Participation in the Qualified Medical Beneficiary Program.

This article has three objectives: to estimate how many eligible elderly beneficiaries are participating in the Qualified Medicare Beneficiary (QMB) program; to determine the characteristics of participating and non participating eligibles; and to identify the most significant barriers to program participation. We used data from the Medicare Current Beneficiary Survey (MCBS) and the Medicare Buy-In file. We found that 41 percent of QMB eligibles are enrolled in the program; participation is higher for poor and less educated beneficiaries, those in poorer health, rural residents, African Americans, and Hispanics. Finally, we found that, in general, eligible beneficiaries are ill informed about the program.

Introduction

The Medicare program's cost-sharing provisions—its premiums, deductibles, and copayments—can present a substantial financial hardship for low-income beneficiaries. To alleviate some of this burden, Congress enacted the QMB program, which requires State Medicaid programs to pay Medicare cost-sharing amounts for low-income Medicare beneficiaries. Since the program began in 1990, however, policymakers and advocates for the elderly have been concerned about low program participation, despite attempts to inform eligible seniors about the benefit.

The objectives of this article are threefold: to estimate how many eligible elderly beneficiaries are participating in the program; to determine the characteristics of participating and non-participating eligibles, and to identify the most significant barriers to program participation. The first section provides background information on the QMB program; the second describes our data and methods; and the third presents our results. The final section discusses the implications of our findings and important areas for further research.

Background

States have always had the option to pay Medicare premiums and deductibles for beneficiaries who qualify for Medicaid. Since 1990, Federal law has required State Medicaid programs to pay the cost-sharing provisions for all Medicare beneficiaries whose incomes do not exceed 100 percent of the Federal poverty level (FPL) and whose resources do not exceed twice the amount established for Supplemental Security Income (SSI) eligibility. Individuals who meet these criteria are termed QMBs.

Since the creation of the QMB program, there have been a number of attempts to enroll QMB-eligible beneficiaries. HCFA has undertaken a number of efforts, including: mailing notices about the QMB program to potentially eligible beneficiaries; mailing Medicare Part A application forms to 250,000 low-income seniors eligible for the QMB program; disseminating copies of a QMB leaflet to supermarkets and other locations; distributing public service announcements; publishing news and feature articles in magazines and newspapers geared toward older Americans, including information in the Medicare Handbook; and advising beneficiaries about a toll-free Medicare hotline for information about the program (Neumann et al., 1994; U.S. General Accounting Office, 1994). Advocacy groups and individual States have also undertaken QMB outreach activities.

Despite these efforts, reports have indicated that many eligible individuals are not participating in the program. A study by Families USA (1992) reported that approximately 2 million of the 4.2 million eligible seniors were not enrolled. A subsequent report by the U.S. General Accounting Office (1994) confirmed the general accuracy of this estimate. Anecdotal reports indicate that some States have not aggressively enrolled eligible individuals, in part because they would rather pay for needed services through the Medicaid program. Little is known about the characteristics or motivations of participating and non-participating eligibles.

Data and Methods

We used the 1992 Income and Assets (I&A) Supplement to the MCBS to identify a sample of elderly, non-institutionalized beneficiaries who met the eligibility criteria for the QMB program. The HCFA-sponsored MCBS is an ongoing survey designed to enable researchers to examine the current status of the Medicare population (Stone, 1993). The survey consists of a series of interviews conducted 3 times a year with a stratified random sample of approximately 12,000 aged and non-aged Medicare beneficiaries, focusing on respondents' health care utilization and expenditures, as well as their health status, family support, living arrangements, and financial resources. The I&A Supplement collects detailed information about beneficiaries' financial resources, including sources of income and assets. We identified respondents as QMB-eligible if their incomes did not exceed 100 percent of the FPL and their assets did not exceed twice the amount established for SSI eligibility.

Next, we developed a questionnaire designed to examine beneficiaries' knowledge of the QMB program, their sources of information, and, for non-enrollees, their reasons for not participating. In the spring of 1993, this questionnaire was fielded as a MCBS survey supplement (the QMB supplement) to the sample of individuals identified as being eligible for the QMB program.

We merged the data from the QMB supplement with two other data bases containing information on our sample of QMB-eligibles—one incorporating data from the MCBS core survey on characteristics of our eligible population and the other containing information from HCFA's 1993 Medicare Buy-In file, which was used to determine whether eligible beneficiaries were actually enrolled in the program. We conducted bivariate analyses using this comprehensive data base to describe the eligible population and beneficiaries' tendencies to enroll based on certain characteristics. We also developed a logistic regression model to predict QMB participation.

Results

As shown in Table 1, approximately 41 percent of eligible elderly beneficiaries (1.9 million individuals) were participating in the QMB program as of 1993.

Table 1

QMB Program Participation: 1993

Participation Status	Number(in Millions)	Percent
Eligibles	4.67	100
Participating	1.93	41
Non-Participating	2.74	59

NOTE: QMB is Qualified Medicare Beneficiary.

SOURCE: (Neumann et al., 1994).

Socioeconomic and Demographic Characteristics

Table 2 shows participation in 1993 among individuals with selected socioeconomic and demographic characteristics. Women, who comprised 73 percent of the eligible population, had higher participation rates (44 percent) than men (35 percent). Participation did not vary greatly with age, though it was slightly higher for the oldest beneficiaries.

Table 2

QMB Program Participation, by Socioeconomic and Demographic Characteristics: 1993

Characteristic	Eligibles	Participating	Non-Participating
	Percent
Gender
Male	27	35	65
Female	73	44	57
Age
65-74 Years	48	42	58
75-84 Years	38	40	60
85 Years or Over	13	43	57
Income
Less Than $1,000	68	48	52
$1,000-1,900	9	12	88
$2,000-3,999	3	12	88
$4,000-5,999	6	47	54
$6,000-7,999	3	36	64
$8,000-9,999	1	28	72
Education
1st Grade or Less	6	69	31
2nd-5th Grade	19	53	47
6th-8th Grade	28	44	56
9th-11th Grade	19	36	64
12th Grade	18	29	71
1-3 Years of College	6	23	77
4 Years of College or More	3	26	74
Race
Native American	1	48	52
Asian-American	3	68	32
African-American	22	51	49
White	69	37	63
Other	4	43	57
Marital Status
Married	28	30	70
Widowed	53	42	58
Divorced	9	58	42
Separated	3	52	48
Never Married	6	55	45
Number of Living Children
0	14	41	59
1	16	35	66
2	18	36	64
3	15	37	63
4	12	33	67
5 or More	25	57	43
Location
Urban	69	40	60
Rural	31	44	56
Region
Northeast	18	36	64
Midwest	15	37	63
South	45	49	52
West	22	41	59

NOTE: QMB is Qualified Medicare Beneficiary.

SOURCE: (Neumann et al., 1994).

Participation varied with income and was somewhat higher for beneficiaries in the lowest income category. Participation fell steadily with increasing level of education. About 69 percent of those in the lowest education category participated, compared with only 25 percent of those with at least some college experience. Across racial groups, Asian-Americans had the highest participation (68 percent), followed by African-Americans (51 percent), Native Americans (48 percent), and white persons (37 percent).

Married beneficiaries had lower participation rates (30 percent) than beneficiaries who were widowed (42 percent), divorced (58 percent), separated (52 percent), or never married (55 percent). Participation varied with the number of beneficiaries' living children, and was highest for those with 5 or more (57 percent). Rural residents had slightly higher participation rates (44 percent) than urban residents (40 percent). Across regions, participation was highest in the South (49 percent), followed by the West (41 percent), the Midwest (37 percent), and the Northeast (36 percent).

Health Status, Utilization, and Insurance

Table 3 shows participation by key health status, health utilization, and health insurance variables. Participation was highest among those responding that they were in fair health (52 percent) or poor health (48 percent), and lowest among those who said that their health was excellent (30 percent) or very good (29 percent).

Table 3

QMB Program Participation, by Health Status, Utilization, and Insurance: 1993

Variable	Eligibles	Participating	Non-Participating
	Percent
General Health1
Excellent	13	30	71
Very Good	20	29	71
Good	29	43	57
Fair	26	52	48
Poor	13	48	53
Hospital Visits2
0	84	41	59
1	13	43	57
2	3	50	50
3	1	52	48
Physician Visits2
0	14	27	73
1	11	33	67
2	15	33	67
3	21	45	55
4	39	50	50
Emergency Room Visits2
0	66	35	65
1	23	53	47
2	8	59	41
3	2	44	56
4	0	71	29
On Medicaid1
Yes	45	88	12
No	55	4	97
SSI
Yes	29	95	5
No	69	19	81
Receive Welfare Income
Yes	18	83	17
No	80	32	68
Other Private Health Insurance
Yes	25	12	88
No	75	51	49

Self-reported.

Maximum of one per MCBS round.

NOTES: QMB is Qualified Medicare Beneficiary. SSI is Supplemental Security Income. MCBS is Medicare Current Beneficiary Survey.

SOURCE: (Neumann et al., 1994).

Participation was higher for beneficiaries who made greater use of health services (i.e., more hospital, physician, and emergency room visits). For example, about 52 percent of those with 3 hospital visits during the previous year and a half participated, as opposed to 50 percent for beneficiaries with 2 visits, 43 percent with 1 visit, and 41 percent with no visits. The pattern was similar for utilization of physician services and emergency rooms.

Participation was much higher among Medicaid recipients (88 percent) than for non-recipients (4 percent). It was also very high among those receiving SSI (95 percent) and welfare income (83 percent). About one-fourth of the respondents indicated that they had other private health insurance; within this group, 12 percent were participating in the program.

Predicting Participation: Regression Results

Table 4 shows the results of multivariate analyses in which we model the probability of QMB participation using logistic regression techniques. We report the normalized logistic coefficients, which can be interpreted as estimates of the change in the probability of program participation for a given change in the independent variables. We also report the standard errors of these marginal effects.

Table 4

Logistic Regression Model of QMB Program Participation

Variable Type and Name	Normalized Coefficients1
Intercept	-0.013(0.069)
Demographic Characteristics
Female	***0.114(0.040)
Age in Years	0.002(0.034)
Age Squared	-0.000(0.000)
African-American	**0.100(0.039)
Hispanic2	***0.164(0.055)
Other Race2	**0.137(0.061)
Family Characteristics
No Living Children3	***-0.221(0.060)
1 or 2 Living Children3	***-0.189(0.044)
3-5 Living Children3	***-0.129(0.044)
Married4	***-0.229(0.054)
Never Married4	0.134(0.082)
Widowed4	***-0.177(0.048)
Education, Income, and Wealth
6 Years of Education or Fewer5	***0.196(0.064)
7-8 Years of Educations	0.100(0.065)
9-11 Years of Education5	0.027(0.066)12
Years of Education5	-0.007(0.069)
Family Income	*-0.033(0.019)
Own Home6	***-0.213(0.040)
Rent Home6	***0.168(0.040)
Health Status
Excellent7	*-0.112(0.063)
Very Good7	***-0.157(0.057)
Good7	-0.075(0.050)
Fair7	0.022(0.049)
1 or More ADL Limitations8	**0.112(0.051)
2 or More ADL Limitations8	***0.209(0.055)
Other
Live in Metropolitan Area	-0.046(0.038)
Midwest9	**0.166(0.055)
South9	***0.254(0.046)
West9	***0.210(0.051)

Significant at the 0.01 level.

Significant at the 0.05 level.

Significant at the 0.1 level.

Numbers in parentheses are standard errors.

Reference category: white, non-Hispanic.

Reference category: 6 or more living children.

Reference category: divorced or separated.

Reference category: more than 12 years of education.

Reference category: living with someone else.

Reference category: poor health status.

Reference category: no ADL limitations.

Reference category: Northeast.

NOTES: QMB is Qualified Medicare Beneficiary. ADL is activity of daily living. The logistic coefficients and standard errors are multiplied by p*(1-p*) where p* is the sample means of the dependent variable. In these models, p* = 0.4132. The normalized coefficients are estimates of the change in the probability of enrolling in the QMB program given a change in the independent variable. All estimation procedures are weighted.

SOURCE: (Neumann et al., 1994).

The regression results indicate that, controlling for other variables, females had significantly higher participation than males (about 11 percentage points higher). Race exerts a strong, independent effect: African-Americans and Hispanic Americans had significantly higher participation than non-Hispanic whites (between 10-16 percentage points higher). Those with less education had significantly higher enrollment rates. For example, participation for those with 6 years of education or fewer was about 20 percentage points higher than for those with at least some college. There is a small but significant negative income effect. Those with higher incomes have lower participation rates.

Other significant variables include home ownership, health status, and region. Compared with those who neither own nor rent (i.e., they live with others), homeowners had lower participation (about 21 percentage points), while renters had higher participation (about 17 percentage points). The health status effect was strong and significant; those in excellent, very good, or good health had considerably lower participation rates than those in poor health. Eligible individuals in Western and Southern States had much higher enrollment rates than those in Eastern States (about 25 and 21 percentage points higher). Midwestern States had enrollment rates 12 percentage points higher than Eastern States. Functional status also affects QMB participation. The results show that those with one or more limitations in activities of daily living had higher enrollment rates than those with no such limitations.

Knowledge of the Program

Table 5 shows how much respondents knew about the QMB program. The table shows that very few eligible beneficiaries (7 percent) had ever heard of the QMB program; participation was higher for those who had (60 percent). Only 5 percent of eligibles believed that they were enrolled as QMBs. Of this group, 94 percent were, in fact, enrolled. Of the 6 percent who reported that they were not enrolled, 19 percent actually were. Of the 88 percent who did not know whether or not they were QMBs, almost 40 percent were.

Table 5

Eligible Beneficiaries' Knowledge About the QMB Program

Responses	Eligibles	Participating	Non-Participating
	Percent
Have You Heard Of The QMB Program?
Yes	7	60	40
No	91	40	61
Don't Know	2	58	42
Are You A QMB?
Yes	5	94	6
No	6	19	81
Don't Know	88	40	60
Have You Applied For The QMB Program?
Yes	1	39	61
No	6	18	82
Inapplicable	93	43	57

NOTE: QMB is Qualified Medicare Beneficiary.

SOURCE: (Neumann et al., 1994).

Reasons for Non-Participation

Finally, Table 6 shows the major reasons eligible non-enrollees provided for not enrolling in the program. Most said that they did not need it (33 percent) or that they did not think they qualified for it (27 percent). Sixteen percent of respondents said that they did not know about it. Others stated that it was too much trouble (7 percent), they just didn't do it (3 percent), and that they didn't want welfare (3 percent).

Table 6

Reasons Given for Not Participating in the QMB Program

Reason	Percent
Don't Need It	33
Don't Think I Qualify	27
Didn't Know About It	16
Too Much Trouble	7
Just Didn't/No Reason	3
Don't Want Welfare	3
Don't Know How	1
Couldn't Get Out To Do It	1
Applied/Didn't Qualify	0
Other	9

NOTE: QMB is Qualified Medicare Beneficiary.

SOURCE: (Neumann et al., 1994).

Discussion

The results reveal 3 major findings: (1) many eligible beneficiaries are not participating in the QMB program; (2) those who do participate tend to be those most in need of QMB benefits; and (3) on the whole, eligible beneficiaries are poorly informed about the program. We discuss each of these in turn.

First, the program is not serving many individuals for whom it is intended. Our analyses indicate that well over 2 million eligible elderly beneficiaries are not participating. Participation remains low even among truly needy individuals. Over 50 percent of those reporting incomes under $1,000 do not participate, for example. Over 50 percent of those who had at least 1 hospital visit over the previous year and a half (and who therefore incurred a $600 deductible per hospitalization) do not participate. Data also suggest that some eligible beneficiaries are purchasing supplemental insurance coverage, despite the fact that the QMB program is designed to cover most of their out-of-pocket health costs.

Second, while participation remains low, the beneficiaries who do participate tend to be those most in need of the program. Beneficiaries enrolled in other government assistance programs, for example, are very likely to be enrolled as QMBs. Over 87 percent of dually eligible beneficiaries (those receiving both Medicare and Medicaid) participate, as do 95 percent of SSI recipients, and 82 percent of those receiving other welfare income.

In general, participation is higher among eligible beneficiaries with the lowest incomes and highest health care utilization. Among eligibles, these are the two subgroups most vulnerable to Medicare out-of-pocket costs—lower income beneficiaries (because they have less money to pay such expenses) and heavy users (because they are likely to incur additional out-of-pocket costs). Non-participating eligibles who are not heavy users of medical services are less burdened with deductibles and copayments (though they are still assessed the monthly Medicare Part B premium).

Participation is also higher among other vulnerable populations—for example, less educated and more socially and geographically isolated beneficiaries. Individuals with less than a sixth grade education are much more likely to participate than those who have completed high school. Widowed, divorced, or never married individuals are more likely to participate than married beneficiaries. Rural residents have higher participation rates than urban residents.

The third finding to emerge from this study is that most eligible beneficiaries are ill-informed about the QMB program. Only 7 percent of eligibles had ever heard of the QMB program; of the 93 percent who have not heard or did not know about the program, approximately 40 percent were actually enrolled. Among non-participants, the most frequently provided reasons for not enrolling were that they did not believe they needed the program (33 percent), they did not think they qualified (27 percent), or they were not aware of the program (16 percent).

These results are consistent in many respects with previous findings on participation in means-tested government programs. Several other studies have noted the problems of low enrollment in such programs. A report by ICF Incorporated (1988) found that only 52 percent of elderly individuals eligible for the SSI program actually participated. Studies by the U.S. Department of Agriculture (Dole and Beebout, 1988) and the U.S. Congressional Budget Office (1988) estimated elderly participation in the Food Stamp program as between 41 and 66 percent.

Previous research has also suggested that the neediest individuals have the highest participation rates. Hollenbeck and Ohls (1984) and Lewin/ICF (1989) found that poor health status and participation in other subsidized government programs had a positive effect on participation in the Food Stamp and SSI programs. Akin, Guilkey, and Popkin (1985) and Blanchard et al. (1982) found that senior citizens' participation in the Food Stamp program rose as income declined. Survey findings reported by Louis Harris & Associates (1986) indicated that SSI participation was higher among elderly Americans who were both poor and living alone.

Previous research has also pointed to informational barriers as an important reason for low enrollment. Coe (1983), for example, found that such barriers were a significant obstacle for potential Food Stamp recipients; more than 40 percent of those meeting the eligibility criteria did not think they were entitled to the benefit. Louis Harris and Associates (1987) reported that among those who were eligible but not participating in the SSI program, 43 percent believed themselves to be ineligible and another 43 percent were unsure of their status.

These findings, taken together with our findings on the QMB program, underscore an important truth about means-tested government programs: Simply legislating that certain individuals are eligible does not ensure their participation. Even with aggressive outreach efforts, many eligible individuals do not enroll. Many do not receive or comprehend outreach information. Some who suspect they are eligible have trouble accessing the system. Others refuse to enroll because of the stigma of welfare. These factors have important implications for policymakers planning and administering State health care reform initiatives. Health reform initiatives to cover previously uninsured populations should be accompanied by carefully targeted outreach to low-income populations not accustomed to utilizing health services on a regular basis. Outreach efforts should seek to educate these populations about newly available benefits and to provide guidance on appropriate ways to access the health care system.

Our results suggest that a number of areas for further research would be fruitful. It would be useful to link data on QMB enrollment to information on beneficiaries' actual out-of-pocket spending for health care; specifically, it would be helpful to know the percentage of after-tax income that enrollees and non-enrollees devote to medical care.

It would also be useful to link information from claims data on the actual utilization and expenditures of QMB eligibles. Doing so would provide a more complete profile of the health experiences of QMB eligibles and enable us to examine, more precisely, the experience of enrollees and non-enrollees. For example, controlling for factors such as age and gender, do enrollees use more medical services and incur higher expenditures than non-enrolled eligibles? Other important questions include whether non-participating eligibles refrain from using medical services in an attempt to avoid cost-sharing requirements, and whether there is any discernible differential between the enrolled and non-enrolled groups in patient outcomes such as mortality or morbidity.

Another area for further investigation involves the Medicaid eligibility status of QMB eligibles. In part because of coding inconsistencies among States (Sparacino, 1994), it was difficult in this study to determine the precise Medicaid eligibility status of those receiving buy-in benefits. The distinction is important because it may shed light on how beneficiaries become enrolled in the QMB program. For example, some Medicare beneficiaries who qualify for Medicaid (and are therefore dually eligible) are receiving the QMB benefit because it comes as part of the Medicaid package, though they do not file a separate QMB application. Others have incomes too high to qualify for Medicaid but low enough to qualify for QMB—these individuals comprise the QMB-only population. Our findings suggest that the system does a good job providing buy-in benefits to dual eligibles, but is less successful at identifying and enrolling QMB-onlys. More precise documentation of this phenomenon is needed.

Better indicators of Medicaid eligibility status would also shed light on beneficiaries' knowledge about the QMB program. Our findings suggested that very few beneficiaries had ever heard of the QMB program. Even some individuals who receive the buy-in benefit may not be familiar with the term QMB because they receive the benefit automatically through their Medicaid eligibility.

Finally, we need to better understand what strategies are most effective in enrolling eligible beneficiaries. Evidence on the successes and failures of past outreach projects is largely anecdotal; more rigorous evaluations are needed. Demonstration projects which test the impact of alternative outreach strategies could provide valuable insights to government agencies and private organizations that seek cost-effective ways of increasing enrollment rates. As we proceed with such projects, we should keep in mind that the most effective strategies will not be one-time efforts. The results presented here highlight the fact that enrolling eligible beneficiaries in the QMB program will be a difficult and ongoing challenge.