Skip to main content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.


The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

ASPE Highlighter Winter 2005

In This Issue: • Unemployment Insurance Among Single Mothers
Best Practices for Cost Management of State Employee Pharmaceutical Benefit Programs
Insurance Coverage in the United StatesQuality of Medicaid Personal Assistance in Arkansas

Message from the Assistant Secretary

Welcome to the ASPE Highlighter! This is a fresh attempt to share some of our research findings that might not otherwise come across your desk. ASPE does applied research focusing on essential policy issues facing the Secretary of the Department of Health and Human Services, the White House, and the Congress. It is intended to inform the debate and deliberation.

This issue of the ASPE Highlighter covers a range of topics that span the breadth of policy at HHS. The first article, Receipt of Unemployment Insurance Among Single Mothers, uncovers a fascinating trend in welfare-to-work. During the recent economic downturn some ex-TANF recipients lost their jobs. ASPE researchers found that low-income single mothers were more likely to rely on the Unemployment Insurance they earned while in the labor force, rather than returning to the TANF rolls. The second article focuses on the best practices State governments have used to competitively purchase prescription drugs. To reduce drug expenditures, some states are looking to Canada, but others are using a variety of tools to acquire safe, high quality drugs at the most competitive prices available. Our third article, Insurance Coverage in the United States, presents the latest data on the uninsured, their essential characteristics, and policy relevant subpopulations. Our fourth article looks at one of the most innovative programs in Long-Term Care policy, the Cash-and-Counseling demonstration. It examines the quality of personal assistance available in Arkansas through both the traditional Medicaid program and the Cash-and-Counseling demonstration.

Over the next few issues, we will be presenting a variety of research findings in different formats. We hope you enjoy the articles and the other information we highlight. We welcome any suggestions for improving future issues.

Michael J. O’Grady, Ph.D.
Assistant Secretary for
Planning and Evaluation

U.S. Department of Health
and Human Services

Unemployment insurance among single mothers

“Receipt of Unemployment Insurance Among Single Mothers,” an Issue Brief1 prepared by staff in the Office of Human Services Policy of the Office of the Assistant Secretary for Planning and Evaluation, presents analysis of data from the Census Bureau’s Current Population Survey (CPS). The study was motivated by the fact that cash welfare caseloads under the Temporary Assistance for Needy Families (TANF) program have not increased during the recent slow down in the economy. The key issue explored in the Issue Brief is whether single mothers are making greater use of traditional unemployment benefits, instead of returning to cash welfare after losing a job. The Issue Brief combines the CPS data analysis with findings from two ASPE-sponsored studies that examine the extent to which former welfare recipients who transition from welfare to work are able to meet the qualification requirements of the unemployment insurance system.

The Issue Brief finds that between 2000 and 2003 receipt of unemployment insurance (UI) benefits increased among low-income single women with children, as shown in the graph below. In contrast, there was very little change in UI receipt among this population of single mothers during the recession of the early 1990s.

What changed is the enactment of welfare reform and the record numbers of welfare recipients who have gone to work. The employment rate of never-married mothers with children under 18—those single mothers who have the highest rate of participation in TANF and are most likely to be affected by changes in welfare policy—has increased, from 49 percent in 1996 to 63 percent in 2003. This dramatic increase in employment among single mothers with children is a likely explanation for the observed increase in use of the unemployment insurance system.

The Issue Brief also summarizes findings from two studies examining the potential eligibility for unemployment insurance among women leaving welfare for work. The low-wages and unstable work histories of many former welfare recipients can make it challenging for them to meet the monetary and non-monetary requirements set by their State unemployment insurance agency. While access of former welfare recipients to UI benefits is not universal, both studies found that former welfare recipients who transition from welfare to work appear to have greater access to unemployment insurance in recent years than was true for their counterparts in the 1980s.

Even though UI benefits generally operate under a tighter time limit than TANF, the payments are higher. Although UI benefits vary from state to state, both studies found potential UI benefits about twice as high as TANF benefits. Thus for the single mother who qualifies for unemployment insurance, as well as for the welfare agency, UI benefits may be a preferred form of social safety net over cash welfare payments.

The Issue Brief concludes that more single mothers have been able to use (and are using) the UI system as a primary safety net than in the past, and that this increased reliance has reduced some of the demand for cash assistance from the TANF program.


Source: ASPE tabulations of the Annual Social and Economic Supplement
of the Current Population Survey.

1 - See for the full Issue Brief.

2 - Note that low-income women with children are defined as woman in households without a husband present (female-heads) in families with income



Best Practices for Cost Management of State Employee Pharmaceutical Benefit Programs


State Governments are employing a variety of mechanisms and practices to manage public employee and retiree prescription drug programs.


Prescription drug costs are one of the fastest growing expenses in the health sector and a major contributor to rising health care costs. Retail prescription drug spending grew almost twice as fast as all other health services in recent years, and is expected to outpace overall health spending growth by an average annual rate of four percent over the next ten years. In response, State officials have taken a number of steps to manage the utilization, costs and value of their public employee prescription drug benefit programs. To understand the nature and features of these public sector cost containment strategies, ASPE has commissioned a study to identify and describe current and emerging public sector practices in drug spending cost containment.

The first phase focused on a report to identify and describe the current cost containment and decision-making environment in the States with specific attention to State government employee benefit programs. The completed report—Best Practices for Cost Management of State Employee Pharmaceutical Benefit Programs—examines relevant information, based on a literature review, public sources, and industry reports; outlines the range and type of proven and emerging cost control strategies; describes specific trends in States’ attempts to assess impact and effectiveness of cost containment approaches; identifies critical gaps in information; and makes relevant recommendations on opportunities for future study.

The report is organized in three thematic sections: 1) State’s Environment—describes what is known about converging pressures of costs, demand, demographics, and state imperatives for action in the government sector environment; 2) Key Learnings-States’ Drug Cost Management Strategies Evolve—provides a description of the various types of cost containment methods reflected in the literature and actually adopted by State employee health benefit programs and/or specific joint purchasing initiatives; and 3) Insights and Opportunities— summarizes key insights on specific states that “stand out” as bellwethers or innovators who merit further study.


State employers, leveraging more than $24 billion in total health care expenditures for 2.4 million active employees and 970,000 retirees, play a significant role in health care financing, but are less often studied in the marketplace. Cost containment strategies and tactics are not new to State governments. Current research, however, is limited in answering the questions of how beneficiaries use prescription drugs; what factors, when controlled by the State, optimize patterns of use, prescribing, absenteeism, and overall health; and how States are flexing their influence as employers and with what consequences to employees/retirees, i.e., compliance and satisfaction.

Most recent survey results of State employers identify the most dominant cost controls and utilization management strategies, including:

  • three-tier co-payment structure;
  • higher cost-sharing and deductibles;
  • incentive-based formularies;
  • mail order pharmacies; and
  • condition-specific disease management initiatives.

Emerging strategies include four-tier copayment structure, step therapy, closed formularies, and intra-State and inter-State purchasing pools.

The interplay of collective bargaining, regulatory, fiscal, and legislative pressures present unique obstacles for States pursuing changes to prescription drug benefit design for State employees and retirees. At least fifteen states stood out in the literature review as trendsetters, but there was little specific empirical State information in the current literature on the actual results from State efforts to reign in the rate of growth of their pharmaceutical spending and the utilization of prescription drugs.

As states continue to gain better understanding of the forces affecting health care costs and utilization in employee and retiree populations, they are becoming more aggressive in pursuing significant changes in benefit design, higher copayments and deductibles, and adopting tactics and incentives that can influence provider and consumer behavior. Further in-depth study is needed to assess States’ decision-making processes, approaches to information gathering, benchmarking, and use of specific metrics to monitor the impact and effectiveness of changing plan design, and adopting incentive-based cost management programs and evidence-based initiatives. Phase II of this study is focusing on in-depth case studies in a number of trendsetter States.

This report was prepared for the Office of the Assistant Secretary for Planning and Evaluation and is available on the ASPE website: For further information, please contact the Project Officer Laina Bush.



According to the Census Bureau’s 2004 Current Population Survey (CPS), there were 45 million uninsured individuals in 2003, or 15.6 percent of the civilian non-institutionalized population. Those that lack insurance represent a diverse group. Understanding the uninsured population is important for policy makers looking to design solutions to the problem. This report describes insurance coverage in the United States and describes the key demographic characteristics of the uninsured.

Before discussing the uninsured, it is important to understand the nature of insurance coverage in the U.S. Health insurance in the U.S. is provided through several major private and public sources (see Figure 1). The majority of Americans have health insurance through either their own, a spouse’s, or a parent’s employment.

  • Employer-sponsored insurance covered 174 million people, or 60 percent of the population in 2003.
  • Directly purchased insurance that people purchase on their own covers 26.6 million, or 9 percent of the population. As a primary source of coverage, directly purchased insurance has an even smaller share of the market. Medicare beneficiaries often purchase direct coverage to supplement their Medicare coverage. In fact, 38 percent of directly purchased private coverage is purchased by the elderly, which is assumed by many analysts to be Medicare supplemental insurance (Medigap), leaving only 14.9 million, or 5 percent of the population, with directly purchased private coverage. Government provided coverage is another important source of insurance.
  • The largest public coverage program is Medicare, with 39.5 million enrollees, or 14 percent of the population. The majority (84 percent) of Medicare beneficiaries are elderly individuals age 65 and older (though some are under 65 and are either disabled or patients with end-stage renal disease).
  • The next largest government program is Medicaid, which on the CPS includes those enrolled in the State Children’s Health Insurance Program (SCHIP). Medicaid and SCHIP covered 35.6 million low-income individuals (12 percent of the population), primarily children, pregnant women, elderly, and disabled people.1
  • Finally, the smallest coverage source was military/veteran coverage, providing insurance to 10 million people, or 4 percent of the population.
Source: ASPE tabulations of the 2004 Current Population Survey


The Uninsured by Income

The 45 million uninsured are more likely to be poor and low income than higher income. Figure 2 shows that over half of the uninsured are below 200 percent of the Federal Poverty Level (FPL), with 25 percent below FPL and 30 percent between 100 percent and 199 percent FPL.3 That the uninsured are concentrated among lower-income individuals is not surprising, given that low-income individuals are less likely to:

  • be working, and if they do work they are less likely to be working full-time;
  • receive an offer of insurance; and
  • be able to afford an offer of coverage.

Not all low-income individuals are eligible for Medicaid. Medicaid eligibility is based on a combination of income and population “category”. The population groups that qualify for Medicaid are generally children, parents of dependent children, pregnant women, the disabled, and the elderly. The income levels at which these groups qualify differs from state to state and group to group, generally with coverage of children and pregnant women being available at higher income levels, followed by the disabled and elderly, then parents of dependent children last. Childless adults who are not disabled or elderly rarely qualify for Medicaid, even at the very lowest income levels.

Source: ASPE tabulations of the 2004 Current Population Survey

Although the income distribution of the uninsured is skewed toward those with lower incomes, Figure 2 shows 28 percent of the uninsured have incomes above 300 percent of FPL, with 11 percent earning above 500 percent FPL.4 That the uninsured comprise non-trivial percentages of middle- and upper-income individuals is surprising. Those with incomes above 300 percent FPL should generally find employer insurance affordable. Data from employers shows that average single coverage premiums for employer-sponsored insurance represent 1.9 percent of income at 300 percent FPL, and average family coverage premiums represent 4.4 percent of income for a family of four at 300 percent FPL (with a higher percentage for smaller families).

The Uninsured by Work Status

The vast majority of the uninsured are working individuals or the children of those who work.5 In 2003, almost half of the uninsured (46 percent) worked full-time, and another 28 percent worked part-time or for part of the year. Many of the uninsured worked for firms that did not offer coverage, or if their employers offered coverage, they either were not eligible or did not accept the offer. Based on data from the 2001 February Supplement to the CPS matched with the 2001 March Supplement to the CPS, 18 million workers were not offered coverage and another 6 million were not eligible for the coverage that their firm offered, representing 54 percent of the uninsured.6 In addition, there are 6.9 million workers and dependents that have declined employer coverage and remain uninsured (19 percent of the uninsured).7 These individuals are most likely to decline employer coverage because it was too costly: 3.8 million, 52 percent, said coverage was too expensive. The February-March match file shows another 2.9 million dependents living with a family member covered by employer-sponsored insurance. While there are no follow-up questions on the February CPS to determine why dependents are uninsured, one can surmise that many of those dependents could have been insured under the covered worker’s employer plan but the worker found it unaffordable to purchase family coverage.8

Part-time workers comprise a disproportionately large percentage of the uninsured because employers often do not offer coverage to part-time workers, and because part-time income may make offered insurance less affordable. The median family income of part-time workers is about $14,000 less than the median family income of full-time workers, $47,500 vs. $61,700.


The CPS-reported figure of 45 million uninsured individuals represents the number of uninsured for a full year.9 However, there are other ways to measure the uninsured, such as those uninsured at a given point-in-time, and those who were ever uninsured for some length of time during the year. The “ever uninsured” figure is of particular policy relevance because it reveals how many individuals faced the significant financial risk of having a medical emergency that would have to be paid for out of pocket. According to the Medical Expenditure Panel Survey (MEPS), there were 64 million people who faced at least one month without
coverage in 2001.10

The MEPS data demonstrate how the uninsured population is not one unchanging group of individuals, but rather a constantly changing group, mirroring the changing nature of employment and income in the economy. While a significant percentage of the uninsured are without coverage for a full year (and longer), an equally significant percentage are uninsured for short periods of time. Of those 64 million who lacked coverage at some point in 2001, 51 percent were uninsured for at least one year. But one-in-five (20 percent) of the uninsured that year were without coverage for three months or less, and one-in-three (34 percent) were uninsured for 6 months or less. Clearly, some people face long-term problems obtaining coverage, either due to their inability to afford coverage or to being employed in jobs that do not offer coverage. Others, by contrast, face short-term periods without coverage as they move between jobs or go through other life transitions.


This analysis has presented an overview of the uninsured population. While the uninsured are concentrated disproportionately in certain subgroups, the uninsured are clearly a diverse population comprised of people from all income levels, racial groups, and employment types. The data presented in this report come primarily from the CPS, which is only one of four major government surveys that include information on the uninsured.11 Each survey has its advantages and disadvantages for purposes of measuring the uninsured. One distinction is that the CPS finds considerably fewer individuals enrolled in public coverage than are found in official program statistics. Perhaps as a result of this public program undercount, the CPS finds far more individuals without coverage for 12 months than other surveys. To obtain the most accurate picture of the uninsured, follow-up analysis is warranted regarding other government surveys, along with analysis that investigates the implications of the public coverage undercount.

Source: ASPE tabulations of the 2004 Current Population Survey

1 - According to official CMS program statistics, Medicaid is actually the larger program. Based on Administrative records, the CMS Office of the Actuary projects 2003 enrollment of 53 million, compared to the CPS 35.6 million. While CMS administrative totals also include some institutionalized individuals and some individuals who only receive aid with Medicare cost sharing, neither of which should report Medicaid on the CPS, the difference between CMS data and CPS data is still substantial. Also note that the CPS estimate for Medicaid includes children covered in SCHIP and a small number in other public programs. Further research is ongoing to refine the estimated number of people covered by Medicaid.

2 - The percentages do not add to 100 percent because individuals can have more than one type of insurance either simultaneously or sequentially during the year.

3 - The FPL in 2003 was $8,980 for a single individual and $18,400 for a family of four.

4 - In 2003, 300 percent FPL was $26,940 for a single individual and $55,200 for a family of four, and 500 percent FPL was $44,900 for a single individual and $92,000 for a family of four.

5 - For this memo, adults were labeled according to their own work status, but children were assigned to the parent with the “most work” during the year. That is, if there were a full-time/full-year worker in the family and a part-time/part-year worker, any children in the household would be labeled “full-time/full-year.”

6 - The numbers are on a base of 38 million uninsured in 2001. Contract workers, part-time workers, and in some cases workers who have not worked for the firm long enough are often not eligible for employer insurance.

7 - Data from a file matching the March and February supplements to the CPS. The March supplement contains detailed demographic and income data for the population. The February supplement contains questions about employer offers and worker take-up of insurance. The Actuarial Research Corporation performed the match.

8 - According to the 1999 Kaiser/HRET Employer Health Benefits Survey, 99 percent of firms that offer workers coverage also offer dependents coverage (though the employer contribution rate may be lower for dependent coverage).

9 - The structure of the CPS questionnaire elicits uninsured status for the entire preceding year. However, there has been considerable debate among researchers for many years as to whether the CPS was actually eliciting uninsured status at the time of the survey. The debate arose because the CPS figure of 45.0 million uninsured is actually far closer to other surveys’ point-in-time counts of the uninsured than those other surveys’ full-year uninsured counts. See ASPE Issue Brief, “Understanding Estimates of the Uninsured: Putting the Differences in Context,”

10 - See ASPE/AHRQ Issue Brief, “Research Note: The Long Term Uninsured,”

11 - See ASPE Issue Brief, “Understanding Estimates of the Uninsured: Putting the Differences in Context”.



An evaluation of the Cash and Counseling Demonstration revealed that when Medicaid beneficiaries of various ages and disabilities have the option to direct their own supportive services (“consumer direction”), their quality of life is improved, satisfaction with services is increased, unmet needs for care are reduced, access to home care is increased, and nursing home usage is reduced—without compromising the beneficiaries’ health or safety.

Medicaid beneficiaries who have disabilities and receive personal assistance services from home care agencies frequently have little control over their care. As a result, some are dissatisfied, have unmet needs, and report diminished quality of life. This report1 describes how consumer direction affects these aspects of care quality relative to agency-directed services.


In 1995, the Robert Wood Johnson Foundation and the Office of the Assistant Secretary for Planning and Evaluation launched Cash and Counseling, a demonstration program in three states—Arkansas, New Jersey, and Florida—to stimulate and strengthen consumer direction and choice in long-term care. The Cash and Counseling program provides a self-directed, individualized budget to recipients of Medicaid personal care services or home- and community-based services. Each person’s allocation is comparable to the value of services that he or she would have received through a traditional agency. Program participants use the allocation to purchase their own care—with the option of hiring friends, family members, or others—instead of receiving it from an agency. They can also use their budgets to modify their homes or vehicles, or to purchase a range of items that will help them live independently. Participants who are unable or unwilling to manage their care themselves may designate a representative, such as a family member, to help them or do it for them. Consulting and bookkeeping services are available to help participants weigh their options and keep up with required paperwork. A number of studies are available on the ASPE website, with more scheduled for release over the coming year. The study described in this article concerns the effects of consumer direction on the quality of Medicaid personal assistance services in Arkansas.




Michael J. O'Grady, Ph.D.
Assistant Secretary for Planning and Evaluation

Barbara Broman
Acting Deputy Assistant Secretary, Office of Human Services Policy

Donald A. Young, MD
Acting Principal Deputy Assistant Secretary for Planning and Evaluation

Ruth Katz
Acting Deputy Assistant Secretary, Office of Disability, Aging, and Long-Term Care

James Scanlon
Acting Deputy Assistant Secretary, Office of Science and Data Policy

Kimberley L. Berlin

Kim Axelrod

Lori S. Hummel

Ling Ban, Susan Marie Belsinger, Jane Hou

Jana Liebermann


Demonstration enrollment, which occurred between December 1998 and April 2001, was open to interested Arkansans who were at least 18 years old and eligible for Personal Care Services (PCS) under the State Medicaid plan. After a baseline survey, the 2,008 enrollees were randomly assigned to direct their own PCS as IndependentChoices participants (the treatment group) or to receive services as usual from agencies (the control group). IndependentChoices participants had the opportunity to receive a monthly allowance, which they or their representatives could use to hire their choice of caregivers (except spouses) and to buy other services or goods needed for daily living. Treatment group members also availed themselves of consultant and fiscal service options. Nine months after baseline, treatment and control group members were asked questions about disability-related adverse events and health problems. They were also asked to rate the following quality indicators: 1) satisfaction with care, 2) unmet needs for assistance with daily activities, 3) quality of life, 4) general health status, 5) self-care, and 6) ability to perform daily activities without help from others.


An independent evaluation by Mathematica Policy Research Inc. found when Medicaid beneficiaries of various ages and disabilities have the option to direct their own supportive services and hire their own caregivers, their quality of life is improved, satisfaction with services is increased, unmet needs for care are reduced, access to home care is increased, and nursing home usage is reduced—without compromising beneficiaries’ health or safety (relative to a randomly assigned control group that received services from agencies). Satisfaction with overall care and quality of life, for example, increased significantly for all treatment group participants, but the results were especially dramatic for the 18-64 year old age group. Consumer-directed Medicaid personal assistance warrants stakeholders’ consideration. Future analyses will examine the effects of IndependentChoices on the use and costs of personal care services and other health care services, the experiences of informal and paid caregivers, and program implementation.


This proven model of consumer-directed supportive services is currently being expanded in 11 more states, allowing thousands more older adults and people with disabilities to have choice and control over the care they receive. Alabama, Iowa, Kentucky, Michigan, Minnesota, New Mexico, Pennsylvania, Rhode Island, Vermont, Washington, and West Virginia have received three-year grants from the Robert Wood Johnson Foundation to replicate and expand upon the successful Cash and Counseling model. The Office of the Assistant Secretary of Planning and Evaluation and the Administration on Aging are partnering with the Foundation to fund technical assistance. Unlike the prior demonstration, the new round of Cash and Counseling will not include control groups. Grantee States will need to secure a Section 1915c or an 1115 waiver in order to implement a consumer-directed individual budget model for Medicaid. The Boston College Graduate School of Social Work serves as the National Program Office to administer the grants and provide oversight and technical assistance for the new program.

This report was prepared by Pamela Doty for the Office of the Assistant Secretary for Planning and Evaluation, Office of Disability, Aging and Long-Term Care Policy. For more information on this and related issues, visit the ASPE website:

1 - “Does Consumer Direction Affect the Quality of Medicaid Personal Assistance in Arkansas?” The full report is available online at