United States General Accounting Office GAO Report to Congressional Requesters January 2003 MEDICAID AND SCHIP States Use Varying Approaches to Monitor Children’s Access to Care GAO-03-222 January 2003 MEDICAID AND SCHIP States Use Varying Approaches to Highlights of GAO-03-222, a report to Monitor Children's Access to Care Representatives Sherrod Brown; John Conyers, Jr.; Diana DeGette; John D. Dingell; Gene Green; William J. Jefferson; Sander M. Levin; and Ted Strickland Over 25 million children have Overall, states imposed more access-related requirements on participating health insurance coverage through providers and more actively monitored children’s use of services in their Medicaid or the State Children’s Medicaid managed care programs than in their Medicaid FFS or SCHIP Health Insurance Program programs. (SCHIP). Coverage alone, however, does not guarantee that services will be available or that children Medicaid managed care: State requirements for managed care plans ranged will receive needed care. GAO was from very broad provisions that health plans must have “adequate” physician asked to evaluate states’ efforts to networks for serving their enrolled members to very specific standards, such facilitate and monitor access to as the number and geographic proximity of physicians and maximum time primary and preventive services for frames within which a new beneficiary receives a first appointment. States children in these jointly funded less often verified data that plans submitted to show compliance with these federal-state programs. The study requirements or independently monitored physicians’ availability. In one surveyed 16 states, covering over instance of verification, a state found that a third of a health plan’s physician 65 percent of the Medicaid and network was not accepting new Medicaid patients, thus limiting access for SCHIP population. GAO analyzed new beneficiaries. The value of plan-submitted data that states used to requirements relevant to managed monitor children’s use of services was often compromised by continuing care and fee-for-service (FFS) delivery systems, including the problems with their completeness and reliability. Furthermore, information number and location of physicians derived from beneficiary satisfaction surveys was not necessarily and their availability to see representative of all Medicaid managed care beneficiaries. beneficiaries, monitoring of health plan or physician compliance with Medicaid FFS: Most states did not set goals for or analyze the availability of these requirements, and collection participating primary care physicians even though a majority of Medicaid- and analysis of beneficiary service eligible children in half of the states reviewed are still served in FFS utilization data. programs. In most FFS programs, beneficiaries may seek care from any providers participating in the Medicaid program and may change providers at any time if they are dissatisfied. However, when FFS payment rates are lower than those paid by other purchasers—which was the case in most states reviewed—providers can be discouraged from participating in Medicaid and thus restrict beneficiaries’ access. States did little to monitor the use of services by Medicaid-eligible children in FFS programs despite having a ready source of data in their claims payment systems. SCHIP: Nine of the 16 states used the same providers, administrative systems, and monitoring approaches for their SCHIP programs as they did for Medicaid. The remaining 7 states, whose SCHIP programs were distinct from Medicaid and used managed care almost exclusively, set few requirements for or monitored providers’ availability to SCHIP-eligible children. States with distinct SCHIP programs also reported fewer efforts to monitor children’s use of services than in their Medicaid programs. Comments on our report from the Department of Health and Human Services highlighted new federal requirements for state oversight of www.gao.gov/cgi-bin/getrpt?GAO-03-222. managed care, and design differences between Medicaid and SCHIP that can To view the full report, including the scope affect monitoring approaches. States we reviewed provided clarifying or and methodology, click on the link above. technical comments regarding their oversight of access, which we For more information, contact Kathryn G. Allen (202) 512-7118. incorporated in the report as appropriate. Contents Letter 1 Results in Brief 4 Background 6 In Medicaid Managed Care, States Focused More on Setting Plan Network Requirements than on Monitoring Plans or Analyzing Service Utilization 14 For Medicaid FFS, State Requirements for Providers and Monitoring of Service Utilization Were More Limited 33 Distinct SCHIP Programs Had Fewer Network Requirements and Less Monitoring of Service Utilization 40 Agency and State Comments and Our Evaluation 44 Appendix I Medicaid HEDIS Measures Related to Service Utilization 50 Appendix II Managed Care Plan Withdrawals from Medicaid in Four States 53 Massachusetts 53 Ohio 54 Tennessee 55 Texas 56 Appendix III Analysis of Medicaid FFS Payment Rates in Selected States 58 Methodology for Comparison of FFS Payment Rates 59 Appendix IV Comments from the Department of Health and Human Services 61 Appendix V GAO Contact and Staff Acknowledgments 65 Related GAO Products 66 Page i GAO-03-222 Medicaid and SCHIP Access Tables Table 1: SCHIP Design Choices for 16 States, as of March 2002 7 Table 2: Share of Children Enrolled in Medicaid and Separate SCHIP Programs, by Service Delivery Method, for 16 States 10 Table 3: Examples of Specific State Standards for Plan Networks and Appointment Waiting Times 17 Table 4: Examples of Medicaid HEDIS Measures Related to Service Utilization for Children 25 Table 5: Estimated Percentage of Medicaid Children Excluded from HEDIS 27 Table 6: Examples of Beneficiary Satisfaction Questions for Children Covered by CAHPS 28 Table 7: Estimated Percentage of Medicaid Children Excluded from CAHPS 29 Table 8: Length of Medicaid Enrollment Required for Selected HEDIS Measures for Children’s and Adolescents’ Use of Services 51 Table 9: Examples of Plan Withdrawal Transition Activities Conducted by Four State Medicaid Programs 57 Table 10: Medicaid FFS Payment Rates, Expressed as a Percentage of Medicare Payments, in 13 States with Traditional FFS or Primary Care Case Manager Delivery Systems That Serve Children 59 Table 11: CPT 4 Codes Used in Comparing Medicaid and Medicare Fees 60 Figures Figure 1: Selected Medicaid Managed Care Plan Network Requirements and Standards in 14 States 16 Figure 2: Variation in 14 States’ Monitoring of Medicaid Managed Care Plans’ Provider Information 19 Figure 3: Selected Requirements for Medicaid PCCM Providers in Seven States 37 Figure 4: Comparison of Seven States’ Requirements and Standards for Providers in Medicaid and SCHIP Managed Care 42 Page ii GAO-03-222 Medicaid and SCHIP Access Abbreviations ACG ambulatory care group AHRQ Agency for Healthcare Research and Quality CAHPS Consumer Assessment of Health Plans CMS Centers for Medicare & Medicaid Services CPT 4 Current Procedural Terminology, 4th edition EPSDT Early and Periodic Screening, Diagnostic and Treatment FFS fee-for-service HEDIS Health Plan Employer Data and Information Set HHS Department of Health and Human Services HRSA Health Resources and Services Administration NCQA National Committee on Quality Assurance PCCM primary care case manager PCP primary care provider SCHIP State Children’s Health Insurance Program Page iii GAO-03-222 Medicaid and SCHIP Access United States General Accounting Office Washington, DC 20548 January 14, 2003 The Honorable Sherrod Brown The Honorable John Conyers, Jr. The Honorable Diana DeGette The Honorable John D. Dingell The Honorable Gene Green The Honorable William J. Jefferson The Honorable Sander M. Levin The Honorable Ted Strickland House of Representatives Over 25 million children have health care coverage through Medicaid or the State Children’s Health Insurance Program (SCHIP), joint federal-state programs that finance health insurance for certain low-income adults and children. Medicaid and SCHIP provide the financial means for low-income children to receive primary, preventive, and specialty care, which are important to ensuring a healthy child and adolescent population. Having a regular provider, or usual source of care, also can help reduce the use of services from high-cost sources such as emergency rooms and inpatient hospital care.1 While health insurance coverage can provide the financial means to obtain care, it does not by itself guarantee that health services will be available and accessible or that beneficiaries will receive needed care. Access to primary care services is significantly affected by local factors that vary across and within states, such as physician supply, location, and willingness to participate in a state’s Medicaid and SCHIP programs. While federal law establishes general requirements to ensure that Medicaid and SCHIP beneficiaries have access to covered health services, the extent to which children actually receive these health care services is influenced by how states implement their programs and monitor access at the state and local levels. The type of service delivery and financing system that states use in their Medicaid and SCHIP programs potentially affects beneficiaries’ ability to 1 A provider may be a physician, a group of physicians practicing together, or an outpatient clinic with physician services. Page 1 GAO-03-222 Medicaid and SCHIP Access locate and obtain services. Managed care, which often entails states making capitation payments to managed care plans to provide or arrange for all services for enrolled beneficiaries, encourages participating plans to offer and coordinate primary and specialty care for beneficiaries. Managed care also may promote efficiency by attempting to ensure that only necessary services are provided in the most appropriate setting. Appropriate safeguards are important, however, as capitation payments can also create an incentive to underserve or even deny beneficiaries access to needed care since plans and, in some cases, providers can profit from not delivering services for which they have already received payment. In contrast, beneficiaries in fee-for-service (FFS) systems, including those receiving care in a primary care case manager (PCCM) system,2 may be at risk for the overprovision of services as providers seek to increase revenue. However, if FFS payment levels are too low, physicians may underserve their patients or be unwilling to participate at all. Our prior work has shown that access to care in Medicaid has been problematic for certain services—such as health screening for children, oral health, and mental health—and for particular populations, such as children with special needs.3 Recent reports that some physicians are unwilling to take more Medicaid patients and that some managed care plans are exiting from the Medicaid program have raised additional concerns about adequate access for eligible children. Now that SCHIP is beginning its sixth year of implementation, a related concern is the experiences of children in accessing care under SCHIP, where states have greater flexibility to decide whom to cover, what services to provide, and how to pay for services, including required beneficiary cost sharing. Accordingly, you asked us to evaluate states’ efforts to routinely monitor access to primary and preventive care services in (1) Medicaid managed care, including actions selected states took when participating health plans withdrew from the program, (2) Medicaid FFS-based delivery systems, including PCCM systems, and (3) SCHIP. 2 FFS systems include traditional FFS, in which a provider bills the program for services provided to an eligible beneficiary, and PCCM systems, in which a physician, physician group practice, or similar entity contracts with the state to locate, coordinate, and monitor primary health services for Medicaid beneficiaries for a nominal monthly, per capita case management fee (usually around $3). 3 See related GAO products listed at the end of this report. Page 2 GAO-03-222 Medicaid and SCHIP Access To examine these issues, we analyzed 16 states’ approaches to monitoring access to primary and preventive health care services in their Medicaid and SCHIP programs. These states were Arkansas, California, Colorado, Florida, Illinois, Louisiana, Maryland, Massachusetts, Michigan, Nevada, New York, Ohio, Pennsylvania, Tennessee, Texas, and Washington. We selected these states to obtain wide representation of geographic regions, managed care and FFS systems, and SCHIP program designs.4 Over 65 percent of all Medicaid and SCHIP beneficiaries resided in these 16 states. To evaluate state approaches to monitoring access to care, we focused our analysis of states’ managed care and FFS delivery systems in three key areas: • specific requirements for participating managed care plans and physicians to help ensure sufficient physician capacity and accessibility for eligible beneficiaries; • actions to independently verify or otherwise monitor provider participation; and • routine data collection and analysis of information on beneficiaries’ actual service utilization, including patient satisfaction surveys. For states’ Medicaid and SCHIP managed care programs, these service utilization data included encounter data, which are individual-level data on service use that plans are required to collect and report to the state; the Health Plan Employer Data and Information Set (HEDIS), developed by the National Committee for Quality Assurance (NCQA) to help purchasers and consumers compare the performance of health plans in providing selected services; and the Consumer Assessment of Health Plans (CAHPS), which is a standardized patient satisfaction survey developed by the federal Agency for Healthcare Research and Quality (AHRQ). We conducted site visits in four states where managed care plan withdrawals had been reported—Massachusetts, Ohio, Tennessee, and Texas—and analyzed information from primary care providers (PCP);5 representatives of advocacy groups; state insurance departments; and managed care plans participating in Medicaid, SCHIP, or both. At the federal level, we interviewed officials at the Centers for Medicare & Medicaid Services 4 States can take three approaches in designing their SCHIP programs: (1) expand Medicaid, (2) construct separate child health programs distinct from Medicaid, or (3) use a combination of both approaches. 5 For purposes of this report, PCPs are usually physicians trained in internal medicine, pediatrics, family medicine, or obstetrics and gynecology who participate in PCCM or managed care programs in Medicaid or SCHIP. Page 3 GAO-03-222 Medicaid and SCHIP Access (CMS), which oversees states’ Medicaid and SCHIP programs, and the Health Resources and Services Administration (HRSA), which has responsibility for analyzing issues related to access to care, as well as joint responsibility for oversight of SCHIP. We reviewed relevant documents, including federal laws, federal regulations, state contracts with managed care organizations, and various federal and state reports related to access. We conducted our work from June 2001 through December 2002 in accordance with generally accepted government auditing standards. Each of the states we reviewed with Medicaid managed care programs set Results in Brief requirements for participating plans’ provider networks, which include the physicians and specialists who have agreed to deliver or arrange for health care services to beneficiaries enrolled in a health plan. These state requirements ranged from broad provisions that health plans must have “adequate” networks for serving their enrolled members, to very specific standards that set, for example, a maximum number of beneficiaries per primary care physician or maximum time frames within which a provider must see a new beneficiary for a first appointment. The states less frequently verified data that plans submitted to them or independently collected or analyzed data to ascertain compliance with the requirements. States that did routinely monitor plans’ compliance with network requirements often identified potential access problems and took steps to address them. For example, a state review of the physicians listed in a plan’s network found that many physicians were not accepting new Medicaid patients, resulting in too few physicians accessible to such patients. Beyond setting requirements for or monitoring plans’ network size and availability, states attempted to assess the extent to which beneficiaries were actually receiving services through three key routine data sources: encounter data that states require plans to submit on individual-level service use, assessments of managed care plans’ performance on specified measures, and periodic beneficiary satisfaction surveys. However, the value of these data was compromised by continuing problems in most states with encounter data’s completeness and reliability; additionally, standardized data on plan performance and beneficiary satisfaction surveys were not representative of all Medicaid managed care beneficiaries. Potential issues of access to care associated with managed care plans withdrawing from Medicaid in four states we visited affected significantly different shares of eligible beneficiaries, ranging from about 1 percent of Medicaid beneficiaries in Texas to almost 50 percent in Tennessee. While these four states had taken various steps to help minimize disruption in access to care for beneficiaries affected by plan withdrawals, it was not clear to what extent these efforts had been Page 4 GAO-03-222 Medicaid and SCHIP Access successful in helping beneficiaries transition smoothly to new health plans and avoid access-to-care problems. States did considerably less in their Medicaid FFS programs—which still serve the majority of children in half of the states we reviewed—to establish requirements for or monitor provider availability or to assess beneficiaries’ utilization of services than in their managed care programs. For traditional FFS programs, beneficiaries may seek care from any providers participating in the Medicaid program and may change providers at any time if they are dissatisfied. However, Medicaid beneficiaries’ ability to easily change providers depends on the number, type, and location of providers willing to take new Medicaid patients, which in turn is strongly influenced by Medicaid payment rates and associated administrative processes. We found that FFS payment rates in most states we reviewed were significantly lower than those paid by other purchasers for comparable services, which can discourage providers from participating in the program and thus restrict beneficiaries’ access to a broad supply of providers. Officials in several of the states we contacted with Medicaid FFS programs said that anecdotal information and complaint data suggested that low payment rates, slow payment, and other administrative issues deterred physicians in primary care or in some specialties from participating in the program. Most of the seven states we reviewed with PCCM programs set certain requirements for participating physicians, such as limiting the number of beneficiaries that a PCCM could enroll in an effort to ensure that physicians had the capacity to serve each beneficiary. However, these states did little to monitor the extent to which beneficiaries were successful in obtaining appointments as needed. In regard to routine data collection and analyses, all but one of these seven states analyzed their FFS claims data and provided PCCMs with comparative data on service utilization patterns for their own practices and for other PCCMs. However, these comparative data often focused on higher-cost services, such as inpatient hospitalization or emergency room use. The majority of the states we reviewed—9 of the 16—designed their SCHIP programs to be an expansion of their Medicaid programs or modeled them after Medicaid, with the same providers and administrative systems. Therefore, in these states, the requirements for, and monitoring of, SCHIP provider participation and beneficiary service utilization mirrored that of their Medicaid programs. In contrast to these states, 7 states chose to serve all or most of their SCHIP beneficiaries through programs that were distinct from Medicaid. These states did significantly less in their distinct SCHIP programs in terms of setting requirements for, Page 5 GAO-03-222 Medicaid and SCHIP Access or monitoring, participating providers or beneficiary service use than they did for their Medicaid programs. We received comments on a draft of this report from the Department of Health and Human Services (HHS), as well as from 13 of the 16 states that were included in our review. In response to our findings, HHS highlighted new federal requirements for state oversight of Medicaid managed care that are to be fully implemented by August 2003. HHS also pointed out that design differences between Medicaid and SCHIP may affect states’ approaches to monitoring access to care. State officials provided clarifying and technical comments regarding their oversight of access to care, which we incorporated as appropriate throughout this report. States’ Medicaid and SCHIP programs are governed by various federal Background requirements regarding eligibility, covered services, and access to care. Under these requirements, states generally have some discretion in determining the amount, duration, and scope of services their programs will provide, and the delivery and financing systems through which beneficiaries will receive care—that is, FFS, managed care, or both. Federal requirements relating to Medicaid beneficiaries’ access to care are established in statute; for managed care service delivery systems, detailed federal regulations regarding access were recently issued.6 SCHIP requirements are also set out in statute but are less specific than those for Medicaid and do not include detailed managed care requirements or regulations comparable to those for Medicaid. Populations Covered and Since 1965, Medicaid has financed health care coverage for certain Program Characteristics categories of low-income individuals—including over 22 million children in 2000. Federal law requires states to extend Medicaid eligibility to children aged 5 and under if their family incomes are at or below 133 percent of the federal poverty level and to children aged 6 through 18 in families with incomes at or below the federal poverty level. At their discretion, most states have set income eligibility thresholds for families with children that expand their Medicaid programs beyond the minimum federal statutory levels. 6 67 Fed. Reg. 40989 (2002). Page 6 GAO-03-222 Medicaid and SCHIP Access In 1997, the Congress established SCHIP, which provides health care coverage to low-income, uninsured children living in families whose incomes exceed the eligibility limits for Medicaid. SCHIP covered over 4.6 million children in fiscal year 2001, generally targeting children in families with incomes up to 200 percent of the federal poverty level.7 Compared with Medicaid, which has specific minimum federal eligibility and benefit requirements, the SCHIP legislation provides states more flexibility in how they choose to structure their programs. States have three options in designing SCHIP: They may expand their Medicaid programs, develop a separate child health program that functions independently of Medicaid, or create a combination of the two approaches. (See table 1 for the program designs of the 16 states in our sample.) While Medicaid expansion programs under SCHIP must use Medicaid’s provider networks and delivery systems, SCHIP separate child health programs may depart from Medicaid requirements particularly with regard to covered benefits and the plans, providers, and delivery systems available to beneficiaries.8 Table 1: SCHIP Design Choices for 16 States, as of March 2002 Design State Medicaid expansion Arkansas,a Louisiana, Ohio, and Tennessee Separate program Colorado, Nevada, Pennsylvania, and Washington b Combination California, Florida, Illinois, Maryland, b Massachusetts, Michigan, New York, and Texas Source: CMS. a In February 2001, Arkansas received approval from CMS to implement a separate SCHIP program; however, this program had not been implemented as of February 2002. b The state’s separate SCHIP portion of its combination program provides coverage either through (1) a premium assistance program for families with access to private insurance coverage or (2) Medicaid providers and services. Premium assistance programs were not included in our study. Medicaid and SCHIP differ in terms of the share of their program expenditures that come from federal funds. No overall federal budget limit exists for the Medicaid program; it is an open-ended entitlement whereby 7 The SCHIP statute allows a state to expand eligibility to 200 percent of the poverty level or up to 50 percentage points above its Medicaid eligibility standard as of March 31, 1997. As of January 2002, states’ upper income eligibility thresholds for SCHIP ranged from 133 to 350 percent of the federal poverty level. 8 Throughout this report, SCHIP beneficiaries enrolled in Medicaid expansion programs are included in the discussion of Medicaid. Page 7 GAO-03-222 Medicaid and SCHIP Access state expenditures for services that are provided under a CMS-approved state Medicaid plan are matched by the federal government using a formula that results in federal shares that currently range from 50 to 76 percent of expenditures, depending on a state’s per capita income in relationship to the national average. The federal share of Medicaid expenditures is about 57 percent. In contrast to Medicaid, federal funding for SCHIP is limited. The Congress appropriated $40 billion over 10 years (from fiscal years 1998 to 2007), with a specified amount allocated annually to each of the 50 states, the District of Columbia, Puerto Rico, and 4 U.S. territories. State SCHIP expenditures are matched by federal payments up to the state’s annual appropriated allotment.9 The SCHIP statute provides for an “enhanced” federal matching rate, with each state’s SCHIP rate exceeding its Medicaid rate. The federal share of each state’s SCHIP expenditures ranges from 65 to 83 percent; the federal share of total SCHIP expenditures is about 72 percent. Delivery Systems States provide Medicaid and SCHIP services through two distinct service delivery and financing systems—managed care and FFS, with the latter including PCCM.10 Under a capitated managed care model, states contract with a managed care organization and prospectively pay the plans a fixed monthly fee per patient to provide or arrange for most health services. Plans, in turn, pay providers either retrospectively for each service delivered on a FFS basis or through prospective capitation payment arrangements. In contrast, in a traditional FFS delivery system, the Medicaid program reimburses providers directly and on a retrospective basis for each service delivered. The PCCM model is similar to a traditional FFS arrangement except that PCCMs are paid a monthly, per capita case management fee, usually around $3, to coordinate care for beneficiaries, in addition to FFS reimbursement for any health care services they provide. PCCMs, which are selected by beneficiaries upon enrollment, are responsible for treating and coordinating the care for those beneficiaries. Coordination may involve referrals to specialists and 9 Annual allotments are made to states for use over a 3-year period. For SCHIP annual allotments that remain unspent after 3 years, the Secretary of HHS is required to determine an appropriate procedure for redistributing any unused SCHIP funds to states that have exhausted their allotments. 10 We included PCCMs as FFS-based arrangements because participating providers are predominately paid on a FFS basis. Thus, throughout this report, the term managed care only refers to capitated managed care arrangements. Page 8 GAO-03-222 Medicaid and SCHIP Access other providers. In some cases, receipt of specialty and other services may require PCCM approval. The 16 states we reviewed often structured their Medicaid and SCHIP service delivery systems differently. As shown in table 2, the exclusive use of managed care was less prevalent in Medicaid than in separate SCHIP programs (3 and 6 states, respectively), with 3 states—Maryland, Michigan, and Tennessee—using managed care for virtually all children in both Medicaid and SCHIP. The states were more likely to use a combination of managed care and FFS approaches for their Medicaid programs than for SCHIP (11 and 5 states, respectively). Despite the recent growth in states’ use of managed care, FFS is still a major component of many states’ programs, especially for Medicaid. Page 9 GAO-03-222 Medicaid and SCHIP Access Table 2: Share of Children Enrolled in Medicaid and Separate SCHIP Programs, by Service Delivery Method, for 16 States Medicaid Separate SCHIP programa FFS-based FFS-based Traditional Traditional State FFS PCCM Managed care FFS PCCM Managed care b c c Arkansas 100% -- -- b California 29% 71% -- -- 100% Colorado 28% 18% 54% -- -- 100% d b e Florida 53% 47% 2% 98% Illinois 87% -- 13% 99% -- 1% c c Louisiana 88% 12% -- -- d f Maryland -- 100% -- -- d Massachusetts 63% 37% -- 65% 35% d Michigan -- 100% -- -- 100% Nevada 41% -- 59% 13% -- 87% b New York 61% -- 39% -- 100% c c Ohio 67% -- 33% -- d Pennsylvania 24% 76% -- -- 100% c Tennessee -- -- 100% -- -- Texas 42% 21% 37% -- -- 100% b b Washington 34% 66% 57% 43% Number of states using system 9 7 14 3 2 11 Source: State data, as of December 2001, except for New York data, which are as of September 2001. a Includes the separate child health programs in states with combination or separate SCHIP programs. b Although this delivery system exists in the state, it includes less than 1 percent of children enrolled in Medicaid or SCHIP and thus was not included in our study. c Not applicable. State’s SCHIP program is a Medicaid expansion; thus, delivery systems are the same as those in the state’s Medicaid program. d Delivery system exists for children with special needs, which is outside the scope of this study. Additionally, state enrolls children in FFS until they transition to a managed care or a PCCM delivery system. For example, families with eligible children in Florida are allowed 90 days in which to select a PCP in managed care or a PCCM; during these 90 days, they are enrolled in Medicaid FFS. e In Florida, delivery systems under SCHIP vary by age. Families with children under age 5 can select between managed care and a PCCM, while older children are limited to managed care service delivery. f Not applicable; Maryland’s separate SCHIP portion of its combination program was not operational when our study began and thus was not included in the study. Page 10 GAO-03-222 Medicaid and SCHIP Access Federal Requirements for A state is required by federal statute to ensure that its payment and Access to Care delivery systems will afford beneficiaries’ access to services similar to that of its general population;11 further, Medical assistance must be provided with reasonable promptness.12 While Medicaid traditional FFS delivery systems have no additional access requirements, managed care and PCCM delivery systems do. States are required to ensure that beneficiaries’ access to care in managed care and PCCM is equal to that available to beneficiaries in traditional FFS. On June 14, 2002, CMS published final rules to implement new provisions the Balanced Budget Act of 1997 set out for states’ Medicaid managed care programs. These new rules address the requirements, prohibitions, and procedures for the provision of different types of Medicaid managed care and PCCM delivery systems. Under these rules, which became effective August 13, 2002, states have until August 13, 2003, to bring all aspects of their state managed care programs into compliance with the new requirements.13 States that wish to use managed care and PCCMs to deliver Medicaid services must have CMS approval to do so. CMS approval is in part intended to ensure that adequate protections are in place to safeguard the interests of beneficiaries enrolled in managed care who may find their freedom to seek the care of any participating provider at any time more restricted than in FFS. In managed care, states may “lock in” beneficiaries to one managed care plan and its network of providers for up to 1 year in order to provide the plan sufficient time and opportunity to manage the care of its enrollees most efficiently and appropriately. States request CMS approval for their managed care programs through one of two methods: (1) as a waiver from certain statutory requirements or (2) as an amendment to the state’s Medicaid plan.14 Fifteen of the 16 states we reviewed received CMS approval to provide managed care through two types of waivers of statutory provisions, program and demonstration waivers, while one state—Nevada—received approval through a state plan amendment. 11 See 42 U.S.C. § 1396a(a)(30); 42 C.F.R § 438.2. 12 See 42 U.S.C. §1396a(a)(8). 13 The new Medicaid managed care rules have more detailed requirements for states than in the past, such as requiring assurances from participating plans concerning the availability of services, adequate capacity and services, coordination and continuity of care, and coverage and authorization of services. 14 Implementing managed care service delivery by amending a state’s Medicaid plan has been an option for states since passage of the Balanced Budget Act of 1997. Page 11 GAO-03-222 Medicaid and SCHIP Access Of the states we reviewed, 12 had approved “freedom-of-choice” program waivers, under section 1915(b) of the Social Security Act, which permitted them to direct beneficiaries to enroll in a managed care system.15 In reviewing and approving program waivers, CMS requires states that wish to limit beneficiaries’ enrollment to managed care to offer a choice of at least two managed care plans or allow beneficiaries to choose between one managed care plan and a PCCM system. CMS also requires states to ensure that (1) managed care plans’ physician networks under the waiver include approximately the same number or more physicians than were available before the waiver’s implementation and (2) services under program waivers are provided within reasonable time frames and are furnished within reasonable distances for the beneficiaries to travel. As a condition of waiver approval, during the period of our review CMS asked states to specify whether they had established access-related requirements for participating plans in areas such as provider capacity, or maximum times frames for beneficiaries to schedule appointments, travel to physicians’ offices, or wait in physicians’ offices to be seen. CMS did not require states to establish specific requirements in these areas, but if they did, they were asked to describe in their waiver applications how they planned to monitor compliance with any established requirements. Initial approval of a program waiver is for a 2-year period, at which time the waiver can be reviewed and approved for renewal by CMS. Waiver renewals can result in changes in specific requirements for states. Six of the states we reviewed—Arkansas, California, Maryland, Massachusetts, New York, and Tennessee—had approved comprehensive research and demonstration waivers, authorized by section 1115 of the Social Security Act, to test concepts likely to further program objectives.16 A demonstration waiver provides a state with greater flexibility to design its Medicaid programs in areas such as eligibility standards, covered benefits, and reimbursement rules. In reviewing and approving demonstration waivers, CMS often establishes terms and conditions with which states must comply that are more prescriptive than requirements for 15 The freedom of choice waiver is established by section 1915(b) of the Social Security Act and is set forth at 42 U.S.C. §1396n(b). The 12 states that we reviewed with program waivers were Arkansas, California, Colorado, Florida, Louisiana, Michigan, New York, Ohio, Pennsylvania, Tennessee, Texas, and Washington. 16 The demonstration waiver is set forth at 42 U.S.C. § 1315(a). In addition to comprehensive waivers, states can also use section 1115 waivers for specific populations or services, such as pharmacy or extending coverage to parents. Four of the 15 states—California, Colorado, Florida, and Illinois—have noncomprehensive section 1115 waivers. Page 12 GAO-03-222 Medicaid and SCHIP Access program waivers. For example, in approving demonstration waivers, CMS has required states to (1) specify ratios that set the maximum number of enrolled beneficiaries per participating PCP, (2) establish the maximum time or distance for beneficiaries to travel to a physician’s office, and (3) limit beneficiaries’ waiting times when scheduling appointments for urgent, routine, or specialty care. Initial approval of a demonstration waiver is for a 5-year period, at which time the waiver can be reviewed and approved for renewal by CMS. In contrast to Medicaid, in their SCHIP programs states may require mandatory beneficiary enrollment in managed care without offering a choice among health plans. Federal SCHIP access-related requirements are also less extensive than those for Medicaid. The SCHIP statute requires that states have methods in place to ensure access to covered services, including emergency services, but does not specify precise requirements.17 States must describe their methods to ensure access to covered services, including any monitoring procedures, in their SCHIP state plans. In addition, the SCHIP statute required each state to submit to the Secretary of HHS a one-time program evaluation in March 2000.18 States must also submit to the Secretary annual reports that show their progress toward reaching their strategic objectives and performance goals, some of which may relate to access to care. 17 See 42 U.S.C. § 1397bb. 18 A state’s SCHIP evaluation was required to address several areas, including (1) the quality of health coverage provided, (2) choices of heath benefits coverage, (3) activities to coordinate SCHIP with other public and private programs, (4) changes in trends in the states that affect the provision of health insurance, and (5) recommendations for improving SCHIP. Page 13 GAO-03-222 Medicaid and SCHIP Access In attempting to ensure access to care in Medicaid managed care, states In Medicaid Managed focused more on setting requirements for managed care plans than on Care, States Focused monitoring compliance with these requirements or on analyzing beneficiaries’ use of services. The 14 states we reviewed with Medicaid More on Setting Plan managed care programs reported varying levels of effort (1) to establish Network certain requirements and standards for participating plans’ physician networks and to monitor their implementation and (2) to collect and Requirements than on analyze data on service utilization, such as encounter data from Monitoring Plans or participating plans and beneficiary satisfaction surveys. State Analyzing Service requirements for plans’ physician networks varied widely in their specificity, from broad statements that health plans must have “adequate” Utilization physician networks serving their enrolled members to very specific standards that set, for example, a maximum average number of beneficiaries per PCP or a maximum time frame for scheduling a first appointment. All but 1 of the 14 states required managed care plans to routinely submit lists of physicians participating in their networks, ranging from weekly to quarterly reporting. However, fewer states independently verified or routinely monitored aspects of the submitted data on managed care plans’ provider networks. For example, 8 of the 13 states receiving plans’ routine lists of participating physicians periodically verified the number of physicians accepting new Medicaid patients, but only 5 states analyzed the number of physicians to identify those participating in multiple plans, which could overstate overall physician capacity. Moreover, only 5 states routinely or independently assessed plans’ compliance with maximum waiting times for beneficiaries’ scheduling appointments. In some cases, states left it to plans to establish time frames for scheduling appointments, rather than setting statewide standards for all plans. Beyond network-related requirements and any associated monitoring, states attempted to assess beneficiaries’ actual use of services through various routine data sources and occasional special studies. Routine data sources included encounter data, where states require health plans to submit data for each service provided to each enrollee, periodic assessments of plans’ performance against standardized measures, and beneficiary satisfaction surveys. But continuing problems with the reliability of encounter data—and the fact that standardized data on plan performance and beneficiary satisfaction surveys were not representative of all Medicaid managed care enrollees—tended to undermine the utility of these data sources in describing the experiences of beneficiaries and their service utilization. The four states we visited that had experienced the withdrawal of managed care plans from their Medicaid programs had taken various steps to help minimize disruption in care for affected Page 14 GAO-03-222 Medicaid and SCHIP Access beneficiaries. However, it is not clear to what extent these states monitored service utilization for beneficiaries affected by such changes and their experiences in transitioning to new plans and physicians. Most States Set Plan To oversee access to care in their Medicaid managed care programs, the Network Requirements but states we reviewed established requirements for participating plans that Less Frequently Monitored most often focused on the size and structure of their physician networks, such as the number and geographic location of PCPs and specialists, and Plans’ Compliance beneficiaries’ ability to schedule appointments. Some states, such as Colorado, Texas, and Washington, had broad requirements that physician networks must be adequate to serve beneficiaries, as shown in figure 1. Among the 14 states that used managed care in their Medicaid programs, broad network requirements were more prevalent for specialists than for PCPs. In contrast, 11 of 14 states set specific standards or ratios relating to the number of enrolled beneficiaries per PCP, and 13 set standards for providers’ geographic proximity to beneficiaries, such as the maximum distance or travel time for a beneficiary to reach a provider’s office. More variation was evident in states’ requirements for plans in terms of appointment scheduling for beneficiaries. All 14 states set maximum time frames to schedule routine and urgent appointments, while 6 states also set maximum time frames for a newly enrolled beneficiary’s first appointment and 8 states set maximum in-office waiting times. Page 15 GAO-03-222 Medicaid and SCHIP Access Figure 1: Selected Medicaid Managed Care Plan Network Requirements and Standards in 14 States State Network size and structure Appointment waiting times PCPs Specialists Geographic First visit Appointment In-office distribution scheduling waiting time California a Colorado Florida Illinois Maryland Massachusetts a Michigan Nevada New York Ohio b Pennsylvania Tennessee Texas a Washington Specific Standard General requirement that health plans’ networks be adequate to serve their members No standard Source: GAO analysis of states’ data, as of December 2001. a State does not have a specific standard but does require plans to monitor this measure. b State only has a standard for selected populations, such as children with special needs. States took varying approaches in setting their requirements for plan networks and appointment waiting times, as shown in table 3. For example, Florida required physicians to certify that their overall practice did not exceed 3,000 patients, whereas other states established specific Medicaid beneficiary-to-PCP ratios ranging from 1,000 to 1 in Pennsylvania to 2,500 to 1 in Tennessee. With regard to appointment waiting times, some states required plans to set their own standards rather than establishing a consistent statewide standard. Page 16 GAO-03-222 Medicaid and SCHIP Access Table 3: Examples of Specific State Standards for Plan Networks and Appointment Waiting Times Availability measure Examples of standards used Plan network PCPs • Florida requires physicians to certify that their overall practice does not exceed 3,000 patients. • Maryland requires each health plan to have enrolled beneficiaries-to-PCP ratios that do not exceed 2,000:1 for adults and PCPs should have no more than 1,500 beneficiaries under age 21. Tennessee uses a maximum 2,500:1 beneficiaries-to-PCP ratio. • Ohio’s contracts with health plans specify a required number of PCPs based on the number of beneficiaries and plans in a county. Specialists • New York requires each participating plan to have 30 specialties: 14 with specific ratios of enrolled beneficiaries to specialists and 16 specialties for which health plans must have at least two providers. • Ohio requires health plans to have a specified number of 6 types of specialists, including dentists, allergists, and general surgeons, per county or service area. • Pennsylvania requires health plans to provide beneficiaries with a choice of at least 2 appropriate specialists within a reasonable geographic distance. Geographic distribution • Ohio requires health plans to ensure that 40 percent of beneficiaries reside within 10 miles of a PCP. • Texas requires health plans to have a PCP within 30 miles of a beneficiary’s residence and specialty care within 75 miles. • Washington requires health plans in urban areas to have two PCPs within 10 miles of 90 percent of beneficiaries; plans in rural areas must have one PCP within 25 miles of most beneficiaries. Appointment waiting times First visit • California requires that the first visit of newly enrolled beneficiaries be within 120 days. • Michigan requires health plans to set a standard for when new beneficiaries should first visit a PCP. • Pennsylvania requires that health assessments, general physical examinations, or first examinations be scheduled within 3 weeks of enrollment. Appointment scheduling • California requires health plans to provide urgent care within 24 hours and to set a standard for routine appointments. • Nevada requires appointments for urgent care within 2 days, and that routine care be scheduled within 2 weeks of request. • New York requires that appointments be scheduled within 24 hours for urgent care, 4 weeks for routine and preventive care, and 4 to 6 weeks for specialist care. In-office waiting time • Florida requires that explanations be given to beneficiaries if they must wait more than 30 minutes; if the wait will exceed an hour, the provider is to reschedule the appointment. • Michigan requires health plans to set an in-office waiting time standard. • Pennsylvania requires that beneficiaries wait no more than 20 minutes on average or 1 hour maximum past their scheduled appointment times. Source: GAO analysis of states’ data, as of December 2001. Routinely monitoring plan performance, especially with established network requirements and standards, is critical because providers can— and do—change their participation in Medicaid managed care, which in turn can affect beneficiaries’ access to care. In some cases, a state may not have set a specific network requirement but nonetheless independently monitors plan performance. States that monitor the extent to which participating plans’ network providers are actually available to beneficiaries are better able to systematically identify and respond to access problems. For example, see the following. Page 17 GAO-03-222 Medicaid and SCHIP Access • In 1999, Tennessee reviewed each managed care plan’s contracts with its providers and contacted providers directly to independently verify their participation with the plan and whether they were open to new Medicaid patients. The state found that, for one health plan, only 44 percent of the participating PCPs accepted new Medicaid patients; of the remaining 56 percent of PCPs, 33 percent had Medicaid patients but would not accept any new ones, and 23 percent either did not accept any Medicaid patients or could not be reached by telephone. Determining that the plan did not comply with requirements for PCP availability, the state required the plan to add providers who would accept new Medicaid beneficiaries before assigning any additional beneficiaries to the plan. State officials also reported that they now conduct a regular telephone survey of providers to verify the provider data that participating plans submit. • Washington has a broad requirement that physician networks be adequate to serve enrolled beneficiaries but does not set as many additional specific standards as do some other states. The state does, however, require participating plans to routinely report which physicians are participating in their Medicaid networks and independently verifies plan reports by periodically placing test calls to physicians. Washington also compiles the physician-level information into a centralized database to review physician participation across health plans in order to better ensure that capacity is not overstated. To monitor plan performance in terms of provider availability to beneficiaries, 13 of the 14 states we reviewed routinely obtained periodic data from participating plans on the number of physicians in their networks, ranging from weekly reports in Maryland to quarterly reports in California, Massachusetts, and New York. As a part of this routine data collection, 12 states also reviewed the geographic distribution of physicians in their networks. Fewer states, however, took additional steps to determine, on an ongoing basis, whether the plan-submitted data adequately reflected network capacity to serve Medicaid beneficiaries. For example, in 9 states the plans’ provider lists identified those physicians who were accepting new Medicaid patients, which would help indicate the extent to which plan networks were open to new public beneficiaries, and 7 states independently verified the accuracy of the submitted provider lists. Five states analyzed information across the plans’ provider lists to help identify the unduplicated number of PCPs available to the Medicaid managed care population and to help avoid overstating overall physician availability. (See fig. 2.) Page 18 GAO-03-222 Medicaid and SCHIP Access Figure 2: Variation in 14 States’ Monitoring of Medicaid Managed Care Plans’ Provider Information State How frequently Do lists identify Does the state Does the state do plans providers verify the account for submit provider accepting new accuracy providers lists to the Medicaid of provider enrolled in state? patients? lists? multiple plans? California Quarterlya b Colorado N/Ac N/A N/A N/A Florida Monthly d d Illinois Monthly b Maryland Weekly Massachusetts Quarterly Michigan Bi-weekly Bi-weekly d d Nevada New York Quarterly Monthly e Ohio Monthly b Pennsylvania Monthly b Tennessee Texas Monthly Washington Monthly Total N/A 9 7 5 Yes No N/A Not applicable Source: GAO analysis of states’ data, as of December 2001. a In some counties in California, plans are required to update their provider lists semiannually. b The state imposed specific standards regarding the number of PCPs in a health plan’s network (such as beneficiaries-to-PCP ratios) but did not account for providers that may be enrolled in multiple plans. c The state plans to reinstitute requirements for health plans to submit provider information quarterly in the next contract period. d The state requires health plans to limit the total number of patients a physician may have across all lines of business (for example, private pay, Medicaid, and other types of insurance coverage), but the state does not monitor compliance with this limit. e The state does not obtain lists from health plans that indicate the number of providers accepting new patients. Instead, it tracks the number of patients each provider is willing to accept through a health plan and compares this information to the number of beneficiaries enrolled with a particular physician. Based on this comparison, the state identifies which physicians should be accepting new patients. Compared to state monitoring of provider network information, even fewer states monitored compliance with their requirements for appointment waiting times. Five of the 14 states with Medicaid managed Page 19 GAO-03-222 Medicaid and SCHIP Access care—California, Massachusetts, New York, Tennessee, and Washington— routinely collected data or otherwise independently verified health plans’ compliance with specific appointment-related standards such as maximum time frames to schedule an initial health assessment (first visit) or routine- care appointment and in-office waiting times. To determine whether beneficiaries newly enrolled in a plan received initial health assessments within 120 days of enrollment, California regularly reviews health plan reports and physician office medical records for a sample of new beneficiaries in each plan. To verify physician compliance with appointment scheduling standards, New York makes random calls to physicians (200 offices per plan service area per year), requesting information on the next available appointment for a specified need, such as routine care, urgent care, or after-hours care. In contrast, Massachusetts directs plans to develop and monitor compliance with their own appointment scheduling requirements, and the state annually reviews and critiques the methodology and results reported by each plan. Absent routine verification or monitoring of plans’ compliance with network and availability requirements, states do not have an adequate assurance that such requirements are having their intended effect on beneficiaries’ access to managed care providers. Officials in one state that did not verify requirements indicated that the standards served as a basis for legal recourse in the event that beneficiaries raised complaints regarding appointment availability. Undertaking additional measures to verify plan compliance, as Tennessee did, can identify more comprehensive network problems that limit access to care for Medicaid beneficiaries and that might otherwise go undetected. New Regulations May Alter The new Medicaid managed care regulations, effective August 13, 2002, States’ Approaches to and to be fully implemented by August 13, 2003, will likely require some Monitoring Managed Care states to alter their approaches to requirements for their participating plans and provider networks. In general, the regulations require that states ensure—through their contracts with managed care plans—that participating plans demonstrate their capacity to serve the needs of their enrollees for any specific standards that states set for access to care. Among other things, the regulations require states to ensure that participating plans • maintain and monitor their networks of providers to provide adequate and timely access to all services covered under their contracts with the states, including monitoring the numbers of network providers who are not accepting new Medicaid patients; Page 20 GAO-03-222 Medicaid and SCHIP Access • ensure that network providers offer hours of operation that are no less than the hours offered to commercial enrollees or comparable to those of Medicaid FFS; • make services included in the contract available 24 hours a day, 7 days a week, when medically necessary; and • establish mechanisms to ensure compliance by providers with state standards for access to care. The regulations require states to certify to CMS—at the time the state enters into a contract with a plan or when there are significant changes that would affect the ability of plans to provide adequate capacity or services—that plans have complied with state requirements for the availability of services covered by managed care contracts. To the extent that states have not made specifications regarding health plan physician network capacity or assurances of access to care, states may need to revise their contracts with plans to comply with this new requirement. States that verify or monitor participating plans’ actual compliance with the terms of their contracts will likely have greater direct and routine information on whether the access-related requirements they have set out for participating plans are achieving their intended benefit for covered beneficiaries. Routine Monitoring of Determining the extent to which Medicaid beneficiaries are utilizing—and Service Utilization Often are satisfied with—covered program services is an important test of the Handicapped by Poor Data effectiveness of any state requirements for managed care plans’ network capacity and accessibility. To assess beneficiaries’ service utilization and satisfaction, the states we reviewed generally required participating plans to routinely provide data from two key sources: encounter data, which are individual-level data on service use, and HEDIS, which provides comparative information across participating plans for designated service measures. Most states also administered CAHPS, which is a standardized beneficiary satisfaction survey. However, for the majority of states we reviewed, the utility of these data for routine monitoring was often handicapped because of the frequent failure of plans to submit reliable encounter data and the exclusion of significant shares of beneficiaries from the HEDIS and CAHPS data. CAHPS survey results were further limited by poor response rates in most states. A few states reported making sufficient progress in their efforts to improve the quality of their encounter data that they could use them to routinely analyze service utilization in their Medicaid managed care programs. In addition to these routine data sources, a few states reported conducting occasional special Page 21 GAO-03-222 Medicaid and SCHIP Access studies that enabled them to identify and focus on access issues pertaining to beneficiaries’ use of services or satisfaction with services received. Encounter Data Encounter data are intended to capture information on beneficiaries’ use of primary and preventive care as well as other services, such as emergency room visits. These data can help states identify patterns of care along several dimensions, such as by type of visit or patient (such as well- child visits by age), by health condition or disorder (such as asthma or diabetes), and by plan. As a condition of their approved federal managed care waivers, states must require Medicaid managed care plans to submit encounter data. But obtaining reliable and useful encounter data has proven to be a difficult undertaking, as we have earlier reported.19 According to CMS and several of the states we reviewed, many states continue to struggle with obtaining reliable and complete encounter data. One state we contacted found that the lack of standardized provider coding and formatting procedures resulted in missing and incomplete data. As a result, only 16 percent of the provider identifiers in the submitted encounter data could be matched to the state’s Medicaid provider master file. Another state noted that its encounter data were of limited use because many health plans were unable to obtain complete data from their providers. Two of the 14 states we reviewed reported that obtaining complete encounter data was more problematic for health plans that paid their physicians a monthly capitated payment that is not linked to the delivery of specific services. For states providing Medicaid services through managed care, encounter data often are the basis for states’ responses to federal reporting requirements under Medicaid’s Early and Periodic Screening, Diagnostic and Treatment (EPSDT) services. EPSDT is designed to provide children and adolescents with access to comprehensive, periodic evaluations of health, developmental, and nutritional status, as well as hearing, vision, and dental services.20 The EPSDT annual reports that states must submit to CMS are designed to capture, by age group, information such as the 19 See U.S. General Accounting Office, Medicaid Managed Care: Challenge of Holding Plans Accountable Requires Greater State Effort, GAO/HEHS-97-86 (Washington, D.C.: May 16, 1997). 20 Federal law requires that EPSDT include services that are necessary to correct or ameliorate defects and physical and mental illnesses and conditions discovered through screening, regardless of whether those services are covered by the state’s Medicaid plan. Page 22 GAO-03-222 Medicaid and SCHIP Access number of children who (1) received EPSDT health screenings,21 (2) were referred for corrective treatment, (3) received dental treatment or preventive services, and (4) were enrolled in managed care plans. However, we have previously reported that managed care plans, particularly those that pay their participating physicians on a capitated basis, often had difficulty collecting and reporting complete and accurate EPSDT data.22 Thus, EPSDT reports that are based on encounter data are often incomplete or inaccurate, compromising the reliability of states’ data on use of these services. Despite these widespread problems with encounter data, a few states we reviewed noted that the reliability and usefulness of their encounter data have improved over time. Maryland, New York, and Michigan, for example, reported sufficient progress with improving the quality of their encounter data that they are now able to use them to analyze service utilization in their Medicaid managed care programs, as indicated below. • Maryland officials noted that after spending several years developing and refining its system for obtaining encounter data, the state is now able to use them as the basis to make risk-adjusted payments to plans and to routinely assess Medicaid managed care beneficiaries’ utilization of well- child, ambulatory, and emergency room visits. The state publicly reports performance information by health plan, creating a strong incentive for health plans to ensure that all encounters are reported. To ensure that the reported encounter data accurately portray services delivered, the state conducts validation studies on the data submitted by health plans. The state also reviews the distribution and frequency of diagnoses reported through the encounter data over time to monitor whether the mix of diagnoses across the population changes. • New York established a data warehouse for Medicaid managed care in 1997. The warehouse includes encounter data submitted by health plans as well as data from other providers’ FFS claims for reimbursement for services provided to managed care beneficiaries outside of their health 21 The components of an EPSDT health screening include a comprehensive health and developmental history, a comprehensive unclothed physical exam, appropriate immunizations, laboratory tests (including a blood lead-level assessment), and health education. 22 See U.S. General Accounting Office, Medicaid: Stronger Efforts Needed to Ensure Children’s Access to Health Screening Services, GAO-01-749 (Washington, D.C.: July 13, 2001). Page 23 GAO-03-222 Medicaid and SCHIP Access plan. A variety of reports on utilization and access data are generated and shared with plans on a restricted access web site. • Michigan is developing a data warehouse that will combine managed care encounter data with FFS claims and public health data, such as vital statistics and immunization records, into a single information system that it will use to analyze beneficiaries’ service utilization. The data warehouse will also be able to create utilization profiles by managed care plan. The state has begun testing the data warehouse that is expected to be operational within the next year. HEDIS HEDIS is a set of standardized performance measures that helps purchasers and consumers compare the performance of managed health care plans.23 HEDIS performance measures are organized into eight categories, four of which include measures directly related to beneficiary service utilization.24 The Medicaid version of HEDIS includes various access-related measures that attempt to capture beneficiaries’ use, often by age, of various preventive and other services from specified providers, as illustrated in table 4. (See app. I for a more detailed list of Medicaid HEDIS measures related to service utilization.) 23 NCQA, an independent foundation, has managed HEDIS since 1992. Originally designed for private employers as purchasers of health care, it has been adapted for public purchasers, regulators, and consumers, including Medicaid. 24 The four general HEDIS categories that directly relate to service utilization are effectiveness of care, access/availability of care, use of services, and satisfaction with the experience of care. The remaining four general categories are health plan stability, cost of care, informed health care choices, and health plan descriptive information. Page 24 GAO-03-222 Medicaid and SCHIP Access Table 4: Examples of Medicaid HEDIS Measures Related to Service Utilization for Children General HEDIS category Specific HEDIS measure Description Effectiveness of Childhood immunization The percentage of enrolled children who turned 2 years old during the care status measurement year, who were continuously enrolled for 12 months preceding their second birthdays and who were identified as having the recommended number of specific immunizations by their second birthdays. Adolescent immunization The percentage of enrolled adolescents who turned 13 during the measurement status year, who were continuously enrolled for 12 months immediately preceding their 13th birthdays and who were identified as having had the recommended number of specific immunizations by their 13th birthdays. Use of appropriate Whether members with persistent asthma are being prescribed medications medications for people acceptable as primary therapy for long-term control of asthma. with asthma Access/ availability Children’s access to PCPs The percentage of enrolled members age 12 months through 24 months, 25 of care months through 6 years, and 7 years through 11 years who had a visit with a network PCP. Annual dental visit The percentage of enrolled members age 4 through 21 who were continuously enrolled during the measurement year and who had at least one dental visit during the measurement year (when dental services are a covered benefit under Medicaid). Use of services Well-child visits in years 3, The percentage of members who were 3, 4, 5, or 6 years old during the 4, 5, and 6 of life measurement year, who were continuously enrolled during the measurement year, and who received one or more well-child visits with a primary care practitioner during the measurement year. Adolescent well-care visits The percentage of enrolled members who were age 12 through 21 years during the measurement year, who were continuously enrolled during the measurement year, and who had at least one comprehensive well-care visit with a PCP or an obstetrician/gynecologist practitioner during the measurement year. Source: NCQA, HEDIS 2000: Technical Specifications (Washington, D.C.: 1999). Twelve of the 14 states we reviewed used HEDIS measures to help assess Medicaid beneficiaries’ utilization of services. Eleven states required participating plans to submit HEDIS performance results, while 1 state— Ohio—conducted its own HEDIS analysis using encounter data submitted by plans.25 Some states used the full set of HEDIS measures, while others used selected measures corresponding to areas of interest. Despite the potential of HEDIS to provide valuable information regarding beneficiaries’ use of managed care services, its narrow focus in identifying beneficiaries to be included in the assessments often limits the ability to generalize results to all beneficiaries within a plan or within a state. Many HEDIS measures require beneficiaries to have 12 months of continuous 25 Tennessee and Texas did not use HEDIS to assess plan performance. Page 25 GAO-03-222 Medicaid and SCHIP Access enrollment in a single managed care plan in the assessment year in order to be included in the measures.26 As a measurement criterion, the continuous enrollment requirement is intended to ensure that comparisons of performance across health plans are made on the basis of sample populations that have been enrolled for similar periods of time. However, because beneficiaries’ average length of time in the Medicaid program can be less than 12 months—ranging from 6 to 9 months in three states—this 12-month enrollment requirement excluded at least one quarter of Medicaid beneficiaries from most of the states we reviewed and more than half in four states, as shown in table 5.27 Consequently, HEDIS measures may not provide a representative measure of service utilization for a significant share of children covered by Medicaid managed care. Another limitation of the HEDIS measures is that they are often based on encounter data and are thus subject to the reliability concerns previously raised. 26 Although some HEDIS measures have a 12-month continuous enrollment requirement, individuals with one gap in enrollment of up to 45 days or less can be included in the sample. However, to be included, individuals must remain with the same health plan after a break in enrollment. 27 According to one report, at least one state—Iowa—analyzed HEDIS measures for individuals that were continuously enrolled in Medicaid for less than 12 months. See NCQA, Medicaid HMO and Fee-For-Service Comparison Strategy: Methodological Issues (Washington, D.C.: NCQA, n.d.). http://www.ncqa.org/Programs/qsg/medicaidcomparison.html (downloaded July 8, 2002). Page 26 GAO-03-222 Medicaid and SCHIP Access Table 5: Estimated Percentage of Medicaid Children Excluded from HEDIS a b State Percentage excluded from HEDIS Colorado 79 Ohio 75 Washington 52 Florida 51 New York 40 Michigan 33 Maryland 32 Pennsylvania 31 Massachusetts 29 California 26 Illinois 24 c Nevada d Tennessee d Texas Source: GAO analysis of states’ data, as of December 2001. a States were asked for enrollment information for the most recent year for which data were available, which was generally 2001. b Percentages represent the portion of the population excluded from the required sample for some of the HEDIS measures because they were enrolled for less than 12 months. c State could not provide exclusion data. d State does not use HEDIS data. Several states we reviewed provided examples of how they used the HEDIS data they received from participating plans. These included using the information to compare each plan’s performance against national Medicaid averages for selected measures and developing report cards to compare results across plans. Given issues we identified with the completeness of the data, however, such uses and comparisons may not be reliable indicators of beneficiaries’ use of services and may render a false impression of beneficiaries’ actual experience in service utilization. CAHPS State Medicaid managed care programs are required to have an internal quality assurance system, which can involve administering beneficiary satisfaction surveys.28 Thirteen of the 14 states we reviewed reported using CAHPS to assess beneficiaries’ experiences with their Medicaid managed 28 42 C.F.R. § 434.34. Page 27 GAO-03-222 Medicaid and SCHIP Access care plans.29 CAHPS is a standardized survey designed to compare the performance of managed care plans on the basis of beneficiaries’ perceptions regarding the care they received through their plans.30 The CAHPS survey covers a range of topics related to service utilization, including appointment scheduling, waiting time in a physician’s office, and the use of specialty services, as shown in table 6. Table 6: Examples of Beneficiary Satisfaction Questions for Children Covered by CAHPS Measure Question First visit • Did you get an appointment for your child’s first visit to a doctor or other health care provider for a checkup, or for shots or drops, as soon as you wanted? Appointment scheduling • In the last 6 months, how often did your child get an appointment for regular or routine health care as soon as you wanted? • In the last 6 months, when your child needed care right away for an illness or injury, how often did your child get care as soon as you wanted? Ambulatory care • In the last 6 months (not counting times your child went to an emergency room), how many times did your child go to a doctor’s office or clinic? In-office waiting time • In the last 6 months, how often did your child wait in the doctor’s office or clinic more than 15 minutes past the appointment time to see the person your child went to see? Referral to specialist • In the last 6 months, how much of a problem, if any, was it to get a referral to a specialist that your child needed to see? Source: CAHPS 2.0, Child Medicaid Managed Care Questionnaire and Child Supplemental Questions, (Rockville, Md.: AHRQ, 1998). Like HEDIS, however, information from CAHPS is only gathered from a subset of beneficiaries. CAHPS has a 6-month continuous enrollment requirement for Medicaid beneficiaries to be included in the survey sample. While this is a shorter minimum enrollment period than for HEDIS, it still resulted in excluding about one quarter or more of covered beneficiaries in five states we reviewed, and nearly half or more in two states, as shown in table 7. Moreover, several states using CAHPS reported that they had low response rates from the sampled population; in some cases, surveys targeted only those beneficiaries with telephones, a 29 Tennessee opted to use a state-designed beneficiary satisfaction survey rather than CAHPS. In most cases, the states we reviewed administered CAHPS directly or through the use of an independent contractor. Three states—Colorado, Illinois, and Pennsylvania— required participating plans to administer CAHPS. 30 CAHPS was developed in 1995 by the federal AHRQ to provide information to help beneficiaries compare health plans. Page 28 GAO-03-222 Medicaid and SCHIP Access practice that has the potential to bias the results for beneficiaries who could not be reached by that method.31 Table 7: Estimated Percentage of Medicaid Children Excluded from CAHPS a b State Percentage excluded from CAHPS Colorado 61 Ohio 49 Texas 31 Florida 30 Washington 24 Michigan 17 Maryland 16 Massachusetts 15 New York 15 Pennsylvania 13 California 11 Illinois 10 c Nevada d Tennessee Source: GAO analysis of states’ data, as of December 2001. a States were asked for enrollment information for the most recent year for which data were available, which was generally 2001. b Percentages represent the portion of the population excluded from CAHPS because this group was enrolled for less than 6 months. c State could not provide data. d Tennessee does not use CAHPS, although it does conduct a state-designed survey of a sample of all state residents about insurance coverage and satisfaction with services, including access to care. Gauging beneficiary satisfaction with services solely through a satisfaction survey is an inherently difficult process, especially when a sample is not representative or response rates are low. To augment information on beneficiary satisfaction, states also had available the results of their complaints and grievances processes, which they are required to have as a condition of their managed care programs. Nearly all of the states we reviewed with Medicaid managed care operated a central hotline or complaint number, where beneficiaries could obtain program information 31 Among the nine states in our sample that reported their response rates, the response rates for the CAHPS survey of families with children ranged from 27 percent in Nevada to 85 percent in Illinois. Page 29 GAO-03-222 Medicaid and SCHIP Access or request assistance locating providers in addition to filing complaints. The states we reviewed generally focused on ensuring that complaints and questions raised by beneficiaries’ calls were addressed. For example, five states—Ohio, Maryland, Michigan, New York and Washington—had information databases that tracked complaints from their inception to their resolution. New York, Ohio, and Washington complaint reports were also analyzed by managed care plan, which allowed officials to identify any trends in beneficiary complaints. As a tool to assess overall problems with access to care, records of complaints and grievances had several limitations. In some cases, states’ hotline or complaint data did not distinguish between requests for assistance and complaints about provider services, thus making it difficult to assess the extent of any systemic access problems. In addition, a small number of complaints could be difficult to interpret at face value; while few complaints or grievances could indicate overall satisfaction with care, it could also indicate a general lack of knowledge about or ability to file a formal complaint or grievance. A small number of complaints also could limit the state’s ability to identify any specific trends of systemic problems with access to care with a specific plan or within a state’s Medicaid managed care program as a whole. Special Studies A few states we contacted reported that they occasionally conducted special studies, in addition to any routine monitoring they did, to assess service utilization issues for their Medicaid beneficiaries. For example, Maryland’s 4-year evaluation of its Medicaid managed care program, published in January 2002, concluded that providers and consumers felt that PCP networks were “under stress” in certain areas of the state, with a notable lack of physicians in rural areas of the state.32 The evaluation also identified significant inaccuracies with plan-submitted data on physician providers, including duplicate provider entries, incorrect provider affiliation status with participating plans, and missing information. As a result, the state took steps to develop and implement more rigorous methods of monitoring plan-submitted data. In particular, the state now monitors plans’ PCP networks, including verification calls to samples of physicians, so that PCP shortage areas can be identified and addressed. In the future, the state also plans to develop more specific standards for 32 Maryland Department of Health and Mental Hygiene, HealthChoice Evaluation, Final Report and Recommendations (Washington, D.C.: Jan. 15, 2002). Page 30 GAO-03-222 Medicaid and SCHIP Access commonly used specialists and monitor plans’ compliance with these standards. In 2001, Washington conducted a survey of new Medicaid managed care beneficiaries in 14 counties as a result of concerns raised by beneficiary advocates in light of managed care plans’ withdrawal from program participation. The study examined these beneficiaries’ experiences with accessing medical care, including emergency room use. The study found that 90 percent of new beneficiaries reported having a PCP after enrollment in Medicaid, compared to 62 percent having a PCP before enrollment in Medicaid managed care. Additionally, there were no significant differences between the experiences of managed care and FFS beneficiaries who had obtained medical, specialist, or emergency room care. However, the study did find that the majority of beneficiaries were unfamiliar with several key processes concerning their managed care plans, such as how to change PCPs within a health plan, contact their health plans when questions or problems arose, and make complaints. Based on these findings, the state plans to work with managed care plans to improve beneficiaries’ awareness regarding PCP selection and communication with their health plans. States Attempted to The managed care industry—in the commercial as well as public sectors— Minimize Impact of has experienced considerable changes in recent years following periods of Managed Care Plan rapid entry of multiple managed care plans in certain markets and subsequent retrenchment based on plans’ willingness or ability to compete Withdrawals on Access to in those markets. Many communities and states have experienced changes Care, but Effect on in the number of managed care plans as a result of numerous health plan Beneficiaries Is Uncertain mergers, acquisitions, and closures as the managed care industry has evolved and matured. State Medicaid programs have often been affected by the withdrawal of some managed care plans from their programs; in some cases, states have intentionally acted to reduce the number of participating plans. The four states we visited—Massachusetts, Ohio, Tennessee, and Texas—had experienced such changes to varying degrees. These states had taken various measures to help minimize any adverse effects on beneficiaries’ access to care due to participating plans leaving the Medicaid program. It is not clear, however, to what extent these states’ efforts had been successful in helping beneficiaries transition smoothly to new health plans and physicians and thus avoid problems with access to care. Page 31 GAO-03-222 Medicaid and SCHIP Access The potential amount of disruption that occurs when a health plan withdraws from a state or community can vary considerably, depending on a number of circumstances. Health plan mergers can result in minimal changes for beneficiaries if they are able to maintain established relationships with their providers and can even strengthen the network of available providers within a plan. In Massachusetts, for example, a merger between two health plans in the early 1990s was considered by state officials to have increased physician availability for Medicaid beneficiaries enrolled in managed care. Officials in Massachusetts also noted that reductions in the number of health plans ensured that participating plans have enough enrolled beneficiaries to spread the costs and risks associated with capitation payments. In other cases, however, the extent of disruption may be more severe, particularly when large numbers of beneficiaries are affected and a significant number of plans struggle to remain financially viable. Each of the four states we visited experienced varying levels of health plan withdrawals from their Medicaid managed care programs. Plan withdrawals over several years have affected almost 50 percent of beneficiaries in Tennessee and over 15 percent in Ohio, raising concerns about the accessibility of care to beneficiaries in these states. The magnitude of health plan withdrawals in Tennessee necessitated state efforts to recruit additional plans, at least one of which was later found by the state to have deficiencies related to failure to pay physicians accurately and promptly. In Ohio, at least one health plan that withdrew from the state program also failed to pay some of its network providers for services already rendered. In such cases, delayed reimbursement by managed care plans can seriously jeopardize providers’ willingness to continue participating in the Medicaid program and provide services to eligible beneficiaries. In contrast to Tennessee and Ohio, plan withdrawals in Massachusetts and Texas have affected a smaller share of beneficiaries. Massachusetts estimated that about 4 percent of its beneficiaries were affected by an early period of plan fluctuation as the state was implementing its mandatory managed care program; since 2000, however, the program has been stable with the same four managed care plans participating. In Texas, approximately 1 percent of beneficiaries have been affected by withdrawals of participating plans since the state implemented managed care in 1996. To avoid disruptions in care for beneficiaries when plans ceased their participation in the Medicaid program, these states had implemented various procedures to help smooth beneficiaries’ transition to other plans or providers. For example, in cases where a withdrawing health plan Page 32 GAO-03-222 Medicaid and SCHIP Access intended to sell its membership to another plan, Ohio first compares the provider network of the withdrawing plan with the health plan that is purchasing the membership. The state does not approve the sale of membership unless most of the providers participating in the withdrawing plan also participate in the purchasing plan. In other cases, a state’s contract with its managed care plans required certain actions. Ohio’s contract, for example, requires a minimum of 75 days advance notice of a plan’s intention to terminate its participation in the program and includes provisions to collect a monetary assurance from the withdrawing plan or to withhold payments until all contractual requirements are completed, including required payments to network providers. Ohio and Texas provided examples of efforts to inform beneficiaries directly affected by a plan’s withdrawal about options available to them to continue care, such as information on other participating plans and how to choose another plan. These four states indicated that they believed their efforts to respond to changes in managed care participation were sufficient to minimize disruption to care for Medicaid-eligible beneficiaries. However, the extent to which the states’ efforts adequately ensured beneficiaries’ access to continuous care was uncertain. Appendix II provides more detail on managed care plan withdrawals in these four states. For states’ FFS-based Medicaid delivery systems, which continue to serve For Medicaid FFS, the majority of children in half of the states we reviewed, requirements for State Requirements participating providers and monitoring of provider availability were significantly more limited than for managed care. State analysis of service for Providers and utilization data to assess the frequency and patterns of care that Monitoring of Service beneficiaries received was also more limited, despite the ready availability of such data through states’ claims payment systems. For traditional FFS Utilization Were More programs, beneficiaries may seek care from any providers participating in Limited the Medicaid program and may change providers at any time if they are dissatisfied. However, Medicaid beneficiaries’ ability to easily change providers is dependent on the number, type, and location of providers willing to take new Medicaid patients, which in turn is strongly influenced by Medicaid payment rates and associated administrative processes. We found that Medicaid FFS payment rates were significantly lower than rates for comparable Medicare services in the majority of states we reviewed, which can discourage provider participation and thus restrict beneficiaries’ access to a broad supply of providers. States that used PCCM programs as part of their FFS service delivery systems were somewhat more prone to set certain requirements for participating PCCMs, such as a maximum number of assigned beneficiaries and their geographic proximity to beneficiaries, than were states with traditional Page 33 GAO-03-222 Medicaid and SCHIP Access FFS systems. States with FFS programs generally did not set requirements for specialists or for physicians’ appointments, such as maximum waiting times to schedule an appointment, as they did for their managed care plans. States were more likely to conduct beneficiary satisfaction surveys for their PCCM programs than for their traditional FFS systems; the survey results, however, had the same constraints as previously discussed for managed care due to the limited share of beneficiaries participating in the surveys and low response rates. Low FFS Payment Rates States are required to ensure that their Medicaid service delivery and Can Reduce Provider payment systems will afford beneficiaries access to services similar to Participation and Restrict those provided to the state’s general population. To do this, states determine which providers may enroll in the Medicaid program to provide Access to Care services, set payment rates for covered services, and pay claims that providers submit for the services they provide. In several of the states we reviewed with Medicaid FFS programs, program officials said that provider survey information and beneficiary complaint data suggested that low payment rates, slow payment, and other administrative issues deterred physicians in primary care or in some specialties from participating in the program. As we reported earlier, if payment rates decline to the point that they cause physicians to leave Medicaid or to reduce the number of beneficiaries they serve, then beneficiary access may be restricted.33 Our analysis of payment rates indicated that Medicaid FFS payments to physicians for primary and preventive services for children were often significantly lower than what Medicare paid for comparable services in many of the states we reviewed. For the 13 states that paid physicians on a FFS basis for Medicaid-eligible children, payment rates ranged from 32 percent to 89 percent of Medicare rates. Nine of these states’ Medicaid rates were two-thirds or less of Medicare rates for comparable services. (See app. III for more detail.) Officials in many of these states said that Medicaid rates were also below those of commercial payers, although they generally had not conducted systematic studies to document these differences. 33 See U.S. General Accounting Office, Medicaid and SCHIP: States’ Enrollment and Payment Policies Can Affect Children’s Access to Care, GAO-01-883 (Washington, D.C.: Sept. 10, 2001). Page 34 GAO-03-222 Medicaid and SCHIP Access Most Traditional FFS Despite the potential for low FFS rates to limit the number of providers Programs Set Few Goals willing to participate in the program, the nine states we reviewed with Regarding the Number of traditional FFS programs did not set specific goals for the number of physicians participating in their Medicaid programs and did not actively Providers, and Conducted monitor the number and location of providers.34 While states had lists of Minimal Monitoring of physicians who were enrolled as Medicaid providers and who submitted Service Utilization claims for services provided, in most cases these lists were not frequently or comprehensively updated and thus did not provide an accurate count of actively participating physicians. Some states’ Medicaid physician databases included physicians who had not provided services to Medicaid patients for years. In one state, the database doublecounted providers who had more than one service location or billing identifier. In addition, although states have claims data that serve as the basis for paying providers for services rendered, only some analyzed this information to identify PCPs, specialists, or other providers who were actively treating Medicaid beneficiaries. Even when they did, states often defined “active” providers to include those who submitted a single claim during the past year. With respect to appointments, such as maximum waiting times to schedule a routine or urgent appointment, none of the states we reviewed with traditional FFS programs had specific standards comparable to those we saw for managed care programs. States also did little to monitor service utilization by Medicaid beneficiaries participating in traditional FFS care despite having a ready source of data in their claims payment systems. Claims data contain the type and frequency of services Medicaid beneficiaries have received and the type of provider delivering the care, which can be used to analyze service utilization. States did report using claims data to develop utilization statistics to meet federal requirements for annual reporting on EPSDT services for children. However, we have reported earlier that state EPSDT reports are often incomplete and unreliable, thus compromising their utility in assessing whether children are receiving required services.35 Beyond EPSDT, only one state with a Medicaid traditional FFS system reported analyzing claims data to evaluate access to care on primary and 34 These nine states were California, Colorado, Illinois, Louisiana, Nevada, New York, Ohio, Texas, and Washington. The share of Medicaid-eligible children participating in these states’ traditional FFS programs ranged from a low of about 30 percent in California and Colorado to about 90 percent in Illinois and Louisiana. (See table 2 for more detail by state.) 35 GAO-01-749. Page 35 GAO-03-222 Medicaid and SCHIP Access preventive services, such as annual well-child and dental visits.36 Rather than evaluate access to primary care, at least three states used claims data to assess inappropriate utilization of higher-cost services, such as emergency room care. For example, Texas collects and analyzes information on beneficiaries who potentially overuse care—defined as those at or above the 90th percentile of use for particular services, including physician, emergency room, and pharmacy services. Patients suspected of misusing services may be restricted to using a specific physician or pharmacy, with the goal of reducing their use of services to a more appropriate level. Four of the nine states with traditional FFS systems reported periodically using beneficiary satisfaction surveys, such as CAHPS, to help assess issues regarding access to care. These states were Colorado, Illinois, Ohio, and Washington. As with Medicaid managed care, however, the utility of these surveys is diminished when there are low response rates and a lack of beneficiary representation in the sample selection. In one state, the survey sample was limited to individuals who had received at least one service in the prior 6 months, thus excluding individuals who may have tried but failed to obtain services. Another state reporting a low beneficiary response rate found that while the cooperation rate was high among those who were reached, many potential respondents in the survey sample could not be contacted because of address or telephone number changes. PCCM Programs Had States’ PCCM programs are a hybrid of FFS and managed care service Some Requirements for delivery approaches. They emulate FFS programs in the sense that the Providers, but Monitoring state has a direct relationship with providers who are enrolled to participate in the program and paid retrospectively for services actually of Service Utilization Was delivered. PCCM programs share characteristics of managed care in the Limited sense that beneficiaries are assigned to a PCCM—a physician, or a practice or other entity—that is responsible for coordinating their care as a case manager. The seven states we reviewed with PCCM programs had more requirements for participating PCCMs than they did for providers in traditional FFS programs, but fewer than PCPs in managed care 36 Since 1995, Ohio has used HEDIS primary care access measures for beneficiaries in its traditional FFS program. Page 36 GAO-03-222 Medicaid and SCHIP Access programs.37 Similar trends were evident in terms of states’ routine monitoring of PCCM availability and beneficiaries’ service utilization: more than FFS, less than managed care. The states we reviewed with Medicaid PCCM programs most often set requirements for the maximum number of beneficiaries that a PCCM could serve and the geographic proximity of PCCMs to their enrolled beneficiaries. None set limits on the number of beneficiaries a specialist could serve, and few set specific standards for appointment waiting times with their PCCMs; overall, PCCM programs had fewer standards than those imposed under managed care, as shown in figure 3. Figure 3: Selected Requirements for Medicaid PCCM Providers in Seven States State Number and location of providers Appointment waiting times PCCMs Specialists Geographic First Appointment In-office distribution visit scheduling waiting time Arkansas Colorado a Florida a a a a Louisiana b Massachusetts a b a Pennsylvania a a a a Texas b Total 5 0 5 2 d 3 d 2 Yes No Source: GAO analysis of states’ data, as of December 2001. a This type of standard exists in this state’s managed care program but not its PCCM program. b This type of standard exists in the state’s PCCM program, but not its managed care program. 37 The seven states we reviewed with PCCM programs were Arkansas, Colorado, Florida, Louisiana, Massachusetts, Pennsylvania, and Texas. Arkansas and Louisiana do not have Medicaid managed care programs other than PCCM, whereas the other five states do. The share of Medicaid-eligible children participating in the seven states’ PCCM programs ranged from a low of 12 percent in Louisiana to a high of 100 percent in Arkansas. (See table 2 for more detail.) Page 37 GAO-03-222 Medicaid and SCHIP Access States’ PCCM capacity requirements were most often based on setting a maximum number of beneficiaries that a PCCM or practice could serve, ranging from 1,000 beneficiaries per PCCM in Arkansas and Pennsylvania to 1,500 in Florida and Massachusetts. Louisiana set a limit of 1,200 beneficiaries per PCCM, or 4,800 for a group practice, and allowed an additional 300 beneficiaries to be enrolled for each nurse practitioner. With regard to geographic requirements, all five of the PCCM programs that had requirements for this standard specified a basic maximum of 30 minutes or 30 miles for beneficiaries to reach their PCCMs. Four of these states set a higher maximum for rural areas—such as 50 miles in Colorado or 60 minutes in Pennsylvania—or allowed general exceptions to the 30- minute standard for beneficiaries living in some rural areas. States typically monitored provider participation in their PCCM programs by compiling weekly or monthly lists of participating PCCMs and the number of beneficiaries each PCCM was assigned, which could serve as the basis for paying the monthly PCCM fee. Monitoring these relative numbers also allowed states to ascertain whether PCCMs could be assigned additional beneficiaries. States therefore had current information on those providers actively participating as PCCMs and the numbers of assigned beneficiaries. This information alone, however, would not yield insights into how easily beneficiaries could see their PCCMs. When states had both managed care and PCCM delivery systems, they less frequently set requirements for PCCM appointment waiting times than they did for managed care.38 Three states that operated both PCCM and managed care programs—Colorado, Florida, and Pennsylvania—did not set any appointment waiting time standards for PCCMs as they did for managed care. In contrast, Massachusetts required its PCCMs to see new patients within a specific time frame in its PCCM program, but not in managed care. Of the four states that did set specific requirements for appointment waiting times, only Texas reported conducting routine monitoring to assess PCCM compliance with those requirements. Texas officials reported conducting audits of a random sample of 20 PCCMs per quarter per service area to evaluate compliance with respect to appointment scheduling and in-office waiting time. 38 Some states’ contracts with PCCMs may include a general requirement that PCCMs provide care on a “timely basis.” Page 38 GAO-03-222 Medicaid and SCHIP Access To monitor service utilization within their PCCM programs, states most often relied on analyses of their FFS claims data. Six of the seven states with PCCM programs provided PCCMs that had a certain minimum number of assigned beneficiaries with periodic data profiles that compare service utilization patterns in their Medicaid practices with those of the overall program or other PCCMs.39 These data profiles often focused on high-cost services or those at risk of overutilization, such as inpatient hospitalization or emergency room use. Three states also included information related to primary and preventive services use. For example, see the following. • Massachusetts provided PCCM practices that had 200 or more enrolled beneficiaries with practice-specific and comparative information about the percentage of children who received a recommended number of well-child visits, by age group. The state further identified, for each practice, individual patients who had not received the recommended number of well-child visits. State program staff members met with each provider twice annually to discuss approaches to address problems identified in these data that may indicate limited access. • Texas provided participating PCCMs with comparative information on selected services per beneficiary, including EPSDT visits, family planning, and immunizations. In contrast, states typically did not monitor the utilization of services provided by specialists, although several state PCCM programs required documentation of PCCM referrals to specialists. Officials in several states were aware of problems with access to some types of providers and specialists in their PCCM programs, including dentists, dermatologists, and pediatric neurosurgeons. In an attempt to address such problems, Arkansas conducted a survey of dentists and Florida conducted a survey of physicians to identify obstacles to their willingness to accept Medicaid patients. While such one-time surveys can provide insightful information about problems and potential solutions in a specific period, they do not take the place of routine or targeted monitoring that can more systematically pinpoint problems for particular specialties, geographic areas, or beneficiaries. 39 Targeting such profiles and analyses to PCPs with a certain minimum volume of beneficiaries allows more meaningful data comparisons with the program and other PCPs than would be possible for PCPs with only a few beneficiaries. Page 39 GAO-03-222 Medicaid and SCHIP Access Each of the states we reviewed with PCCM programs conducted beneficiary satisfaction surveys. In addition, Colorado administers its CAHPS survey to individuals participating in all three of the states’ Medicaid service delivery systems—managed care, traditional FFS, and PCCM—in order to help assess experiences of program beneficiaries relative to one another. However, given the shortcomings identified earlier—low response rates and exclusions of certain beneficiaries from sample selection—states could not with confidence generalize the results of these beneficiary surveys to the larger population. States have used the flexibility provided by SCHIP to take varying Distinct SCHIP approaches for their service delivery systems for eligible children. Of the Programs Had Fewer 16 states we reviewed, 9 states chose to serve their SCHIP beneficiaries through programs that were primarily designed as expansions of Medicaid Network or modeled on their Medicaid programs in terms of benefits and provider Requirements and networks.40 These 9 states used the same health plan contracts for Medicaid and SCHIP managed care, and the same provider lists for both Less Monitoring of programs’ FFS-based delivery systems. In these cases, the extent of SCHIP Service Utilization monitoring would mirror that of the states’ Medicaid programs. On the other hand, 7 states designed at least part of their SCHIP programs to be distinct from Medicaid. These programs relied almost exclusively on managed care to deliver services. Although most of these states also had significant shares of their Medicaid beneficiaries in managed care, they set significantly fewer provider network requirements for their distinct SCHIP programs than for Medicaid and did less monitoring of providers enrolled in their SCHIP programs and of children’s use of services in SCHIP. In general, few states with distinct SCHIP programs routinely collected and analyzed data to ensure that SCHIP-eligible children were receiving covered services. 40 These states were Arkansas, Illinois, Louisiana, Maryland, Massachusetts, Nevada, Ohio, Tennessee, and Washington. Page 40 GAO-03-222 Medicaid and SCHIP Access States with SCHIP The seven states that chose to serve all or most of their SCHIP Programs Distinct from beneficiaries through programs that were distinct from Medicaid used Medicaid Set Few Provider managed care delivery systems almost exclusively.41 These states were not bound by access-related requirements comparable to those for Medicaid Requirements PCCM or managed care programs. As such, they set provider network requirements and monitored service utilization less often in their distinct SCHIP managed care programs than they did in their Medicaid managed care programs. As shown in figure 4, only two of these seven states set specific beneficiary-to-PCP ratios for SCHIP, compared to five states for Medicaid, and no state set specific requirements for specialists, compared to three states for Medicaid. Similarly, only one of the seven states with distinct SCHIP programs set a maximum waiting time for a first appointment with a PCP and none had a requirement for in-office waiting times; in contrast six of these states’ Medicaid managed care programs set specific requirements for one or both of these access measures. Only four of the distinct SCHIP programs in these states set any specific standards for appointment scheduling, compared to all seven of the states’ Medicaid managed care programs. 41 These states were California, Colorado, Florida, Michigan, New York, Pennsylvania, and Texas. With the exception of Florida, all of the states used managed care delivery systems for all of their SCHIP programs; Florida enrolled a small number of SCHIP children into a PCCM program. Page 41 GAO-03-222 Medicaid and SCHIP Access Figure 4: Comparison of Seven States’ Requirements and Standards for Providers in Medicaid and SCHIP Managed Care State Network size and structure Appointment waiting times PCP Specialists Geographic First Appointment In-office distribution visit scheduling waiting time California a Colorado Floridab Michiganc New York Pennsylvania a Texas Number of states with standards for 5 3 7 5 7 4 Medicaid Number of states with standards for 2 0 6 1 4 0 SCHIP Medicaid managed care provider availability standard SCHIP provider availability standard No standard for Medicaid or SCHIP Source: GAO analysis of states’ data, as of December 2001. Note: Table does not include Medicaid and SCHIP programs that have only a general requirement that health plans’ networks be adequate to serve their members. a Although the state did not have a specific standard, it did require plans to monitor this measure. b Florida’s separate SCHIP programs vary by beneficiary age category. The SCHIP column in this table refers to the program for older children, as the program for children under age 5 is modeled after Medicaid and thus has the same standards as the Medicaid program. c Michigan’s data reflect its arrangement with all participating health plans except for one plan, which operates under different requirements. The seven states with distinct SCHIP programs also monitored the availability of PCPs in plan provider networks less frequently than in Medicaid. In contrast to Medicaid managed care where nearly all states monitored providers at least quarterly, just three states required plans to submit provider lists periodically throughout the year—Colorado, New York, and Texas. To confirm provider information submitted by plans participating in SCHIP, only New York systematically contacted physicians Page 42 GAO-03-222 Medicaid and SCHIP Access to verify information about whether network PCPs were accepting new SCHIP patients.42 Four states required SCHIP plans to submit physician data annually or every several years during state licensure reviews or for the contract renewal process. Among these, California’s SCHIP program required plans to indicate the number and percentage of PCPs and specialists accepting new patients and also to notify the state when there was a change in the provider network that resulted in disruption of 25 or more beneficiaries. The extent of states’ monitoring of participating plans’ SCHIP provider networks did not appear to be related to whether SCHIP-eligible beneficiaries had access to commercial or noncommercial networks within the plans. Some states—such as New York and Texas—did not know whether SCHIP-eligible beneficiaries had access to the same providers as were participating in plans’ commercial networks. Other states—such as Florida, Michigan, and Pennsylvania—stated that most if not all of their SCHIP populations did have access to the same providers as in the plans’ commercial networks. However, without direct monitoring of PCPs enrolled in SCHIP plan networks, states had little or no direct knowledge of the extent to which PCPs would see SCHIP beneficiaries, including whether enrolled PCPs would accept new SCHIP patients at all or limited their practice to only a small number. Distinct SCHIP Programs States with SCHIP programs distinct from Medicaid reported fewer efforts Monitored Service to monitor children’s utilization of services than in their Medicaid Utilization Less than managed care programs. This held true for their use of encounter data as well as for HEDIS measures and CAHPS beneficiary satisfaction survey Medicaid data. CMS does not require states to collect encounter data from managed care plans participating in SCHIP, as it does in Medicaid managed care. Of the states we reviewed with distinct SCHIP programs, we found that two states—Florida and Texas—were attempting to collect as well as analyze encounter data for SCHIP-eligible children in order to assess the type and frequency of services they received. Florida’s distinct SCHIP program uses encounter data to compare the number of ambulatory visits made by SCHIP beneficiaries to the number of visits that would be expected for 42 To achieve this purpose, the state contacted a sample of 50 to 200 providers for each plan participating in Medicaid and SCHIP, twice a year. Page 43 GAO-03-222 Medicaid and SCHIP Access those children based on their diagnoses.43 Texas’ distinct SCHIP program, which was initiated in 2000, has used encounter data to compare immunization rates by plan with rates in commercial plans. Four of the seven states—California, Michigan, New York, and Pennsylvania—required plans to submit HEDIS data so that the states could assess plans’ performance with respect to access to various preventive and other services.44 Compared to Medicaid, these HEDIS data may be more complete in three of these states—California, Michigan, and Pennsylvania—because they had opted to provide SCHIP-eligible children with continuous eligibility for a 12-month period, thus increasing the likelihood that a more representative share of eligible children and their families would be included in the assessments. Five of the seven states—California, Florida, Michigan, Pennsylvania, and Texas—used CAHPS to assess beneficiaries’ satisfaction with care. Compared to Medicaid, the CAHPS data for four of these states’ SCHIP programs may be more complete than for their Medicaid programs because these states provide continuous eligibility for a 12-month period. We provided a draft of our report for comment to HHS, as well as to Agency and State Medicaid and SCHIP officials in the 16 states included in our analysis. We Comments and Our received comments from HHS and from 13 states. Three states did not respond with comments. Evaluation HHS Comments With regard to states’ Medicaid managed care programs, HHS highlighted new requirements included in CMS’s June 2002 regulation implementing Medicaid managed care provisions of the Balanced Budget Act of 1997. HHS commented that, among other things, the regulation requires states to develop a quality strategy setting access standards for network adequacy and timeliness of access to care. HHS described this new regulation as also making clear the states’ responsibility to continually monitor plans’ compliance with these standards. While many states, including 13 of the 14 43 Florida’s distinct SCHIP program uses the Ambulatory Care Groups (ACG) Case-Mix Adjustment System to assign beneficiaries to 1 of 53 ACG categories for the purpose of this analysis. 44 Although these four states used HEDIS in both their separate SCHIP and Medicaid programs, only New York reported comparing the results across the two programs. Page 44 GAO-03-222 Medicaid and SCHIP Access states we reviewed with Medicaid managed care delivery systems, were already subject to certain access requirements as a condition of receiving waivers of federal Medicaid requirements to operate their managed care programs, these requirements were not consistent from state to state. This new regulation, which must be fully implemented by August 13, 2003, has the potential to bring a more systematic approach to access requirements. More importantly, its emphasis on state monitoring could better ensure that such requirements are achieving their intended purposes. For states’ Medicaid FFS delivery systems, HHS acknowledged the relationship between reimbursement rates and provider participation, noting that states can increase payment rates in geographic areas and specialties where access has been demonstrated to be a problem. Beyond reimbursement rates, HHS commented that our draft report pointed out a lack of data to quantify whether there is an access problem in Medicaid FFS. To the contrary, our report indicates that despite a ready source of information—claims data—for evaluating access to care in a FFS environment, states generally did not do so. HHS agreed that our placement of PCCM programs in the FFS category was accurate from a reimbursement standpoint, but stated that PCCM should be considered a managed care delivery system because PCPs are expected to coordinate care. We continue to believe that a PCCM program is better described as an FFS-based delivery system because the differences between PCCM and managed care reimbursement approaches can differentially affect provider incentives in providing covered services. Our report does distinguish, however, the degree to which managed care, traditional FFS, and PCCM programs employ access standards and monitoring. Overall, states with PCCM programs tended to establish more standards and conduct somewhat more monitoring than for their traditional FFS programs, but less than for their managed care programs. With regard to our finding that states with distinct SCHIP programs did significantly less to monitor access to care than for their Medicaid managed care programs, HHS stated there was a key difference in design and intent by the Congress between SCHIP and Medicaid. HHS commented that SCHIP allows states to have the flexibility to design programs that mirror private insurance and rely on private insurance mechanisms to ensure access to and quality of care, rather than laying out specific requirements. Acknowledging that states may not have comparable requirements for SCHIP and Medicaid monitoring provider participation and beneficiary service utilization, HHS said that states are monitoring enrollment, health access, and outcomes in their SCHIP Page 45 GAO-03-222 Medicaid and SCHIP Access programs. However, with regard to access, we found that few states with distinct SCHIP programs monitored provider network participation or routinely collected and analyzed data to ensure that SCHIP-eligible children were receiving covered services. We did not intend to suggest that states should use the same processes for their SCHIP and Medicaid programs, but rather simply to contrast states’ monitoring of access to care for low-income children eligible for these two programs. HHS’s comments are reprinted in appendix IV. Additionally, HHS provided technical comments, which we incorporated as appropriate. State Comments Several states provided clarifying comments regarding their oversight of access to care in Medicaid and SCHIP. These comments generally pertained to additional factors affecting access to care, the relationship between monitoring and access, and the extent of monitoring in traditional FFS and distinct SCHIP programs. Two states identified factors that affect access to care within their Medicaid and SCHIP programs but are not easily controlled by the states. One state noted that the supply of physicians is severely limited in some states and in some regions of states, affecting all payers, including commercial payers as well as Medicaid and SCHIP. Another state raised the point that the extent to which children receive health care services is influenced by how well their parents or guardians understand and comply with recommended levels of health care set by providers or by the Medicaid program. We agree that provider supply and parental decision making are important determinants in children’s access to care and can be difficult factors for state programs to address. However, the type of monitoring activities addressed in this report can help to identify such factors and areas or locations where problems may be more pronounced, thus leading to more targeted solutions. Four states identified certain activities that they believed facilitated access to care, but were not addressed in the report. One state, for example, noted that its Medicaid program helped beneficiaries locate a source of medical care, and another state described an initiative to send letters to parents of beneficiaries reminding them to schedule medical appointments. Although we recognize that these activities may help promote access to care for the Medicaid and SCHIP populations, this report did not address activities that primarily facilitated access, such as providing outreach to beneficiaries or offering provider payment incentives. Instead, we focused on states’ efforts to (1) establish and Page 46 GAO-03-222 Medicaid and SCHIP Access monitor requirements for provider availability and (2) gather and analyze data on receipt of care. In this regard, one state commented that the report had a “narrow perspective” on what constitutes monitoring in managed care and cited a range of indicators that it used, including beneficiary complaints, grievance reports, state fair-hearing requests, utilization data, and immunization rates. While such sources of data and activities hold strong potential for providing information concerning access to care, this report identified certain shortcomings of some of these indicators as programwide measures of access. For example, complaint and grievance system data can yield important information about problematic providers or services, but are not reliable measures of programwide access. Four states cautioned against what they saw as a correlation made in the report between the amount of monitoring that a state does and the degree of access to care for program beneficiaries. For example, one state said that the report suggested that if monitoring is limited, access is also limited, and disagreed that this is necessarily the case. We did not intend to present such a direct correlation. However, if a state does not monitor data related to its access standards and to utilization of services, it may not know the extent to which beneficiaries encounter problems locating and obtaining services. During the course of our work, we identified instances where state data collection and monitoring revealed access problems that were then addressed to improve beneficiary access. A few states emphasized that they considered HEDIS and CAHPS important tools that had helped them monitor health plan performance or achieve improvements in quality of care for Medicaid and/or SCHIP beneficiaries. One state noted that HEDIS was important in identifying and helping to reduce gaps between commercial and Medicaid plan performance. Another state questioned whether the continuous enrollment requirements for HEDIS (12 months) and CAHPS (6 months) would in fact bias the results of any analysis of beneficiaries’ access to care because it excludes some beneficiaries. In particular, this state believed that the benefits of improvements made by health plans are not limited to individuals enrolled for the full 6- or 12-month period. We agree that HEDIS and CAHPS are important tools in monitoring and comparing performance across plans, which necessitates that the sample population be defined by a comparable enrollment period. However, we do not believe that states can assume that all beneficiaries have access to care on the basis of HEDIS and CAHPS results that exclude a significant portion of the program population from their samples. Page 47 GAO-03-222 Medicaid and SCHIP Access Two states discussed the extent to which they monitored access in Medicaid traditional FFS compared with Medicaid managed care delivery systems. One state said it analyzes data on key health outcomes for children, such as ambulatory-sensitive hospital admissions and trends in health care utilization. Both states specifically noted their efforts to comply with federally required reporting of EPSDT utilization for their FFS programs. Nevertheless, most of the states in our sample had few or no goals regarding the number of providers available to FFS beneficiaries and, with the exception of federally required EPDST reporting, few analyzed data related to access to primary care. Similar to HHS’s view, one state noted that the report did not account for the fact that distinct SCHIP programs may choose approaches to program design and monitoring that differ from Medicaid, including approaches used in monitoring states’ commercial managed care plans. For example, this state and others reported relying on state insurance department licensure of health plans as the means of monitoring provider network adequacy, rather than imposing additional SCHIP-specific requirements. We acknowledge in our report that states’ SCHIP programs may rely on different design and monitoring options than Medicaid. Overall, however, states with distinct SCHIP programs reported fewer efforts to monitor children’s access and use of services than in their Medicaid managed care programs. Several states also provided technical comments, which we incorporated as appropriate. As arranged with your offices, unless you release its contents earlier, we plan no further distribution of this report until 30 days after the issue date. At that time, we will send copies of this report to the Administrator of the Centers for Medicare & Medicaid Services and the Administrator of the Health Resources and Services Administration. We also will make copies available to others upon request. In addition, the report will be available at no charge on the GAO Web site at http://www.gao.gov. Page 48 GAO-03-222 Medicaid and SCHIP Access If you or members of your staffs have any questions regarding this report, please contact me on (202) 512-7118. Other contributors to this report are listed in appendix V. Kathryn G. Allen Director, Health Care—Medicaid and Private Health Insurance Issues Page 49 GAO-03-222 Medicaid and SCHIP Access Appendix I: Medicaid HEDIS Measures Appendix I: Medicaid HEDIS Measures Related to Service Utilization Related to Service Utilization Four of the eight general categories of the Health Plan Employer Data and Information Set (HEDIS) measures for Medicaid managed care plan performance relate directly to beneficiary service utilization. These four categories include effectiveness of care, access/availability of care, use of services, and satisfaction with the experience of care.1 Many of the measures in these categories require beneficiaries to be continuously enrolled for some period, often 12 months, in order to be assessed. Table 8 details selected HEDIS measures that pertain to service utilization for children and adolescents enrolled in Medicaid managed care programs and the length of continuous enrollment required. 1 The remaining four general categories are health plan stability, cost of care, informed health care choices, and health plan descriptive information. Page 50 GAO-03-222 Medicaid and SCHIP Access Appendix I: Medicaid HEDIS Measures Related to Service Utilization Table 8: Length of Medicaid Enrollment Required for Selected HEDIS Measures for Children’s and Adolescents’ Use of Services Length of continuous a Category Measure name enrollment Effectiveness of care Childhood immunization status 12 months Adolescent immunization status 12 months Cervical cancer screening 12 months Chlamydia screening 12 months Prenatal care in first trimester About 9 months prior to delivery Checkups after delivery About 2 months after delivery Comprehensive diabetes care 12 months Use of appropriate medications for people with asthma 24 months Follow-up after mental illness hospitalization 1 month after discharge Antidepressant medication management 12 months Advising smokers to quit 6 months Access/availability of Children’s access to primary care providers 12 monthsb care Initiation of prenatal care From 1 to 9 months prior to c delivery Annual dental visit 12 months Availability of language interpretation services None Use of services Frequency of ongoing prenatal care None Well-child visits in the first 15 months of life From 1 to 15 months of aged Well-child visits in the third, fourth, fifth and sixth year of life 12 months Adolescent well-care visits 12 months Inpatient utilization—general hospital/acute care None Ambulatory care None Inpatient utilization—nonacute care None Discharges and average length of stay—maternity care None Cesarean section rate None Vaginal birth after cesarean section rate None Births and average length of stay, newborns None Mental health utilization—inpatient discharges and average length of stay None Mental health utilization—percentage of members receiving inpatient, None day/night care, and ambulatory services Chemical dependency utilization—percentage of members receiving None inpatient, day/night care, and ambulatory services Outpatient drug utilization None Satisfaction with the experience of care Consumer Assessment of Health Plans (adults and children) 6 months Source: National Committee for Quality Assurance, HEDIS 2000: Technical Specifications (Washington, D.C.: 1999). a For measures listed with a continuous enrollment requirement, HEDIS guidelines indicate that the managed care entity must assess on a measure-by-measure basis whether the measure may be reported in the current measurement year. Partial year reporting for the measures in this table was considered acceptable or possible by the HEDIS guidelines. Page 51 GAO-03-222 Medicaid and SCHIP Access Appendix I: Medicaid HEDIS Measures Related to Service Utilization b Older age groups (7 to 11 years) require 24 months enrollment. c Measure requires continuous enrollment of at least 43 days prior to delivery but no more than 279 days. d Measure requires that child is enrolled from 31 days through 15 months of age. Page 52 GAO-03-222 Medicaid and SCHIP Access Appendix II: Managed Care Plan Withdrawals Appendix II: Managed Care Plan Withdrawals from Medicaid in Four States from Medicaid in Four States Four states we visited—Massachusetts, Ohio, Tennessee, and Texas—had varying experiences in terms of the number and impact of managed care plan withdrawals from their Medicaid managed care programs. In some cases, as in Massachusetts, the changes occurred early in the states’ implementation of their programs and the number of plans has been stable in recent years; in other cases, as in Ohio and Tennessee, the changes in participating plans continued over time and presented ongoing challenges to the states in managing their programs and ensuring appropriate access to care for their beneficiaries. The proportion of Medicaid beneficiaries affected by withdrawals of participating managed care plans ranged from about 1 percent in Texas to almost 50 percent in Tennessee. Following is a brief description of managed care plan withdrawals in each of the four states and examples of some of the measures states took to minimize disruption to beneficiaries’ care as a result of the changes. Health plan participation in Massachusetts’ Medicaid managed care Massachusetts program has slowly stabilized, with four plans participating in the program since 2000. Earlier fluctuations occurred, however, as the state shaped its program to limit the number of participating plans and as some health plans decided to consolidate or leave the market. These early changes in participating plans affected about 4 percent of the state’s Medicaid population. Massachusetts began its current Medicaid managed care program in July 1997 with nine participating health plans.1 Two of the health plans, created by hospital systems that had traditionally provided services to lower income individuals, were formed specifically for this program. Of the remaining seven plans participating in the state’s Medicaid program, many were commercially available. Within the first 2 years of the program, however, the number of participating health plans declined to five. This reduction was partially a result of the state’s decision to contract with fewer health plans and to provide each health plan with a greater volume of beneficiaries. As a result of this decision, contracts were not renewed with two health plans and approximately 42,000 beneficiaries (about 4 1 Prior to 1997, Massachusetts had a managed care program with voluntary enrollment for most Medicaid beneficiaries. As many as 13 health plans participated in the state’s Medicaid managed care program during the early 1990s. However, since enrollment was not mandatory, only a small number of Medicaid beneficiaries joined health plans. These low enrollment figures, coupled with health plan consolidations, resulted in some plans leaving the Medicaid program. Page 53 GAO-03-222 Medicaid and SCHIP Access Appendix II: Managed Care Plan Withdrawals from Medicaid in Four States percent of the state’s Medicaid population) had to select other plans. In addition, during this period several health plans merged and at least one plan left the Massachusetts health care market altogether. State officials reported that some plans lost interest in participating because of Medicaid’s administrative and reporting requirements. Additionally, commercial plans found that the Medicaid benefit package included certain services—such as behavioral health services—that the health plans did not provide their other members. This meant that health plans had to establish networks specifically for Medicaid beneficiaries; without a “critical mass” of Medicaid beneficiaries, however, health plans had difficulty remaining financially viable in the program. When participating plans withdrew from the state’s Medicaid program, state officials said that beneficiaries enrolled in the affected plans were informed that the plans would no longer be participating in the program and were provided an opportunity to choose other plans or enroll in the state’s primary care case manager program. Since the inception of its mandatory managed care program in 1996, Ohio Ohio has faced a large number of health plan withdrawals. As of January 2002, 10 plans had completely withdrawn from program participation, while 3 additional plans had withdrawn from specific counties in the state. Over 224,000 Medicaid beneficiaries—over 15 percent of the state’s Medicaid population—were affected by plan withdrawals. As a result of these withdrawals and providers’ growing reluctance to participate in managed care, Ohio changed from mandatory to voluntary managed care enrollment in some counties and fee-for-service (FFS) in others. As of April 2002, Ohio had 7 managed care plans serving Medicaid beneficiaries in 15 counties, with mandatory enrollment in only 4 of the counties. Ohio Medicaid officials expected to see some fluctuations in plan participation in the early years of its program. They anticipated that some plans would withdraw due to the state’s requirement that plans that did not have significant enrollment—from 10 to 15 percent of the eligible population—within 2 years of the program’s inception would be required to leave the program. Several reasons were provided for the number of plans that eventually withdrew from the program, including voluntary withdrawal and court-ordered liquidations. In many cases, health plans sold their Medicaid membership to other plans. State officials acknowledged that the relatively large number of plan withdrawals affected individuals’ perception of the program and led to changes in the state’s managed care enrollment policy, with some counties switching from mandatory to voluntary managed care enrollment. Concerns about Page 54 GAO-03-222 Medicaid and SCHIP Access Appendix II: Managed Care Plan Withdrawals from Medicaid in Four States the program’s viability and stability were increased when the state insurance department liquidated one Medicaid health plan in 1998 and some of its network providers did not receive compensation from the plan. State officials did not believe that beneficiaries’ access to care was affected by these plan withdrawals. In cases where a health plan’s membership was sold to another plan, the state attempted to ensure continuity of care by requiring that at least 90 percent of the current plan’s primary care providers (PCP) were included in the provider network of the purchasing plan.2 In other cases, we were told, beneficiaries were notified of their health plan’s impending withdrawal and provided an opportunity to select another plan if available. If a beneficiary did not select a health plan, or there was no alternative plan available, then the beneficiary returned to the state’s FFS program. In areas with mandatory managed care enrollment, however, beneficiaries were not allowed to remain in FFS indefinitely; they were required to select another plan or be automatically assigned to one. In establishing its mandatory managed care program in January 1994, Tennessee Tennessee expanded Medicaid eligibility to hundreds of thousands of previously uninsured individuals and enrolled them into 1 of 12 capitated managed care plans. Four plans left the program or were sold from 1994 through 1999.3 Since 2001, plan withdrawals have increasingly been an area of concern, with large numbers of Medicaid beneficiaries affected by changes to the state’s 2 largest health plans. For example, in 2001, almost 580,000 beneficiaries, or 41 percent of the state’s Medicaid population, were affected when 1 plan withdrew from the western and central portions of the state and a second plan’s contract was terminated due to solvency issues. In response to the first of these two withdrawals, the state took two actions: (1) it recruited two new health plans to join the market and (2) it created a self-insured plan to serve as a backup in areas of the state where beneficiaries could not be adequately served by other health plans. 2 Of the beneficiaries affected by plan withdrawals in Ohio from 1996 to January 2002, nearly half were involved in withdrawals that were the result of a plan selling its membership to another plan. 3 Plan withdrawals during this period affected approximately 105,000 beneficiaries, about 7 percent of the state’s Medicaid population. Page 55 GAO-03-222 Medicaid and SCHIP Access Appendix II: Managed Care Plan Withdrawals from Medicaid in Four States As of April 2002, 10 health plans were participating in Tennessee’s Medicaid managed care program although 2 plans, covering 21 percent of the state’s Medicaid beneficiaries, were considered to be at financial risk. The state announced its intention in March 2002 to terminate its contract with 1 health plan, which would necessitate the transfer of approximately 135,000 beneficiaries to other health plans.4 A second plan, with over 160,000 beneficiaries, was under rehabilitation by the state’s insurance department. In view of the instability of the program and participating plans, Tennessee has taken several steps to help ensure continuous access to care for Medicaid beneficiaries. In order to provide time to plan ahead in the event of plan withdrawals, the state’s contract with participating plans requires 6 months of advance notice of an intended withdrawal and a transition plan to assure uninterrupted care to beneficiaries. When plans stopped participating in the program, beneficiaries were either provided the option to select new health plans or were assigned to health plans. Texas began its capitated Medicaid managed care program in1996 in four Texas areas of the state. Since 1996, managed care was expanded to three additional service areas and now exists in 46 of the state’s 254 counties. Since the rollout of managed care began, only three plans have withdrawn from participation in Texas’ Medicaid managed care program, affecting less than 20,000 beneficiaries, or approximately 1 percent of the state’s Medicaid population.5 Two of these withdrawals were from the same service delivery area, leaving three plans participating in that area.6 However, the state contends that prior to the withdrawals there was a saturation of health plans in that service delivery area. As of July 2002, 11 plans were participating in the Medicaid managed care program. 4 According to the state, the decision to terminate the contract was based on problems including the plan’s financial solvency and failure to pay accurate and timely claims. As of May 1, 2002, the state was working with the health plan in an attempt to resolve these problems. 5 Over 50,000 additional beneficiaries were affected when their health plan was acquired by another participating health plan. 6 There are seven service areas, each consisting of multiple counties, in Texas’ Medicaid capitated managed care program. Health plans are contracted by service area, with some health plans having contracts in multiple service areas. Page 56 GAO-03-222 Medicaid and SCHIP Access Appendix II: Managed Care Plan Withdrawals from Medicaid in Four States In one instance, a participating plan gave the state less than 3 weeks’ notice of its intent to leave the program. Because of the limited notice, beneficiaries were automatically assigned to other plans in order to minimize disruption in their access to care. Although these assignments were initially made without direct input from the affected beneficiaries, their prior PCP and specialist utilization patterns were taken into account during this assignment process and beneficiaries were later given an opportunity to change plans. The state paid particular attention to the number of complaints during these transition periods and did not see a dramatic change. As such, state officials believe that the transitions went smoothly. Texas has a number of other measures in place to facilitate beneficiaries’ enrollment in alternative plans when their plans leave the program, as illustrated in table 9 along with additional examples from other states we visited. Table 9: Examples of Plan Withdrawal Transition Activities Conducted by Four State Medicaid Programs Type of action Examples of action Contractual requirements of managed • Massachusetts takes responsibility for notifying beneficiaries of the health plan’s care plans withdrawal from the program and the process beneficiaries must undergo to continue to receive services; however, health plans must continue to provide services until the beneficiary is disenrolled and participating in another plan. • Ohio’s contract requires the collection of monetary assurance or the withholding of payments from withdrawing health plans until all contract requirements are completed. • Texas requires health plans to provide the state 90 days notice of their intention to terminate participation, Ohio requires 75 days notice, and Tennessee requires 6 months advanced notice. • Tennessee’s contract requires withdrawing health plans to submit transition plans to ensure uninterrupted care to beneficiaries. Notification of beneficiaries and other • Ohio, Tennessee and Texas send letters to beneficiaries informing them of the health stakeholders plan’s withdrawal. The letters may include a list of the other health plans, important telephone numbers, and actions beneficiaries must take. • Texas notifies stakeholders, including the enrollment broker and other health plans, of the impending withdrawal. Remaining health plans in the area are provided with a list of PCPs that are only participating with the exiting plan. Additionally, the state or health plan notifies providers of the plan’s intention to withdraw from the program. Coordination between plans • In Texas, the withdrawing plan identifies individuals with special needs and a dialogue between the current and future case managers begins. In addition, the withdrawing health plan provides instructions for providers on seeking authorization for continued services from new health plan. Source: GAO analysis of states’ data, December 2001. Page 57 GAO-03-222 Medicaid and SCHIP Access Appendix III: Analysis of Medicaid FFS Appendix III: Analysis of Medicaid FFS Payment Rates in Selected States Payment Rates in Selected States Nationally, low Medicaid physician fees have been a long-standing area of concern because they can affect the degree to which physicians participate in Medicaid, and thereby affect beneficiaries’ access to care. The relative fees paid by different insurers—Medicare, Medicaid, and SCHIP—can also affect providers’ willingness to participate in these programs. Since many children in Medicaid remain in fee-for-service (FFS)-based programs, we compared Medicaid fees for selected office visit and pediatric preventive medical care services to the corresponding Medicare fees. While Medicare is a federal health insurance program primarily for the elderly and persons with disabilities, some children do receive Medicare benefits and thus its fee schedule includes fees for pediatric medical services. Among the 13 states we reviewed that used FFS-based delivery systems as a key care delivery system for Medicaid children,1 Medicaid fees for primary and preventive care ranged from 32 percent to 89 percent of what Medicare would pay for similar services. (See table 10.) Concerns with the adequacy of Medicaid physician payment levels were also identified in studies of Medicaid physician payment in California, Washington, and Maryland.2 1 Tennessee enrolls nearly all beneficiaries in managed care; therefore, we did not collect a Medicaid FFS payment schedule that can be compared to Medicare rates. 2 See PriceWaterhouseCoopers, Comparing CPT Code Payments for Medi-Cal and Other California Payers (Oakland, Calif.: June 2001) and University of Washington, State Primary Care Provider Study, Health Policy Analysis Program (Seattle, Wash.: February 2001). A study was also conducted in Maryland because, even though most beneficiaries are served through managed care, the state Medicaid program’s FFS payment rates for some groups of beneficiaries are considered to affect what managed care plans pay physicians. See State of Maryland Department of Health and Mental Hygiene, Report on the Maryland Medical Assistance Program and Maryland Children’s Health Program – Reimbursement Rates Fairness Act (Baltimore, Md.: September 2001). Page 58 GAO-03-222 Medicaid and SCHIP Access Appendix III: Analysis of Medicaid FFS Payment Rates in Selected States Table 10: Medicaid FFS Payment Rates, Expressed as a Percentage of Medicare Payments, in 13 States with Traditional FFS or Primary Care Case Manager Delivery Systems That Serve Children Medicaid FFS payments as a percentage Statea of Medicare FFS payments (weighted) b Massachusetts 89 Arkansas 71 Florida 71 Texas 71 Nevada 66 Ohio 64 b Illinois 61 Washington 60 Colorado 57 Louisiana 57 New Yorkb 54 California 48 Pennsylvania 32 Source: GAO analysis of Medicare data and states’ data, as of December 2001. a Other study states not shown on this table include the following: Tennessee enrolls nearly all beneficiaries in capitated managed care, and therefore, we did not collect a Medicaid FFS payment schedule that can be compared to Medicare rates. Maryland uses a FFS-based delivery system for less than 5 percent of children and includes only those children requiring case management for rare and expensive conditions, or who are technology dependent. Michigan uses FFS-based care only for children in an eligibility category for special needs. b Illinois, Massachusetts, and New York provide payment enhancements for some services, in addition to the regular fee for the service; where appropriate, these enhancements were included in the analysis. For our comparative analysis of Medicaid and Medicare FFS payments, we Methodology for obtained fee schedules from 13 of the 16 states we reviewed, compiling Comparison of FFS fees for 12 medical services using selected codes from a commonly used procedural coding system—the standard Physicians Current Procedural Payment Rates Terminology, 4th edition (CPT 4). (See table 11.) For each state, we weighted the Medicaid and corresponding lowest Medicare fees3 for that state by the relative utilization of the service among pediatricians, identified from a 1999 American Academy of Pediatrics survey.4 The sum 3 States can have more than one Medicare payment rate for a service, varying by locality. 4 Monique Morris and Suk-fong Tang, Pediatric Service Utilization, Fees and Managed Care Arrangements: 2001 Report Based on 1999 Data (Elk Grove Village, Ill.: American Academy of Pediatrics, 2001). Page 59 GAO-03-222 Medicaid and SCHIP Access Appendix III: Analysis of Medicaid FFS Payment Rates in Selected States of the weighted Medicaid fees was then expressed as a percentage of the sum of the Medicare payments in order to develop a single, weighted payment rate. Table 11: CPT 4 Codes Used in Comparing Medicaid and Medicare Fees CPT 4 code Description Office or other outpatient visit 99201 New patient, 10 minute visit 99202 New patient, 20 minute visit 99203 New patient, 30 minute visit 99213 Established patient, 15 minute visit 99214 Established patient, 25 minute visit Preventive medical services 99381 New patient, under 1 year 99382 New patient, 1 to 4 years 99383 New patient, 5 to 11 years 99391 Established patient, under 1 year 99392 Established patient, 1 to 4 years 99393 Established patient, 5 to 11 years 99394 Established patient, 12 to 17 years Source: CPT 4. Page 60 GAO-03-222 Medicaid and SCHIP Access Appendix IV: Comments from the Department Appendix IV: Comments from the of Health and Human Services Department of Health and Human Services Page 61 GAO-03-222 Medicaid and SCHIP Access Appendix IV: Comments from the Department of Health and Human Services Page 62 GAO-03-222 Medicaid and SCHIP Access Appendix IV: Comments from the Department of Health and Human Services Page 63 GAO-03-222 Medicaid and SCHIP Access Appendix IV: Comments from the Department of Health and Human Services Page 64 GAO-03-222 Medicaid and SCHIP Access Appendix V: GAO Contact and Staff Appendix V: GAO Contact and Staff Acknowledgments Acknowledgments Carolyn L. Yocom, (202) 512-4931 GAO Contact Catina Bradley, Karen Doran, Laura Sutton Elsberg, Mary Giffin, Michelle Acknowledgments Rosenberg, and Ann Tynan made key contributions to this report. Page 65 GAO-03-222 Medicaid and SCHIP Access Related GAO Products Related GAO Products Mental Health Services: Effectiveness of Insurance Coverage and Federal Programs for Children Who Have Experienced Trauma Largely Unknown. GAO-02-813. Washington, D.C.: August 22, 2002. Children’s Health Insurance: Inspector General Reviews Should Be Expanded to Further Inform the Congress. GAO-02-512. Washington, D.C.: March 29, 2002. Medicaid and SCHIP: States’ Enrollment and Payment Policies Can Affect Children’s Access to Care. GAO-01-883. Washington, D.C.: September 10, 2001. Medicaid: Stronger Efforts Needed to Ensure Children’s Access to Health Screening Services. GAO-01-749. Washington, D.C.: July 13, 2001. Medicaid Managed Care: States’ Safeguards for Children With Special Needs Vary Significantly. GAO/HEHS-00-169. Washington, D.C.: September 29, 2000. Oral Health: Factors Contributing to Low Use of Dental Services by Low- Income Populations. GAO/HEHS-00-149. Washington, D.C.: September 11, 2000. Medicaid and SCHIP: Comparisons of Outreach, Enrollment Practices, and Benefits. GAO/HEHS-00-86. Washington, D.C.: April 14, 2000. Oral Health: Dental Disease Is a Chronic Problem Among Low-Income Populations. GAO/HEHS-00-72. Washington, D.C.: April 12, 2000. Children’s Health Insurance Program: State Implementation Approaches Are Evolving. GAO/HEHS-99-65. Washington, D.C.: May 14, 1999. Medicaid Managed Care: Challenge of Holding Plans Accountable Requires Greater State Effort. GAO/HEHS-97-86. Washington, D.C.: May 16, 1997. (290046) Page 66 GAO-03-222 Medicaid and SCHIP Access The General Accounting Office, the investigative arm of Congress, exists to GAO’s Mission support Congress in meeting its constitutional responsibilities and to help improve the performance and accountability of the federal government for the American people. GAO examines the use of public funds; evaluates federal programs and policies; and provides analyses, recommendations, and other assistance to help Congress make informed oversight, policy, and funding decisions. GAO’s commitment to good government is reflected in its core values of accountability, integrity, and reliability. The fastest and easiest way to obtain copies of GAO documents at no cost is Obtaining Copies of through the Internet. GAO’s Web site (www.gao.gov) contains abstracts and full- GAO Reports and text files of current reports and testimony and an expanding archive of older products. The Web site features a search engine to help you locate documents Testimony using key words and phrases. You can print these documents in their entirety, including charts and other graphics. Each day, GAO issues a list of newly released reports, testimony, and correspondence. GAO posts this list, known as “Today’s Reports,” on its Web site daily. The list contains links to the full-text document files. To have GAO e-mail this list to you every afternoon, go to www.gao.gov and select “Subscribe to daily E-mail alert for newly released products” under the GAO Reports heading. Order by Mail or Phone The first copy of each printed report is free. Additional copies are $2 each. A check or money order should be made out to the Superintendent of Documents. GAO also accepts VISA and Mastercard. Orders for 100 or more copies mailed to a single address are discounted 25 percent. Orders should be sent to: U.S. General Accounting Office 441 G Street NW, Room LM Washington, D.C. 20548 To order by Phone: Voice: (202) 512-6000 TDD: (202) 512-2537 Fax: (202) 512-6061 Contact: To Report Fraud, Web site: www.gao.gov/fraudnet/fraudnet.htm Waste, and Abuse in E-mail: firstname.lastname@example.org Federal Programs Automated answering system: (800) 424-5454 or (202) 512-7470 Jeff Nelligan, managing director, NelliganJ@gao.gov (202) 512-4800 Public Affairs U.S. General Accounting Office, 441 G Street NW, Room 7149 Washington, D.C. 20548
Medicaid and SCHIP: States Use Varying Approaches to Monitor Children's Access to Care
Published by the Government Accountability Office on 2003-01-14.
Below is a raw (and likely hideous) rendition of the original report. (PDF)