Kimball Lewis, Marilyn Ellwood, and John L. Czajka
Mathematica Policy Research Inc.
Office of Health Policy
Office of the Assistant Secretary for Planning and Evaluation
U.S. Department of Health and Human Services
It became clear early on in the research for this report that many complex issues surround the process of estimating the health insurance status of America's children. We are indebted to various individuals and organizations who helped us to sort through and understand these issues. First and foremost, we would like to thank Laura Brice and Christy Schmidt of ASPE who supported this project and provided thoughtful guidance throughout its course. We would also like to thank the members of the advisory panel, who provided extremely helpful comments on the first draft of this report. The advisory panel members were: Linda Bilheimer of the Congressional Budget Office; Paul Fronstin of the Employee Benefit Research Institute; John Holahan of The Urban Institute; Gene Lewit of the Packard Foundation; and Kathy Swartz of the Harvard School of Public Health. Dave Baugh and Roger Buchanan of HCFA(now known as CMS) provided us with useful notes from their interviews of the various persons who have estimated the number of uninsured children. Shruti Rajan and Beth Kessler of The Urban Institute conducted simulations for us of Medicaid eligibles using the TRIM2 model. Jim Reschovsky of the Center for Studying Health Systems Change provided information about the Community Tracking Study data. At MPR, Jan Watterworth did the library research, Julie Sykes provided valuable comments, Marsha Gold reviewed an initial draft, Larry Radbill helped us with our description of the SIPP, and Melanie Lynch prepared the manuscript. The authors gratefully acknowledge the contributions of these individuals but accept full responsibility for any errors that remain.
Chapter I. Introduction
Policymakers are currently considering proposals aimed at reducing the number of children without health insurance. The debate over various proposals could benefit from better information about the uninsured child population. To start, there is a lack of consensus about the number of uninsured children and the extent to which some of the uninsured are eligible for Medicaid but not participating. This study reviews the literature on children's health insurance pattens and Medicaid program participation. Specifically, this literature review focuses on (1) uninsured children -- how many are there, what are their characteristics, how long are they uninsured, and why are they uninsured; and (2) Medicaid eligibility, enrollment, participation rates, program dynamics, and measurement issues.
This literature review is the first task in an eight month research contract awarded by the Department of Health and Human Services, Office of the Assistant Secretary for Planning and Evaluation (ASPE), to Mathematica Policy Research, Inc. (MPR). The overall study objective is to improve understanding of the issues involved in analysis of children's health insurance patterns. This literature review was designed to identify key analytic questions that are not fully answered from current research. Subsequent tasks will include the design and implementation of further analyses of uninsured children using the Survey of Income and Program Participation (SIPP) and possibly other data as well. These analyses will benefit from the literature review, particularly with regard to the identification of methodological issues in measuring children's health insurance patterns. This review also provides a basis of comparison for key estimates produced in the additional tasks of this effort.
The first step in the literature review was to develop general criteria for inclusion of studies. These criteria were as follows:
- the analyses had to focus on one of the following: (1) uninsured children -- how many are there, how long are they uninsured, and why are they uninsured; (2) Medicaid eligibility, enrollment, participation rates, program dynamics, and measurement issues; and (3) participation rates of low-income children in related government programs, such as the former Aid to Families with Dependent Children (AFDC) and Food Stamps.
- the selected studies had to include empirical data analysis
- the studies had to be recent (since 1990) except for earlier works that were considered seminal
We obtained studies through an automated database search, reference lists from identified articles, the contents of current health policy journals, and the current research reviews in health policy newsletters. For the automated database search, we used the National Library of Medicine's Medical Literature Analysis and Retrieval System (MEDLARS). We also consulted with the ASPE Project Officer and other researchers involved in studying the uninsured to identify additional published and unpublished works. Although we obtained nearly seventy articles that initially met the stated criteria for this review, we limited the total number of articles reviewed by excluding articles that used older data that have since been updated or that simply presented overviews of the work done by other researchers that were already included in the review.
We begin this review in Chapter II by addressing the question, How many children are uninsured? We present key estimates of the uninsured using the most recent data available. We compare and contrast the estimates and discuss methodological issues involved in using survey data to make the estimates. In Chapter III, we present the characteristics of uninsured children as described in the literature. In Chapter IV, we analyze the literature on the question, How many uninsured children are eligible for Medicaid and what is the Medicaid participation rate for children? In this chapter we also describe the various ways that researchers have used survey data to estimate eligibility for the Medicaid program. As a point of comparison for Medicaid participation rates, we present literature on the participation rates for the AFDC and Food Stamp programs. In Chapter V, we analyze the literature on the characteristics of uninsured children who are eligible for Medicaid. In Chapter VI, we summarize our findings and discuss the next steps in this project.
Chapter II. How Many Children are Uninsured?
Although most researchers agree on the general trends of health insurance coverage -- that is, that the rate of private or employer-sponsored coverage is falling while public sponsored coverage is rising -- they often disagree on the actual number of uninsured, or even how the uninsured should be defined and measured. For example, although the most widely cited estimate of the number of uninsured children in America is about 9.8 million, which is based on the March 1996 Current Population Survey (CPS), there is debate over whether this estimate is closer to the number of uninsured at a point in time or the number uninsured throughout the year. Moreover, estimates of the uninsured using alternative data sources, or using CPS data that have been edited because of problems with the reporting of Medicaid, vary from the basic CPS estimate by as much as one-third.
This chapter reviews the literature on the recent estimates of the uninsured and why the estimates from various databases differ. We review the estimates of the uninsured based on the following databases: the March CPS, the SIPP, the National Health Interview Survey (NHIS), the National Medical Expenditure Panel Survey (MEPS), the Community Tracking Study (CTS), and the Kaiser/Commonwealth Survey of Americans.
A. MEASURING THE UNINSURED
Estimates of the number of uninsured may differ for a number of reasons, such as which data are used for the estimate, how the data are interpreted, and the time-period over which the uninsured are defined. Monheit (1994) reviewed how estimates of the uninsured population are obtained and reached four general conclusions. First, a consistent, long-term series of estimates is not available. For example, the CPS, the most commonly used source for health insurance coverage, changed the content of its health insurance questions at various times, which has resulted in artificial changes in the number who are uninsured from one year to the next. Second, analysts using the same data may obtain varying estimates for the same time periods because the data can be interpreted in different ways. For example, estimates of the uninsured done by The Urban Institute using CPS data are lower than estimates by other groups using the same data because The Urban Institute adjusts their estimates for the known underreporting of Medicaid in the CPS. Third, as mentioned above, estimates across alternative data vary. And fourth, estimates vary depending upon the time-frame around which the estimate is made. For example, the number of persons uninsured throughout a given year will be less than the number of uninsured at a point in time which, in turn, will be less than the number of uninsured at any time during a given year.
B. CPS ESTIMATES OF THE UNINSURED
The most commonly cited estimates of the number of uninsured children -- those produced by the Census Bureau, the Congressional Budget Office (CBO), the U.S. General Accounting Office (GAO), the Employee Benefit Research Institute (EBRI), and The Urban Institute -- use the March CPS as their source. The CPS, which is the source of the official Government statistics on employment and unemployment, is a nationally representative monthly survey of approximately 57,000 households in the United States. The sample is based on the civilian noninstitutionalized population of the United States, which includes persons living in households and group quarters (for example, college dormitories and rooming houses), but does not include residents of institutions (for example, homes for the aged) and persons living abroad. As well as being nationally representative, the sample is also representative of each of the 50 states and the District of Columbia, although for most states the samples are too small for precise state-level estimates.
The main purpose of the survey is to collect, by means of personal interviews, information on the employment status of the population during the survey month. In addition, supplemental questions are regularly added to the core questionnaire on topics such as health, education, income, and previous work experience. The March CPS contains supplemental questions on the health insurance status of each person in the household in the prior calendar year. Specifically, respondents are asked whether they had any of various types of private or public health insurance in the previous year. Respondents are permitted to report more than one type of health insurance coverage, although it is impossible to tell from the data whether persons with multiple types of coverage had the coverage concurrently or at different times during the previous year. The health insurance portion of the March 1995 CPS questionnaire is presented in Appendix A.
Respondents are never asked directly whether they were uninsured in the previous year. Instead, estimates of the uninsured are calculated as a residual -- that is, the uninsured are all those who do not report having some type of coverage in the previous year. As a result, the uninsured are those without any coverage throughout the previous year. However, some researchers believe that the CPS estimates of the uninsured are too high and, thus, that respondents may be reporting their health insurance status as of the interview date. This and other issues pertaining to estimates of the uninsured according to the CPS are discussed below.
The two principal classes of estimates of the uninsured using the CPS are (1) those done by the Census Bureau, CBO, GAO, and EBRI; and (2) those done by The Urban Institute. These estimates are summarized in Table II.1. The Urban Institute's estimates of the uninsured differ from the other estimates because they adjust for the underreporting of Medicaid in the CPS.
1. Census Bureau, CBO, GAO, and EBRI Estimates
Beginning with the March 1995 CPS, when the health insurance questions were revised to eliminate the possibility of respondent inconsistencies, the Census Bureau (Bennefield 1996a), CBO
CPS ESTIMATES OF THE UNINSURED BY SOURCE
|Census Bureau, EBRI, CBO, and GAO, and others||1996 CPS||1995||Uninsured throughout 1995 (or point-estimate, depending on interpretation of CPS definition of uninsured)||Children age 17||9.8||13.8%|
|Children age 18||10.5||14.0%|
|Adults age 18-64||30.5||19.0%|
|Nonelderly age 0-64||40.3||17.4%|
|Urban Institute||1996 CPS||1995||Uninsured throughout 1995 (or point-estimate, depending on interpretation of CPS definition of uninsured). Adjusted for the Medicaid undercount in the CPS using the TRIM2 model.||Children age 17||6.9||9.8%|
|Children age 18||7.6||10.3%|
|Adults age 18-64||28.8||18.0%|
|Nonelderly age 0-64||35.7||15.5%|
(Bilheimer 1997), GAO (1997), and EBRI (Fronstin 1996) began publishing identical estimates of the number of uninsured.(1) Using the March 1996 CPS, they found the following:
- Children age 0 to 17: 9.8 million uninsured (13.8 percent of all children)
- Children age 0 to 18: 10.5 million uninsured (14.0 percent of all children)
- Adults age 18 to 64: 30.5 million uninsured (19.0 percent of all adults)
- All persons age 0 to 65: 40.3 million uninsured (17.4 percent of all persons)(2)
None of these organizations adjusted their estimates for the underreporting of Medicaid in the CPS.
2. The Urban Institute's Estimates
The Urban Institute's estimates of the uninsured differed from others because the Institute adjusted for the underreporting of Medicaid in the CPS. The Institute used its Transfer Income Model (TRIM2), a microsimulation model, to test for Medicaid eligibility among non-reporters of Medicaid and then selected individuals to participate so that the size of the resulting Medicaid population in the model matched Health Care Financing Administration (HCFA(now known as CMS)) administrative data according to age and disability status of all persons ever enrolled in Medicaid in a given year. Using the TRIM2 model with March 1996 CPS data, the Institute found the following:
- Children age 0 to 17: 6.9 million uninsured (9.8 percent of all children)
- Children age 0 to 18: 7.6 million uninsured (10.3 percent of all children)
- Adults age 18 to 64: 28.8 million uninsured (18.0 percent of the population)
- All persons age 0 to 65: 35.7 million uninsured (15.5 percent of all persons)
The estimate of 6.9 million uninsured children in 1995 is 30 percent lower than the CPS estimates that include no adjustment for the underreporting of Medicaid. In all, the Institute simulated 2.9 million children to participate in Medicaid who reported no health insurance coverage in the CPS.
Researchers debate whether the Institute's adjustment for the underreporting of Medicaid yields improved estimates of the uninsured. One potential problem is that the Institute's adjustment may overcompensate for the underreporting because the adjustment is based on administrative estimates of the number of persons ever enrolled in Medicaid during the year, while CPS estimates of the uninsured may reflect those uninsured at a point in time. Another potential problem is that although the Institute adjusts for Medicaid underreporting, it makes no adjustment to reported private employment coverage, which could be either over- or underreported. This is important because the uninsured are calculated as a residual and, therefore, accurate estimates of the uninsured require accurate estimates of coverage for all other types of insurance. Despite these potential problems, the fact remains that Medicaid is substantially underreported in the CPS and, therefore, will affect most estimates of the uninsured in one way or another. The issues of underreporting of Medicaid and whether the CPS estimates may reflect those enrolled at a point in time are described in more detail below.
3. CPS Health Insurance Measurement Issues
When interpreting estimates of the uninsured done by researchers using the CPS, it is important to understand that these estimates are affected by various measurement issues that specifically pertain to the CPS data. These issues include the following: the time-frame of the CPS measures of health insurance, Medicaid underreporting, and survey undercoverage of the population.
If respondents answer the CPS health insurance questions as intended -- that is, as coverage at any time during the previous year -- then estimates of the uninsured should be interpreted as those without coverage throughout the previous year. However, some researchers believe that the CPS estimates of the uninsured are too high and, thus, that respondents may be reporting their health insurance status as of the interview date. Swartz (1986) compared CPS estimates of the uninsured with estimates from three other surveys that asked respondents about their health insurance coverage as of the interview date. The three other surveys were: the National Medical Care Expenditure Survey (1977), the Health Interview Survey (1978), and the National Medical Care Utilization and Expenditure Survey (1980). She found that the CPS estimates more closely resembled the point-in-time estimates of these surveys. CBO agreed with Swartz and considers its own CPS-based estimates of the uninsured to be closer to a point-in-time estimate rather than an estimate of those uninsured throughout the previous year (Bilheimer 1997).
Although the CPS estimates may resemble point-in-time estimates of the uninsured, there is evidence that some respondents interpret the questions correctly and report their status as of the previous year. For example, in 1995, 15 percent of children enrolled in Medicaid according to the CPS also reported coverage by private health insurance (Fronstin 1996). These children are probably not reporting their current status, since it is unlikely that this many children would be covered by Medicaid and private insurance at the same time. Instead, they are probably reporting their status as of the previous year when they were covered by private insurance for part of the year and Medicaid for part of the year. Other researchers provide additional evidence that many respondents interpret the questions correctly. For example, Kronick (1989)(3) found that private employer-sponsored health insurance coverage in the CPS is more consistent with employment status in the previous year than in the interview month. In addition, the first round of the Medical Expenditure Panel Survey (MEPS), which asked respondents whether they were uninsured continuously from January 1, 1996 to their interview date 3 to 6 months later (and links their responses to employment related data), provided estimates that were strikingly similar to the CPS (Beauregard et al. 1977).(4)
In a more recent analysis, Bennefield (1996c) compared longitudinal data from the SIPP with the standard health insurance data from the CPS and with data from experimental questions on the March 1995 CPS that asked about current health insurance status. Bennefield's results indicated that CPS respondents interpreted the standard health insurance questions correctly and provided their health insurance status as of the previous year. However, he found that respondents may have had recall problems and failed to report some coverage and, as a result, the CPS estimates of the uninsured looked more like point-in-time estimates. Some researchers, though, doubt the usefulness of the experimental health insurance questions on the CPS because they yielded extremely large numbers of uninsured.(5)
Long and Marquis (1996) compared the 1993 March CPS estimates of the uninsured in 10 states with the findings from the Robert Wood Johnson Foundation (RWJF) Family Health Insurance Survey. The RWJF survey was administered to approximately 2,000 families each in Colorado, Florida, Minnesota, New Mexico, New York, North Dakota, Oklahoma, Oregon, Vermont, and Washington during 1993. The uninsured and those covered by Medicaid were oversampled. The content includes considerable detail on insurance status -- both current and throughout the previous year. Across the 10 states included in the RWJF survey, the CPS estimate of the uninsured for all persons (14.7 percent) fell between the RWJF estimate of the currently uninsured (15.7 percent) and the uninsured throughout the previous year (12.2 percent). Long and Marquis also examined each state individually and found that for 9 of 10 states, the CPS measure fell between the RWJF current and throughout the previous year measures; in the remaining state, the CPS estimate was above the RWJF estimate of the currently uninsured by only 0.2 percentage points. Long and Marquis concluded that using the CPS as if it were a measure of the currently uninsured generally will understate estimates of the uninsured at a point in time. However, the CPS measure was considerably closer to the RWJF currently uninsured estimate than the uninsured throughout the previous year. The CPS estimate was only 6 percent below the RWJF currently uninsured estimate, but 20 percent above the RWJF estimate of the uninsured throughout the previous year.
Overall, though, most researchers tend to agree that the CPS probably contains a mixed bag of reporting -- that is, some respondents report health insurance status during the previous year, some report it as of the interview date, and some fail to report it altogether -- which, in the end, yields estimates that are probably best interpreted as health insurance status at a point in time.
b. Medicaid Underreporting
One weakness of the CPS is that the number of persons reporting Medicaid is lower than the number of persons enrolled in Medicaid in a given year according to administrative data from HCFA(now known as CMS) -- the agency that administers the Medicaid program. This problem is often referred to as "underreporting." Underreporting is thought to occur because survey respondents may not admit to being covered because of the stigma associated with public assistance programs, because they are not currently receiving health services, or because they may not realize they are covered by Medicaid.(6)
Most estimates of the number of uninsured use CPS data that are not first adjusted for Medicaid underreporting. This could lead to overestimates of the uninsured if many of those that appear uninsured are actually enrolled in Medicaid. It is important to recall, though, that CPS estimates of the uninsured are calculated as a residual and, therefore, accurate estimates rely on the accurate reporting of both Medicaid and private coverage. Unlike Medicaid, though, it is not known whether private insurance is reported accurately using the CPS.
The evidence for underreporting on the CPS is usually based on comparisons with HCFA(now known as CMS) administrative data, which show the number of persons ever enrolled in Medicaid during a given year. This comparison is valid if CPS health insurance questions are answered as intended -- that is, health insurance status during the previous year. However, if the CPS provides a point-in-time estimate, as many researchers believe it does, then comparing CPS Medicaid enrollment estimates with HCFA(now known as CMS) administrative data will tend to exaggerate the problem of Medicaid underreporting, since the number of Medicaid enrollees at a point in time will always be less than those ever enrolled during a year.(7) Comparisons with HCFA(now known as CMS) data will also be exaggerated because persons in institutions, who often receive Medicaid, are not within the sampling frame of the CPS. However, most institutionalized Medicaid enrollees are elderly (in nursing facilities), while the estimates of the uninsured in this paper focus on the nonelderly. Despite the definitional differences between HCFA(now known as CMS) and CPS data, the apparent underreporting of Medicaid on the CPS is probably too large to be explained by time-period and sample frame differences alone.
Holahan et al. (1995) estimated that, in 1990, 19.6 million nonelderly individuals were enrolled in Medicaid according to the CPS versus 24.7 million according to HCFA(now known as CMS) administrative data -- a 21 percent difference. To be comparable to the underreporting estimate of Holahan et al., underreporting should be calculated in relation to what HCFA(now known as CMS) refers to as "Medicaid eligibles," which is defined by HCFA(now known as CMS) as those ever enrolled during the year.(8) In 1993, 34.3 million nonelderly individuals were enrolled in Medicaid according to HCFA(now known as CMS), a 15.5 percent difference from the CPS estimate. As shown in the top panel of Table II.2, Medicaid underreporting then rose to 19.8 percent in 1994 and 21.0 percent in 1995 (Fronstin 1997b; HCFA(now known as CMS) 1996).
Medicaid underreporting for children age 0 to 17 follows the same general trend as that for all nonelderly individuals, although the underreporting rate is slightly higher. In 1995, for example, 16.5 million children were enrolled in Medicaid according to the CPS versus 21.4 million according to HCFA(now known as CMS) data -- 22.9 percent underreporting (bottom panel of Table II.2; Fronstin 1997a; HCFA(now known as CMS) 1996). Therefore, the underreporting of Medicaid among children could be as high as 4.9 million
children in 1995, suggesting that many of the 9.8 million children thought to be uninsured according to the CPS could actually be enrolled in Medicaid.(9) However, some of the Medicaid underreporting could be due to (1) children with both Medicaid and private coverage in the previous year reporting only private coverage, and (2) the CPS survey systematically failing to obtain interviews from population subgroups that may receive Medicaid disproportionately.
CPS AND HCFA(now known as CMS) ESTIMATES OF MEDICAID ENROLLMENT,
NONELDERLY AND CHILDREN, 1992 TO 1995
(Numbers in Millions)
|Nonelderly Enrollees (Age 0-64)|
|HCFA(now known as CMS)||31.4||34.3||35.8||36.7|
|Children Enrollees (Age 0-17)|
|HCFA(now known as CMS)||18.4||20.2||21.0||21.4|
|Source: CPS enrollment numbers from EBRI (1997a and 1997b)
HCFA(now known as CMS) enrollment numbers from 2082.
Notes: Child enrollees = number of children age 0-14 plus one-half of the children age 15-20.
HCFA(now known as CMS) data represent those ever enrolled during the year. CPS data are best interpreted as those enrolled at a point in time.
That the number of children with Medicaid fell from 1993 to 1994 in the CPS may be an artifact of the mid-decade shift in the sample framework for the CPS (Swartz, 1997).
Underreporting of Medicaid on the CPS may actually be worse than is indicated by the rates in Table II.2 because HCFA(now known as CMS) data sometimes underreport the number of Medicaid enrollees. A few states submit to HCFA(now known as CMS) only their data on recipients, or persons who receive services in a given year, rather than enrollees, or all persons who were enrolled during the year. This is evidenced by the fact that the number of enrollees is either zero or almost the same as the number of recipients. For example, the difference between the number of enrollees and the number of recipients is 3 percent or less in 1995 in the following states: Connecticut, District of Columbia, Hawaii, Louisiana, New York, and Tennessee.(10)
Overall, little research has been done on the problem of Medicaid underreporting and its effect on the estimates of the uninsured, partly because, until recently, HCFA(now known as CMS) Medicaid enrollment data were not considered reliable. Recent improvements to the HCFA(now known as CMS) data, though, make this type of research more feasible. Such research is warranted since studies on the characteristics of the uninsured will be affected to the extent that those who appear to be uninsured may actually be enrolled in Medicaid.
c. Undercoverage of the Population
According to the Census Bureau (Bennefield 1995), all demographic surveys, including the CPS and the SIPP, suffer from undercoverage of the population. Undercoverage results from missed housing units in the sampling frame and missed persons within sampled households. The Census Bureau estimated that the overall CPS and SIPP undercoverage rate is about seven percent and that undercoverage varies with age, sex, and race. They reported that for some groups, such as 20 to 24 year old black males, the undercoverage rate is as high as 27 percent. The Census Bureau noted that even though their weighting procedures partially correct for the bias due to undercoverage, the final impact of undercoverage on estimates is unknown. This problem could bias estimates of the uninsured if the groups that are missed in the survey are either disproportionately insured or disproportionately uninsured.
C. SIPP ESTIMATES OF THE UNINSURED
The value the SIPP adds to analyses of the uninsured is that it allows researchers to examine the dynamic aspects of the uninsured that are not apparent in the oft-cited point-in-time estimates. For example, Swartz and McBride (1990) pointed out that data collected at a point in time from a population with dynamic movements are more likely to contain people who are in long spells without health insurance even though most people have fairly short spells (this phenomenon is described in more detail below).(11) In short, using point-in-time data to describe the characteristics of the uninsured presents a myopic picture of all individuals who lose health insurance. The SIPP's longitudinal data, though, can answer questions such as:
- How many are uninsured in at least one month of a given year?
- How many are uninsured throughout a given year?
- How do the number of uninsured in a one-year period compare with that of a 2-year or more period?
- What is the average duration of all spells of uninsurance? How does this compare to the average duration for all those uninsured at a point in time?
1. Overview of the SIPP
The SIPP is a multipanel longitudinal survey of adults in a sample of approximately 20,000 households selected to be representative of the noninstitutionalized resident population of the United States. The research reviewed below is based on data from either the 1990, 1991, or 1992 SIPP panels. These panels followed sampled adults for approximately two-and-a-half years, interviewing them either in person or by telephone every four months.(12) During each SIPP interview (called a wave), household-, family-, and person-level information is collected for each of the previous four months on income, labor force activity, program participation (such as AFDC, Food Stamps, and Medicaid), and health insurance status.
Over the life of a given SIPP panel, the Census Bureau produces a separate data file for each wave. At the conclusion of the panel, the Census Bureau produces a full panel file containing one record for each person who was ever a member of a SIPP household. This file differs in two important ways from what users would create were they to link each of the wave files for the panel. First, an entirely separate set of edit and imputation procedures is used by the Census Bureau when creating the full panel file, including the imputation of missing waves when a single missing wave is bounded on both sides by reported information for that person. Because these data processing procedures take advantage of the reported longitudinal information in the file, the full panel files are generally believed to have superior data to those contained in the wave files. Second, the full panel files contain weights that are not available on any of the wave files. The Census Bureau generally creates two types of weights for its longitudinal files: (1) calendar year weights for all persons present throughout a given calendar year of the panel, and (2) panel weights for persons present throughout the entire panel. All other persons do not receive weights and, thus, are not counted in longitudinal estimates.(13)
Like all longitudinal surveys, not all respondents remain in the survey for all the interviews. This is known as attrition and may bias estimates using the SIPP to the extent that those who attrite are systematically different from those that do not. Research on this topic has indicated that for many estimates, there is no detectable bias that can be attributed to attrition. Still, more research needs to be conducted before firm conclusions can be drawn on the effects of attrition in the SIPP.
2. Measuring Health Insurance Status Using the SIPP
The SIPP asks respondents whether they were covered by employer- or union-sponsored insurance, other private health insurance, Medicare, military health care, or Medicaid. Like the CPS, estimates of the uninsured using the SIPP are calculated as a residual -- that is, the uninsured are those who do not report receiving coverage of any type. Unlike the CPS, though, SIPP respondents are asked about health insurance coverage in each month of the 4-month reference period. The health insurance portion of the 1993 SIPP survey instrument is presented in Appendix B.
The SIPP also suffers from underreporting of Medicaid. For example, HCFA(now known as CMS) administrative data show that 35.7 million persons were ever enrolled in Medicaid in 1992. In comparison, Bennefield (1996c) calculated that 12.3 percent of all persons, or approximately 30.5 million persons, reported Medicaid for at least one month in 1992 based on the SIPP -- an underreporting of about 15 percent, or roughly the same magnitude of the underreporting on the CPS. Therefore, the number of uninsured based on the SIPP may be overestimated somewhat, assuming that private health insurance is reported accurately (or, at least, not overreported). Also like the CPS, the SIPP suffers from undercoverage of the population in general. According to the Census Bureau, though, the final impact of undercoverage on estimates is unknown.
Various SIPP estimates of the uninsured, even those for the same time period, may not be comparable because there are a number of different alternatives for analyzing a given time period based on the SIPP. Some examples of these alternatives are as follows:
- Because SIPP panels overlap, researchers often have a choice of SIPP panels for a given time period, or researchers can combine SIPP panels.
- The weights researchers use will depend on the length of the time period analyzed and the specific SIPP file used. Researchers may use calendar year weights, panel weights, or wave-specific weights.
- Some researchers may have used sophisticated duration estimates, while others may have used simple slice-in-time analyses.
This last point deserves further explanation. Researchers' estimates of the uninsured using the SIPP can vary substantially depending on whether they use duration estimates for all spells observed over a period of time or whether they simply examine spells in progress at a point in time. Swartz and McBride (1990) pointed out that studies of the dynamics of health insurance coverage that concentrate on uninsured spells already in progress are misleading because spells in progress at a point in time are disproportionately long spells, whereas most spells are actually fairly short. Using data from the 1984 panel of the SIPP, Swartz and McBride demonstrated this phenomenon by comparing the distribution of spell lengths for all persons for whom they could observe a spell beginning in the SIPP with the distribution of spell lengths among persons whose spells were in progress at a point in time. Using a survival analysis technique, they found that half of all observable spells ended within 5 months, and another 16.5 percent ended within 5 to 8 months. Only 15 percent of all spells lasted more than 2 years. In contrast, among spells in progress at a point in time, 58 percent lasted more than 2 years and only 13 percent ended within 5 months.
Although researchers have calculated estimates for a wide variety of time periods, this section discusses only a few estimates simply to give an overall picture of the uninsured according to the SIPP. In addition, we do not attempt to compare and contrast estimates because, as stated above, estimates can vary based on the SIPP files and methodology used and most researchers do not publish their precise methodology. A more detailed set of estimates that researchers have produced is presented in Tables II.3 (for children) and II.4 (for all persons).(14) First, though, we compare SIPP estimates of the uninsured with the corresponding CPS estimates.
3. SIPP Versus CPS Estimates of the Uninsured
Bennefield (1996c) compared the SIPP and CPS estimates of the uninsured and offered explanations as to why they seem to differ. Bennefield compared the CPS estimates of the uninsured for 1991, 1992, and 1993 with two types of estimates from the SIPP: (1) the SIPP first quarter average monthly estimates for 1992, 1993, and 1994, which can be considered point-in-time estimates; and (2) the SIPP estimates of those uninsured throughout the year for 1991, 1992, and 1993.
SIPP ESTIMATES OF UNINSURED CHILDREN BY SOURCE
|Time Period||Universe||Estimate Definition||Number
|1990||Children age 18||Average monthly uninsured||11.1||16.2%||The Lewin Group (1997, Draft)||not cited|
|Children age 17||Point-estimate of uninsured in wave 1 of 1990 panel (10/89 to 4/90)||-||13.3%||Urban Institute (Blumberg et al. 1997)||1990|
|1991||Children age 18||Average monthly uninsured||11.5||16.5%||The Lewin Group (1997, Draft)||not cited|
|1992||Children age 17||Point-estimate of uninsured in wave 8 of 1990 panel (10/89 to 4/90)||-||13.3%||Urban Institute (Blumberg et al. 1997)||1990|
|Children age 18||Average monthly uninsured||12.4||17.2%||The Lewin Group (1997, Draft)||not cited|
|1993||Children age 18||Uninsured throughout||-||6.5%||CBO (Bilheimer, 1997)||1992|
|Uninsured at any given time||-||13.5%|
|Uninsured at least one month||-||15.5%|
|Average monthly uninsured||13.0||17.9%||The Lewin Group (1997, Draft)||not cited|
|24-month period from February 1991 to January 1993||Children age 17||Uninsured throughout||3.0||-||Families USA 1997||1991|
|Uninsured at least one month||20.5||-|
|Children age 17 uninsured at least one month||Uninsured 12 months or longer||9.6||47%|
|32-month period from early 1991 through mid-1993||Children age 17||Uninsured throughout||2.2||3.2%||Census (Bennefield 1995)||1991|
|Uninsured at least one month||19.6||29.0%|
|28-month period from early 1992 through 1994||Children age 17||Uninsured at least one month||-||30%||Census (Bennefield 1996b)||1992|
|Median number of months uninsured|
SIPP ESTIMATES OF UNINSURED FOR ALL PERSONS BY SOURCE
|Time Period||Universe||Estimate Definition||Number
|1991||All persons||Uninsured throughout||-||7.0%||Bennefield 1996c||1991|
|1992||All persons||Uninsured throughout||-||7.6%||Bennefield 1996c||not cited|
|Uninsured throughout||18.1||7.2%||Census (Bennefield 1995)||1991|
|Uninsured first quarter (point-estimate)||-||14.8%||Bennefield 1996c||not cited|
|Uninsured at least one month||50.7||20.3%||Census (Bennefield 1995)||1991|
|All persons||Uninsured throughout||19.4||-||Census (Bennefield 1996b)||1992|
|Uninsured throughout||-||7.7%||Bennefield 1996c||not cited|
|Uninsured first quarter (point-estimate)||-||14.6%||Bennefield 1996c||not cited|
|Uninsured at least one month||53.6||21.2%||Census (Bennefield 1996b)||1992|
|1994||All persons||Uninsured first quarter (point-estimate)||-||14.5%||Bennefield 1996c||not cited|
|32-month period from early 1991 through mid-1993||All persons||Uninsured throughout||9.7||4%||Census (Bennefield 1995)||1991|
|Uninsured at least one month||64||26.5%|
|28-month period from early 1992 through 1994||All persons||Uninsured throughout||11.9||4.8%||Census (Bennefield 1996b)||1992|
|Uninsured at least one month||66.6||27.0%|
|Median number of months uninsured||5.7||-|
He chose the SIPP first quarter average monthly estimates for his SIPP point-in-time estimates because they correspond with March, the month in which the CPS collects data about the previous year. Bennefield found that the CPS estimates are more similar to the SIPP point-in-time estimates than the annual estimates, suggesting that CPS respondents were reporting their current health insurance status ( Table II.5). He found uninsurance rates of 14 to 15 percent for all persons for both the CPS estimate and the SIPP point-in-time estimate.(15) In contrast, he found uninsurance rates of 7 to 8 percent for the SIPP annual estimates.
Bennefield showed that the estimate of the uninsured throughout a given year using the SIPP is substantially lower than CPS estimates because the SIPP has substantially more persons reporting private health insurance coverage. For example, in 1993 the SIPP showed an 81 percent annual coverage rate for private health insurance versus 70 percent for CPS; in comparison, the SIPP point-in-time coverage rate was 72 percent. Unlike estimates of private insurance, estimates of government-sponsored health insurance were generally consistent across timeframes and surveys -- the CPS Medicaid coverage rates were 11 to 12 percent for the periods analyzed, and both the annual and point-in-time SIPP coverage rates were 9 to 11 percent for the periods analyzed. It is not clear what conclusions should be drawn from the fact that private health insurance coverage accounted for much of the difference between the CPS and SIPP annual estimates. On the one hand, if recall problems were to blame for higher CPS estimates of the uninsured compared with SIPP annual estimates, then respondents seemed to be more likely to fail to recall private insurance than public insurance. Such an explanation is plausible if those publicly insured are more likely than those privately insured to have coverage throughout the year. On the other hand, CPS respondents may simply be reporting their health insurance status as of the interview date.
HEALTH INSURANCE STATUS OF ALL PERSONS: CPS VERSUS SIPP FOR VARIOUS YEARS
(or Q1 1994)
(or Q1 1993)
(or Q1 1992)
|Percent w/ Private Coverage|
|Percent w/ Medicaid|
|Source: Bennefield (1996c).|
Even though the CPS estimates of the uninsured are more widely cited, Census Bureau officials suggest that SIPP may be better suited to measure health insurance information for a number of reasons.(16) First, the SIPP may have less recall error than the CPS because it has a shorter recall period (4 months for the SIPP versus over 1 year for the CPS). Second, respondents may be more likely to answer the SIPP health insurance questions because the questions are more detailed and are better positioned at the beginning of the interview. Third, the SIPP attempts to interview each person in the household age 15 and over, whereas the CPS interviews only one person, who may not obtain accurate information on all household members. Finally, the SIPP is especially designed to measure program participation (such as Medicaid), whereas the CPS is primarily a labor force survey.
4. Other SIPP Estimates of the Uninsured
As one might expect, as the reference period for SIPP estimates of the uninsured lengthens, the percent uninsured throughout decreases while the percent uninsured in at least one month increases. Estimates of uninsured children in 1993 versus the 32-month period from early 1991 through mid-1993 illustrate this point:
- 6.5 percent of children age 0 to 18 were uninsured throughout 1993 (Bilheimer 1997) while only 3.2 percent were uninsured throughout the 32-month period (Bennefield 1995).(17)
- 15.5 percent of children age 0 to 18 were uninsured at least one month in 1993 (Bilheimer 1997) while 29.0 percent were uninsured at least one month throughout the 32-month period (Bennefield 1995).(18)
Thus, for a given reference period, the percentage of children uninsured throughout is considerably less than the percentage uninsured in at least one month. This simply suggests there is substantial churning among uninsured children. From the examples above, 6.5 percent were uninsured throughout 1993 versus 15.5 percent uninsured at least one month.(19) The evidence of churning is even greater as the reference period increases: 3.2 percent were uninsured throughout the 32-month period versus 29.0 percent for at least one month. In short, although a substantial number of children are uninsured at a point in time (about 14 percent according to the CPS) the SIPP data tell us that the problem of uninsured children is even more widespread -- over a two-and-a-half year period almost one-third of all children will be uninsured at some point (Swartz 1994).(20)
D. OTHER ESTIMATES OF THE UNINSURED
The CPS and the SIPP are the most commonly used surveys to measure the health insurance status of individuals, primarily because of their large sample sizes, rich economic and demographic data, and repetition on a regular basis. Nevertheless, other surveys, which are generally smaller than the CPS and the SIPP, have been conducted that measure the health insurance status of individuals. One of the principal values of these other surveys is that they help to validate results from the CPS and the SIPP. Below, we present the estimates of the uninsured using four other data sources: (1) the National Health Interview Survey, (2) the Medical Expenditure Panel Survey, (3) the Community Tracking Study, and (4) the Kaiser/Commonwealth Survey of Americans. A summary of these estimates is presented in Table II.6.
1. National Health Interview Survey
The NHIS is the principal source of information on the health of the civilian noninstitutionalized population of the United States. The NHIS, conducted by the National Center for Health Statistics, is an annual cross-sectional survey of approximately 43,000 households and 106,000 persons. The survey is conducted on a rolling basis throughout the year so that one-twelfth of the annual sample is interviewed each month. The main objective of the NHIS is to monitor the health of the United States population through the collection and analysis of data on a broad range of health topics. The health insurance questions on the NHIS define uninsured as lacking health coverage in the previous month. The health insurance portion of the NHIS questionnaire is presented in Appendix C.
According to the NHIS, there was a monthly average of 11.5 million uninsured children age 0 to 17 in 1994 (NHIS 1996). This estimate is 17 percent higher than the CPS estimates, which is expected if we view the CPS as an annual estimate, or at least one covering a longer period of time than "last month."
OTHER ESTIMATES OF THE UNINSURED
|Source||Data||Time Period||Universe||Estimate Definition||Number
|NHIS Correspondence (1996)||National Health Interview Survey (NHIS)||1994||Children age 17||Uninsure defined as lacking coverage in previous month. Estimate is a 12 month average of survey responses.||11.5||-|
|Beauregard et al. (1997)||1996 Medical Expenditure Panel Survey (MEPS)||First half of 1996||Nonelderly age 0-64||Without insurance throughout the first half of 1996.||44.5||19.2%|
|Children age 17||11.0||15.4%|
|Reschovsky et al. (1997)||Community Tracking Study (CTS)||Late 1996 / early 1997||Children age 18||Point-estimate||8.8||12.1%|
|Davis et al. (Summer 1995)||Kaiser/Commonwealth Survey of Americans||2-year period from 1991 to 1993||Adults age 18-64||Without insurance at least one month||53||33%|
2. Medical Expenditure Panel Survey
Beauregard et al. (1997) estimated the number of uninsured in the U.S. using the 1996 MEPS.(21) The MEPS is co-sponsored by the Agency for Health Care Policy and Research (AHCPR) and the National Center for Health Statistics. The sample of 9,400 households is a subsample of the households responding to the 1995 NHIS and is representative of the civilian noninstitutionalized population of the U.S. The survey was designed to yield comprehensive data that estimate the level and distribution of health care use and expenditures, monitor the dynamics of the health care delivery and insurance systems, and assess health care policy implications. MEPS is the third in a series of national probability surveys conducted by AHCPR on the financing and utilization of medical care in the United States. The National Medical Care Expenditure Survey (NMCES, also known as NMES-1) was conducted in 1977, and the National Medical Expenditure Survey (NMES-2) in 1987.
The MEPS collects data from a nationally representative sample of households through a rotating panel design. In this design, data are collected through a pre-contact interview followed by a series of five rounds of in-person interviews over a two year period of time. As a rotating panel survey, this series of data collection rounds is begun again each subsequent year on a new sample of households drawn from the NHIS sampling frame to provide overlapping panels of survey data, which when combined with other ongoing panels will provide continuous estimates of health care expenditures at both the person and household level.
Each MEPS interview collects information pertaining to a specific time period called the "reference period." The reference period for the first interview began January 1, 1996 and ended on the date of each responding unit's first round interview, conducted from March through June 1996. The health insurance section of the MEPS collects information about private and public health insurance programs. It identifies the household members covered by health insurance and various details about their plans. For employer sponsored coverage, a link is created to job characteristics collected in the employment section of the questionnaire. For individuals who are uninsured at the beginning of the year, information is collected on the length of time they have been uninsured. Additional questions clarify whether each person identified by each policy was covered throughout the reference period. Information for public insurance is collected on Medicare, Medicaid, Medicaid waiver programs, military sponsored plans, and other government programs. A questionnaire section on managed care determines whether household members that are either publicly or privately insured are covered under a managed care plan; additional questions ask about the characteristics of their managed care plans. The entire MEPS health insurance questionnaire is too large to include in this document. Instead, portions of the health insurance questions pertaining to Medicaid and the uninsured are presented in Appendix D.
Beauregard et al. used MEPS to develop a national estimate of the uninsured population based on sample persons who were uninsured continuously from January 1, 1996 to their first-round interview date 3 to 6 months later.(22) By this measure they found that 19 percent of nonelderly adults age 0 to 64 (44.5 million persons) and 15 percent of children age 0 to 17 (11 million children) were uninsured. Although these estimates appear higher than those of the CPS (9.8 million uninsured children) and other data sources, Beauregard et al. concluded that once time-period and definitional
issues are considered, their estimates are consistent with the findings of the CPS.(23) In reaching this conclusion, though, they assumed the CPS measured the uninsured throughout the previous year rather than at a point in time. If, instead, the CPS is viewed as a point-in-time estimate of the uninsured, then the MEPS and CPS findings are not consistent, since the MEPS estimate of those uninsured throughout a 3 to 6 month period should be considerably less than a point-in-time estimate. Additional research comparing MEPS and CPS estimates of the uninsured is warranted before firm conclusions can be drawn.
3. Community Tracking Study
Reschovsky et al. (1997) estimated the number of uninsured children age 0 to 18 using the CTS Household Survey. The CTS, sponsored by the Robert Wood Johnson foundation and conducted by the Center for Studying Health System Change, consisted of telephone interviews for 33,000 families, 11,600 of which had children. Information was gathered on all adults and one randomly chosen child in each household. Altogether, the survey has information on about 60,000 individuals. The data are weighted to be representative of the United States and are adjusted for nonresponse. Interviews took place, primarily via telephone, between July 1996 and July 1997. Like the MEPS, the CTS health insurance questionnaire is too large to include in this document; therefore, only the portions pertaining to Medicaid and the uninsured are presented in Appendix E.
Reschovsky et al. estimated that at any point in time from late 1996 to early 1997, there were approximately 8.8 million uninsured children, or about 12 percent of all children age 0 to 18. This is substantially lower than the estimate of 10.5 million uninsured children in 1995 produced by Fronstin and others using the March 1996 CPS. Reschovsky et al. pointed out that the difference between the two numbers most likely reflects methodological differences in how the two surveys asked about health insurance coverage. Reschovsky et al. asserted that the CTS estimate is lower in part because of methodological innovations in how insurance coverage is measured. With the CTS, persons who reported none of the various types of coverage when asked about each one individually, were then asked directly whether they were, in fact, uninsured. Some respondents at that point did report coverage. The CPS, in contrast, never directly asks respondents whether they are uninsured. Reschovsky et al. acknowledged, though, that the debate over whether the CPS is a point-in-time or period-of-time estimate confounds comparisons between the CTS and the CPS. If the CPS is a period-of-time estimate, then the CTS estimate, which is clearly a point-in-time estimate, would be expected to be higher rather than lower than the CPS estimate.
Reschovsky et al. did not adjust their estimate of uninsured children for possible underreporting of Medicaid in the CTS. Medicaid underreporting in the CTS appears to be even more substantial than in the CPS. Reschovsky et al. found that 9.8 million children age 0 to 17 were enrolled in Medicaid in the CTS versus 16.5 million in the CPS. Although the underreporting of Medicaid is quite high in the CTS, it is difficult to determine exactly the extent to which Medicaid may be underreported because the survey did not ask about Medicaid coverage in households where everyone reported private coverage. Therefore, the CTS missed Medicaid enrollees who were also covered by private insurance.(24) In any case, CTS estimates of the uninsured may be inaccurate because Medicaid enrollment appears to be underreported.
4. Kaiser/Commonwealth Survey of Americans
Davis et al. (1995) estimated the number of uninsured adults age 18 to 64 using the 1993 Kaiser/Commonwealth Survey of Americans. The survey consisted of telephone interviews with a nationally representative sample of 2,000 adults in August 1993. Davis et al. found that one-third of all adults, or 53 million adults, lacked health insurance at some point during the 2-year period from 1991 to 1993. They also found that at the time of the survey in August 1993, 18 percent of all adults were uninsured, which is very close to EBRI's estimate, based on the CPS, that 19 percent of adults age 18 to 64 in 1993 were uninsured (Fronstin 1995). The survey had a response rate of only 53 percent (compared with CPS and SIPP response rates in excess of 90 percent) raising concern about potential response bias. The possible effect of Medicaid underreporting on these findings is unknown because the percentage of respondents with Medicaid was not reported.
E. STATE-LEVEL ESTIMATES
Several researchers have combined CPS surveys to increase the sample sizes enough to produce state-level estimates of the uninsured. Below, we give an overview of two of these studies, one by The Urban Institute and one by Families USA.
1. The Urban Institute
Winterbottom et al. (1995) combined data from the 1991, 1992, and 1993 March CPS surveys to obtain state-level estimates of the health insurance status of individuals. Because CPS households are interviewed for two consecutive years and Winterbottom et al. only wanted to include each household once, they included all the observations from the 1993 CPS plus approximately half of the observations from the 1991 and 1992 surveys. Thus, combining three years of CPS data doubles the sample size, which reduces the sampling variance.(25) Winterbottom et al. then used The Urban Institute's TRIM2 model to adjust for underreporting of Medicaid.
Winterbottom et al. found that the rate of uninsured among children age 0 to 17 varied by state and region. For example, in the West South Central region -- the region with the highest rate of uninsurance -- 18.5 percent of children were uninsured.(26) In contrast, in the East North Central region -- the region with the lowest rate of uninsurance -- 6.8 percent of children were uninsured.(27) Winterbottom et al. pointed out that uninsurance rates vary by region and state for a number of reasons, including the rate of employer-sponsored insurance coverage and the rate of Medicaid coverage. Winterbottom et al. used the following example of the uninsurance rates of all persons age 0 to 64 to illustrate their point:
"The Middle Atlantic region has the lowest rate of employer coverage among its poverty population -- only 11.5 percent have employer-sponsored coverage -- significantly lower than the 15.8 percent coverage in the Mountain states. However, because the Middle Atlantic region has a high rate of Medicaid enrollment in the poverty population -- 53 percent of the poor get their primary coverage through the program -- its uninsured rate of 25.1 percent is not the highest. The Mountain States, with greater employer coverage among the poor, have a higher uninsured rate (32.6 percent) than the Middle Atlantic region because Medicaid covers fewer of the poor in the Mountain States region (40 percent)."
2. Families USA
Families USA (Families USA 1997) used 1995 and 1996 March CPS data in combination with imputation equations developed from the 1991 SIPP panel to estimate the number of children age 0 to 17 who were without health insurance in one or more months over the 2-year period from 1995 through 1996.(28) They estimated that 23.1 million children, or 33 percent of all children, were without health insurance in at least one month of the two-year period from 1995 to 1996. Families USA noted that the proportion of children with gaps in health insurance varied significantly from state to state due to differences in state economies and residents' income, the prevalence of jobs that offer employer-based coverage, the scope of public insurance programs (especially Medicaid), and the existence of other state health reforms. They found the highest proportions of uninsured children in southern and southwestern states. This supports the finding of Winterbottom et al. that the three regions with the highest proportion of uninsured children are (1) the West South Central, (2) the South Atlantic, and (3) the East South Central. According to Families USA, the following ten states had the highest percentage of children who experienced gaps in their health insurance during the period 1995 through 1996: Texas (46 percent); New Mexico (43 percent); Louisiana (43 percent); Arkansas (42 percent); Mississippi (41 percent); District of Columbia (39 percent); Alabama (38 percent); Arizona (38 percent); Nevada (37 percent); and California (37 percent). Families USA did not report confidence limits for these estimates, however, and readers are cautioned that the estimates for smaller states are not nearly as precise as those for larger states.
1. Because of these revisions and a change in the sample framework for the survey in 1995, it is difficult to compare estimates done before and after that year. Researchers believe that these revisions, coupled with the change in the sample framework for the survey in 1995, may have increased the number of persons reporting that they were insured (Swartz 1997). As a result, estimates before the 1995 CPS are not fully comparable to more recent estimates. However, at least one researcher has made adjustments to the CPS data in order to make more valid comparisons of health insurance coverage before and after the 1995 revisions (Fronstin, 1997b).
2. Including the elderly increases this estimate by only 300,000 individuals, since nearly all elderly are covered by Medicare. If the elderly are included, then 15.4 percent of the total population was without health insurance in 1995.
3. As cited in Monheit (1994).
4. Of course, the MEPS could have similar reporting problems to the CPS. However, the MEPS health insurance questions are much more detailed than the CPS questions, and interviewers are trained specifically on asking health related questions. The MEPS study is discussed in more detail later in this chapter.
5. Personal correspondence with Linda Bilheimer of CBO and Kathy Swartz of the Harvard School of Public Health, September 15, 1997.
6. One reason suggested by researchers for why people may not realize they are covered by Medicaid is that they are enrolled in a managed care program and, therefore, report that they are covered by the managed care company. If, indeed, this is occurring, then the problem of Medicaid underreporting could get worse as more states adopt Medicaid managed care programs.
7. For example, if most persons are enrolled for only 10 of 12 months, then the number of persons ever enrolled in the year should be adjusted downward by almost 17 percent to be comparable to a point-in-time estimate. With additional HCFA(now known as CMS) data, it is possible to calculate the average monthly number of persons enrolled in Medicaid.
8. HCFA(now known as CMS) also reports the number of Medicaid "recipients." Recipients, though, are defined by HCFA(now known as CMS) as the subset of enrollees, or eligibles, that utilized services during the year
9. Some researchers (for example, see Thorpe 1997a and 1997b) cite the number of children enrolled in Medicaid in 1995 according to the CPS as 13.8 million rather than 16.5 million. These researchers are reporting the number of children covered only by Medicaid during the previous year, thus excluding children that were covered by both Medicaid and private insurance during that period.
10. Nationally, enrollees exceeded recipients by 14 percent (HCFA(now known as CMS) 2082 tables for 1995).
11. They further add that analyses of people in poverty, people receiving welfare (AFDC), and people experiencing unemployment have been consistent in finding that most people who experience any of these situations, do so for short periods of time (see, for example, Bane and Ellwood 1986; O'Neill, Bassi, and Wolf 1987; Akerlof and Main 1980).
12. The 1992 panel was extended about one year, to 40 months.
13. One exception to this rule is that persons who die are given weights. Also, an important implication of this weighting design is that almost all infants born during a calendar year will have a zero-weight for that calendar because they are not part of the population on January 1. As a result, infants will be underrepresented in the SIPP data, which is of particular concern because infants have been a key target group for Medicaid expansions over the last decade.
14. Sometimes only a count or a percent is presented in the tables, but not both, because that is all a researcher reported. We do not attempt to calculate implied percents or counts because the denominator over which they are calculated is not always reported and not always obvious.
15. Bennefield did not cite the denominator for his calculation of the SIPP estimate of all persons who were uninsured. Based on other SIPP literature, the universe of all persons represents about 250 million persons. This would suggest that there were about 37 to 38 million uninsured at a point in time, an often cited figure.
16. Interview of Chuck Nelson and Bob Bennefield of the Census Bureau, conducted by Dave Baugh and Roger Buchanan of HCFA(now known as CMS) (July 23, 1997).
17. The comparable estimates for all persons were 7.7 percent (Bennefield, 1996c) and 4 percent (Bennefield 1995), respectively.
18. The comparable estimates for all persons were 21.2 percent (Bennefield, 1996b) and 26.5 percent (Bennefield 1995), respectively. That the differences between the 1-year and the 32-month estimates for adults was less than that for children suggests that there was less churning of the uninsured for adults.
19. Point-estimates of the uninsured using the SIPP fell somewhere between the uninsured throughout and the uninsured in at least one month estimates. In 1993, 13.5 percent of children were uninsured at any given time (Bilheimer 1997).
20. Swartz (1994) provides a useful overview of the dynamic of people without health insurance, along with references to earlier works on this topic.
21. A similar analysis of the MEPS with the same findings can be found in Vistnes and Monheit (1997).
22. Full-year 1996 MEPS estimates of the uninsured were not available as of the writing of this paper, but are planned once all interviews pertaining to 1996 are completed.
23. One of the definitional differences they cite is that the CPS assumes all AFDC recipients and most SSI recipients are enrolled in Medicaid, whereas MEPS usually does not. In addition, CPS counts children of adults covered by Medicaid as insured, whereas for these preliminary estimates, MEPS does not (unless their families reported them as covered). Finally, unlike MEPS, CPS counts as uninsured military veterans whose source of coverage is the Department of Veterans Affairs.
24. Missing Medicaid coverage for this group -- that is, those covered by private insurance and Medicaid -- would not affect the estimates of the uninsured because this group is reported as receiving private insurance.
25. The two samples are not independent, in that they tend to be pulled from the same neighborhoods; therefore, the doubled sample size yields something less than a proportionate reduction in variance.
26. The West South Central region includes the following states: Arkansas, Louisiana, Oklahoma, and Texas.
27. The East North Central region includes Illinois, Indiana, Michigan, Ohio, and Wisconsin.
28. The imputation equations predict which children will be uninsured in one or more months over the 2-year period on the basis of their demographic and economic characteristics. Researchers generally do not use the SIPP alone for state-level estimates because not all states in the SIPP are uniquely identified and because SIPP's relatively small sample sizes make resulting estimates imprecise. Even the CPS, with its larger sample sizes, does not support precise estimates in most of the smaller states.
Chapter III. Characteristics of the Uninsured
This chapter examines the literature on the static and dynamic characteristics of uninsured children. Even though estimates of the number and percentage of uninsured children vary by data source and research methodology, the estimates of the characteristics of uninsured children are fairly consistent. Therefore, unless researchers strongly disagree, we simply present the overall findings from one source for key characteristics.
A. STATIC CHARACTERISTICS OF THE UNINSURED
The characteristics presented in this section are all static -- that is, they describe a defined population of the uninsured at a defined time. Although static characteristics provide a useful overall picture of the uninsured, static characteristics can also be deceiving because they can mask the fact that the uninsured are not a homogenous population. For example, the long-term uninsured and the short-term uninsured probably have very different characteristics. Examining these two groups together may provide a picture of the uninsured that does not look like its component parts. For this reason, much of the current literature on the characteristics of the uninsured could be improved by using longitudinal data and examining the short-term and long-term uninsured groups separately.
Most of the discussion which follows is based on EBRI's recent analysis of the March 1996 CPS (Fronstin 1996 and 1997a) because it is timely and presents detailed characteristics of uninsured children. Twelve other sources we reviewed also reported findings on the characteristics of uninsured children: Beauregard et al. (1997); Bennefield (1995); Bennefield (1996a); Bennefield (1996b); Families USA (1997); Holahan (1997); Holahan et al. (1995); Monheit (1994); Reschovsky et al. (1997); Summer et al. (1997); U.S. GAO (1997); and Winterbottom et al. (1995). Their findings were generally consistent with the EBRI results unless otherwise noted.
A summary of EBRI's findings is presented in Table III.1. In this table, the characteristics of the uninsured are presented alongside the characteristics of the privately insured and the publicly insured in order to show how these groups differ.(1) Children with both private and public coverage are included in both the private and public columns. The following characteristics are examined: age, race and ethnicity, citizenship, family structure and poverty level, parents' employment and health insurance status, and parents' education. Except for age characteristics, where the literature includes research done by the Urban Institute using the CPS with the TRIM2 model, all of the estimates we examine in this section are based on the March 1996 CPS and are not adjusted for the underreporting of Medicaid in these data. Therefore, the estimates presented could be biased to the extent that some of the uninsured who are analyzed may actually be receiving Medicaid but not reporting so.
As shown in Table III.1, EBRI found that of all uninsured children age 0 to 17 in 1995, 6.7 percent were infants, 26.5 percent were age 1 to 5, 38.4 percent were age 6 to 12, and 28.4 percent were age 13 to 17. As expected, the publicly insured contained a larger percentage of infants and children age 1 to 5 because the poverty-related Medicaid expansions are most generous to these groups.
The Urban Institute also estimated the percentage of children in various age groups that were uninsured (Holahan 1997). The estimates were adjusted for the underreporting of Medicaid in the CPS using the TRIM2 microsimulation model by selecting Medicaid eligible individuals to participate in the program even though they did not report doing so.
DISTRIBUTION OF ALL PRIVATELY INSURED, PUBLICLY INSURED,
AND UNINSURED CHILDREN BY VARIOUS CHARACTERISTICS
|(percentage of children within coverage category)|
|Age 0 to 1||4.5||7.3||6.7|
|Age 1 to 5||26.8||35.0||26.5|
|Age 6 to 12||39.5||36.9||38.4|
|Age 13 to 17||29.2||20.8||28.4|
|(percentage of children within race category)|
|(percentage of children within coverage category)|
|Families with incomes below 200% of poverty||25||79||70|
|At least one parent employed||98||68||89|
|At least one parent employed full-time, full-year||88||38||64|
|At least one uninsured parent||-||-||80|
|At least one parent with employment-based insurance||92||-||16|
|At least one parent with a college degree||32||8||11|
|Source: Fronstin (1996); Fronstin (1997a).|
The Institute's findings are compared with EBRI's findings in Table III.2. The Institute found that a smaller percentage of children in all age groups were uninsured, but particularly among younger children.(2) They found that 8.3 percent of children age 1 to 5 and 10.2 percent of children age 6 to 12 were uninsured, versus 12.7 percent and 13.8 percent, respectively, according to EBRI. The most likely reason the Institute found a relatively smaller percentage of younger uninsured children compared with EBRI is that younger children were more likely than older children to be eligible for Medicaid under the poverty-related criteria (in 1995, states were required to cover children up to age 6 with family incomes below 133 percent of poverty and children age 6 to 11 in families with incomes below poverty). Thus, the adjustment for underreporting of Medicaid among younger children accounted for much of the 2.9 million difference between The Urban Institute's and EBRI's estimates of the number of uninsured children (6.9 million versus 9.8 million, respectively). Nevertheless, the difference between The Urban Institute's and EBRI's estimates of uninsured older children (age 13 to 17), are still quite high in absolute terms -- about 0.3 million children.
2. Race and Ethnicity
EBRI found that even though most of the uninsured were white (49 percent), the overall uninsurance rate for whites (10.5 percent) was lower than the national average (13.8 percent) and lower than any other race and ethnicity group. Hispanics were, by far, the most likely to be uninsured and the least likely to be covered by private health insurance. Overall, more than one-quarter (26.8 percent) of all Hispanic children were uninsured in 1995, according to EBRI's analysis of the CPS.
DISTRIBUTION OF UNINSURED CHILDREN IN 1995 BY AGE
|1 to 5||20.5||8.3||12.7|
|6 to 12||27.5||10.2||13.8|
|13 to 17||19.3||13.0||14.5|
|Source: Holahan (1997) and Fronstin (July 1997).
Note: The Urban Institute's findings are adjusted for Medicaid undercount using the TRIM2 microsimulation model.
Blacks fared quite a bit better than Hispanics -- the uninsurance rate among blacks was 15.3 percent. The lower rate of uninsurance among blacks in comparison to Hispanics was due to the fact that blacks were both more likely to be privately insured and more likely to be publicly insured. The rate of private and public insurance for blacks was 44 percent and 49 percent, respectively, versus 38 percent and 39 percent for Hispanics. Blacks have the highest rates of public coverage.
EBRI found that 10 percent of uninsured children were noncitizens versus 4 percent of the publicly insured and 2 percent of the privately insured. That such a high percentage of the uninsured children were noncitizens suggests that their parents disproportionately work at jobs without health benefits or, if they do not work, are less likely than citizens either to be eligible for or to participate in Medicaid.
4. Family Structure and Poverty Level
EBRI found that uninsured children were more likely than the privately insured to be in single-parent families (38 versus 20 percent) but less likely than the publicly insured (56 percent). Similarly, the uninsured were more likely than the privately insured to be in families with incomes below 200 percent of poverty (70 versus 25 percent) but less likely than the publicly insured (79 percent). Thus, in terms of family characteristics, the uninsured seem to represent a middle ground between the privately and publicly insured.
EBRI made the point that the underlying data can explain how so many uninsured children appear to be in families with incomes well above poverty:
"Families with two workers can easily earn $40,000 or more if both parents earned $20,000. However, most parents earning $20,000 do not have access to health insurance. In many cases, their employer does not offer insurance because of the nature of the job. In addition, workers are often asked to pay the full cost of family coverage, which could be very expensive and amount to a relatively high percentage of a worker's $20,000 salary."
5. Parents' Employment and Health Insurance Status
In terms of parents' employment status, EBRI once again found that uninsured children represented a middle ground between the privately and publicly insured. Uninsured children were less likely than the privately insured to have at least one employed parent (89 versus 98 percent) but more likely than the publicly insured (68 percent). EBRI also examined the full-time, full-year employment status of individuals (because private health insurance benefits are typically offered only to full-time, full-year employed) and again found that the uninsured represented a middle ground between the privately and publicly insured -- 64 percent of uninsured children had parents who were employed full-time and full-year, versus 88 percent and 38 percent for the privately and publicly insured, respectively. In terms of health insurance status, EBRI found that 80 percent of uninsured children had at least one uninsured parent and 16 percent had at least one parent with employment-based insurance.
6. Parents' Education
EBRI found that 11 percent of uninsured children's parents had a college degree compared with 32 percent of privately insured children and 8 percent of publicly insured children.
B. DYNAMICS OF THE UNINSURED
In this section, we examine available research on the dynamics of uninsured children, including the length of time they are uninsured, the events leading to loss of coverage, and the events causing them to regain insurance. When examining the dynamics of the uninsured, it is important to understand that researchers' findings may differer markedly because of methodological differences in analyzing longitudinal data. Recall from the discussion on SIPP estimates of the uninsured in Chapter 2 that estimates of the duration of spells without health insurance can vary substantially depending on whether they are based on all spells or spells in progress at a point in time. Spells in progress at a point in time contain a disproportionate number of long spells. Because of this, estimates for all spells, which are often produced using survival analysis techniques, generally give a more accurate picture of the dynamics of the uninsured.
Families USA (1997) used the 1991 SIPP panel to examine how many children were uninsured and how long they were uninsured (see Table II.3A). Families USA did not use survival analysis for these estimates, electing instead to examine all the spells in progress during the 24-month period from February 1991 to January 1993. Of the 20.5 million children uninsured at least one month during that period, 47 percent were uninsured for 12 months or more, 15 percent were uninsured throughout, and only 7 percent were uninsured for less than 3 months.
Bennefield (1996b), using a survival analysis technique on data from the 1992 SIPP panel over a 28-month period, found that the median spell of noncoverage for those under age 18 was only 4.0 months (also shown in Table II.3A).(3) This was considerably shorter than the median spell of 5.8 months or longer for all other age groups. Bennefield's report does not give any further results on length of enrollment for children.
We also looked at research on the events triggering uninsurance for both adults and children. Insurance loss for most individuals is employment-related, according to analysis of SIPP data focusing on calendar year 1994 by The Lewin Group (1997, Draft). Of the two million Americans, on average, who became uninsured each month in 1994, about 58 percent cited changes in employment as their primary reason for losing coverage. Lewin defined a change in employment as loss of employment, loss of employment for a spouse or parent, termination of an employer plan, or a shift from full-time to part-time worker status. In addition, Lewin's analysis of SIPP data from 1991 through 1993 found that only about 8 percent of all persons lost their coverage due to a change in occupation for the same employer or a shift from full-time to part-time status. They also found that those with the lowest incomes were more than twice as likely as those with higher incomes to cite a job change as the reason for losing coverage.
The events triggering uninsurance for children appear to be somewhat different from that of all persons. Using data from a special coverage supplement to the 1993 NHIS, Lewin found that individuals under age 22 who lost their health insurance coverage were less likely to cite job related reasons than all people who lost coverage (44 versus 58 percent). But, they noted that an additional 18 percent indicated that they lost private coverage because they became ineligible as dependents due to age. Therefore, 62 percent of these children became uninsured due to some break in employer coverage. The reasons that children under age 22 lost their health insurance coverage according to Lewin's analysis of the 1993 NHIS are presented in Table III.3 (reproduced from the Lewin report).
Little research has been done on the events that cause uninsured children to regain their health insurance. Blumberg et al. (1997), using data from the 1990 panel of the SIPP, examined theinsurance status of children in wave 8 of the SIPP who were uninsured in wave 1.(4) Blumberg et al. found that of those uninsured in wave 1, 52 percent were uninsured in wave 8, 29 percent were privately insured, and 19 percent were enrolled in Medicaid.
REASONS CHILDREN UNDER AGE 22 LOST HEALTH INSURANCE COVERAGEa
|Reasons for Losing Employer Coverage||Percentage|
|Laid off, lost job, or unemployed||11.7|
|Spouse or parent laid off, lost job or unemployed||28.8|
|Employer stopped offering coverage||1.4|
|Cut back to part-time status||0.6|
|Benefits from employer/former employer ran out||1.1|
|Subtotal: Job Related Reasons||43.6|
|Death of a spouse||0.6|
|Divorce or separation||1.6|
|Became ineligible because of age||17.7|
|a Includes all persons currently without health insurance coverage who lost their coverage within the past three years
b Voluntary termination includes: dissatisfied with previous insurance, do not believe in insurance, and free/inexpensive care available
Source: The Lewin Group (1997, Draft) tabulations of the 1993 NHIS Health Insurance Supplement.
Only somewhat dated research has been done comparing the characteristics of the long-term uninsured with the short-term uninsured. Swartz, Marcotte, and McBride (1993a, 1993b), and Swartz and McBride (1990) measured various distributions of uninsured spell lengths in the 1984 SIPP panel using survival analysis. Swartz, Marcotte, and McBride (1993a) used a hazard model of spell durations to estimate the relative effects of socio-economic and demographic characteristics on the duration of a spell without health insurance. They found that monthly family income, educational attainment, and industry of employment in the month prior to losing health insurance are the characteristics that have the greatest impact on the exit rate from being without health insurance. In particular, a low exit rate is positively correlated with low family income, low educational attainment, and employment in specific industrial sectors (agriculture/forestry/fishing and mining combined, construction, personal services and entertainment services combined, and public administration).
Monheit and Schur (1988) used the 1984 SIPP panel to examine various cohorts of the uninsured population. They did not, however, use survival analysis techniques. They found that the uninsured were heterogenous, consisting of many persons who lost coverage for relatively short periods of time, others who experienced periodic spells without coverage, and those who were persistently uninsured. The persistently uninsured, compared with all persons who lost coverage, were a much more economically disadvantaged group with far less labor market attachment and less access to employment related insurance. Monheit and Schur pointed out that longitudinal analyses of the uninsured are useful because the characteristics of the uninsured differ by spell length.
1. Almost all publicly insured children are insured by Medicaid. A small proportion of children are covered by state only programs.
2. We do not present the uninsured rate for infants because the Urban Institute reported uninsured rates for infants and pregnant women together.
3. That the median is 4 months is due in part to the pronounced "seam effect" in the reporting of health insurance status in the SIPP. Four months is the length of the reference period for each interview. Changes in insurance status are reported to occur disproportionately between interviews. In effect, most spells of fewer than 4 months duration and probably some with 5 or 6 month durations are reported in such a way that they appear to have exactly 4 month durations.
4. Wave 1 of the SIPP covered reference months between November 1989 and April 1990; wave 8 included months between February and August 1992.
Chapter IV. How Many Uninsured Children Are Eligible for Medicaid and What Is the Medicaid Participation Rate for Children?
This chapter examines the literature on two related estimates: (1) how many uninsured children appear to be eligible for Medicaid, but are not participating, and (2) how many children overall are eligible for Medicaid and how many participate.(1) More published research has addressed the first estimate than the second, using the family income and age of the uninsured child to calculate the number or proportion of uninsured children eligible for Medicaid, but not participating. In most instances, family income and the child's age are compared to a uniform set of national income standards based on the poverty-related Medicaid expansions.
In reality, though, Medicaid eligibility is much more complex, and it varies substantially by state. State income and asset thresholds for different age groups vary significantly (some state income thresholds now reach 250 percent of poverty for all children under age 18), the majority of states have medically needy programs (with the spend-down component), and the majority of states also extend coverage to so-called Ribicoff children (children in families not meeting AFDC dependent child requirements).(2) Moreover, children with private insurance coverage can be eligible for Medicaid if family income is low enough, if coverage is limited, or if uncovered medical expenses are high enough to reduce family income to medically needy levels. As a result of these complexities, few researchers have attempted to use simulation models to estimate overall Medicaid eligibility for children, accounting for state variation, family structure, and uncovered medical expenses. Even when they have attempted to simulate these more complex provisions, the literature is sparse, largely unvalidated, and tends to lack details on the eligibility simulation algorithms used.
Estimates of the number of uninsured eligible for Medicaid may tend to be too high because of the Medicaid underreporting problem that plagues most of the survey data used for the estimates. Medicaid underreporting suggests that some of the Medicaid-eligible uninsured may actually be enrolled in Medicaid but not reporting it during the survey. Only one organization, The Urban Institute, attempts to adjust for Medicaid underreporting when estimating the number of uninsured eligible for Medicaid. As a result, the Institute's estimate is about half that of the estimates of other organizations using the same data. To complicate matters further, recall from Chapter II that there is some evidence that The Urban Institute may over-adjust for Medicaid underreporting, making their estimate of the Medicaid-eligible uninsured too low. In conclusion, the combined effects of Medicaid underreporting and the difficulty of simulating Medicaid eligibility make estimates of the Medicaid-eligible uninsured variable and inexact.
This chapter first examines the literature on the estimates of the number of uninsured children who are eligible for Medicaid. Then, this chapter examines the more general question of how many children overall are eligible for Medicaid and how many participate.
A. HOW MANY UNINSURED CHILDREN ARE ELIGIBLE FOR MEDICAID?
Estimates of the number of uninsured children eligible for but not participating in Medicaid have been undertaken by the following researchers and organizations: The Urban Institute, Reschovsky et al., Thorpe, GAO, and the Center for Budget and Policy Priorities (CBPP). All but Reschovsky et al. used the March CPS for their estimates; Reschovsky et al. used the CTS Household Survey data. These estimates are presented below and summarized in Table IV.1.
ESTIMATES OF THE NUMBER OF UNINSURED CHILDREN WHO ARE ELIGIBLE FOR MEDICAID BUT NOT PARTICIPATING, BY SOURCE
|Source||Data||Time Period||Estimate Definitions and Eligibility Criteria||Uninsured Children Eligible for Medicaid|
|Pct. of All Uninsured|
(which adjusts for Medicaid undercount)
|CPS 3/96||1995||Estimate definition: Children age 17
State-specific poverty related
Medically needy children
|Reschovsky et al. (1997)||CTS||late 1996 /
|Estimate definition: Children age 18.
State-specific poverty-related criteria only.
|Thorpe (1997b)||CPS 3/96||1995||Estimate definition: Children age 18
|GAO (1996)||CPS 3/95||1994||Estimate definition: Children age 11
Poverty-related only. Not state-specific. Only includes children age 5 with family incomes below 133% of poverty and children 6-11 with family incomes below 100% of poverty.
|Center for Budget and Policy Priorities (Summer et al. 1997)||CPS 3/95||1994||Estimate definition: Children age 10.
Poverty related only. Not state-specific. Only the federal minimum poverty-related eligibility criteria for children age 10.
Generally, these estimates showed that the overall number of uninsured children eligible for, but not participating in Medicaid, ranged from 1.6 to 3.3 million during 1995 (24 to 45 percent of all uninsured children). Differences in the estimates appear to vary by data source used, by whether or not adjustments were made to reconcile Medicaid enrollment with Medicaid administrative data, and by the complexity of the simulation method that was used.
1. The Urban Institute
The Urban Institute used March 1996 CPS data to estimate the number of uninsured children age 0 to 17 that were eligible for Medicaid in 1995 but not participating (personal communication with Beth Kessler of The Urban Institute, August 12, 1997). To make the estimate, they used their TRIM2 microsimulation model, which simulates Medicaid eligibility for children on the basis of the following criteria: state-specific poverty related criteria, AFDC and SSI participation, state medically needy programs, and asset eligibility. In addition, the TRIM2 model adjusts for the CPS underreporting of Medicaid enrollment and AFDC and SSI participation, thereby increasing the number of Medicaid enrollees.(3) They found that in 1995, at least 1.6 million (24 percent) of the 6.9 million children classified as uninsured were eligible for Medicaid, but not enrolled.
2. Reschovsky et al.
Reschovsky et al. (1997) used the CTS data to determine how many uninsured children age 0 to 17 were eligible for Medicaid but not enrolled. Reschovsky et al. determined Medicaid eligibility only on the basis of age and family income in relation to each state's poverty-related eligibility standards. Reschovsky et al. cautioned that their eligibility estimates were approximations because of the following limitations in the data and their eligibility algorithm: (1) eligibility under Medicaid is usually based on monthly income while the CTS asked respondents for annual income only; (2) annual income may be subject to recall error; (3) asset eligibility was not taken into account; and (4) medically needy and some other Medicaid provisions were not taken into account. Reschovsky et al. found that 3.2 million (38 percent) of the 8.5 million children uninsured according to the CTS were eligible for Medicaid. In making this estimate, Reschovsky et al. did not adjust the CTS data to account for Medicaid underreporting.
Thorpe (1997b), using March 1996 CPS data, estimated that 3.3 million (31 percent) of the 10.5 million uninsured children age 0 to 18 in 1995 were eligible for Medicaid. Thorpe did not describe in any detail his methodology for determining Medicaid eligibility, so it is not clear whether he used state specific income thresholds. Thorpe did not adjust the CPS data to account for Medicaid underreporting.
4. Estimates for Poverty-Related Expansion Children Only
Two organizations, the GAO and CBPP, used the March 1995 CPS to estimate a subset of younger children eligible for but not participating in Medicaid -- only those eligible for Medicaid by federal mandate. This included all uninsured children age 0 to 5 in families with incomes below 133 percent of poverty and uninsured children born after September 30, 1983 with family income below 100 percent of poverty. Both organizations used only the federal minimum income thresholds and did not account for states using higher thresholds.
GAO (1996) estimated that there were 2.9 million uninsured children under age 12 in 1994 who were eligible for Medicaid by federal mandate. CBPP (Summer et al. 1997) also used 1995 CPS data, but their estimate was computed for children under age 11. CBPP found 2.7 million uninsured children under age 11 in 1994 were eligible for Medicaid (45 percent of the 5.9 million uninsured children under age 11).
B. WHAT IS THE MEDICAID PARTICIPATION RATE FOR CHILDREN?
Estimates of the overall Medicaid participation rate for children have been undertaken by the following researchers and organizations: The Urban Institute, CBO, GAO, and CBPP. All used the March CPS for their estimates These estimates are presented below and summarized in Table IV.2. Estimates of the overall Medicaid participation rate varied depending on whether Medicaid underreporting adjustments were used, the complexity of the eligibility simulation, and the universe of children examined.
1. The Urban Institute
The Urban Institute used March 1996 CPS data with its TRIM2 microsimulation model to simulate the average yearly Medicaid participation rate for children age 0 to 17 in 1995. Urban Institute researchers found that an average of 26.4 million children were eligible for Medicaid and 22.1 million enrolled, for a participation rate of 83.6 percent.(4) This high participation rate incorporates the TRIM2 model's adjustment for Medicaid underreporting. As far as we can determine, none of the other estimates reported in Table IV.2 includes an adjustment for Medicaid underreporting. As a result, these other participation rates are likely to be depressed.
ESTIMATES OF MEDICAID PARTICIPATION RATES FOR CHILDREN, BY SOURCE
|Source||Data||Time Period||Estimate Definitions and Eligibility Criteria||Eligibles (millions)||Enrollees (millions)||Part. Rate|
(which adjust for Medicaid undercount)
|CPS 3/96||1995||Estimate definition: Participation rate for all children age 17
State-specific poverty related
Medically needy children
|Reschovsky et al. (1997)||CTS||late 1996 /early 1997||Estimate definition: Participation rate for all children age 17
Eligibility criteria: State-specific poverty-related criteria only
|CBO (Bilheimer 1997)||CPS 3/96||1995||Estimate Definition: Participation rate for children who met expansion eligibility criteria.
Children receiving welfare cash assistance or covered by private insurance are excluded from all calculations. No other criteria specified.
|GAO (1996)||CPS 3/95||1994||Estimate definition:
Children age 11 eligible for Medicaid by federal mandate because of age and income eligibility.
Poverty-related only. Not state-specific. Only includes children age 5 with family incomes below 133% of poverty and children 6-11 with family incomes below 100% of poverty.
|Center for Budget and Policy Priorities (Summer et al. 1997)||CPS 3/95||1994||Estimate definition:
Participation rate for children age 10.
Poverty related only. Not state-specific. Only the federal minimum poverty-related eligibility criteria for children age 10.
Participation rate for children age 10 who do not receive welfare cash assistance. This approximates the participation rate among children in poverty-related eligibility groups only.
2. Congressional Budget Office
Using 1996 CPS data, CBO (Bilheimer 1997) estimated that in 1995 the Medicaid participation rate was 60 percent for children age 0 to 18 who did not receive welfare cash assistance and did not have private insurance. Their report did not provide details on their methodology for estimating Medicaid eligibility. It is our understanding that CBO does not routinely adjust its CPS estimates for Medicaid underreporting.
3. SIPP Estimates of Medicaid Eligibility
To date, there have not been any Medicaid eligibility estimates published using the SIPP. Blumberg et al. (1997) adapted The Urban Institute's TRIM2 microsimulation model to run on SIPP data for their study of crowd-out, but they have not yet reported any Medicaid eligibility or participation rate estimates.
4. Estimates for Poverty-Related Expansion Children Only
Both the GAO (1996) and the CBPP (Summer et al. 1997) used the March 1995 CPS to estimate Medicaid participation for the subset of younger children eligible for Medicaid by federal mandate. As stated in the previous section, this included all uninsured children age 0 to 5 in families with incomes below 133 percent of poverty and uninsured children born after September 30, 1983 with family income below 100 percent of poverty. GAO did not calculate a Medicaid participation rate per se. However, it estimated that 14.3 million children under age 11 in 1994 were eligible for Medicaid by federal mandate because of age and family income. Of those, 11.4 million had private or public insurance coverage and 2.9 million were uninsured. Thus, they estimated insurance coverage (public or private) of 79.7 percent for this Medicaid eligible subset of children.
CBPP found that approximately two-thirds of the children age 0 to 10 who were eligible for Medicaid were enrolled in the program in 1994. When they excluded all children receiving welfare cash assistance from their calculations in order to isolate the participation rate among those eligible for Medicaid under the expanded poverty-related groups, they found that the participation rate dropped to 38 percent.(5)
C. PARTICIPATION RATES IN OTHER PUBLIC ASSISTANCE PROGRAMS
In this section, we examine the participation rates in other public assistance programs as a point of comparison to the Medicaid participation rates presented above. Specifically, we present estimates of participation rates for the Aid to Families with Dependent Children (AFDC) program and the Food Stamp Program (FSP). Comparisons of AFDC and FSP participation rates are not exactly comparable to Medicaid participation rates because one can be participating in Medicaid without actually receiving services, where as participants in AFDC and the FSP all currently receive benefits.
Ruggles and Michel (1987) estimated AFDC household participation rates for 1973 through 1984 using CPS data and The Urban Institute's TRIM2 microsimulation model, which adjusts for underreporting of AFDC participation in the CPS. During those years, they found that the participation rate ranged from 75 percent to 83 percent.
Stavrianos (1997), using SIPP data and the MATH® SIPP microsimulation model, which adjusts for the underreporting of FSP participation in the SIPP, found that the household participation rate for the FSP declined from 59 percent to 56 percent between 1985 and 1988 and then rose to 69 percent in 1992 and 1994. Stavrianos noted that in order to understand participation rate trends, it is necessary to examine the trends in eligibility and participation -- the two component parts of the participation rate. For example, the decline in the participation rate from 1985 to 1988 was largely due to legislative changes authorized by the 1985 Food Security Act. Although the act substantially increased the number of eligible households, most of the newly eligible households did not participate initially. Therefore, even though the number of FSP participants remained steady from 1985 to 1988, the participation rate fell.
Most of the Medicaid participation rates cited above are lower than the estimated participation rates in the AFDC and Food Stamp programs. Devaney, Ellwood, and Love (1997) cite four reasons that Medicaid may have a low participation rate. First, they note that there is a general lack of awareness that children can now qualify for Medicaid even if both parents are present in the home or one parent is working full time. Second, the time-consuming, sometimes difficult application process is an obstacle to many people. Third, families may not apply for Medicaid because of its stigma as a welfare program. Fourth, because most children are healthy, their parents may not feel there is a compelling reason for them to apply for Medicaid.(6) One additional reason that Medicaid may have a low participation rate is that the uninsured may have access to free care, either from physicians who choose not to deal with the Medicaid program, or from free clinics.
1. Throughout this chapter, the term "Medicaid eligible" will be used to refer to those that are eligible for Medicaid regardless of whether they are actually enrolled. This differs from HCFA(now known as CMS)'s definition of the term "Medicaid eligible," which denotes those that are enrolled but may or may not be receiving services.
2. Ribicoff coverage continues to be important for children in two-parent families born before October 1, 1983 (ages 14 to 18) who are not covered by the mandatory poverty-related expansions for children or by optional state expansions.
3. The Urban Institute is the only organization that accounts for Medicaid and cash welfare underreporting in their estimates of Medicaid participation rates.
4. The participation rate among children whose eligibility was tied to welfare participation was higher than that for those whose eligibility was tied to poverty related expansions. Using March 1994 CPS data and the TRIM2 model, Dubay and Kenney (1996) found that the Medicaid participation rate among children whose eligibility was tied to welfare participation was 90 percent versus 69 percent for children eligible under the poverty-related expansions.
5. Comparing Medicaid participation rates among the cash and non-cash eligibility groups is problematic using the CPS because the Census Bureau assigns Medicaid to children in all families that receive AFDC and in most families that receive SSI. In other words, the Medicaid participation rate for the cash groups in the CPS is, by definition, 100 percent.
6. Related to this fourth reason is the argument that many parents may feel their children are, in fact, covered by Medicaid because their children can be enrolled once they get sick. Of course, these children would not be receiving preventive health services.
Chapter V. Characteristics of Uninsured Children Who Are Eligible for Medicaid
The last aspect of this literature review focuses on what researchers have discovered about the group of uninsured children who appear to be eligible for Medicaid, but not participating. In particular, how do they differ from children who participate in Medicaid? Two overall questions are involved:
- How do the financial and demographic characteristics of these children and their families differ from those of children participating in Medicaid?
- What are the longitudinal characteristics of eligible nonparticipants? How long are they uninsured, and how long are they eligible for Medicaid?
The first question is important in helping determine which groups of children need to be targeted for federal and state outreach efforts to increase health insurance coverage. The second question is important in understanding the dynamics of the uninsured. Family financial situations may fluctuate considerably from month to month, so that children may be uninsured only for short periods of time. Or, some months families may have income so near the eligibility thresholds that they may not realize they are eligible. The Balanced Budget Act of 1997 gave states the option of extending 12-month continuous eligibility to children, but little information has existed to determine what the impact of this option might be. It would be helpful to understand better the events that trigger uninsurance and potential Medicaid eligibility, and how long the spells of uninsurance and Medicaid eligibility last. Unfortunately, little research has focused on these questions.
An article by Blumberg et al. (1997) provided some information on the dynamics of uninsured children eligible for Medicaid but not participating. They used the 1990 SIPP panel to examine the distribution of transitions into Medicaid by previous insurance status and found that 53 percent of those enrolling in Medicaid were previously uninsured while 46 percent were covered by private health insurance. They did not, though, examine whether and how long the previously uninsured that transitioned into Medicaid were eligible for Medicaid.
Chapter VI. Summary and Next Steps
Although there is more agreement about the number and proportion of uninsured children than in the past, there is still a rather substantial difference of opinion. For the most part, estimates differ depending on whether adjustments for Medicaid enrollment, based on HCFA(now known as CMS) administrative data, are made to national survey data. This difference of opinion caused the 1995 estimates of uninsured children under age 18 using CPS data to range from 6.9 million (9.8 percent) according to the Urban Institute (which adjusts for Medicaid under reporting in the CPS) to 9.8 million (13.8 percent) according to EBRI, the Census Bureau, CBO, and others (who do not adjust for Medicaid under reporting in the CPS) -- a difference of about 30 percent, or nearly 3 million children. This issue of Medicaid under reporting exists with both the CPS and the SIPP, the most widely used sources for estimates. For example, unadjusted CPS data showed 4.9 million fewer children (23 percent) enrolled in Medicaid in 1995 than HCFA(now known as CMS) administrative data showed. This difference represents more than half of the children thought to be uninsured according to the CPS, although under reporting of Medicaid can occur among the privately insured as well as the uninsured.
Other issues are also involved in counting uninsured children. There is increased awareness of how critical the time period definition is to estimates. The CPS is designed to measure the number of children uninsured throughout a given year. Yet, most researchers believe the CPS estimates of the uninsured are closer to point in time estimates -- meaning the commonly used annual estimates of the number of uninsured children may be overstated. The extent of churning among uninsured children is not well-understood yet. Some analyses of SIPP data suggest that almost one-third of all children will experience some period of noncoverage over a 2 ½ year period. However, research on the length of spells of noncoverage for children is still in the early stages, and there appear to be some inconsistent results.
Another concern is how the national surveys word their questions about insurance coverage. With both the CPS and the SIPP the number of uninsured must be defined as a residual -- that is, the uninsured are those who do not report receiving coverage of any type. The recent CTS survey asked respondents directly whether they were in fact uninsured, and the study authors believe this was a factor in their estimate of the number of uninsured children nationwide being lower than the CPS. In any event, the problem of asking about coverage, especially public coverage, may get worse before it gets better. With the moves to managed care, state insurance programs for children, and other state health reform efforts, there may be even more confusion in the future as to what types of coverage low-income persons have. This underscores the value of the CTS survey's approach of asking respondents directly whether they are uninsured if they fail to indicate coverage.
In response to the recently passed expansions in insurance coverage for low-income children, states would like more information on counting uninsured children at the state level. However, the national surveys generally lack a sufficient sample size to support state-specific estimates of uninsured children, although some researchers are now combining multiple years of data in order to produce state level estimates.
Most research on the characteristics of the uninsured has used cross-sectional data, without taking into account the potential heterogeneity of the population by length of uninsurance. Nevertheless, most researchers agree on the following overall profile:
- The uninsured are found in every age group of children, even among the very young supposedly covered by past expansion efforts under Medicaid
- Over two-thirds of uninsured children live in families with family income less than 200 percent of poverty
- About 70 percent live in two-parent families; further, 64 percent have one parent working full-year, full-time
- The vast majority (80 percent) of uninsured children have one parent who is also uninsured
- Even though a majority of uninsured children are white, minorities (especially Hispanics) are disproportionately represented
- Uninsured children also include a disproportionate number of noncitizens (10 percent versus 4 percent in the general population of children)
- There are inconsistencies in the research about how long children are uninsured and the extent to which there is churning among the population
As would be expected (given disagreements about the number of uninsured), researchers do not agree on the number of children who are Medicaid eligible or Medicaid participation rates, although most agree that Medicaid participation rates are lower than participation rates in the AFDC and Food Stamp programs. Some differences in estimating Medicaid eligibles result from the extent to which the simulation models used to develop estimates reflect Medicaid programmatic and state-specific eligibility rules, which can be very complex. However, this source of differences will probably diminish in the future, as expansions to all poverty-related children are implemented.
There is little research about the characteristics of uninsured children who are eligible for Medicaid but not participating. This confounds the development of outreach efforts to increase public coverage for uninsured children.
This literature review was intended to assist the project team in understanding what research has been done to date on uninsured children and help them refine topics for further study. Clearly, many questions remain unanswered about the extent of uninsurance among children and the characteristics of those who are uninsured. Five topics were identified by ASPE as needing further investigation:
- Insured versus uninsured children. What is distribution of both the number and lengths of spells of uninsurance for children? What factors are associated with short-term and long-term spells? What is the extent of churning or turnover? What are the trigger events leading to uninsurance and then coverage? What are the branching probabilities (i.e. what type of insurance did the child have prior to becoming uninsured and what type of insurance does the child obtain after being uninsured)?
- Medicaid eligibility among children: population dynamics. What are the implications of churning for Medicaid eligibles? To what extent are there short-term and long-term eligibles and how do these subgroups differ? What are the events triggering Medicaid eligibility and do they differ between short-term and long-term eligibles? What are the characteristics of children leaving Medicaid and what is their insurance status?
- Medicaid eligibility among children: information from examining alternative data bases and trends over time. What are the alternative ways of defining Medicaid eligibility when examining various data sources? For example, would examining two SIPP waves several years apart provide useful insights (1983-1984 and 1993-1994, or 1986-1988 to include a recession), or is this too complicated? Have participation decisions/rates and the characteristics of Medicaid enrollees changed over time?
- Medicaid participation among children: rates and dynamics. What are the issues involved in determining the Medicaid participation rate and how do rates differ, dependent on the methodology? Are there events which seem to trigger Medicaid eligibility and participation?
- Medicaid participation among children: characteristics of participants and comparison with eligible non-participants and other program populations. What factors correlate with Medicaid participation? How do participants and eligible non-participants compare? How do children enrolled in Medicaid compare with children participating in the AFDC and Food Stamp programs?
In addition to these five topics, MPR is responsible for suggesting two additional topics for study. Based on the findings from the literature review, two alternatives include:
- Medicaid Enrollment Patterns Using HCFA(now known as CMS) Administrative Data. What is the distribution of Medicaid enrollment spells, using linked years of Medicaid person-based data in selected states? What are the characteristics of children with short and long spells (with adjustments for age)?
- Medicaid Enrollee Characteristics: Comparisons of HCFA(now known as CMS) Administrative Data and SIPP. What are the differences in enrollee characteristics between HCFA(now known as CMS) person-based Medicaid administrative data (in selected states) and SIPP data? This analysis would explore type of eligibility (receiving cash assistance, medically needy, expansion groups), age, race, citizenship status (where available), county of residence. Do any differences provide insights regarding the problem of Medicaid under reporting in the national surveys?
Although beyond the scope of this project, the literature review also points to some other possible avenues for reconciling the CPS and SIPP data on Medicaid participation with HCFA(now known as CMS) administrative data. If resources were available, would it be feasible for the Census Bureau to link CPS and/or SIPP person-based data with person-based Medicaid enrollment data in selected states? Without this type of effort, it seems likely that estimates of uninsured children using national survey data will continue to be problematic. With the recent legislation providing funding to states to expand their coverage of children, more attention than before is likely to be focused on counting children who remain uninsured. If the problem of under reporting enrollment in public insurance program persists, how will these efforts be evaluated?
Akerlof, G. and B. Main. "Unemployment Spells and Unemployment Experience." American Economic Review, December 1980, pp. 885-93.
Bane, M. and D. Ellwood. "Slipping Into and Out of Poverty: The Dynamics of Spells." Journal of Human Resources, vol. 21, Winter 1986, pp. 1-23.
Beauregard, Karen M., Susan K. Drilea, and Jessica P. Vistnes. "The Uninsured in America -- 1996." MEPS Highlights, no. 1, May 1997.
Bennefield, Robert, L. "Health Insurance Coverage 1995." Current Population Reports, P60-195. Washington, DC: Census Bureau, 1996a.
Bennefield, Robert L. "Who Loses Coverage and for How Long?" Current Population Reports, P70-54. Washington, DC: Census Bureau, May 1996b.
Bennefield, Robert L. "A Comparative Analysis of Health Insurance Coverage Estimates: Data from CPS and SIPP." Presentation at the 1996 Joint Statistical Meetings of the American Statistical Association, 1996c.
Bennefield, Robert L. "Dynamics of Economic Well-Being: Health Insurance, 1991-1993." Current Population Reports, P70-43. Washington, DC: Census Bureau, 1995.
Bilheimer, Linda T. "CBO Testimony on Proposals to Expand Health Coverage for Children." Testimony before the Subcommittee on Health. U.S. House of Representatives, Committee on Ways and Means, Washington, DC, April 8, 1997.
Blumberg, Linda J., Lisa Dubay, and Stephen Norton. "Did the Medicaid Expansions for Children Displace Private Insurance? An Analysis Using SIPP Data." Presentation slides. Washington, DC: The Urban Institute, June 1997.
Davis, Karen, and others. "Health Insurance: The Size and Shape of the Problem." Inquiry, vol. 32, no. 2, Summer, 1995, 196-203.
Devaney, Barbara L., Marilyn R. Ellwood, and John M. Love. "Programs that Mitigate the Effects of Poverty on Children." The Future of Children, vol. 7, no. 2, Summer/Fall 1997, pp. 88-112.
Dubay, Lisa, and Genevieve Kenney. "Did Medicaid Expansions for Pregnant Women Crowd Out Private Coverage?" Health Affairs, vol. 16, no. 1, January/February 1997, pp. 185-193.
Dubay, Lisa, and Genevieve Kenney. "Revisiting the Issues: The Effects of Medicaid Expansions on Insurance Coverage of Children." The Future of Children, vol. 6, no. 1, Spring 1996, pp. 152-161.
Fronstin, Paul. "Expanding Health Insurance for Children: Examining the Alternatives." EBRI Issue Brief no. 187. Washington, DC: EBRI, July 1997a.
Fronstin, Paul. "Trends in Health Insurance Coverage." EBRI Issue Brief no. 185. Washington, DC: EBRI, May 1997b.
Fronstin, Paul. "Sources of Health Insurance and Characteristics of the Uninsured: Analysis of the March 1996 Current Population Survey." EBRI Issue Brief no. 179. Washington, DC: EBRI, November 1996.
Fronstin, Paul. "Sources of Health Insurance and Characteristics of the Uninsured: Analysis of the March 1994 Current Population Survey." EBRI Issue Brief no. 158. Washington, DC: EBRI, February 1995.
Fronstin, Paul. "Children Without Health Insurance: An Analysis of the Increase in Uninsured Children Between 1992 and 1993." Inquiry, vol. 32, Fall 1995, pp. 353-359.
Families USA. "One Out of Three: Kids Without Health Insurance 1995-1996." Washington, DC: March 1997.
Holahan, John. "Expanding Insurance Coverage for Children." Washington, DC: The Urban Institute: 1997.
Holahan, John, Colin Winterbottom, and Shruti Rajan. "A Shifting Picture of Health Insurance Coverage." Health Affairs, vol. 14, no.4, Winter 1995.
Lewin Group. "Exploring the Question of Substitution: Will Expansion of Health Insurance Coverage for Children Result in Shifts From Private to Public Coverage?" The Lewin Group, Inc., July 1997.
Lewit, Eugene M., and Linda Schuurmann Baker. "Health Insurance Coverage." The Future of Children, vol. 5, no. 3, Winter 1995, pp. 192-204.
Long, Stephen H. and M. Susan Marquis. "Some Pitfalls in Making Cost Estimates of State Health Insurance Coverage Expansions." Inquiry, vol. 33, Spring 1996, pp. 85-91.
Monheit, Alan C. "Underinsured Americans: A Review." Annual Review of Public Health, vol. 15, 1994, pp. 461-85.
Monheit, Alan C., and Claudia L. Schur. "The Dynamics of Health Insurance: A Tale of Two Cohorts." Inquiry, vol. 25, no. 3, Fall 1988, pp. 315-327.
National Health Interview Survey (NHIS). Copy of correspondence with Centers for Disease Control and Prevention, National Center for Health Statistics, October 1996.
O'Neill, J., L. Bassi, and D. Wolf. "The Duration of Welfare Spells." Review of Economics and Statistics, vol. 69, May 1987, pp. 241-248.
Reschovsky, James D., Peter J. Cunningham, and Jeremy D. Pickreign. "New Evidence on Uninsured Children in the U.S. from the Community Tracking Survey: Implications for the Implementation of CHIP." Washington, DC: Center for Studying Health System Change, 1997 (Draft).
Ruggles, Patricia, and Richard C. Michel. "Participation Rates in the Aid to Families with Dependent Children Program: Trends for 1967 through 1984." Washington, DC: The Urban Institute, April 1987.
Stavrianos, Michael. "Food Stamp Program Participation Rates: January 1994." Washington, DC: Mathematica Policy Research, Inc., March 1997.
Summer, Laura, Sharon Parrott, and Cindy Mann. "Millions of Uninsured and Underinsured Children are Eligible for Medicaid." Washington, DC: Center on Budget and Policy Priorities, 1997.
Swartz, Katherine. "Changes in the 1995 Current Population Survey and Estimates of Health Insurance Coverage." Inquiry, vol. 34, no. 1, Spring 1997, p. 70-79.
Swartz, Katherine. "Dynamics of People Without Health Insurance: Don't Let the Numbers Fool You." Journal of the American Medical Association, vol. 271, no. 1, January 5, 1994, pp. 64-66.
Swartz, Katherine. "Interpreting the Estimates from Four National Surveys of the Number of People Without Health Insurance." Journal of Economic and Social Measurement, vol. 14, 1986, pp. 233-243.
Swartz, Katherine and Timothy D. McBride. "Spells Without Health Insurance: Distributions of Durations and Their Link to Point-in-Time Estimates of the Uninsured." Inquiry, vol. 27, no. 3, Fall 1990, pp. 281-288.
Swartz, Katherine, John Marcotte and Timothy D. McBride. "Personal Characteristics and Spells Without Health Insurance." Inquiry, vol. 30, Spring 1993a, pp. 64-76.
Swartz, Katherine, John Marcotte and Timothy D. McBride. "Spells Without Health Insurance: The Distribution of Durations When Left-Censored Spells are Included." Inquiry, vol. 30, Spring 1993b, pp. 77-83.
Thorpe, Kenneth E. "Incremental Approaches to Covering Uninsured Children: Design and Policy Issues." Health Affairs, vol. 16, no. 4, July/August, 1997a, pp. 64-78.
Thorpe, Kenneth E. "Strategies for Covering Uninsured Children: Financial and Coverage Implications." Washington, DC: American Hospital Association, July 1997b.
U.S. General Accounting Office. "Uninsured Children and Immigration, 1995." Publication no. GAO/HEHS-97-126R. Washington, DC: GAO 1997.
U.S. General Accounting Office. "Health Insurance for Children: Private Insurance Coverage Continues to Deteriorate." Publication no. GAO/HEHS-96-129. Washington, DC: GAO, June 1996.
Vistnes, JP and AC Monheit. "Health Insurance Status of the Civilian Noninstitutionalized Population: 1996." MEPS Research Findings No. 1. (AHCPR Pub. No. 97-0030). Rockville, MD: Agency for Health Care Policy and Research, 1997.
Winterbottom, Colin, David W. Liska, and Karen M. Obermaier. State-Level Databook of Health Care Access and Financing. Second Edition. Washington, DC: Urban Institute Press, 1995.