Unsettling Scores

A Ranking of State Medicaid Programs

April 2007

Annette B. Ramírez de Arellano, Dr.PH.
Sidney M. Wolfe, M.D.
Public Citizen Health Research Group

This version of the report appears without graphs, figures and tables. For a printable full report, download the full report (PDF).
Acknowledgments
Executive Summary

I. Introduction
II. Methods
III. Eligibility
IV. Scope of Services
V. Quality of Care
VI. Provider Reimbursement
VII. National Results
VIII. State Reports
IX. Conclusions

Appendix (PDF)
Glossary
Bibliography (PDF) 

Acknowledgments

The data on which this report is based comes primarily from the Kaiser Family Foundation online database, which is maintained and updated as part of the Foundation’s Commission on Medicaid and the Uninsured. This source has proved invaluable in our analyses, and is essential to all who are interested in whom and what Medicaid covers.

Several persons served as sounding boards and graciously shared their understanding of Medicaid with us. We would like to thank the following for their wisdom and generosity in sharing their knowledge: Mary Gabay of Gabay Consulting LLC; Andy Schneider, Principal, Medicaid Policy LLC; Bruce C. Vladeck, Ph.D., Interim President, University of Medicine and Dentistry of New Jersey; and Doris Lotz, M.D., M.P.H.. All conclusions and interpretations, however, are those of the Health Research Group.

The staff of Public Citizen, both within and outside the Health Research Group, handled much of the coding, weight-adjusting, and computation. Asa Tapley helped devise the scoring protocol, and Sylvia Park, M.D., M.P.H., reviewed it. Without the statistical and technological know-how and editorial and communication skills of Peter Lurie, M.D., M.P.H., Elizabeth Barbehenn, Ph.D., Shiloh Stark, Kate Resnevic, Ed Uechi, and Gleb Radutsky, the data entry and computational processes would have been infinitely more cumbersome and time-consuming. These colleagues were also essential in ensuring that the data could be easily accessed online, and provided much insight on how best to use and display the data. Barbara Holzer, Rachel Pleatman, and Robert Yule provided editorial judgment and sound copyediting. Kristy Jackson designed the cover, and Candice Hayes was instrumental in monitoring the printing process. All deserve our most sincere thanks.

Executive Summary

Enacted over 40 years ago, Medicaid has evolved with changing demographic, technological, and health needs; political priorities; and fiscal realities. The program has been called many things: the cornerstone of the nation’s health system, a safety net for the neediest, a workhorse that goes into action when help is needed, and an inflexible and flawed system, among others.

Whatever its shortcomings, Medicaid continues to grow, reflecting and refracting many of the trends affecting health care as a whole. At present, the program

• covers approximately 55 million Americans;
• accounts for one out of every five health care dollars spent in the nation;
• pays for 41 percent of all U.S. births and 50 percent of all long-term care costs;
• supports tens of thousands of health care providers throughout the country; and
• represents the largest source of federal grant support to states.   

Purpose and scope of study

While there are abundant data on Medicaid, these tend to avoid making value judgments. This report therefore seeks to fill the existing gap. We feel that it is not enough to say “this is the way things are;” instead, we should assess and say “this is the way things should be.”

Almost 20 years ago, the Public Citizen Health Research Group published a report on Medicaid, Poor Medicine for Poor People, ranking state Medicaid programs. The current report seeks to update that report. But because programmatic mandates have changed and states now have considerably more latitude in how they run their programs, the indicators are different, as are the sources of data. As a result, there is greater variety among states, as well as greater differences within states.

Each state program has been evaluated in terms of four categories: eligibility, scope of services, quality of care, and reimbursement. These were in turn measured by 55 indicators, and the resulting scores were weighted according to the relative value given to each category by experts. The ranking system gives a state a score for each category as well as an overall score.

Major findings

Nationally, the state Medicaid programs are severely challenged: even the best state scores only 645.9 points on a scale of 1000. And the worst state rates a score of only 317.8, i.e., less than a third of the total maximum points.

The state-by-state breakdowns reveal marked disparities between and among states. The top 10 states, ranking #1 to #10, tend to cluster in the Northeast but also include three states in the Midwest and two in the Northwest. The following states occupy the first 10 ranks, in descending order: Massachusetts, Nebraska, Vermont, Alaska, Wisconsin, Rhode Island, Minnesota, New York, Washington, and New Hampshire.

The 10 most deficient state programs have overall scores ranging from between 317.8 and 379.1 of the total 1000 points. The worst, in order from 50th to 41st, are in Mississippi, Idaho, Texas, Oklahoma, South Dakota, Indiana, South Carolina, Colorado, Alabama, and Missouri. 

The overall score of top-ranked Massachusetts is more than twice that of bottom-ranked Mississippi. A breakdown of scores by category further highlights the existing disparities: the scores vary 2.5-fold for scope of services, and more than threefold for eligibility. In the remaining two categories, which have fewer indicators and are therefore more volatile, variations among states are even more dramatic: in quality of care, the difference is more than 17-fold; in reimbursement, it is more than 20-fold.

The overall ranks are followed by state-specific summaries with the breakdown of scores by category. This allows states to pinpoint their areas of weakness, and to more successfully target their interventions. It also highlights states that have achieved success in one or more areas and can therefore serve as models for other jurisdictions.

I. Introduction

Medicaid has been called many things:

• the cornerstone of the nation’s diverse and complex system of financing health and long-term services;[1]
• a safety net for those who need assistance with health care;[2]
• a workhorse, expected to aid populations when no other source will help;[3]
• an amalgamation of responses to different problems over 40 years;[4]
• a surprisingly flexible program;[5] and
• a rigidly inflexible and inefficient system that is not financially sustainable.[6]

Whatever the prevailing contradictory views of the system and its operation, there is consensus that the program has an “overwhelming level of diversity and complexity.”[7] Lofty in its goals but often miserly in its actual impact on people, Medicaid mirrors changing economic circumstances, conflicting political pressures, and fluctuating demographic and medical needs. A complicated partnership between states and the federal government has yielded more than 50 different programs, each with its own distinctive features and idiosyncrasies. Indeed, state variation in eligibility, covered care, program administration, and reimbursement for services is now the rule rather than the exception.[8] These allowable state-by-state variations are a major weakness in Medicaid.

Twenty years ago, Public Citizen Health Research Group conducted a comprehensive state-by-state assessment of the Medicaid Program. That report ranked all states on the basis of five criteria: eligibility, services, provider availability, quality, and reimbursement, each of which was measured through an array of operational indicators. Public Citizen’s 1987 report was used by states to examine their status vis-à-vis other states and the nation as a whole. The report prompted states to confront their deficiencies and improve their programs. It also provided leverage to those states that were getting less than their fair share in federal funds or had not fared well in the monies allocated by their state legislatures. In addition, the report underscored the disparities that exist among states, thereby revealing the significant differences in access to care that Americans face simply because of where they happen to live.

The current report seeks to update, though not replicate, the previous one. An update is particularly timely and necessary because many changes have taken place over the course of two decades. Much federal legislation has either focused directly on Medicaid or enacted changes in other health and welfare services that have had important implications for Medicaid. These changes have affected each of the five criteria that we focused on in the previous report, as well as some of the indicators that were used to measure them. Furthermore, Medicaid has come of age. As the program enters its fifth decade, it is going through a programmatic “midlife crisis.” It is not surprising that the program’s advocates as well as its critics are taking stock of where the program is at present in order to point out areas where changes are needed. Although much of the concern revolves around program costs, we feel that this focus fails to address more fundamental aspects of the program, including equity in access to care and the quality of services rendered. 

In 2005, U.S. Department of Health and Human Services (DHHS) Secretary Michael Leavitt declared that the program was no longer meeting its potential. He named a bipartisan commission to plan for an improved Medicaid that would “provide quality health care in a financially sustainable way.”[9] The commission was charged with preparing two reports. The first, which was submitted September 2005, outlined recommendations for Medicaid to achieve $10 billion in savings over the next five years. The second, submitted December 29, 2006, sought to address the following issues:

• How to expand coverage to more Americans while being fiscally responsible;
• Ways to provide long-term care to those who need it;
• A review of eligibility, benefits design, and delivery; and
• Improved quality of care, choice, and beneficiary satisfaction.

Scope and purpose of the current report

This report neither substitutes nor supplements the reports prepared by the federally-mandated commission. Its principal audience includes policymakers as well as advocates and individual consumers. Nevertheless, it does not consider issues related to political appeal or cost-effectiveness as much as it addresses the scope of the program, its access to those in need, and its monitoring of the services delivered. Like its predecessor, this report tries to answer the question: “If I were a poor, sick person, in which state would I have the best chance of becoming eligible for Medicaid and getting comprehensive, quality health care?” The evaluation criteria and the indicators used to measure them therefore reflect the consumer’s perspective and aim to answer the following questions:

1. Am I eligible to receive Medicaid services in this state? Given the variety of pathways to determine eligibility, this is not an easy question to answer. An extensive list of requirements may enter the decision. Criteria for eligibility include the following: age, income, citizenship status, assets, work status, marital status, enrollment in school, medical condition, and improvement potential, among others. The permutations and combinations of eligibility requirements make for a complicated patchwork of enrollees. As a result, a “protected” population in one state may very well be entirely expendable in another. 
2. If I am eligible, does this state cover the particular services I need? While there are certain services that all states cover, there are more than 30 optional services that may be included in a state’s offerings. And some of these may be covered only when given by certain providers; have limitations in terms of populations covered, frequency, duration, and scope; require cost-sharing; or may be provided only under certain conditions.
3. If the services are covered, will they be of adequate quality? Although Medicaid programs generate an inordinate number of statistics, and information technology has greatly facilitated the collection and analysis of data, most states lack reliable, measurable criteria to assess the quality of care they provide, or even to establish profiles of who is getting what care, when, and at what cost. Instead, the focus has been on billing and fraud detection. As a result, data on quality are very spotty. Unlike Medicare data, which are collected centrally and processed nationally, information on Medicaid depends on the capabilities and priorities of each state. As a result, much of the information may not be comparable across jurisdictions, and state initiatives cannot be properly evaluated.
4. Will the state pay for my services in a way that encourages access, equity, and quality? States have been experimenting with ways to expand their coverage while keeping costs down and increasing efficiency. In some cases, these objectives are at cross-purposes, and involve trade-offs that are not always explicit to the consumer. As more and more states have been granted waivers from the originally mandated services, they have been given greater leeway in coverage, reimbursement policies, and organization of services. An increasing proportion of Medicaid beneficiaries is now enrolled in managed care. And more states are experimenting with cost-sharing, ostensibly as a way of making consumers more “prudent purchasers” of health care. 

There are relatively few indicators for quality of care and reimbursement, but we have used those that reflect both a commitment to patient care and an interest in treating Medicaid providers equitably vis-à-vis practitioners who serve other populations.

Organization of the report

The report is organized by topic as well as by state. Following a chapter on Methods (II), Chapters III through VI focus on one of the four questions listed above, which correspond to the four categories we examined: eligibility, scope of services, quality of care, and provider reimbursement. In each case we present the state scores and rankings for the specific category, thus allowing comparisons among and between states. These are followed by a chapter on national results (VII), a summary of both the overall scores and the category-specific ranks. Chapter VIII presents state-by-state data, highlighting each state’s scores and ranks. Chapter IX summarizes our main conclusions. 

II. Methods

Our study assesses four aspects of the Medicaid program: eligibility, scope of services, quality of care, and reimbursement. Each of these categories is measured through different indicators. The choice of indicators is understandably contingent upon the availability of data. We therefore relied on data that are routinely collected and published, broken down by state. For each of the categories studied, we took what are basically qualitative characteristics and transformed them into quantitative values. This allowed us to give each state specific points for each indicator, depending on its performance for those indicators. These scores allow us to rank all states except Tennessee[10], both by category and as a whole.  States can then be compared to each other, as well as compared against the maximum possible number of points for each category. 

Basic guidelines

Certain principles underlie the selection of indicators, their interpretation, and the points assigned to them. While the Scoring Protocol included in the Appendix describes each indicator and the unadjusted points assigned to it, there were general principles governing the assignment of points, and these cut across categories and indicators. Our scoring methodology is based on the following guidelines:

Scoring process

The scores on which our rankings are based began with the four categories assessed — eligibility, scope of services, quality of care, and reimbursement. Each of these categories was broken down into a number of indicators, which ranged from three to 36 per category. In some cases, the indicators were composites and were further broken down, as indicated in the Scoring Protocol in the Appendix.

The indicators were evaluated by an expert on Medicaid, who suggested adding, eliminating, or combining certain indicators. Each indicator was given a maximum number of points, ranging between one and 11. These points were considered “raw scores” which were then adjusted to reflect their relative value.

To determine the relative weight of each category and indicator, two other recognized experts in the field of Medicaid were consulted. They were asked to distribute 100 points among the four categories. The mean of the points assigned to each category was then computed, and divided by 100 to determine the relative weights. The relative weights for the four categories are as follows:

Eligibility, .35
Scope of services, .20
Quality of care, .20
Reimbursement, .25

The same experts were asked to further distribute 100 points among the indicators in each of the categories, and these values were also averaged, then divided by 100. A listing of all category and indicator weights can be found in the Appendix (PDF).

The final score for each indicator for each state was therefore the fraction of the total maximum points obtained by the state for each given indicator, multiplied by both the category weight and the weight assigned to each individual indicator. Because the resulting numbers were very small, these products were then multiplied by 1000. As a result, all scores are based on a theoretical total of 1000 points overall.

The scores for each indicator within a category were then added, the sum being the score for that particular category. The sum of the scores for all four categories constitutes the overall score for each state.

Because the data were entered into a spreadsheet, the computation of the adjusted scores allowed for aggregations of indicators within a given category. The final scores were then sorted by order of magnitude, thereby allowing rankings by category and overall.

In a few cases of three categories (scope of services, quality of care, reimbursement) and overall, two or more states had the same score and therefore shared a rank. In these cases, the subsequent ranks were adjusted so that there were as many ranks as programs scored. For example, two states tied for 5th place received the same rank, 5th, but the rank following that was 7th rather than 6th. 

The many indicators used and the weighting of the raw scores yielded final scores that were rounded off to one decimal point. This is not intended to overstate the degree of precision, but rather to allow us to draw distinctions among state programs. We have therefore emphasized how states rank generally with respect to each other instead of focusing on any differences that may distinguish, say, a state ranking #22 from one ranking #23 overall as well as in any one category.

The detailed Scoring Protocol (including the definition, rationale, and source of each indicator and its components) is described in the Appendix (PDF). 

Limitations

The Medicaid Program has been called “resilient” because it is constantly re-inventing itself to meet new circumstances. One scholar has indicated that Medicaid’s infrastructure provides “a base from which almost any health matter can be addressed.”[14] Despite its deficiencies, the program has been responsive to new technologies (e.g., drugs and devices), emerging disease entities (e.g., HIV/AIDS), natural events (Hurricane Katrina), ideological trends (assumption of personal responsibility, lifestyle modifications), and modalities of care (managed care). To take a snapshot in time of the program is therefore to miss some of the changes and adaptations that are occurring continually. This is particularly the case at present, when many states are looking to modify their programs or significantly alter their involvement in health care delivery.

Several states (including California, Massachusetts, New Jersey, New York, and Vermont) are seeking to reduce or even eliminate the number of uninsured within their jurisdictions. Others (including West Virginia, Florida, and Kentucky) are incorporating incentives for health-seeking behaviors among their enrollees. Yet others are modifying the way care is given and the incentives for participation in the program. Because of the fact that we were dealing with a moving target, we had to establish a cut-off point beyond which no new data would be incorporated. This was January 15, 2006. Any time-limited data will therefore not capture the current fluidity in the state policy arena.

A second limitation of our study refers to the need for comparability between and among states. The efforts of those states that have attempted to “break the mold” in covering their Medicaid-eligible population may therefore not be fully represented in our scoring protocols. This limitation is also tied into a significant constraint: the availability of data. Large-scale data collection efforts require consistency in definitions. As a result, some of the variations that occur from state to state may not be fully evident in tables that adopt a uniform system of coding and description.[15]

Because our rankings entail an analysis of secondary sources, they rely on data collected for different purposes and at varying points in time. We have relied primarily on the online data on Medicaid made available through the Kaiser Family Foundation’s Commission on Medicaid and the Uninsured. In some cases, these data have been supplemented with other sources, as indicated in the Scoring Protocol. Moreover, when the data for a particular state were not available on the Kaiser database, we consulted directly with the state, thus filling in gaps to complete the national picture.

The data on provider reimbursement are limited by the fact that they reflect only those Medicaid payments that are made under fee-for-service systems. While all states other than Tennessee have a fee-for-service component, this is limited in states that have adopted capitated systems under managed care. At present, more than 60 percent of all Medicaid enrollees in the United States are enrolled in managed care.[16] When the data are broken down by state, the proportion varies between zero and 100 percent.[17] Our indicators on provider reimbursement thus have greater validity and reliability for those states that rely more on fee-for-service and have a lower proportion of users enrolled in managed care. Because the indicators do not apply to Tennessee, we have not computed a score for this state under the reimbursement category. This omission precludes the state being ranked with the other states in this category, and therefore in the overall category as well. As a result, the overall ranks and those in reimbursement range between one and 50 rather than between one and 51.   

[see full text PDF for tables, figures, and appendix]

III. Eligibility

Consumers who want to know if they are eligible for Medicaid have to work their way through a complicated list of demographic categories to see if they qualify under any of these. Indeed, federal law describes more than 50 eligibility pathways.[18] Many demographic variables — e.g., age, sex, marital status, family composition, income, disability, and disease — help define current eligibility criteria. 

In keeping with the program’s labyrinthine complexity, there are some populations that must be covered by all states, and there are others who may be covered as well. To add another layer of confusion, not everyone is covered under the same circumstances, nor for the same services.

While Medicaid is more frequently known as a “program for the poor,” and the program has always targeted low-income individuals, not all the poor are eligible and not all the eligible are poor.[19] To be covered, the poor must meet financial requirements (regarding income, assets, and expenses) as well as categorical requirements (regarding age, family circumstances, employment status, blindness, disability, and other factors). These requirements exclude many people from Medicaid. Indeed, it is estimated that approximately 60 percent of poor Americans are not covered by Medicaid.[20]

The Medicaid program varies greatly from state to state. Eligibility rules vary from one state to another, although there are guidelines that govern local options. While federal regulations require all states to cover certain groups and limit the additional groups that states may cover, each state can elect to include other groups falling somewhere between the federal “floor” and “ceiling.” As a result of these differences among states, the same person may be eligible in one state but ineligible in another. Moreover, many states have taken advantage of Medicaid “waivers” which exempt them from eligibility and coverage requirements as long as they are budget-neutral and do not cost the federal government more than prior coverage.

Mandatory groups

When first enacted, Medicaid was linked to beneficiaries of the federally-assisted income maintenance program Aid to Families with Dependent Children (AFDC). After 1972, the program also included those covered by Supplemental Security Income (SSI), a program which provides cash assistance to help aged, blind, and disabled people who have little or no income. Since then, changes in Medicaid and SSI have created additional groups of beneficiaries whose financial eligibility is based solely on income and resources rather than on cash assistance.[21] The inclusion of these “poverty-related” groups expanded Medicaid to include pregnant women and children by separating Medicaid eligibility from receipt of AFDC. At present, these groups represent a growing proportion of Medicaid beneficiaries, accounting for over one-quarter of the total.[22]

As a result of these and other changes, all states must provide Medicaid coverage to the following eligibility groups:

AFDC-eligible individuals as of July 16, 1996: States must provide Medicaid to individuals who qualified for AFDC as of that date.

Poverty-related groups: States must cover all pregnant women and children below age 6 with incomes up to 133 percent of the federal poverty level (FPL).

All children born after September 30, 1983 with incomes up to 100 percent FPL. This requirement covers poor children under the age of 19.

Current and some former recipients of SSI: Despite this requirement, states may use more restrictive eligibility standards for Medicaid than those used for SSI if states were using those standards prior to the enactment of SSI in 1972.

Foster care and adoption assistance: States must cover all recipients of foster care and adoption assistance under Title IV-E of the Social Security Act.

Certain Medicare beneficiaries: All Medicare beneficiaries with incomes below the poverty level are eligible for Medicaid assistance to pay for Medicare premiums, deductibles and cost-sharing. In addition, individuals at the lowest income levels are entitled to full Medicaid benefits, which provide “Medigap” services (i.e., services not covered by Medicare). The latter individuals are most often referred to as “dual eligibles.”

The determination of Medicaid eligibility is two-tiered: First, individuals must fall within one of the previously listed “categorical” groups. Once the individual is found to meet the categorical restrictions, financial tests are applied.[23] States have some latitude concerning the latter, which further adds to the administrative complexity of the program and to the inter-state variation. While states have limited flexibility to modify income standards, they have greater discretion concerning how “countable income” is defined. By excluding certain types of income from their definitions of “countable income,” states can liberalize their eligibility criteria without violating income standards. The following are among the “income disregards” that can be excluded from countable income: a certain portion of earned income during the first few months of employment, a given dollar amount as a child care allowance, and a set amount for married couples.[24]

Optional groups

States can provide Medicaid coverage to other groups. These optional groups fall within the mandated defined categories, but the financial eligibility standards are more liberally defined. Optional eligibility groups include:

Poverty-related groups: States may choose to cover pregnant women and infants with family incomes up to 185 percent of the FPL.

Medically needy: States may choose to cover individuals who do not meet the financial standards for program benefits but fit into one of the mandated groups and have income and resources within special “medically needy” limits established by the state. Individuals whose resources are above the “medically needy” standards may qualify by “spending down” — i.e., incurring medical bills that reduce their income and/or resources to the necessary level.

Recipients of state supplementary payments: States may opt to include individuals who do not receive SSI but qualify for other state cash payments.

Long-term care: States may cover residents in medical institutions or those receiving certain long-term care services in community settings if their incomes are less than 300 percent of the SSI payment level.

Working disabled: States have the option of covering those who are disabled (as defined in Social Security Administration guidelines) but who do not qualify for Medicaid under any statutory provision due to their income. States opting to cover this group may also cover those who lose their Medicaid eligibility as a result of losing SSI due to medical improvement.

Persons with specific diseases: Persons with specific medical diagnoses may be covered by Medicaid under certain conditions. All states and the District of Columbia have chosen to cover women who need treatment for breast or cervical cancer if they are under 65, uninsured, and otherwise not eligible for Medicaid. Benefits are limited to the period during which treatment is provided.[25]

A total of 13 states and the District of Columbia cover persons with tuberculosis (TB) who are uninsured, but coverage is limited to services related to the treatment of TB.

Some parents of disabled children: The Deficit Reduction Act of 2005 provides states the option of allowing parents of disabled children to “buy in” to the Medicaid program if they have a family income below 300 percent of the federal poverty level. This option is subject to a premium.

Other groups: States may extend eligibility beyond these groups. The use of specific waivers allows states to diverge from certain provisions of the Medicaid Act. Waivers granted under Section 1915(b) of the Social Security Act are called “freedom of choice” waivers because they permit a state to limit the providers of Medicaid services and require beneficiaries to obtain services through a managed care organization.[26] Section 1115(a) of the Social Security Act provides even greater leeway to the states. That legislation allows states to carry out experimental, pilot, or demonstration projects that, in the judgment of the Secretary of DHHS, are likely to assist in promoting the objectives of the Act, including those of the Medicaid statute.[27] These waivers permit a state to alter the scope of services and to expand eligibility to persons who would not otherwise be eligible for the Medicaid program. Using section 1115 waivers, states can adopt less restrictive methodologies for calculating income and resources. This discretion allows states to institute broader coverage, and hence has the potential to address the needs of otherwise uninsured populations. But the waivers also allow states to reduce benefits, increase cost sharing, and cap enrollment for some beneficiaries. Projects authorized under Section 1115 are usually approved for a five-year period and may be extended under certain circumstances. Demonstration projects must be budget neutral over their life. Several states have used the authority conferred under Section 1115 to launch “health care reform demonstrations” that include restructuring the delivery of services.[28]

Medicaid coverage is very much in a state of flux as a result of these waivers. In fiscal year 2007, 12 states reported planning to implement new Section 1115 waivers. These vary by size and scope, and hence by expected impact. Moreover, they are designed to meet different objectives, the three top goals being increasing private or employer-based coverage, encouraging personal responsibility, and expanding eligibility.[29] The state-specific information that is part of this report will describe some of the initiatives that are now in effect or under consideration.

In addition, states are addressing a Medicaid law that went into effect on July 1, 2006, restricting benefits to those who can provide proof of citizenship. This measure, which was part of the Deficit Reduction Act of 2005, requires that beneficiaries and applicants to Medicaid present a birth certificate, passport, or another form of identification in order to apply. This documentation replaces the previous requirement that applicants to Medicaid attest in writing that they are citizens, under penalty of perjury.[30]

The legislation has both ideological and fiscal roots. Originally designed to prevent undocumented immigrants from gaining to access to care, the measure was also touted as a cost-saving device, estimated to save the federal government some $220 million over five years and $735 million over 10 years.[31] The Congressional Budget Office calculates that Medicaid enrollment will decrease by 35,000 because of loss of coverage, and some states are already feeling the effects of the new requirements. 

Findings

Given the multiple pathways into Medicaid, states exhibit much variety in how they score in the eligibility category. Of the four categories examined, eligibility is the one weighted most heavily, accounting for 350 of the total 1000 points. States that rank high in this category are therefore more likely to score high overall.

Rhode Island, the highest-ranking state in eligibility, earned a total score of 296.8, while Indiana, with a score of 90.6, had the lowest eligibility value. There is therefore a more than threefold difference between the two ends of the eligibility spectrum.

In addition to Rhode Island, the other states ranking among the “Top 10” in terms of eligibility include, in descending order of rank:

Vermont, 283.7 points
New York, 264.8 points
Washington, 260.9 points
California, 258.9 points
Minnesota, 254.5 points
District of Columbia, 248.5 points
Massachusetts, 247.6 points
Wisconsin, 246.6 points
Hawaii, 245.0 points

Given the high relative weight of this category, it is not surprising that seven of these 10 states are also among the “Top 10” overall.

The 10 states with the lowest ranks in eligibility are:

Indiana, 90.6 points
Alabama, 91.6 points
Mississippi, 92.6 points
Arizona, 95.5 points
South Dakota, 101.1 points
Nevada, 108.5 points
Texas, 110.3 points
Idaho, 117.1 points
Delaware, 127.1 points
Virginia, 131.0 points

As with their higher-ranked counterparts, most of these states (Indiana, Alabama, Mississippi, South Dakota, Texas, and Idaho) are also among the 10 programs with the worst scores overall.

Eligibility indicators regarding children [extending services to children above the federal poverty level, and coverage under the State Children’s Health Insurance Program (SCHIP)] accounted for 24.8 percent of the 350 possible points under eligibility. Eligibility for women’s services (care provided to pregnant women and services provided to those with breast/cervical cancer) accounted for an additional 16.8 percent of the total points. Coverage of these children and women’s groups thus accounted for 41.5 percent of all points in the eligibility category. Looked at another way, these two categories of indicators accounted for 14.5 percent of all points for all categories. It is therefore useful to see how those states with the highest and lowest rankings in eligibility treated these two important groups.

For the two children’s and the two women’s eligibility subcategories described above, the five states with the lowest eligibility ranks had the following scores out of a possible 145.3 points:

South Dakota, 24.6 points
Alabama, 30.4 points
Arizona, 34.2 points
Indiana, 60.0 points
Mississippi, 61.9 points

Conversely, the five states with the highest eligibility rankings had the following scores for these children’s and women’s subcategories:

Rhode Island, 115.7 points
Vermont, 115.3 points
Washington, 79.8 points
New York, 75.4 points
California, 69.5 points

As the above breakdown shows, there is no overlap between the top and bottom five states in their scores for these populations. The state with the fewest points, South Dakota (with 24.6 points), has only 21.3 percent as many points as Rhode Island, the state with the most points for eligibility for children’s/women’s services.   

Ultimately, eligibility is the most important category. If a person is deemed ineligible for Medicaid, it matters little what services are available, how good they are, or how equitably the providers are paid. Yet widely divergent eligibility requirements continue to plague the Medicaid program. For example, a pregnant woman in family of three needs to have an annual income of less than $22,128 in order to qualify for Medicaid in Wyoming, while her Minnesota counterpart can be covered with an income of up to $45,650. Similarly, an infant’s family’s income would have to be less than $22,128 in Virginia for the baby to be covered, but less than $49,800 in Missouri. These are disparities that reflect local political decisions but have a ripple effect throughout the Medicaid program, undermining the very concepts of “one nation,” equal opportunity, and equal protection. 

[see full text PDF for tables, figures, and appendix]

IV. Scope of Services

Scope of services is the category exhibiting the most variety, complexity, and nuances. Over time, states have modified the optional services they provide under Medicaid in response to need, federal financial incentives, and political imperatives.

Subcategories of service

Because this category has the largest number of indicators, we have grouped them into seven major, mutually-exclusive subcategories:

• Services by type or target group
• Women’s services
• Services delivered by specific providers
• Rehabilitation services
• Devices and equipment
• Drugs
• Transportation

Most of these are in turn broken down into a number of discrete services, which were scored using different point values and weights before being reaggregated into the seven categories.

Services by type or target group
This broad category includes the following 16 subcategories: targeted case management; free-standing ambulatory surgery; diagnostic, screening, and preventive services; home and community-based services; home health services; hospice care; in-patient psychiatric services for those under 21; in-patient institutions for mental diseases and other institutions for mental diseases for those 65 and over; intermediate care facility services for persons with mental retardation; nursing facility services other than for mental diseases; Program of All-Inclusive Care for the Elderly (PACE); personal care services; sickle cell services; private duty nursing services; rehabilitation services for those with mental illness and substance abuse; and tobacco-dependence treatments.

Women’s services
These services relate to pregnancy and reproductive health and are reimbursed at a higher matching federal rate. States therefore have an added incentive to cover them, and most do. Nevertheless, some are more generous than others in their coverage, and this is reflected in their scores.

Services delivered by specific providers
These include non-physician providers who provide a wide range of primary and specialized services to Medicaid beneficiaries. These providers agree to “accept assignment,” which means that they accept the state’s payment as payment in full for the services rendered and cannot bill the patient for an additional amount. Because some of the services under this rubric are quite broad, states may choose to impose restrictions by type of patient or service, or limit the duration or frequency of the service provided. In many cases, these services are provided as part of an institutional stay.

Services in this category include the following: chiropractor services, dental services, nurse anesthetist services, nurse practitioner services, optometrist services, podiatrist services, and psychologist services.

Rehabilitation services
These services include the following: occupational therapy; physical therapy; and speech, hearing, and language services. They are subject to much variation, and are often limited by type of beneficiary, trigger condition, rehabilitation potential, frequency and duration of service, and other variables.

Devices and equipment
This category comprises dentures; eyeglasses; hearing aids; medical equipment and supplies; and prosthetic and orthotic devices.

Drugs
While drugs are a covered service, they are subject to restrictions that vary by state. Our indicators take this into account, reflecting the variations in scope that emerge even within covered services.

Transportation
Here, we are including this service only for states that include it under their State Medicaid Plan. Some states include this as an administrative expense, and are not represented here.

Measuring scope of service

In general, we are ranking states only in terms of the non-mandated services they provide. Most of the services listed above are optional. In the case of mandated services, we have taken into account only those characteristics that affect scope and that exceed or refine the mandated minimum levels. Over time, optional services have increased their share of Medicaid expenditures. In 1998, for example, Medicaid spending on optional services accounted for 65 percent of the total spent by the federal and state governments.[32]

Rankings are based on the following criteria:

Coverage: States offering an optional service receive credit in their scores, regardless of how limited the scope or how restricted the eligible population. The total number of points, however, may reflect the scope of service, as indicated below.

Population covered:   Some states cover only the categorically needy, while others extend services to the medically needy as well. The latter receive more points than the former in our scoring scheme.

Comprehensiveness: In general, the wider the scope of services, the higher the score. Limitations in terms of amount, frequency or duration will be taken into account in applying this criterion.

Lack of a financial barrier: Services that do not depend on cost-sharing on the part of the consumer are rewarded in our rankings. When co-pays are required, a distinction may be made between a nominal fee that is unlikely to deter access to services, and a more significant amount that may constitute a barrier to prompt care.  

The rationale for the indicator and the way in which each indicator was scored is described in detail in the Scoring Protocol included in the Appendix. Our scoring in this category leans toward the conservative, and we assign points to any state reporting that it covers a given service. In practice, however, the service may be seriously curtailed by the fact that: it is not available everywhere within the state; there may not be sufficient practitioners to provide the service promptly and effectively; or Medicaid providers offering the service may have capped their clientele and may not be taking new Medicaid patients. 

Findings

Table 2 (PDF) and Figure 5 (PDF) present all states and their ranks with respect to scope of services. The range in scores runs from top-ranked New York (with 168.3 points or 84.2 percent of the total score) to Mississippi (with 66.8 points or 33.4 percent): a more than 2.5-fold difference. The average score is 117.7, or 58.9 percent of the total points.

The “Top 10” Medicaid programs in terms of scope of services are as follows:

New York, 168.3 points
Minnesota, 158.1 points
Oregon, 155.0 points
Washington, 145.8 points
Illinois, 145.1 points
North Dakota, 145.1 points
Maine, 142.6 points
Arizona, 142.5 points
Tennessee, 141.6 points
California, 141.0 points

Because this was by far the category with the most indicators, the states’ overall scores tend to be evenly distributed throughout the spectrum, with only one tie among states (between North Dakota and Illinois, who share the 5th rank). Although the two top-ranked states in scope of services — New York and Minnesota — rank among the top 10 overall, only one other state (Washington) also falls within both the overall and the category-specific top 10 ranks. 

The following 10 states place at the bottom in scope of services, ranking from 51st to 42nd:

Mississippi, 66.8 points
Oklahoma, 71.7 points
Alabama, 71.9 points
Georgia, 76.3 points
Wyoming, 81.9 points
South Carolina, 82.9 points
Delaware, 86.2 points
Idaho, 91.6 points
Arkansas, 94.4 points
Connecticut, 98.9 points

Although this category is not weighted as heavily as eligibility in our adjusted scores, it nevertheless reflects overall program performance, and half of the states ranking in the bottom 10 in scope of services also rank at the bottom in the overall score. These states are Mississippi, Oklahoma, Alabama, South Carolina, and Idaho.

When services are broken down by type of care, it becomes evident that different states have different priorities in deciding the package of services they offer their Medicaid populations.

With respect to services by type or target group, the following states rank in the top five, in descending order:

New York, 73.9 points
Oregon, 68.4 points
Maryland, 65.0 points 
Rhode Island, 63.7 points
Michigan, 62.9 points
Tennessee, 62.9 points

The bottom rankings in services by type or target group are occupied by:

Georgia, 30.0 points
Wyoming, 30.0 points
Mississippi, 30.9 points
Delaware, 34.5 points
Alabama, 34.7 points
Oklahoma, 34.7 points

In terms of women’s services, the following five states occupy the top ranks:

California, 14.6 points
Minnesota, 14.6 points
New York, 14.6 points
Vermont, 14.6 points
Maryland, 13.5 points

The bottom rankings in women’s services are occupied by:

Idaho, 2.8 points
South Dakota, 2.8 points
Utah, 3.3 points
Nevada, 3.8 points
North Dakota, 3.8 points
Wyoming, 3.8 points

In provider-specific services, the following states earn the top ranks:

West Virginia, 31.2 points
Minnesota, 30.4 points
Arizona, 29.3 points
Iowa, 29.1 points
New Jersey, 28.6 points
North Dakota, 28.6 points

The bottom rankings in provider-specific services are occupied by:

Alabama, 10.2 points
Delaware, 11.7 points
Alaska, 11.8 points
Nevada, 12.8 points
Georgia, 13.8 points

In devices and equipment, the top rankings are occupied by:

New York, 32.2 points
Rhode Island, 32.2 points
Oregon, 30.8 points
New Jersey, 29.9 points
Illinois, 29.9 points
North Dakota, 29.9 points

The bottom rankings of devices and equipment are occupied by:

Delaware, 9.7 points
Mississippi, 9.7 points
Oklahoma, 9.7 points
Texas, 10.6 points
Maryland, 12.4 points
West Virginia, 12.4 points

Of all the categories, scope of services presents the most options for the states. Services cover the lifespan (from prenatal care to hospice), involve a broad range of facilities and providers, and can expand or contract as a function of need and budgetary possibilities. Even when two states offer the same package of services, they can do so under very different conditions. States can impose cost-sharing, or limit the frequency, duration, or amount of service provided to a given beneficiary. For this reason, this is the category with the most indicators and the most finely-calibrated scores.

[see full text PDF for tables, figures, and appendix]

V. Quality of Care

Given the large number of beneficiaries and the expenses involved in the program, Medicaid is under pressure to prove that it can deliver quality care. Up to now, however, the focus on quality has been primarily on avoiding fraud. Some states appeal to consumers to be careful about divulging their Medicaid card number, and urge their beneficiaries to avoid seeking medical care they do not need. For their part, Medicaid providers are told to watch for “upcoding” of procedures (billing for a more complex and costly procedure than what was actually delivered); to monitor attempts to “unbundle” a single medical event into its component parts in order to increase the fees; to be cautious of cost reports that do not reflect hours worked; and to be suspicious of anyone getting excess prescriptions that they may be reselling.[33]

While these measures may be necessary to protect the fiscal integrity of the program, they are not directly related to the quality of care. In fact, because Medicaid comprises more than 50 different programs, there are no overall indicators of quality that all states maintain. Our comparisons are therefore based on measures that serve as markers of quality.

Markers of quality

The data on quality vary a great deal and are a lot more complete for some services, such as nursing home care. Because this type of care was notoriously and dangerously neglected for many years, it has been subjected to greater oversight and more complete data collection. Since 1987, the Centers for Medicare and Medicaid Services (CMS) has defined the protocol that all states must use to survey their nursing care facilities and report their findings.

In cases such as nursing home care or services for children, where a significant proportion of a given service or target population is covered by Medicaid, we have used the quality indicators available for each state for all patient populations as a proxy for quality of care for the service covered by Medicaid. While these data have the limitation of not being specific enough, they provide a close approximation of the quality available to Medicaid recipients.

In the case of nursing home care, the rationale for using statewide data, even when not Medicaid-specific, includes the following: 

• The overwhelming majority of nursing homes (93.9 percent) accept Medicaid patients.[34]
• In 2003, the most recent year for which data are available, Medicaid paid for 46 percent of all nursing home expenditures, and this proportion is likely to have increased since then. An additional 12 percent was paid for by Medicare.[35]  
• Because nursing home care is so expensive, 56 percent of nursing home residents eventually “spend down” their resources and qualify for help from Medicaid.[36]

In the case of services for children, the rationale is that Medicaid covers a significant portion of their medical care: the program covers more than one in four children in this country.[37] Moreover, what is adopted as the standard of care under Medicaid is often reflective of what providers do for the pediatric population as a whole, regardless of payer. Additionally, this is one of the few populations for which data on results are available.

The indicators used under the quality of care category cover structure, process, and outcomes.[38] Indicators of structure include those ingredients or elements that facilitate or promote quality of care. Process measures include whether proper procedures were used in delivering care. Outcome measures include both improvements in health status and the avoidance of adverse results.

Findings

In part because states have not been held accountable for the quality of their Medicaid programs, they earn the lowest scores in this category. The median score for this category is a meager 28.2 percent of 200 points. 

Because states have so many deficiencies in this area, even those ranking at the top have low scores, boosted only by the fact that many others do even worse in this category.

The following states score in the “Top 10” in this category:

Massachusetts, 143.0 points
Rhode Island, 109.0 points
Ohio, 106.7 points
Florida, 106.4 points
Nebraska, 105.4 points
Kentucky, 105.1 points
Alabama, 97.1 points
Alaska, 95.5 points
Virginia, 94.0 points
Maine, 92.7 points

The 10 states with the lowest scores all earn less than 12 percent of the maximum points in this category. They are as follows, ranking between #51 and #42:

Idaho, -4.4 points
Oklahoma,  -3.8 points
Nevada, 8.4 points
Louisiana, 10.2 points
Kansas, 18.0 points
Maryland, 18.8 points
Arkansas, 19.5 points
South Carolina, 20.1 points
Georgia, 22.4 points
Colorado, 22.4 points

Unlike the previous two categories, quality of care shows a very broad spread in scores, with a more than 17-fold difference between the states with the highest and lowest positive scores (Massachusetts, with 143.0 points; Nevada, with 8.4 points). 

To a large extent, much of the difference can be accounted for by differences in the quality of their nursing home facilities. Because some of the indicators used rely on evidence-based benchmarks for adequacy in nursing home care, states that fall short of the acceptable minimum standards earn negative points. As a result, quality of care is the only category in which two states (Oklahoma and Idaho) have negative scores.

The distribution of scores has two “tails” representing statistical outliers on either side of the spectrum: one state that scores considerably higher than the rest, and the two that are at the very bottom, with negative scores. When these three states are omitted, the differences in scores are significantly reduced, although they still vary by a very large factor of 13.0.

Despite its top rank, Massachusetts earns only 71.5 percent of the total points in this category. It is followed at a distance by Rhode Island, with only 54.5 percent of the total points. 

These findings suggest that “quality control” needs to be drastically redefined within the Medicaid program. At present, the term is used to refer to the CMS’ statutory responsibility to monitor state and local Medicaid eligibility determinations. However, the sifting and sorting of people to see if they are indeed eligible for services is more of an accounting procedure than a quality assessment process. Accountability therefore needs to supplement the current emphasis on accounting. Only then will the public be served and the government be assured that it is getting value for the monies invested in the program.

[see full text PDF for tables, figures, and appendix]

VI. Provider Reimbursement

Medicaid is financed by the states and the federal government. Federal funding for the program comes from general revenues.  As an entitlement program, Medicaid’s federal spending levels are pegged to the number of people participating in the program and the services provided; spending is therefore open-ended and subject to fluctuations that are difficult to budget. As costs have risen over time, the program has become an important arena in which issues related to resource allocation have played out.

Even when states may be reluctant to commit an increasing share of their revenues to the program, the political and economic reality is that they need to leverage their share of the costs to maximize what they get from the federal government. The stakes for all participants are high. At present, Medicaid:

• covers over 55 million Americans;[39]
• is a major budget item for the states, averaging 16 percent of all state spending;[40]
• represents the largest source of federal grant support to states;[41]
• accounts for eight percent of all federal spending[42] and one of every five health care dollars spent in the U.S.;[43]
• is the nation’s main source of coverage for long-term care;[44]
• supports tens of thousands of health care providers throughout the country; and 
• has a significant multiplier effect on the U.S. economy as a whole.[45]

It is therefore not surprising that the financing of Medicaid is a topic that is often debated, defused, reframed, or circumvented, depending on who is affected and who is doing the debating. 

Medicaid financing

The federal government contributes between 50 percent and 76 percent of the payments for services provided under each state Medicaid program.[46] This contribution, known as the Federal Matching Assistance Percentage (FMAP), varies from state to state and from year to year because it is based on the average per capita income in each state. States with lower per capita incomes receive a higher federal matching rate. The federal matching rate for administrative costs is uniform for all states and is generally 50 percent.

Although the sliding FMAP was intended to have a redistributive effect and therefore sought to reduce disparities between states, it does this only partially because of the constraints imposed by the statutory minimum FMAP. The funding formula is also problematic for additional reasons.[47] First, the cost of coverage is substantial for both federal and state governments, and is difficult to predict. In addition, Medicaid’s matching payments do not automatically adjust to changing economic conditions. The program’s scope may therefore be forced to contract during an economic downturn, thus having a negative effect on both the beneficiaries and those who are newly uninsured.

Furthermore, states have used “Medicaid maximization” or “revenue enhancement” strategies to increase federal spending in the program; in some cases, these payments may constitute up to one-sixth of a state’s Medicaid expenditures.[48] Because the monies obtained through such strategies enter the states’ coffers without earmarking, they are often used for purposes unrelated to the population and services for which Medicaid was created. As a result, these strategies have been the target of measures to insure greater accountability. These measures have included legislation, regulation, greater federal oversight, and moral suasion. Changes in intergovernmental transfer rules would reduce federal payments to states by almost $24 billion over 10 years.[49] States are therefore poised to adjust to a significant shortfall in federal revenues, and many are restructuring their services in anticipation of lost funds.

Few indicators relate directly to reimbursement. We have therefore relied on those that cover three aspects of Medicaid finances: payments per enrollee, by demographic group; physician fees; and Medicaid fees compared to Medicare fees. Because the data on fees are restricted to payments made under fee-for-service and do not reflect payments made to managed care organizations, they capture a decreasing proportion of Medicaid enrollees, particularly in some states where a vast majority of program beneficiaries are in managed care. Nevertheless, fee-for-service reimbursement rates also have an impact on what managed care organization rates pay physicians, as many states peg their capitation rates to what they pay under fee-for-service.[50] Because TennCare, Tennessee’s Medicaid program, does not use fee-for-service, that program has not been included in our calculations under reimbursement.

Findings

Of the four categories examined, reimbursement is the one with the fewest indicators. It is therefore subject to much fluctuation between and among states. At the same time, it is the “lumpiest” category, with several states sharing the same rank in some cases.

States have wide discretionary authority concerning the methods and amounts of fees.[51] Medicaid fees have lagged in comparison with other physician fees, including those paid under Medicare, and many states face physicians who are reluctant to see Medicaid patients or who place limits on the number or proportion of Medicaid patients in their practices, thus closing off options for new entrants. Physician reimbursement is therefore a proxy for access to care, as research has shown that acceptance of new Medicaid patients is higher in states that have higher Medicaid fees relative to Medicare than in states with lower Medicaid fees.[52]

Unlike the fairly even distribution of scores that characterizes some of the other categories assessed in this report, reimbursement has states with very high and very low scores. At the high end is Alaska, which pays Medicaid providers much more than the national average in order to attract and retain them. As a result, Alaska earns the maximum number of points allotted to this category, 250 points, the only case in which a state does so.

The other states within the top 10 ranks are the following:

Delaware, 200.4 points
Nevada, 185.3 points
Arizona, 184.2 points
Nebraska, 161.3 points
Wyoming, 160.1 points
Iowa, 160.1 points
Maryland, 152.7 points
Wisconsin, 152.0 points
Montana, 144.6 points
Connecticut, 144.6 points

At the other end of the scoring scale, the states occupying the bottom 10 ranks in reimbursement are the following:

New Jersey, 12.2 points
New York, 44.0 points
Rhode Island, 59.5 points
Missouri, 66.9 points
Pennsylvania, 68.0 points
District of Columbia, 68.8 points
Florida, 75.4 points
California, 75.4 points
Oklahoma, 75.4 points
Texas, 79.5 points
Illinois, 79.5 points
Michigan, 79.5 points
Maine, 83.1 points
Indiana, 83.5 points

Because New Jersey ranks so low, the scores between the highest- and the lowest-ranking states vary 20.5-fold. But even when the two states representing the extreme values are omitted, the difference in scores between the second-highest state (Delaware) and the next-to-last state is still approximately 4.5-fold.

These are differences that make a difference. States have understandably attempted to keep their Medicaid costs low by paying providers lower fees, and this has had an impact on access to care. Low payment rates deter physician participation in the program, or lead providers to cap their Medicaid clientele. This is especially the case among physicians in solo practice or working in small groups.[53] As a result, an increasing proportion of Medicaid patients are relying on physicians who practice in larger groups, hospitals, or community health centers.

[see full text PDF for tables, figures, and appendix]

VII. National Results

Overview

As summarized in Table 5, the state Medicaid programs show much variation between and within the categories assessed.

Table 5. Summary of state Medicaid program scores

Almost all state Medicaid programs are doing poorly in meeting all of their basic objectives.
The best overall score is only 645.9 (64.6 percent) and the average score is 472.3 out of 1000 points. The median overall score of 471.1 means that half of all states have scores lower than this. Further, 31 states have scores of less than 500 (50 percent of possible points). Highlighting the problem of very widespread and uneven performance is the fact that a total of 30 states (over one-half of states) were in the bottom 10 in one or more of the four categories. These 30 states include: Alabama, Arizona, Arkansas, California, Colorado, Connecticut, Delaware, District of Columbia, Florida, Georgia, Idaho, Illinois, Indiana, Kansas, Louisiana, Maryland, Michigan, Mississippi, Missouri, Nevada, New Jersey, New York, Oklahoma, Pennsylvania, Rhode Island, South Carolina, South Dakota, Texas, Virginia, and Wyoming.

The 10 highest-scoring states earn between 645.9 and 548.9 points of the maximum 1000.
The following states occupy the first 10 ranks, in descending order: Massachusetts (645.9), Nebraska (625.5), Vermont (616.1), Alaska (609.9), Wisconsin (606.8), Rhode Island (600.0), Minnesota (591.2), New York (560.2), Washington (550.0), and New Hampshire (548.9).

The 10 most deficient state programs have overall scores ranging from 317.8 to 379.1 of the total 1000 points.
The worst programs, in order from 50th to 41st, are in Mississippi (317.8), Idaho (325.2), Texas (335.5), Oklahoma (336.7), South Dakota (352.6), Indiana (357.2), South Carolina (364.0), Colorado (375.7), Alabama (376.3), and Missouri (379.1).

Even the top-ranking programs fall short in some categories and have ample room for improvement.
When the data are broken down by category, there are gaps between the scores of even those at the top and the maximum scores. For example, in eligibility, Rhode Island, the highest-ranking state in that category, earns only 84.8 percent of the maximum points in that category. Similarly, in scope of services, the state ranked first, New York, gets 84.2 percent of the total points in that category. In quality of care, even the best-scoring state, Massachusetts, receives a modest 71.5 percent of the total. Reimbursement is the only area in which one state, Alaska, gets the maximum number of points. This can be attributed to the fact that the state pays its Medicaid providers much more than the rest of the country in order to attract and retain practitioners. This is a clear anomaly, as suggested by the fact that the second-ranking state in this category earns only 80.2 percent of the total points and the average for the other 48 states was only 44.3 percent of the total.

Emphasizing the spotty performance of some of the top-ranking states is the fact that two states in the “Top 10” overall, New York and Rhode Island, were in the bottom 10 in one of the four categories; both states had poor reimbursement policies. This poor showing confirms that even the states with the most resources, best intentions, and higher overall scores are failing in one or more of the categories we examined: eligibility, scope of services, quality, and reimbursement.

Quality of care is the category in which states earn the lowest scores.
The absence of national benchmarks and uniform data collection and reporting systems makes quality difficult to assess across states. As a result, states are in the anomalous position of having much data, but limited information and practically no intelligence that can be used in decision-making.

There are marked inter-state differences, with some categories showing greater disparities than others.
Because uniform federal guidelines have been undermined by waivers and the federal financial matching formula does not fully make up for differences in the resources each state devotes to Medicaid, disparities between states are quite dramatic. Not surprisingly, the gradients between the best and worst Medicaid programs can be quite steep. For example:

• the difference in overall scores between the highest- and the lowest-ranking states is twofold (Massachusetts with 645.9 points, Mississippi with 317.8 points);
• the difference in scores between the highest- and lowest-ranking states in the eligibility category is 3.3-fold (Rhode Island with 296.8 points, Indiana with 90.6 points);
• the difference in scores between the highest- and lowest-ranking states in the scope of services category is 2.5-fold (New York with 168.3, Mississippi with 66.8 points);
• the difference in scores between the highest- and lowest-ranking states with positive scores in the quality of care category is more than 17-fold (Massachusetts with 143.0 points, Nevada with 8.4);
• the difference in scores between the highest- and lowest-ranking states in the reimbursement category is more than 20-fold (Alaska with 250 points, New Jersey with 12.2 points). Even when the two extremes, which are statistical outliers, are omitted, the difference in scores between the second-scoring state (Delaware) and the state ranked #49 (New York) is still more than 4.5-fold.

There is significant intra-state variation in scores, with only a handful of states ranking consistently (i.e., within 15 ranks) across categories.
Some states have made a conscious decision to trade off populations covered for breadth of services, or vice-versa. In other cases, the variation reflects the result of many discrete decisions, each one a response to a time-bound, local situation. Only four states — Nebraska, New Hampshire, Texas, and Wisconsin — show overall consistency (placing within 15 ranks) across all four categories.  Interestingly, three of these (Nebraska, New Hampshire, and Wisconsin) rank in the “Top 10,” which indicates that some succeed in all or most categories, without making significant trade-offs that weaken the program.

The 10 best states and the 10 worst states tend to cluster geographically.
Half of those that rank in the top 10 are in the Northeast: Massachusetts, Vermont, Rhode Island, New York, and New Hampshire. Another three — Nebraska, Wisconsin, and Minnesota — are in the Midwest. An additional two are in the Northwest: Washington and Alaska. The 10 worst states also show a geographic pattern, with three of them in the South (Mississippi, South Carolina, and Alabama) and four in the south central part of the country: Colorado, Texas, Oklahoma, and Missouri.

States whose Medicaid programs rank the lowest also tend to fare poorly in overall health rankings.
In order to test if our Medicaid rankings were in any way associated with the national health rankings published in America’s Health Rankings[54], we computed Spearman’s correlation coefficient. The resulting measure showed an association between the two ranks [rho (a measurement of correlation) = .5480] and it is highly significant (p= <.0001). Thus the states whose Medicaid programs rank the lowest also tend to have the worst health indicators overall.

States’ ranks tend to correlate with median household income.
The programs that do better are those in states that have higher overall median household incomes. Conversely, those states whose median household income is low tend to have lower-ranking Medicaid programs. The Spearman’s correlation coefficient in this case showed a correspondence between the two rankings (rho= .4480, p < .0012). Unfortunately, this means that the poorest states, which have the most need for the program, have the worst services, the Medicaid version of what has been codified as the “inverse care law” which says that “the more you need, the less you get.”[55]

Although the federal matching formula is designed to mitigate existing inequalities, it does this only to an extent. All states receive at least a 50 percent match, and those with lower per capita incomes receive a larger percentage. In fiscal year 2007, the Federal Matching Assistance Percentage (FMAP) for Medicaid ranged from 50.0 percent (in 12 states) to 75.9 percent (for Mississippi).[56] While the 12 states that are at the 50 percent FMAP get a federal dollar for every state dollar they spend on the program, those that have a higher FMAP get more.  Mississippi, for example, receives approximately $3.00 from the federal government for every state dollar it devotes to Medicaid. If Mississippi chooses to reduce its Medicaid expenditures, it also forgoes its corresponding share of the federal match.  Reducing its Medicaid spending by $1 will therefore “cost” Mississippithe $3 in matching funds and result in a total reduction of $4 in its Medicaid budget.[57] The political and financial stakes in the program are therefore higher for the poorer states. But these states also have competing needs, and health spending may be sacrificed to other pressing priorities.

Specific populations fare much better in some states than others
Moreover, even within a given state, those with certain characteristics or conditions are likely to be more successful than others in gaining access to care.   Therefore the question, “As a Medicaid enrollee, where am I most likely to gain access to the most comprehensive and best care?” can only be answered by, “It depends.” The answer is contingent on the patient’s demographic characteristics and medical requirements. For example, a woman in need of reproductive services is clearly at an advantage in four states — California, Minnesota, New York, and Vermont — all of which offer more expansive coverage with fewer restrictions than the rest of the nation. Conversely, she would do well to avoid seeking care in South Dakota, Idaho, and Utah, where there may be eligibility restrictions, limited services, or others barriers to care.

Similarly, patients who need devices and equipment are better off in New York or Rhode Island, while those requiring rehabilitation services (e.g., a stroke victim or someone recovering from an accident) are likely to be more successful in gettig comprehensive care in New York or Tennessee.

In short

In summary, this evaluation of Medicaid demonstrates a bleak picture for millions of people in many states.

The first barrier, eligibility, is difficult to get past for millions of uninsured people. The wide variation in eligibility scores, more than threefold between the best and worst states, reflects this, as does the fact that 23 states had eligibility scores less than 50 percent of the total possible (350 points), thus keeping people out who would be eligible were they to live in certain other states.

But even for those eligible for Medicaid, the scope of services is extremely uneven. In addition to the 2.5-fold difference between the best and worst states, 10 states had scope of services scores of less than 50 percent of the possible 200 points.

Similarly, even if people are eligible for Medicaid in their state and the program provides those services needed by particular patients, the miserly reimbursement policies in many states make it less likely that they will be able to find a physician who can provide these services. There was a 20.5-fold difference between the best and the worst scores on reimbursement; in this important category, 31 states had scores that were less than 50 percent of the total possible 250 points.

Despite limitations on measuring more indicators of quality because such data are not uniformly collected, this category demonstrated very poor results for almost all states. With a maximum score of 200 points in this category, only six states had scores of more than 50 percent of this point total and 18 states had scores of less than 25 percent of 200 points. 

Overall, and in many ways, Medicaid is failing to deliver care to millions of people desperately in need of good quality health services.

[see full text PDF for tables, figures, and appendix]

VII. State Reports

Alabama state report (PDF) 
Alaska state report (PDF)
Arizona state report (PDF) 
Arkansas state report (PDF) 
California state report (PDF)
Colorado state report (PDF)
Connecticut state report (PDF)
Delaware state report (PDF)
District of Columbia report (PDF)
Florida state report (PDF)
Georgia state report (PDF)
Hawaii state report (PDF)
Idaho state report (PDF)
Illinois state report (PDF)
Indiana state report (PDF)
Iowa state report (PDF)
Kansas state report (PDF)
Kentucky state report (PDF)
Louisiana state report (PDF)
Maine state report (PDF)
Maryland state report (PDF)
Massachusetts state report (PDF)
Michigan state report (PDF)
Minnesota state report (PDF)
Mississippi state report (PDF)
Missouri state report (PDF)
Montana state report (PDF)
Nebraska state report (PDF)
Nevada state report (PDF)
New Hampshire state report (PDF)
New Jersey state report (PDF)
New Mexico state report (PDF)
New York state report (PDF)
North Carolina state report (PDF)
North Dakota state report (PDF)
Ohio state report (PDF)
Oklahoma state report (PDF)
Oregon state report (PDF)
Pennsylvania state report (PDF)
Rhode Island state report (PDF)
South Carolina state report (PDF)
South Dakota state report (PDF)
Tennessee state report (PDF)
Texas state report (PDF)
Utah state report (PDF)
Vermont state report (PDF)
Virginia state report (PDF)
Washington state report (PDF)
West Virginia state report (PDF)
Wisconsin state report (PDF)
Wyoming state report (PDF)

X. Conclusions

We have titled our report Unsettling Scores because it is indeed disturbing that, after more than four decades, the Medicaid program has clearly failed to achieve its objectives and to therefore fully meet the needs of those it serves or is supposed to serve.

[see full text PDF for tables, figures, and appendix]

Glossary

Aid to Families with Dependent Children (AFDC) – Program operating between 1935 and 1996. Enacted in 1935 as part of the original Social Security Act and rescinded as part of welfare reform, the program sought to support needy children deprived of at least one parent’s presence and full support. 

Beneficiary ­–Person eligible for and enrolled in the Medicaid program in the state in which he or she resides. Also referred to as enrollee.

Budget neutral –Having no impact on a budget. States may be granted authorization to deviate from certain Medicaid mandates if the proposed changes are budget neutral and therefore do not affect the bottom line.

Capitation– Modality of payment whereby a state Medicaid program pays a plan or provider based on the number of Medicaid beneficiaries under its care. Payment is usually a fixed amount per person per month.

Categorically needy– Persons eligible for Medicaid because they fall under specific categories or groups, i.e., children, the aged, individuals with disabilities. These qualify for the basic mandatory package of Medicaid benefits.

Co-pay, co-payment– A fixed dollar amount that a Medicaid beneficiary may have to pay when receiving services. This varies by state, service, and eligibility category.

Cost-sharing– Group of measures requiring Medicaid beneficiaries to bear part of the cost of a service. These can be co-payments (a fixed monetary amount), deductibles (a fixed amount the patient must pay before coverage begins), or co-insurance (a percentage of the total cost of a service). While cost-sharing is often advocated as a way of controlling costs by making beneficiaries “prudent purchasers,” it may also act as a barrier or deterrent to needed care.

Eligibility– Determination of who is covered by Medicaid. Although some populations are covered throughout the nation, others vary from state to state.

Federal Matching Assistance Percentage (FMAP) –Share of the costs of Medicaid borne by the federal government. This varies from one state to another, and currently ranges between 50 percent and 76 percent. States with lower per capita incomes have a higher percentage of their Medicaid costs covered by the federal government. The FMAP may be higher for some services than for others. For example, family planning services are matched at a higher rate than other services as an incentive for states to provide these services.

Federal Poverty Level (FPL) –  Income level below which an individual or family is considered poor in the United States. The federal poverty threshold is determined annually by the U.S. Census Bureau and is based on increases in general inflation. The U.S. Department of Health and Human Services has adapted the census poverty thresholds as guidelines for use in Medicaid. At present (2007), these guidelines establish the poverty level for the 48 contiguous states and the District of Columbia at $10,210 for a single person and $17,170 for a family of three. The guidelines are somewhat higher for Alaska and Hawaii because of their higher cost-of-living. State Medicaid programs establish their eligibility thresholds as a fraction or multiple of the federal poverty guidelines.

Indicator –Measurement of an aspect of the Medicaid program. These have been grouped by category (eligibility, scope of services, quality of care, and reimbursement). Some indicators are composites of several measures, as described in the Scoring Protocol in the Appendix.

Managed care –Modality of service delivery under which an organization or health plan provides a specific set of services to an enrolled population for a fixed, prepaid annual fee. Currently, most Medicaid beneficiaries are in managed care plans.

Mandated services– Services all states are required to provide to their Medicaid beneficiaries.

Medically needy –Optional Medicaid eligibility group comprising individuals who qualify for coverage because of high medical expenses. These individuals must meet Medicaid’s categorical requirements but have incomes that are too high to qualify under “categorically needy” coverage. In some cases, services provided to the medically needy are not as comprehensive as those provided to other Medicaid beneficiaries.

Program of All-Inclusive Care for the Elderly (PACE)– Optional benefit that some states provide to Medicaid beneficiaries 55 years or older who require the level of care usually provided by a skilled nursing facility. This program, originally begun in California, allows beneficiaries to live at home, attend a day treatment center, and get other supplementary services. 

Quality of Care– Degree to which Medicaid programs are performing in accordance with accepted standards of care. This category has a limited number of indicators, and includes markers that suggest better monitoring of services or better health outcomes.

Reimbursement– For purposes of this report, this category includes per capita spending by state Medicaid programs for specific groups, and physician payments.

State Children’s Health Insurance Program (SCHIP)– Federal-state matching program which provides health coverage for uninsured low-income children. In a few cases, the program may also cover their parents. States have the option of administering SCHIP through their Medicaid program or through a separate program. The average federal matching rate for SCHIP is higher than that for Medicaid, but SCHIP allocations to states are in the form of a block grant, capped at a specific amount, rather than open-ended.

Section 1115 waiver– Legislative measure through which states can receive authorization from the Secretary of Health and Human Services to waive compliance with many of the requirements of the Medicaid statute. This waiver, named for the section of the Social Security Act under which it was enacted, allows states to experiment with different approaches to delivering care under the Medicaid program. These waivers have become increasingly popular, and many states have used them to expand or restrict eligibility, coverage of services, and payment. Waivers are granted for a five year period.

Scope of Services –As used in this report, this refers to coverage of care provided by the different state Medicaid programs. Although there are mandated services that all states must provide, the focus in this report is on optional services or those that exceed the required minimum. Because different populations are eligible for different services, both the what and the who are taken into account in determining the points assigned to most indicators in our Scoring Protocol.  

Social Security Act –Legislation enacted in 1934 to provide economic and other support to specific groups. Medicaid was enacted as Title XIX of the Social Security Act in 1965. 

Spend Down – The process of using up all income and assets on medical care costs to become eligible for Medicaid.

Supplemental Security Income (SSI) –Federal program for the poor, aged, and disabled that provides its beneficiaries a monthly amount. In most states, those who are eligible for SSI are also deemed eligible for Medicaid.

Waiver – Provision that allows Medicaid programs to depart from eligibility or benefit rules for temporary periods. While Section 1115 waivers are the best-known, other waivers may be approved to address specific events, such as the disruption of services and population displacement that occurred following Hurricane Katrina.