Throughout the 1990s the number

MarketWatch Provider Risk Sharing In Medicaid Managed Care Plans Medicaid risk-sharing arrangements are not on the decline, as is risk sharing in other types of health insurance. by Debra A. Draper and Marsha R. Gold ABSTRACT: Provider risk sharing was common throughout the 1990s. Recent evidence suggests waning interest, although no information exists that is specific to Medicaid. This paper examines risk-sharing arrangements in Medicaid managed care through a survey of participating plans in eleven states conducted during 2001. Risk sharing is prevalent among Medicaid-participating plans and often involves traditional providers. The flight from risk that others describe is not yet apparent in Medicaid, but Medicaid s idiosyncrasies might mean that trends appearing in other lines of business do not apply. Throughout the 1990s the number of risk-sharing arrangements between health plans and providers grew rapidly. 1 More recently, however, there is evidence that plans and providers in some markets are losing interest in developing new or continuing existing risk-sharing arrangements. Several factors appear to be contributing to this change: the financial instability of risk-bearing organizations; consumer and provider backlash against managed care; and an increasingly restrictive regulatory environment, which may extend to plan-provider contracting arrangements. 2 In 2000, for the first time, there was a decline in the number of plans reporting capitation arrangements with providers primary care physicians, specialists, and hospitals. 3 Conceptually, risk sharing is attractive on several fronts. For plans, it provides a mechanism for controlling costs; for providers, it preserves their autonomy by shifting to them responsibilities for managing service use, costs, and quality. 4 However, it also has a down side, particularly if the amount of risk transferred is large or payments are not commensurate with expected costs. Many provider organizations have little experience managing risk and lack the necessary infrastructure to manage it effectively. 5 If the degree of risk transferred to providers is more than they can absorb, their continued viability, as well as that of plans with whom they contract, may be jeopardized. This in turns raises the likelihood that the level and quality of care available to enrollees could be negatively affected. There is limited information on the use and form of risk-sharing arrangements, and the information that does exist is not specific to state-based public programs such as Medicaid and the State Children s Health Insurance Program (SCHIP). The most detailed national information comes from a study sponsored by the Medicare Payment Advisory Commission (MedPAC), which involved a telephone survey of a multistage stratified random sample of Debra Draper is a researcher at Mathematica Policy Research in Washington, D.C., where Marsha Gold is a senior fellow. HEALTH AFFAIRS ~ Volume 22, Number 3 159 2003 Project HOPE The People-to-People Health Foundation, Inc.

Health Tracking health maintenance organizations (HMOs) in the sixty markets with the highest HMO penetration. A key focus of that study was the aggregate transfer of risk from plans serving commercial markets to some form of intermediate entity such as an independent practice association (IPA), a large medical group, or a physician-hospital organization (PHO). 6 The survey excluded many plans serving only the Medicaid market. In 1999, when the survey was conducted, approximately half of the HMOs interviewed reported some form of provider risksharing arrangement. These arrangements accounted for half of the HMOs commercial and Medicare enrollees. Plans had similar arrangements for their commercial and Medicare products, and where differences occurred, they typically involved more risk transfer in Medicare. 8 No information was captured on the survey that was specific to risk-sharing arrangements in Medicaid managed care. This paper focuses on provider risk sharing in plans participating in Medicaid managed care in eleven states. It examines the use of risk-sharing arrangements by these plans; the role of traditional safety-net providers, a factor that is particularly relevant in Medicaid; and the scope of responsibilities that risk-sharing arrangements create for both plans and providers. Policy Importance Since the 1990s states have increasingly relied on managed care as the vehicle to organize and deliver health care services to their Medicaid beneficiaries. By 2001, 5 percent of the nearly thirty-seven million Medicaid beneficiaries nationally were enrolled in a managed care arrangement three-quarters were in HMOs, with the remainder in a primary care case management (PCCM) program. 9 States capacity to operate HMO-based Medicaid managed care programs is largely dependent on their ability to attract and retain plan participants. This is made more challenging because regulatory prohibitions in Medicaid preclude participating plans from using, to the same extent, some of the same mechanisms, such as cost sharing, commonly used on the commercial side to manage care and control costs. Consequently, plans participating in Medicaid must place greater reliance on other mechanisms, such as risk sharing, to foster care management while also safeguarding their financial viability, thereby making Medicaid an attractive business in which to continue to participate. 10 A better understanding of how Medicaid plans use risk-sharing arrangements provides important insight not only into ways in which plans approach cost control but also into factors that may influence plan participation more broadly. Even more so in the current context, in which states are facing difficult and mounting fiscal pressures, this understanding is critical as states make decisions about reductions in Medicaid payments, benefits, and eligibility, whichmustbebalancedwiththeirabilityto maintain enough plan participants to keep the HMO-based Medicaid program viable. 11 Study Methods Data for this paper were derived from a telephone survey of health plans participating in Medicaid managed care (and SCHIP) in eleven states with a combined Medicaid and SCHIP enrollment of at least 5,000 members. 12 Our survey states were Arizona, California, Florida, Maryland, Michigan, Missouri, New Jersey, New York, Pennsylvania, Texas, and Washington, which together represent more than half of the Medicaid managed care enrollment nationally. 13 Because California operates various models of Medicaid managed care, which are customized to the state s individual counties, the survey was limited to plans in Los Angeles and Orange Counties. The telephone survey was conducted between April and August 2001, with an overall response rate of 82 percent. We used a descriptive approach to the analysis of the survey data. Given the survey s design and the characteristics of the responses, we conducted data analysis without weighting plans. Weights are typically used to adjust for the sample design to support unbiased estimates of the population and correct for biases in response. Because the survey was struc- 160 May/June 2003

tured as a census of plans in the selected states, weights are unnecessary to account for the use of sampling techniques. In an analysis of nonresponse, response patterns were relatively consistent across plan characteristics, although they varied by state. We considered state-based weighting but rejected it because it would have complicated the analysis and could have introduced other errors of at least equal importance. Questions about provider risk-sharing arrangements were limited to plans participating in Medicaid managed care. 15 Among the 102 Medicaid-participating plans, 65 were Medicaid-dominant (more than 5 percent Medicaid enrollment) and 3 were commercial. Characteristics of these plans varied, but the majority were for-profit; operational since 1990; affiliated with another organization (including a sizable share of provider-sponsored plans); not accredited by the National Committee for Quality Assurance (NCQA); and small to medium in size, with total plan enrollment of 100,000 members or less. Risk-sharing arrangements. The interviews with Medicaid-participating plans covered two types of provider risk-sharing arrangements: global capitation and professional services capitation. Definitions of these arrangements are drawn from the MedPAC work. 16 Global capitation is defined as arrangements in which all or most of the risk for professional and hospital services is transferred to an entity that assumes that risk. Professional services capitation is defined as arrangements that involve capitated rates for physician services including primary care, specialty care, and related professional care such as specialty referrals, nonphysician providers, and related laboratory and radiology services. Prevalence Of Risk-Sharing Arrangements By type of Medicaid plan. Medicaid plans actively engaged in risk sharing with the providers with which they contracted in 2001. Forty-two percent of these plans made use of global capitation, and 66 percent used professional services capitation, with at least some of their providers (Exhibit 1). In aggregate, 6 percent of Medicaid plans used at least one of these types of risk-sharing arrangements, and percent used both. Medicaid-dominant plans were more likely than commercial plans were to engage in risk sharing with providers, particularly when the arrangement was professional services capitation. Further, commercial plans reported that they were somewhat less likely to use risk-sharing arrangements for their Medicaid product than for their commercial lines of business. When used, risk-sharing arrangements EXHIBIT 1 Use Of Risk-Sharing Arrangements By Medicaid Plans, 2001 Plan type All (N = 102) Medicaiddominant (N = 65) Commercial (N = 3) Global capitation Percent of plans using any arrangement Mean percent of Medicaid enrollment covered 42% 53 45% 51 38% 5 Professional services capitation Percent of plans using any arrangement Mean percent of Medicaid enrollment covered 66 4 8 6 68 Percent of plans using either type of arrangement Percent of plans using both types of arrangements 6 85 38 62 19 SOURCE: Survey of Medicaid managed care plans in eleven states, 2001. HEALTH AFFAIRS ~ Volume 22, Number 3 161

Health Tracking covered a large percentage of plans Medicaid enrollment (Exhibit 1). For plans with global capitation arrangements in 2001, the mean percentage of Medicaid enrollment covered under these arrangements in a plan was 53 percent, compared with 4 percent for professional services capitation arrangements. The mean percentage of Medicaid enrollment that risk-sharing arrangements covered, regardless of type, was similar for Medicaid-dominant and commercial plans. The survey findings show no clear trends over time as to the share of Medicaid enrollment that risk-sharing arrangements covered. Among the forty-three Medicaid plans that used global capitation, 63 percent said that the share of Medicaid enrollees these arrangements cover was staying about the same; 21 percent said that it was increasing; and 16 percent said that it was decreasing. Similarly, among the sixty-seven plans that used professional services capitation, 2 percent reported that the share of Medicaid enrollees covered by these arrangements was staying about the same; 19 percent reported an increase; and 9 percent reported a decrease. Commercial rather than Medicaid-dominant plans were more likely to report an increase in the share of Medicaid enrollment covered by global capitation arrangements, whereas the reverse was true when professional services capitation arrangements were used. Among the study s eleven states. Risk-sharing arrangements were common in all eleven of the states studied in 2001, although the form and prevalence of these arrangements varied (Exhibit 2). In each of these states at least half of the plans participating in Medicaid engaged in some form of risk contracting. In addition, at least one plan in each state except Arizona used both global and professional services capitation arrangements. Although global capitation was most prevalent in Medicaid plans in California, it was also common for plans in Washington State, Missouri, and Texas. With the exception of Florida, at least half of the plans participating in Medicaid in each state used professional services capitation. This type of risk-sharing arrangement was most prevalent among plans in New Jersey, Washington State, California, and Pennsylvania. In most of the states the majority of plans indicated generally stable enrollment under both capitation arrangements. However, plans in Washington State reported the most dra- EXHIBIT 2 Variation In The Use Of Risk-Sharing Arrangements Among Medicaid Plans In Eleven Study States, 2001 Percent of Medicaid plans using risk-sharing arrangements State Number of plans Global capitation Professional services capitation Either type Both types Arizona California Florida Maryland Michigan Missouri New Jersey New York Pennsylvania Texas Washington 6 10 5 13 8 6 24 10 6 0% 0 40 38 63 1 60 6 % 80 60 62 63 83 6 1 60 83 % 90 1 80 69 5 83 5 1 80 100 All Medicaid plans 102 42 66 6 SOURCE: Survey of Medicaid managed care plans in eleven states, 2001. 0% 60 20 1 25 40 162 May/June 2003

matic change in their global capitation arrangements: Three of four plans that used this type of arrangement reported declining Medicaid enrollment covered by these arrangements. Plans in California reported the most noteworthy changes in professional services capitation arrangements. Among the eight plans there that used these types of arrangements, five reported an increase in the share of Medicaid enrollment covered by providers that contracted under professional services capitation arrangements. Traditional safety-net providers. Plans participating in Medicaid in 2001 often used both global and professional services capitation arrangements with traditional safety-net providers, including federally qualified health centers (FQHCs) and hospital outpatient departments/clinics (Exhibit 3). Medicaid plans tended to use global capitation arrangements with hospital outpatient departments/ clinics and professional services capitation arrangements with FQHCs. Half of plans used global capitation with hospital outpatient departments/clinics, while percent used it with FQHCs. Further, 53 percent of plans using global capitation arrangements used them with either type of provider, while 28 percent used them with both types. In comparison, 55 percent of plans used professional services capitation with FQHCs, while 46 percent used it with hospital outpatient departments/ clinics. Moreover, 58 percent of plans with professional services capitation arrangements used them with either of these provider types, while 40 percent used them with both. Overall, Medicaid-dominant rather than commercial plans were more likely to use some type of risk-sharing arrangement with traditional safety-net providers, which likely reflects, in part, the comparatively higher number of Medicaid-dominant plans that are provider sponsored. Scope Of Risk-Sharing Arrangements Risk-sharing arrangements in 2001 varied depending on the amount of risk that plans transferred to providers, the mechanisms plans used to limit risk, and the functions plans delegated to providers. 1 Amount of risk transferred by plans. EXHIBIT 3 Traditional Safety-Net Providers In Risk-Sharing Arrangements, 2001 Plan type All Medicaiddominant Commercial Global capitation Number of Medicaid plans using global capitation arrangements Percent using capitation with Federally qualified health centers Hospital outpatient departments/clinics Either provider type Both provider types % 51 53 28 36% 56 59 28 % Professional services capitation Number of Medicaid plans using professional services capitation arrangements Percent using capitation with Federally qualified health centers Hospital outpatient departments/clinics Either provider type Both provider types 6 55% 46 58 40 51 62% 6 16 % SOURCE: Survey of Medicaid managed care plans in eleven states, 2001. HEALTH AFFAIRS ~ Volume 22, Number 3 163

Health Tracking On average, plans using global capitation arrangements transferred to providers nearly three-quarters of the state capitation payment they received. The share was slightly higher for Medicaid-dominant plans than for commercial plans (5 percent versus 69 percent). Under professional services capitation arrangements, plans transferred to providers a lower share of the state capitation payment, which averaged 44 percent. Similar to the findings for global capitation, the share of the state capitation payment that plans transferred to providers was also slightly higher for Medicaiddominant plans, 45 percent versus 40 percent for commercial plans. Despite escalating prescription drug costs, the financial risk that plans transferred to providers did not typically include risk for these costs. 18 However, when plans did transfer the risk for prescription drug costs, it was more likely to occur in global rather than professional services capitation arrangements. Forty-one percent of plans with global capitation arrangements shifted some risk for drug costs to providers in their largest such arrangement; 20 percent of plans transferred the risk to providers in their largest professional services capitation arrangement. Commercial plans were more likely than Medicaiddominant plans were to transfer risk for drug costs under global capitation arrangements (46 percent versus 39 percent); Medicaiddominant plans were more likely to do so under professional services capitation arrangements (24 percent versus percent). Plans not transferring risk for prescription drug costs to providers were likely to be at risk themselves with their respective states. This was true for approximately half of plans with global capitation arrangements and nearly three-quarters with professional services capitation arrangements in 2001. Some states, however, carved out prescription drugs from their capitation payments to plans. Mechanisms used by plans to limit provider risk. Plans used a number of mechanisms to limit provider risk in their largest risk-sharing arrangements (Exhibit 4). These mechanisms were more common in global than in professional services capitation arrangements. This pattern reflects the fact that providers assume a greater degree of risk in global capitation arrangements. The mechanisms plans use most often to limit risk include stop-loss and provider reinsurance, but some plans also limit risk by excluding selected high-cost cases and by risk-adjusting the capitation payment. Overall, 95 percent of plans with global capitation arrangements in 2001 used at least one such mechanism in their largest of these arrangements, as did 6 percent of plans with professional services capitation arrangements. Commercial plans were more likely than Medicaid-dominant plans were to use stop-loss mechanisms and to exclude selected high-cost cases, regardless of the specific type of risk-sharing arrangement. In contrast, Medicaid-dominant plans used provider reinsurance and risk-adjusted capitation payments more often than commercial plans did. Functions delegated by plans. Plans delegated to providers a number of functions under their largest risk-sharing arrangements in 2001 (Exhibit 4). The function most frequently delegated in both global and professional services capitation arrangements was credentialing. Although the extent varied, plans also delegated utilization review, provider selection, provider payment, and provider services. However, plans were reluctant to delegate to providers the responsibility for quality assurance and member services. Overall, 88 percent of plans with global capitation arrangements delegated at least one of these functions, as did 6 percent of plans with professional services capitation arrangements. The delegation patterns of Medicaiddominant and commercial plans using global capitation arrangements were similar, but Medicaid-dominant plans were more likely than commercial plans were to delegate quality assurance and member services. In comparison, the delegation patterns of Medicaiddominant and commercial plans differed in professional services capitation arrangements: Commercial plans were considerably more likely to delegate functions in general. 164 May/June 2003

EXHIBIT 4 Scope Of Medicaid Plans Largest Risk-Sharing Arrangements, 2001 Plan type All Medicaiddominant Commercial Global capitation Number of Medicaid plans using any global capitation arrangements Percent limiting risk using Stop-loss Provider reinsurance Exclusion of selected high-cost cases Risk-adjusted capitation Other (such as performance bonds) None of the above 68% 66 41 21 5 62% 68 35 35 3 83% 62 54 23 0 Percent delegating Credentialing Utilization review Provider selection Provider payment Provider servicing Quality assurance Member services None of the above 68 61 49 49 46 24 1 12 0 59 48 48 48 30 22 64 64 Professional services capitation Number of Medicaid plans using any professional services capitation arrangements 6 51 16 Percent limiting risk using Stop-loss Provider reinsurance Exclusion of selected high-cost cases Risk-adjusted capitation Other (such as performance bonds) None of the above 48% 44 8 24 4% 4 2 6 25 53% 36 20 19 Percent delegating Credentialing Utilization review Provider selection Provider payment Provider servicing Quality assurance Member services None of the above 52 28 16 11 20 26 30 10 3 5 5 36 21 19 SOURCE: Survey of Medicaid managed care plans in eleven states, 2001. Conclusions And Policy Implications Although the scope and characteristics of the arrangements vary, plans participating in Medicaid managed care actively engage in risk sharing with providers, many of which are traditional safety-net providers. Based on the evidence presented in this paper, there is nothing to suggest a flight from risk in Medicaid managed care that some describe in commercial and other markets, where the decline in HEALTH AFFAIRS ~ Volume 22, Number 3 165

Health Tracking risk-sharing arrangements appears to be largely associated with the migration of enrollment out of risk-based HMOs into nonrisk products such as preferred provider organizations (PPOs). 19 It may be that the flight from risk phenomenon is too recent to have filtered through to Medicaid. However, it also may mean that Medicaid managed care may be more of a niche product with idiosyncrasies that keep the trends that appear in other lines of business from applying to it. 20 Provider participation in Medicaid has traditionally been more limited than that for commercial and even Medicare lines of business, because of the historically low payment rates in Medicaid and other factors including extensive administrative requirements. 21 States solicitation of mainstream providers to participate in Medicaid managed care has met with mixed results, and traditional safety-net providers continue to provide the foundation for many plans Medicaid networks. 22 This might reflect the fact that traditional providers are heavily dependent on Medicaid revenue to complement supplemental funding such as disproportionate-share hospital (DSH) payments and special funding for community clinics. It also might reflect the special preferences that Medicaid plans often extend to traditional providers such as through the autoassignment process. 23 The potential commonality of interest between Medicaid-dominant plans and traditional safety-net providers, which is unique to Medicaid, also might foster stronger collaborative relationships leading to more sustainable risk-sharing arrangements. 24 Consequently, the provider contracting pressures evident in other lines of business might not apply to Medicaid managed care in the same way. Findings discussed in this paper, however, are based on research conducted during a period of relative strength and expansion of public programs. States are now confronting serious fiscal crises. Medicaid is a major component of state spending that faces inflationary pressures because of expanded eligibility and rising costs such as for prescription drugs. Provider risk sharing provides an important cost containment mechanism, but it could be severely jeopardized if states respond to budget pressures by limiting payments to plans, leading plans and providers to decide that risk sharing is no longer tenable. Plans decisions about participating in Medicaid and SCHIP are highly dependent on their financial experience with these programs. 25 The recent experience of the Medicare+Choice (M+C) program illustrates the problems that can arise when payment rates to plans are constrained. 26 If states respond to cost pressures by severely limiting the growth of payments to plans, plans may be unable to shift these limits to providers. As a result, the current goodwill and collaboration many states have with plans and providers and that Medicaid-dominant plans and traditional safety-net providers have developed may erode, taking plan and provider participation levels down with them. The likely outcome is a severely compromised program infrastructure, resulting in reduced access for Medicaid and SCHIP beneficiaries. This paper is based on research funded by the Henry J. Kaiser Family Foundation. An earlier version of the paper was presented as a poster session at the AcademyHealth annual meeting in Washington, D.C., June 2002. The authors thank Bob Hurley for his review of an earlier draft. All views expressed in this paper are those of the authors and do not necessarily represent those of the Kaiser Family Foundation or Mathematica Policy Research Inc. NOTES 1. M. Gold et al., Financial Risk Sharing with Providers in Health Maintenance Organizations, Inquiry (Spring 2002): 34 44; InterStudy, Competitive Edge 11.1, Part II: HMO Industry Report (St. Paul, Minn.: InterStudy, 2001); T. Lake et al., Health Plans Selection and Payment of Health Care Providers, 1999 (Washington: Medicare Payment Advisory Commission, May 2000); J. Robinson and L. Casalino, The Growth of Medical Groups Paid through Capitation in California, New England Journal of Medicine 3, no. 25 (1995): 1684 168; and S. Shortell, R. Gillies, and D. Anderson, The New World of Managed Care: Creating Organized Delivery Systems, Health Affairs (Winter 1994): 46 64. 166 May/June 2003

2. R. Hurley et al., A Longitudinal Perspective on Health Plan-Provider Risk Contracting, Health Affairs (July/Aug 2002): 4 153; D. Draper et al., The Changing Face of Managed Care, Health Affairs (Jan/Feb 2002): 11 1; and J. Robinson and L. Casalino, Reevaluation of Capitation Contracting in New York and California, 1 May 2001, www.healthaffairs.org/webexclusives/ Robinson_Web_Excl_05101.htm (25 February 2003). 3. InterStudy reports the following capitation use rates by HMOs using this type of reimbursement method for any portion of reimbursement: primary care physicians, 63 percent; specialists, 42 percent; and hospitals, 25 percent. The data are as of 1 July 2000. InterStudy, Competitive Edge 11.1. 4. Gold et al., Financial Risk Sharing ; and Robinson and Casalino, The Growth of Medical Groups. 5. Gold et al., Financial Risk Sharing ; and Lake et al., Health Plans Selection and Payment. 6. Lake et al., Health Plans Selection and Payment.. Gold et al., Financial Risk Sharing. 8. T. Lake, M. Gold, and R. Hurley, HMO Provider Networks in Medicare+Choice: Comparing Medicare and Commercial Lines of Business, Managed Care Quarterly 9, no. 4 (2001): 16 22. 9. Centers for Medicare and Medicaid Services, 2001 Medicaid Managed Care Enrollment Report, www.cms.gov/medicaid/managedcare/ mmcss01.asp (25 February 2003). 10. R. Hurley and D. Draper, Medicaid Confronts a Changing Managed Care Marketplace, Health Care Financing Review (Fall 2002): 11 25. 11. V. Smith et al., Medicaid Spending Growth: Results from a 2002 Survey (Washington: Kaiser Commission on Medicaid and the Uninsured, September 2002). 12. For additional details about the study methodology, see M. Gold et al., Medicaid and SCHIP Managed Care: Plan Participation, Provider Networks, and Quality Improvement (Washington: Mathematica Policy Research, March 2002). 13. The eleven states selected were among thirteen states with the largest Medicaid managed care enrollment nationally. Oregon and Tennessee were excluded because of the extensive research that has been conducted on Medicaid managed care in these states.. D. Draper and M. Gold, Customizing Medicaid Managed Care California Style, Health Affairs (Sep/Oct 2000): 2 238. A total of 109 health plans were interviewed, but seven of these plans participated only in SCHIP. Because questions about risk sharing were only asked of the Medicaid-participating plans, the seven SCHIPonly plans are not included here. 15. Lake et al., Health Plans Selection and Payment. 16. Questions about the scope of risk-sharing arrangements were asked of plans for their largest global capitation and largest professional services capitation arrangements. 1. Considering the national policy focus on rising prescription drug costs, we wanted to determine whether plans shift risk for these costs to providers. Otherwise, we did not ask about specific medical services for which plans shift risk to providers. 18. Hurley et al., A Longitudinal Perspective ; Draper et al., The Changing Face of Managed Care ; and California Medical Association, The Coming Medical Group Failure Epidemic (San Francisco: CMA, 1999). 19. Hurley and Draper, Medicaid Confronts a Changing Managed Care Marketplace. 20. The literature on provider participation in Medicaid generally predates the transition to managed care. See, for example, R. Deprez, M. Agger, and L. McQuinn, Access to Physicians, Obstetric Care Use, and Adequacy of Prenatal Care for Medicaid Patients in Maine: 1985 1989, Obstetrics and Gynecology 88, no. 3 (1996): 4 4; J. Perloff, P. Kletke, and J. Fossett, Which Physicians Limit Their Medicaid Participation and Why, Health Services Research 30, no. 1 (1995): 26; E. Adams, Effect of Increased Medicaid Fees on Physician Participation and Enrollee Service Utilization in Tennessee, 1985 1988, Inquiry, no. 2 (1994): 13 18; and F. Sloan, J. Mitchell, and J. Cromwell, Physician Participation in State Medicaid Programs, Journal of Human Resources 13, Supplement (198): 211 245. 21. M. Gold et al., Participation of Plans and Providers in Medicaid and SCHIP Managed Care, Health Affairs (Jan/Feb 2003): 230 240. 22. Hurley and Draper, Medicaid Confronts a Changing Managed Care Marketplace. 23. See Gold et al., Medicaid and SCHIP Managed Care, for further details on special preferences and protections that Medicaid plans extend to traditional providers. 24. B. Gray and C. Rowe, Safety-Net Health Plans: A Status Report, Health Affairs (Jan/Feb 2000): 185 193. 25. Gold et al., Participation of Plans and Providers. 26. M. Gold, Medicare+Choice: An Interim Report Card, Health Affairs (July/Aug 2001): 120 138. HEALTH AFFAIRS ~ Volume 22, Number 3 16