
85 Years or Over	12,091	128,493	144,002	3,418	2.4
Sex
Male	68,439	748,597	765,120	-51,916	-9.2
Female	125,267	1,160,689	1,143,753	-142,203	-11.0
Race
White	184,757	1,829,284	1,817,786	-196,255	-9.7
Black	3,094	37,941	48,612	7,577	18.5
Asian	81	5,202	3,553	-1,730	-32.7
Hispanic	223	12,091	16,146	3,832	31.1
Native American	18	1,827	1,297	-548	-29.7
Other	378	22,941	21,479	-1,840	-7.9
Specialty Group
All Psychiatrists	21,208	47,424	46,919	-21,713	-31.6
Psychiatry	19,857	27,882	27,048	-20,691	-43.3
Clinical Psychiatry	966	12,229	12,281	-914	-6.9
LCSW	355	7,313	7,590	-78	-1.0
All Primary Care	88,606	236,888	267,720	-57,774	-17.7
General Practice	30,910	22,533	25,225	-28,218	-52.8
Family Practice	39,145	58,845	70,545	-27,445	-28.8
Internal Medicine	18,551	155,510	171,950	-2,111	-1.2
Obstetrics/Gynecology	3,766	13,720	13,405	-4,081	-23.3
Dermatology	23,100	49,864	56,340	-16,624	-22.8
Plastic Surgery	1,761	2,698	2,480	-1,979	-44.4
Anesthesiology	2,931	459,710	367,712	-94,929	-20.5
Cardiology	0	85,122	92,256	7,134	8.4

NOTE: LCSW is licensed clinical social worker.

SOURCE: Centers for Medicare & Medicaid Services: Data from the Medicare Beneficiary Enrollment and Part B Claims Files, 1997-1998.

Fifty Two percent of beneficiaries seen in 1997 by opt-out providers in the 1998 cohort were from age 65 to 74 (n=14,146), while beneficiaries between age 75 to 84 were 30.1 percent of the study cohort (n=8,194). Beneficiaries age 85 or over comprised 7.0 percent of the study cohort (n=1,897). Non-elderly beneficiaries (age 64 or under) were 10.9 percent of the study cohort (n=2,966). When compared to the distribution of Medicare beneficiaries, beneficiaries age 65 to 74 were more likely to have been treated by providers who later chose to private contract while non-elderly beneficiaries and beneficiaries age 85 or over were less likely to have been treated by a provider who opted out in 1998.

In the 1998 opt-out cohort there were 37.4 percent males (n=10,429) and 62.6 percent females (n=16,774) in 1997. Compared to the population of Medicare beneficiaries in 2001 (43.4 percent male, 56.6 percent female), females were slightly more likely to be treated in 1997 by a 1998 opt-out provider. While white beneficiaries represented 85 percent of Medicare beneficiaries in 2001 (Centers for Medicare & Medicaid Services, 2002), 94.9 percent of beneficiaries treated by opt-out providers in the 1998 cohort were white persons. Of the beneficiaries treated, 2.5 percent were black, 0.8 percent were Hispanic, 0.3 percent were Asian, 0.1 percent were Native Americans, and 1.4 percent were from other backgrounds. Beneficiaries from these racial groups were less likely to be treated by a provider who chose to private contract in 1998.

Services Provided

Opt-out providers delivered 193,706 services to these beneficiaries prior to opting-out (Table 7). However, opt-out providers were not the only providers of Medicare Part B services to these beneficiaries. Non-opt-out providers provided 1,909,286 services to these beneficiaries during this period. Thus, opt-out providers provided only 9.2 percent of all Medicare services to this beneficiary cohort in the year prior to opt-out. After these providers had opted-out, providers remaining active in Medicare provided 1,908,873 covered services to this cohort of Medicare beneficiaries. Although total Medicare covered services declined by 9.2 percent during the 12 months after providers opted-out, the Medicare beneficiaries that saw opt-out providers in 1997 were able to obtain care from providers who remained active in Medicare in 1998, many of whom may have treated these Medicare beneficiaries during the previous year. Again, some of the drop off observed may be because no Part B services were needed during this period.

As shown in Table 7, non-elderly beneficiaries showed the highest percent of drop off in Medicare service use after provider opt-out in 1998. Beneficiaries in this cohort (age 85 or over) actually had an increase in Medicare Part B services during this period. Although the drop off in services used was higher for females, there is probably not a significant difference by sex. While drop off in Part B service use by white persons differed little from the overall trend, large percent increases were noted for black persons and Hispanics and a large percent decrease for Asians. While Part B service use by black persons in this group was not affected by the introduction of private contracting, the small number of beneficiaries in the other non-white groups limits the conclusions one can draw concerning any observed changes in Part B service use.

In the 12 months after opt-out (Table 7), services to beneficiaries served in 1997 by opt-out providers delivered by psychiatrists, general practice physicians, family practice physicians, obstetricians/gynecologists, dermatologists, plastic/reconstructive surgeons, and anesthesiologists declined sharply. In contrast, services delivered by clinical psychologists, licensed clinical social workers and internal medicine specialists showed little drop off and cardiologists show an increase in services delivered. This reflects the differences in specialty private contracting rates (Table 3) and may indicate that some beneficiaries may have chosen to enter into private contracts with psychiatrists, general practice physicians, and plastic/reconstructive surgeons who they received care from during 1997. One caution to be noted is that changes observed in billed services from year to year may reflect changes in billing patterns (e.g., unbundling of services) between 1997 and 1998. As such, annual changes in number of services are likely to exhibit greater change than covered charges from year to year.

Medicare Covered Charges

Covered charges for Medicare services provided by opt-out providers totaled $7,003,370 (Table 8). In comparison, covered charges for services provided by non-opt-out physicians totaled $59,440,497. Covered charges for services provided by opt-out physicians represented only 10.5 percent of total Medicare covered charges for services provided to these beneficiaries. In the year after opt-out, Medicare covered charges totaled $64,110,532, representing a drop-off of 3.5 percent. Non-elderly beneficiaries were the only age group for which a major drop off in covered charges (11.2 percent) was noted. Although the drop-off in covered charges for females was higher than for males, neither change differs very much from the overall decline in covered charges during this period. Medicare covered charges for black persons, Hispanics, and Native Americans increased from 7.8 to 9.3 percent while covered charges for Asians declined by 13.6 percent. Again, except for black persons, the number of beneficiaries in the remaining groups is small and inferences concerning change in service use should be made with caution.

Table 8. Covered Services, by Beneficiary Demographics and Specialty Group: 1997-1998.

Variable	1997 Opt-Out (n=27,203)	1997 Non-Opt-Out (n=26,701)	1998 Non-Opt-Out (n=26,584)	Change 1998–1997	Percent Change
Total	$7,003,370	$59,440,497	$64,110,532	-$2,333,335	-3.5
Age
0-64 Years	1,032,121	7,903,930	7,931,427	-1,004,624	-11.2
65-74 Years	3,502,190	27,293,116	29,967,929	-827,377	-2.7
75-84 Years	2,031,784	19,777,597	21,303,395	-505,986	-2.3
85 Years or Over	437,275	4,465,854	4,907,782	4,653	0.1
Sex
Male	2,654,990	23,880,692	26,187,084	-348,598	-1.3
Female	4,348,380	35,559,806	37,923,449	-1,984,737	-5.0
Race
White	6,667,756	56,917,591	61,273,951	-2,311,396	-3.6
Black	134,032	1,158,812	1,413,604	120,760	9.3
Asian	18,111	140,901	137,426	-21,586	-13.6
Hispanic	60,240	425,924	524,352	38,188	7.8
Native American	3,642	30,882	37,742	3,218	9.3
Other	109,319	766,387	723,458	-152,248	-17.4
Specialty Group
All Psychiatry	1,537,546	3,382,382	3,475,734	-1,444,194	-29.4
Psychiatry	1,443,875	1,954,327	1,976,406	-1,421,796	-41.8
Clinical Psychiatry	72,946	974,433	995,448	-51,931	-5.0
LCSW	20,725	453,622	503,880	29,533	6.2
All Primary Care	2,110,451	7,758,126	9,108,673	-759,904	-7.7
General Practice	629,944	681,792	833,095	-478,641	-36.5
Family Practice	849,613	1,741,168	2,169,052	-421,729	-16.3
Internal Medicine	630,894	5,335,166	6,106,524	140,464	2.4
Obstetrics/Gynecology	184,116	765,742	771,097	-178,761	-18.8
Dermatology	906,491	177,593	2,421,602	1,337,518	123.4
Plastic Surgery	312,718	350,173	307,177	-355,714	-53.7
Anesthesiology	200,305	1,692,049	1,687,145	-205,209	-10.8
Cardiology	0	4,822,517	5,479,113	656,596	13.6

NOTE: LCSW is licensed clinical social worker.

SOURCE: Centers for Medicare & Medicaid Services: Data from the Medicare Beneficiary Enrollment and Part B Claims Files, 1997-1998.

As shown in Table 8, large decreases in covered charges after opt-out were noted for care from psychiatrists, general practice physicians, and plastic surgeons. Some substitution across specialties may be occurring. Licensed clinical social workers may be substituting for psychiatrists, internists for general practice physicians, and dermatologists for plastic/reconstructive surgeons. This may also account for the unusually large increase in covered charges for dermatologists during this period.

Discussion

Despite the many anecdotal reports of widespread discontent and dissatisfaction with Medicare payment rules and administrative procedures among physicians and other providers paid under the Medicare Part B RBRVS, very few providers found private contracting attractive. Although private contracting can be considered a method for boycotting the Medicare Program (Hirschman, 1970), providers who have opted-out represent a very small percentage of active eligible providers nationally as well as within States and three-digit ZIP Code areas. These providers tended to have small-scale practices and were not the major source of Medicare services for beneficiaries affected by providers choosing private contracting. Indeed, it appears that many beneficiaries chose to obtain care under Medicare from other providers after the providers that they were receiving care from elected to opt-out of Medicare. Thus, the impact of private contracting on access to Medicare Part B services is very small, and it appears unlikely that opt-out providers were the only available source of needed care. Private contracting also appears to have had minimal impact on Medicare Part B service use by minorities and beneficiaries age 85 years or over.

Providers who opted-out of Medicare are very different from providers remaining in Medicare with respect to specialty, practice characteristics, and Medicare Program activity. Providers who opted-out are primarily drawn from a small number of specialty groups. Even when compared only to active providers within their specialty, providers who opted out of Medicare stood out as markedly different. These providers are much less likely to have been Medicare participating providers, are more likely to be in solo practice, and have markedly fewer Medicare billings for fewer allowed charges than providers remaining active in Medicare. One can suggest that these providers have less of a reason to remain affiliated with the Medicare Program.

These data imply that many physicians and other providers treat Medicare patients despite serious specialty-specific MFS payment policy concerns or other economic incentives that may appear to make private contracting attractive. Providers who have opted-out of Medicare appear to have little incentive for interaction with the Medicare Program and, thus, may be viewed as peripheral to the program. They often are from specialties with longstanding differences with Medicare physician payment policy and have consistently chosen to buffer themselves from the impact of changes in Medicare payment policy by maintaining low participating provider rates.

The specialty groups with the highest numbers of opt-out providers, however, were not major losers affected by Medicare physician fee schedule cuts from 1994 to 2000. Rather, these specialty groups have had lower than average assignment and participating provider rates throughout the history of the program. Concern with medical record documentation requirements, oversight and enforcement, and audits of physicians by the Department of Justice for Medicare fraud may be a valid concern, but is clearly not the major reason why providers opted-out as these results indicate. Further monitoring will, of course, be necessary to determine changes in the number and type of providers opting-out of Medicare in future years, especially since the 2002 to 2004 reductions in payment levels.