Participation Of Plans And Providers In Medicaid And SCHIP Managed Care

Transcription

1 Participation Of Plans And Providers In Medicaid And SCHIP Managed Care While eleven large states report that they have been able to attract enough plans and providers, the current economic climate will create challenges. by Marsha Gold, Jessica Mittler, Debra Draper, and David Rousseau ABSTRACT: For Medicaid and SCHIP managed care programs to succeed, they must attract enough and the right kinds of plans and providers to meet access and care goals. In 2001 we analyzed practices and perceptions that bear on these goals by surveying managed care plans participating in Medicaid or SCHIP, or both, in eleven states. Participating plans appear supportive of both programs and are largely able to secure providers to participate, too. To date, SCHIP has not attracted many plans not already participating in Medicaid. While perceptions were positive in 2001, maintaining current plan and provider relationships in an environment that has become much more budget constrained will be challenging. Medicaid and the State Children s Health Insurance Program (SCHIP) are important sources of health coverage and access to care for low-income and other vulnerable Americans without job-based insurance. 1 In recent years many states, at least in urban areas, have structured their Medicaid programs around capitated (prepaid) managed care models. 2 Such models also are a key component of most SCHIP programs, and some states have used the authority in SCHIP legislation to depart from traditional Medicaid models to encourage broader plan participation and access to providers. For managed care to succeed, states must attract enough participating plans and providers, which in turn need to be able to provide access to high-quality care. Access has long been an issue in the traditional Medicaid program, with its low payment levels, unfavorable attitudes among providers about public programs, and geographic maldistribution, among other factors, which limit providers participation. 3 In some states and programs, mainstreaming giving beneficiaries access to providers that treat commercial enrollees has been an important goal. Marsha Gold is a senior fellow, Jessica Mittler a health researcher, and Debra Draper a senior health researcher with Mathematica Policy Research. David Rousseau is a senior policy analyst with the Kaiser Commission on Medicaid and the Uninsured. 230 January/February Project HOPE The People-to-People Health Foundation, Inc.

2 Medicaid & SCHIP But states also are concerned about protecting the safety net of providers that traditionally serve the low-income population, whether or not they are insured. 4 In this paper we report on findings from a 2001 survey of Medicaid and SCHIP managed care plans in the eleven states with the largest Medicaid managed care enrollment excluding Tennessee and Oregon, which have been extensively surveyed. The surveyed states are not necessarily representative of practices across the nation, but they capture the experience of a sizable share of participating plans and enrollees. We report here findings that are relevant to four key questions. First, does SCHIP attract a different set of plans than Medicaid? Second, how do plans perceive these programs relative to one another? Third, do plans now in Medicaid and SCHIP perceive their participation to be viable in the long run, and what factors determine this? And fourth, to what extent do plans provide access to a medical home and mainstream care, and does this differ by plan type? Study Methods Survey parameters. The survey was fielded largely by telephone between April and August 2001, in Arizona, California, Florida, Maryland, Michigan, Missouri, New York, New Jersey, Pennsylvania, Texas, and Washington. In California we limited the survey to plans under contract in Los Angeles and Orange Counties, because California s managed care models vary extensively across counties and we wanted to apply the available survey resources to cover more states. 5 The survey includes all capitated plans in these states that participated in Medicaid or SCHIP, or both, with at least 5,000 enrollees in both programs combined. The areas included in the survey covered half of all full-risk enrollment in Medicaid and half of SCHIP enrollment nationwide in Response rates and plan characteristics. From a sample frame of 155 plans, 133 were eligible for the survey, and 109 responded. 7 The response rate averaged 82 percent. Response rates were relatively similar for plans with different characteristics, with no major differences between for-profit and nonprofit plans, commercial versus Medicaid-dominant plans, and plans of different sizes (Exhibit 1). 8 Response rates varied across states, from a high of percent (New Jersey, Texas, and Washington) to a low of 7 percent (Arizona) and 50 percent (Florida). 9 Because of the low response, Florida s results may not be representative of that state. We considered weighting the data to account for biases in responding plans by state, but we did not. The key reason is that the number of plans is too small to support development of appropriate adjustments that introduced less error than they removed. Small numbers are a problem for any survey of managed care plans, but this is a particular problem in studying state-based programs such as Medicaid and SCHIP, which contract with a few plans in each state. SCHIP models. States typically follow one of two SCHIP models: establishing a stand-alone program or running their program through Medicaid using policies and procedures that are relatively consistent with those in Medicaid. Six states HEALTH AFFAIRS ~ Volume 22, Number 1 231

3 EXHIBIT 1 Characteristics Of Responding And Nonresponding Plans In The Medicaid/SCHIP Survey, 2001 All eligible plans Responding Response rate All plans % Commercial Medicaid-dominant Unknown For-profit Nonprofit Unknown Plan age 10 years or more 5 9 years Less than 5 years Unknown Affiliation National firm Blue Cross None Total enrollment Under 50,000 50,000 99,999,000 or more Unknown Medicaid enrollment Under 10,000 50,000 99,999,000 or more Unknown SCHIP participant Yes No State Arizona California a Florida Maryland Michigan Missouri New Jersey New York Pennsylvania Texas b Washington SOURCE: Mathematica Policy Research, survey of State Children s Health Insurance Program (SCHIP) plans for the Henry J. Kaiser Family Foundation, NOTE: Based on data available to draw the sample. a In California, we surveyed eligible plans participating in the programs in Los Angeles and Orange Counties. We did not survey the partially capitated physician-hospital consortia in Orange County. b Seven SCHIP-only plans were excluded because their eligibility for the study could not be determined at the time. California, Florida, Michigan, New York, Pennsylvania, Texas have stand-alone SCHIP programs. The remaining five do not vary their managed care models or policies for Medicaid and SCHIP. The comparisons of Medicaid to SCHIP are based 232 January/February 2003

4 Medicaid & SCHIP around these six stand-alone states and focus on comparing Medicaid and these distinct SCHIP programs. Distinct SCHIP plans are those that either (1) only have a SCHIP, not a Medicaid product; or (2) have both, but plans said that the SCHIP product is only somewhat or not very similar to Medicaid. 10 Study Findings Does SCHIP attract a different set of plans? Most commonly, the same plans participate in both Medicaid and SCHIP (Exhibit 2). This pattern appears to represent voluntary decisions made by the plans rather than mandates by the states. Only one plan in states with both Medicaid and a stand-alone SCHIP program reported that it was required to participate in SCHIP as a condition of doing business in Medicaid. 11 Of the twenty-five plans participating in Medicaid only, 3 percent said they had the opportunity to participate in SCHIP but did not because rates were too low (four plans), anticipated enrollment was too small (three plans), there was not an adequate provider network (one plan), SCHIP contract terms did not meet plan requirements (three plans), or the SCHIP product did not fit into the company s strategy (four plans). There is no clear profile of a Medicaid or SCHIP participating plan (Exhibit 3). Total enrollment varies greatly, with close to an equal share having enrollments above and below,000 enrollees. Medicaid/SCHIP enrollment made up 75 percent or more of total plan enrollment in 4 percent of Medicaid plans and 53 percent of plans with distinct SCHIP products, a fact that probably explains the low rate of National Committee for Quality Assurance (NCQA) accreditation. 12 EXHIBIT 2 Participating Plans, By Medicaid/SCHIP Product Mix, 2001 SCHIP only (7 plans) % Separate SCHIP (42 plans) 38% Medicaid only (25 plans) 23% SCHIP through Medicaid, with similar policies (29 plans) 27% SCHIP through Medicaid, but productisdistinct(plans) a % SOURCE: Mathematica Policy Research survey of Medicaid/State Children s Health Insurance Program (SCHIP) plans for the Henry J. Kaiser Family Foundation, NOTE: Plans with SCHIP-only products are understated because these products typically were in Texas, and at the time of the survey, available information was insufficient to assess whether seven SCHIP-only plans met the eligibility requirements. a These plans said that their SCHIP product was only somewhat or not very similar to Medicaid. These plans are analytically treated the same as plans in separate SCHIP programs, a category we term distinct SCHIP. See Note 10 in text. HEALTH AFFAIRS ~ Volume 22, Number 1 233

5 EXHIBIT 3 Key Characteristics Of Medicaid And Distinct SCHIP Plans, 2001 For profit Nonprofit Government Plan operational Before or later Ownership Blue Cross/Blue Shield Commercial insurance National managed care company a Provider sponsored b Other Part of multistate organization Yes No NCQA accredited Yes Scheduled No Total enrollment c 25,000 or less 25,000 50,000 50,000,000, ,000 Over 250,000 Medicaid/SCHIP dominant (75+ enrollment) Commercial Any Medicaid (n=102) 52% Any distinct SCHIP (n=55) 4% SOURCE: Mathematica Policy Research, survey of State Children s Health Insurance Program (SCHIP) plans for the Henry J. Kaiser Family Foundation, NOTE: NCQA is National Committee for Quality Assurance. a National managed care companies are not licensed as a commercial insurance company. b Provider-sponsored plans have 50 percent or more of plan ownership held by providers or provider organizations. c For this variable, the valid number of plans responding is 94 for Medicaid and 48 for SCHIP. SCHIP enrollment levels are much lower, on average, than those in Medicaid. Among distinct SCHIP plans, 34 percent had fewer than 5,000 enrollees; 7 percent had under 25,000 enrollees. In contrast, combined Medicaid/SCHIP enrollment in plans run through Medicaid was 25,000 or more in 58 percent of plans, and fewer than 5,000 in only 3 percent of plans. These differences reflect Medicaid s longer history, its greater enrollment (almost ten times larger than SCHIP), and SCHIP s more limited eligibility criteria. (At the time of the survey, enrollees generally were children in families whose incomes were too high to qualify for Medicaid.) SCHIP s low enrollment levels could be responsible for the extensive overlap in plans participating in both programs, since plans often feel they need a minimum number of enrollees to make participation fiscally feasible, and at 234 January/February 2003

6 Medicaid & SCHIP the time of the survey this minimum may only have been possible through participation in both programs. However, by the time the survey was fielded in 2001, SCHIP enrollment had grown substantially only in Washington was it very small (7,21). Enrollment exceeded,000 in all but three other states surveyed: Arizona (8,83), Michigan (7,181), and New Jersey (99,847). 13 How do plans perceive Medicaid and SCHIP? In each program, plans perceptions of the state as a business partner are generally positive. At least 70 percent of plans in each program agree that the state understands managed care; actively solicits information from plans; wants a long-term commitment with plans through multiyear contracts; and is easy to work with in terms of complying with reporting, quality assurance, or utilization review requirements. The predictability and stability of requirements are issues for plans in each program, however, as is the equitability of how plans enrolling those with differing health status risk are treated. Perceptions across programs differ most on administrative burden and encounter-reporting requirements, with plans participating in Medicaid reporting greater burden. State-by-state perceptions. We also examined data on a state-by-state basis. This comparison did not show dramatic or consistent differences by state, except in Arizona, where plans characterized their relationship with the state as overwhelmingly positive. 14 In the six stand-alone states, comparisons of Medicaid versus SCHIP were mixed. In California, SCHIP is clearly viewed as more attractive, and this is also true, but to a lesser extent, in Florida and Pennsylvania. In all three of these states SCHIP is administered completely apart from Medicaid, with different personnel, administrative systems, and authority. In Michigan, New York, and Texas plans views were more mixed, and plans in Michigan actually viewed Medicaid more favorably than they viewed SCHIP. 15 Ability to serve children. In addition, we asked plans participating in both (in the six stand-alone states) to compare the two programs on eight largely administrative features related to serving children (Exhibit 4). The results show substantial variability. When differences exist, plans view SCHIP more favorably on these features than they do Medicaid. In particular, plans see SCHIP as more attractive because they perceive SCHIP to have better patient compliance, improved enrollment stability, higher payment rates, and increased willingness of providers to participate. Plans with other lines of business, especially commercial plans (78 percent of commercial plans, data not shown), find SCHIP more attractive because SCHIP s requirements are more similar to those of their other lines of business. Perceptions on these topics vary across states, however. Plans in Texas, New York, and Florida were particularly likely to view SCHIP more favorably than they viewed Medicaid. Is long-run plan participation viable? When asked in mid-2001, most plans participating in Medicaid or SCHIP expected to participate in the programs the following year. Seventy-eight percent of Medicaid-participating plans and 83 percent of SCHIP-participating plans said they would definitely continue the next year; HEALTH AFFAIRS ~ Volume 22, Number 1 235

7 EXHIBIT 4 Programs For Children By Plans In Both Medicaid And Distinct SCHIP, 2001 Percent of plans reporting Programs are the same SCHIP is better Medicaid is better Payment rates for enrollees Patient compliance Enrollment stability and turnover Provider willingness to participate System demands by patient health needs Ease of marketing and enrolling State performance requirements Similarity of requirements to commercial or other lines of business a 31 34% 5 57% 13 9% SOURCE: Mathematica Policy Research, survey of State Children s Health Insurance Program (SCHIP) plans for the Henry J. Kaiser Family Foundation, 2001 NOTES: Plans were asked how Medicaid and SCHIP compare in serving children for the attributes listed above. N=48. a The item was not applicable for plans without any other lines of business. Among those meeting our definition of commercial plans, 78 percent said SCHIP is better and 22 percent said the programs are the same. most of the others said they probably would. Those for which this was not the case were not concentrated in any one state. Plans make decisions on Medicaid and SCHIP participation separately but see advantages in participating in both, including better coordination of care for people who may shift among programs. Withdrawal from Medicaid would be more likely to cause a withdrawal from SCHIP, plans said, than the reverse. Specifically, 48 percent of plans that participated in both Medicaid and a distinct SCHIP program said they would very likely stay in Medicaid if they left SCHIP; only 31 percent said the reverse. This distinction probably reflects Medicaid s larger enrollment and revenue base, which affects the contribution to fixed costs and a plan s bottom line. Importance of specific factors to continued participation. Because of Medicaid s dominance, we asked plans in Medicaid how important various factors would be to their continued participation. Virtually all (9 percent) said that the level of state capitation rates is very important, followed by the ability to maintain the existing base of affiliated providers (89 percent), community mission (84 percent, including 92 percent of provider-sponsored plans), effect of Medicaid on the bottom line (78 percent), and providers interest and willingness to participate (74 percent). Strategic considerations were important but less dominant: 55 percent said that the ability to position the Medicaid product was very important; 51 percent said that the ability to grow Medicaid enrollment to offset costs was very important. Fewer plans cited the ability to maintain a diversified set of products as important 23 January/February 2003

8 Medicaid & SCHIP (35 percent) (53 percent of plans that were not Medicaid-dominant). 1 From a business perspective, both minimum enrollment levels and financial performance are important to continued participation. Eighty-three percent of Medicaid plans said that they need to meet a minimum Medicaid enrollment level to remain viable (10,000 25,000 members for 52 percent of plans). Reported minimums are lower for commercial than for Medicaid-dominant plans, probably because their greater enrollment base in other product lines allows these plans to spread fixed costs more easily. Sixty-eight percent reported a needed minimum profit margin. Of those reporting a figure, most commonly this was 3 5 percent (38 percent) or lower (58 percent). Among those that are not Medicaid-dominant, more than half said they don t apply the same criteria to Medicaid as they do to their commercial business and expect more profit on the commercial line. Among the ninety plans that reported their 2000 profitability across all lines of business, 71 percent made some profit, 5 percent broke even, and 23 percent reported a negative return overall. Confidence in the future. At the time of the survey, plans were generally confident of the future. Sixty-six percent of Medicaid and 71 percent of distinct SCHIP plans strongly agreed that they were confident the plan would still be operational under the same ownership in three to four years. Half of Medicaid plans and 57 percent of distinct SCHIP plans strongly agreed that they were confident the plan could handle the cost pressures of the market. Sixty percent each of Medicaid plans and SCHIP plans said they strongly agreed that they were confident that their plan s Medicaid/SCHIP plans were viable. (Most of the rest at least somewhat agreed with these three statements.) As might be expected, when asked to name the one thing they would change about the state program to make their plan s continued participation more likely in the future, improved payments were most likely to be named, followed by addressing complaints about program requirements that added to plan costs. To what extent do plans provide a medical home and mainstream care? Both Medicaid and SCHIP plans create a prominent role for primary care physicians in their provider networks. 17 These physicians generally must authorize referrals to specialists (89 percent of plans of each type), but exceptions typically are built in for some high-volume or sensitive services (such as obstetric/ gynecological, mental health/substance abuse, or eye care). 18 Retaining a provider network. Retaining a provider network is critical to plans ability to continue participating in Medicaid and SCHIP. The strength and stability of the network also is important for beneficiaries access and continuity of care. The survey results provide a mixed picture of how easy it is for plans to retain a network (Exhibit 5). On the one hand, most plans say that developing and retaining a network is not generally a problem. Most also describe their relationship with physicians and hospitals as collaborative rather than adversarial. But many report problems renegotiating physician and hospital contracts. More problems exist HEALTH AFFAIRS ~ Volume 22, Number 1 237

9 EXHIBIT 5 Medicaid And SCHIP Programs Ability To Secure Providers, By Program, 2001 All Medicaid (n=102) Distinct SCHIP (n=55) Agree a Disagree b Agree Disagree It is easy to develop and retain a network of Primary care physicians Specialists Hospitals 8% % % % 30 1 Problem c Not a problem Problem Not a Problem Renegotiation of contracts has been a problem for Physicians FQHCs d Hospitals 53% % % % 9 32 Maintaining an adequate network has been a problem for Specialists overall Pediatric subspecialists Obstetrical subspecialists 51% % % % Collaborative e Adversarial f Collaborative Adversarial Plan has a collaborative relationship with Physicians Hospitals 89% 71 11% 29 9% 75 4% 25 SOURCE: Mathematica Policy Research, survey of State Children s Health Insurance Program (SCHIP) plans for the Henry J. Kaiser Family Foundation, 2001 a Includes plans that somewhat or strongly agree. b Includes plans that somewhat or strongly agree. c Includes plans answering somewhat or major problems. d The remaining 13 percent and 18 percent of plans for Medicaid and SCHIP, respectively, have no federally qualified health center (FQHC) contracts or otherwise indicated that this was not applicable. e Includes plans answering somewhat or strongly collaborative. f Includes plans answering somewhat or strongly adversarial. with contracting with specialists, particularly pediatric subspecialists, than with primary care physicians, and problems contracting with physicians are more likely to occur in Medicaid than in distinct SCHIP programs. Mainstreaming. Despite mainstreaming goals, attracting commercial plans to Medicaid or SCHIP does not necessarily result in enrollees seeing the same physicians that commercial enrollees see. Among commercial plans participating in Medicaid, only 19 percent require Medicaid participation from their commercial network, and 81 percent pay different rates for Medicaid than for commercial enrollees. SCHIP has made more progress in mainstreaming: 44 percent of commercial plans require commercial network providers to participate in the plan s SCHIP network, and 48 percent pay different rates for SCHIP enrollees than for commercial enrollees. In fact, all seven of the plans participating in SCHIP only do not use providers other than those in their commercial networks. In contrast, 57 percent of commercial plans in Medicaid go outside of their com- 238 January/February 2003

10 Medicaid & SCHIP mercial network to develop Medicaid provider networks. This strategy is most commonly viewed as a way of maintaining existing care arrangements, including access to safety-net providers, improving geographic access, and ensuring cultural and linguistic competency. While Medicaid provider networks may not be identical to commercial networks even in commercial plans, there appear to be some other favorable changes in care delivery. Plans said that Medicaid managed care has shifted the primary sources of care away from the hospital emergency room, which 71 percent said was the first or second most common source of care before Medicaid managed care. However, data to confirm such a shift are lacking. Conclusions And Policy Implications The findings from this survey are somewhat encouraging. They show that the plans states have attracted to SCHIP and Medicaid appear committed to them. In contrast to common perceptions, plans that participate in both programs view the states with which they contract relatively positively in terms of performance. A key challenge for states will be to maintain these relationships now that economic conditions are less favorable and state budgets are tight. As the survey shows, plans ability to meet both minimum enrollment and profitability criteria is critical to retaining plans in Medicaid, with capitation rates, provider participation, and administrative requirements all important to plans. If states respond to fiscal stress by cutting back on rates or limiting enrollment, plans may have trouble meeting these targets. Lower rates also may make it more difficult to retain the support and continued participation of affiliated providers. Medicaid and SCHIP face many of the same challenges in securing adequate plan and provider participation; SCHIP has benefited to some extent by its unique population (children from families with incomes above Medicaid levels) and more flexible federal requirements, which allow states to structure programs to be more like those of commercial products if they wish. SCHIP does not appear to have done much better than Medicaid has in attracting more commercial plans. Having a managed care program in and of itself does not eliminate disparities in provider access between people covered in public programs like Medicaid and SCHIP and the privately insured. Program differences reflect variation in patient mix, provider payment levels, and geography across public and private programs. 19 The experience captured in the survey is based on largely favorable economic conditions and support for a new program (SCHIP). Policymakers interested in Medicaid and SCHIP should pay special attention to the way states are responding to the current economic climate and budgetary restrictions, which could have a considerable effect on plans and providers participation in Medicaid and SCHIP and ultimately on beneficiaries access to care. This paper is based on research that was funded by the Henry J. Kaiser Family Foundation. Mathematica Policy Research Inc. provided support to prepare this paper. All views expressed in the paper are those of the authors alone. HEALTH AFFAIRS ~ Volume 22, Number 1 239

11 NOTES 1. Institute of Medicine, Coverage Matters: Insurance and Health Care (Washington: National Academy Press, 2001); and S. Rosenbaum et al., The Children s Hour: The State Children s Health Insurance Program, Health Affairs (Jan/Feb 1998): Kaiser Commission on Medicaid and the Uninsured, Medicaid and Managed Care, Medicaid Fact Sheet (Washington: Kaiser Commission, 2001). 3. J. Mitchell, Medicaid Participation by Medical and Surgical Specialists, Medical Care 21, no. 9 (1983): ; J.D. Perloff, P. Kletke, and J.W. Fossett, Which Physicians Limit Their Medicaid Participation, and Why, Health Services Research 30, no. 1 (1995): 7 2; and F. Sloan, J. Mitchell, and J. Cromwell, Physician Participation in State Medicaid Programs, Journal of Human Resources 13 (Supplement 1978): For additional background on these issues, see M.E. Lewin and S. Altman, eds., America s Health Care Safety Net: Intact but Endangered (Washington: National Academy Press, 2000). 5. D. Draper and M. Gold, Customizing Medicaid Managed Care California Style, Health Affairs (Sep/Oct 2000): In Orange County we surveyed only the health plans contracting with Cal-Optima.. Centers for Medicare and Medicaid Services, Medicaid Managed Care Plan Type Breakout Enrollment by State, June 30, 2000 (Baltimore: CMS, 2002); CMS, The State Children s Health Insurance Program Annual Enrollment Report, Federal Fiscal Year 2001 (Baltimore: CMS, 2002); and Medi-Cal Policy Institute, Medi-Cal County Data Book (Oakland, Calif.: MCPI, 2002). 7. Of the twenty-two ineligible plans, seven were sold or merged with another plan in the sample frame (two of which dropped or did not offer Medicaid), three were out of business, three duplicated another plan in the frame with a different name, and two did not serve the state or area we were calling about. 8. See M. Gold et al., Medicaid and SCHIP Managed Care: Plan Participation, Provider Networks, and Quality Improvement (Washington: Henry J. Kaiser Family Foundation, forthcoming) for further discussion of these issues. 9. Seven SCHIP-only plans in Texas were excluded because their eligibility could not be determined. 10. While the concept and terminology distinct SCHIP program is not commonly used, we felt it was an important term to use in a study focusing on managed care. To gain flexibility, a number of states with programs that qualify under federal law as a separate SCHIP program actually administer the program through Medicaid and may use the same managed care contracting vehicles and policies. To help distinguish similar from dissimilar programs (although this is inherently a subjective determination), we asked plans how similar their SCHIP practices are to Medicaid. Only six of the forty-eight distinct SCHIP plans included in the study qualified through this criterion, however; most were stand-alone SCHIP programs. 11. We are uncertain why only one plan would report this versus others in the same state. However, Medicaid is a complex program, and there could be unique circumstances associated with the plan in question. 12. Large private employers typically encourage NCQA review; states do not typically require NCQA accreditation, although some may encourage it or accept it to meet various contract standards. N. Kaye, Medicaid Managed Care: A Guide for States, 5th ed. (Portland, Maine: National Academy for State Health Policy, 2001). Specific data or requirements on specific states are not readily available. Further, Medicaid-dominant plans may not have the resources or incentives to mount the effort needed for such a review. 13. The State Children s Health Program Annual Enrollment Report. 14. The long history of Medicaid managed care in Arizona (the program began in 1982) probably has much to do with these favorable ratings. 15. We cannot extrapolate too much from Texas, because seven SCHIP-only plans were not included. 1. In Medicaid-dominant plans, Medicaid and SCHIP account for 75 percent or more of the enrollment. 17. Plan policies on who can be a primary care physician are relatively consistent in Medicaid and SCHIP, reflecting some flexibility as to the types of providers that can serve in this capacity. In most plans, internal medicine subspecialists can be selected as the primary care physician, at least for enrollees with special health needs. In addition, many plans allow nurse practitioners, physician assistants, and OB/GYNs to function as primary care physicians. 18. See Gold et al., Medicaid and SCHIP Managed Care. 19. B. Smedley, A. Stith, and A. Nelson, eds., Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care (Washington: National Academy Press, 2002); and IOM, Coverage Matters. 240 January/February 2003