Medicaid/SCHIP as Primary Source of Health Insurance During the Year

Order Code RL32555 CRS Report for Congress Received through the CRS Web Medicaid/SCHIP as Primary Source of Health Insurance During the Year August 31, 2004 Chris L. Peterson Analyst in Social Legislation Domestic Social Policy Division Congressional Research Service ˜ The Library of Congress Medicaid/SCHIP as Primary Source of Health Insurance During the Year Summary Although most Americans receive health insurance through an employer, 40.8 million individuals were enrolled in Medicaid or the State Children’s Health Insurance Program (SCHIP) at some point in 2001, representing 14% of the U.S. population. However, not all of these individuals relied on Medicaid/SCHIP as their primary source of health insurance. Medicaid/SCHIP enrollees are heterogeneous, particularly in terms of the duration of their enrollment and the extent to which they rely on Medicaid/SCHIP as their sole source of coverage. This heterogeneity is best understood by analyzing Medicaid/SCHIP enrollees in a given year as three distinct groups rather than a single group: ! ! ! enrollees who relied on Medicaid/SCHIP as their sole source of health insurance for the entire year (18.8 million); enrollees who relied on Medicaid/SCHIP as their sole source of health insurance for only part of the year (12.3 million); and enrollees who were enrolled in Medicaid but who never relied on it as their sole source of health insurance during the year; that is, when they had Medicaid, they had other coverage as well (9.7 million). Among enrollees who relied on Medicaid/SCHIP as their sole source of health insurance for the entire year, more than two-thirds (69%) were children (under age 19) and none were aged, since virtually all the aged had Medicare. On the other hand, among the Medicaid enrollees who never relied on it as their sole source of insurance, less than a quarter (23%) were children but nearly half (46%) were aged. Enrollees who relied on Medicaid/SCHIP as their sole source of health insurance for only part of the year were uninsured for two-thirds of the months they spent without solely Medicaid/SCHIP; job-based health insurance coverage comprised nearly all of the remaining months. Of these enrollees with part-year reliance on solely Medicaid/SCHIP, 1.8 million experienced a loss of job-based coverage during the year. Looking at the characteristics of the policyholders linked to these individuals (usually a parent or spouse), one finds substantial disruption in their employment and insurance status. For example, 61% of individuals were linked to policyholders who experienced a drop in their wages or had some period during the year in which they did not work. Those Medicaid enrollees who never relied on it as their sole source of health insurance always had some form of health insurance in addition to Medicaid. Twothirds (67%) of these individuals were enrolled in Medicare all year. The remainder were mostly covered by job-based coverage. The estimates in this report are based on monthly Medicaid/SCHIP enrollment reported in the 2001 Medical Expenditure Panel Survey (MEPS). This report will be updated to reflect new data or to expand on the analysis. Contents Medicaid/SCHIP As A Source of Health Insurance . . . . . . . . . . . . . . . . . . . 1 Enrollees with Part-Year Reliance on Medicaid/SCHIP . . . . . . . . . . . . . . . . 6 Loss of Job-Based Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Data Sources and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Methodology and Variable Definition . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 List of Figures Figure 1. Primary Source of Health Insurance, Among Those Ever Enrolled in Medicaid/SCHIP, By Months of Primary Coverage in Medicaid/SCHIP, 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 Figure 2. Characteristics of Medicaid/SCHIP Enrollees in 2001, by Portion of Year Enrolled . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Figure 3. Source of Health Insurance Prior to Medicaid/SCHIP Enrollment, Among Those Who Gained and Ever Relied on Medicaid/SCHIP as Sole Source of Health Insurance in 2001 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Figure 4. Changes in Policyholder’s Work Status and Wages, Among Individuals Who Ever Relied on Medicaid/SCHIP as Sole Source of Health Insurance and Who Lost Job-Based Coverage in 2001 . . . . . . . . . 7 List of Tables Table 1. Selected Characteristics of U.S. Noninstitutionalized Population and Medicaid/SCHIP Enrollees, with 90% Confidence Intervals, 2001 . . . 10 Table 2. Selected Characteristics of Those Who Relied on Medicaid/SCHIP as Their Sole Source of Health Insurance for Part of the Year and Who Lost Job-Based Coverage in 2001, with 90% Confidence Intervals . . . . . . 11 Medicaid/SCHIP As Primary Source of Health Insurance During the Year Medicaid is a joint federal/state entitlement program that makes payments for the health care and long-term care of low-income children and families, disabled individuals, and the aged (those 65 and older). The State Children’s Health Insurance Program (SCHIP), created through the Balanced Budget Act of 1997 (P.L. 105-33), is also a joint federal/state program and generally provides health insurance to uninsured children in families with income just above the state’s Medicaid eligibility levels. Under SCHIP, states can cover children either under their existing Medicaid programs or create separate programs modeled after private insurance. This report examines the characteristics of individuals enrolled in Medicaid or SCHIP at some point in 2001, the most recent year in which month-by-month data are available on health insurance enrollment. The source of data used for this report is the 2001 Medical Expenditure Panel Survey (MEPS), which contains estimates of monthly Medicaid/SCHIP enrollment along with other important information on the noninstitutionalized U.S. population. For this report, all of the estimates refer to the noninstitutionalized U.S. population. Like most nationally representative general population surveys, MEPS does not include the institutionalized (for example, those in nursing homes). Because the findings in this report are based on survey data, they are subject to sampling error and other limitations, which are discussed in greater detail in the “Data Sources and Limitations” section.1 Medicaid/SCHIP As A Source of Health Insurance Although most Americans receive health insurance through an employer, 40.8 million individuals were enrolled in Medicaid/SCHIP at some point in 2001, representing 14% of the U.S. noninstitutionalized population. However, not all of these individuals relied on Medicaid/SCHIP as their primary source of health insurance or had Medicaid/SCHIP for the entire year. Figure 1 illustrates the extent to which enrollees relied on Medicaid/SCHIP throughout the year as their sole source of health insurance. The first group of particular interest consists of those who relied on Medicaid/SCHIP as their sole source of health insurance for the entire year. As illustrated in the bottom section of the figure, these 18.8 million individuals account for less than half of those enrolled in Medicaid/SCHIP in 2001. 1 Because MEPS is the source of data for this report, the estimates may differ from those using other sources of data, including administrative enrollment data and other surveys, such as the Current Population Survey (CPS). CRS-2 The second group is the 12.3 million individuals who relied on Medicaid/SCHIP as their sole source of health insurance for only part of the year. This group is illustrated in the middle of Figure 1. On average, two-thirds of months in which these individuals were not covered solely by Medicaid/SCHIP were spent uninsured; job-based health insurance comprised nearly all of the remaining months. The third and final group is the 9.7 million individuals who were enrolled in Medicaid but who never relied on it as their sole source of health insurance. These enrollees always had some form of health insurance in addition to Medicaid. As shown in upper section of the figure, two-thirds (67%) of these individuals were enrolled in Medicare all year.2 The remainder were mostly covered by job-based coverage. Current law does not prohibit Medicaid enrollees from having other forms of health insurance; whether a person has health insurance is not a factor in determining one’s eligibility for Medicaid. However, according to current law, individuals are not eligible for SCHIP if they have any other source of health insurance. The “dual eligible” individuals (those covered by both Medicare and Medicaid) qualified for Medicare when they turned age 65 or because of a disability. These individuals also qualified for Medicaid because of their low income or high healthcare spending relative to their income. For dual eligibles, Medicare is the primary payer of their acute healthcare costs; Medicaid covers much of these individuals’ cost-sharing under Medicare as well as benefits not currently covered by Medicare, such as prescription drugs. The three groups of Medicaid/SCHIP enrollees discussed above differ from each other in ways besides their sources and duration of health insurance coverage. The groups are compared in Figure 2 based on commonly used characteristics — age, race/ethnicity, income-to-poverty ratio. 2 When people have multiple sources of health insurance at the same time, analysts often create a hierarchy to label one of those sources “primary.” For this report, Medicare was at the top of the hierarchy. In other words, Medicare was considered the primary coverage for Medicare enrollees, regardless of what other coverage they may have had. From there, the hierarchy proceeds as follows: job-based; all other private coverage (e.g., nongroup) and military or veterans coverage; and Medicaid. Medicaid is the last among the sources of coverage because it is the “insurer of last resort”; Medicaid pays for services only after all other health insurance sources have paid. If a person had none of these sources of coverage, they were considered uninsured. CRS-3 Figure 1. Primary Source of Health Insurance, Among Those Ever Enrolled in Medicaid/SCHIP, by Months of Primary Coverage in Medicaid/SCHIP, 2001 Othera 40 Medicaid enrollees who never relied on just Medicaid, always had other coverage (9.7 million) Job-based Number of Medicaid/SCHIP Enrollees (millions) Medicare All non-Medicaid/ SCHIP coverageb 30 Uninsured Medicaid/SCHIP was sole source of coverage for part of the year (12.3 million) Medicaid/SCHIP 20 10 Medicaid/SCHIP was sole source of coverage for c entire year (18.8 million) Medicaid/SCHIP 0 1 2 3 4 5 6 7 8 9 10 Average months enrolled in primary source of health insurance 11 12 Source: Congressional Research Service (CRS) analysis of the Household Component of the 2001 Medical Expenditure Panel Survey (MEPS). Note: Horizontal slivers showing three or more types of coverage represent the primary coverage (in person-months) for those who did not comprise a large enough group to otherwise be visible in this figure. When people have multiple sources of health insurance at the same time, analysts often create a hierarchy to label one of those sources “primary.” For this report, Medicare was at the top of the hierarchy. In other words, Medicare was considered the primary coverage for Medicare enrollees, regardless of what other coverage they may have had. From there, the hierarchy proceeds as follows: job-based; all other private coverage (e.g., nongroup) and military or veterans coverage; and Medicaid. If a person had none of these sources of coverage in a month, they were considered uninsured. a. Nongroup coverage, group coverage not related to a person’s employment, and military or veterans coverage. b. Usually job-based coverage, but also includes Medicare, nongroup coverage, group coverage not related to a person’s employment, and military or veterans coverage. c. Includes 1 million infants born in 2001 who were covered solely by Medicaid/SCHIP from birth. CRS-4 Figure 2. Characteristics of Medicaid/SCHIP Enrollees in 2001, by Portion of Year Enrolled 80% Race/Ethnicity Age 100% 100% Adults<65 31% Adults<65 42% Aged 46% 80% 60% 40% 40% Children 58% Adults<65 31% 20% 20% Children 23% 0% Hispanic 25% Other, 5% Hispanic 26% Other, 6% Hispanic 14% Black 29% Black 19% Part of year Never Portion of year with Medicaid/SCHIP as sole source of coverage 200%+ 17% 200%+ 23% 100%-199% 38% 100%-199% 34% 40% White 40% White 51% White 57% 200%+ 39% 100%-199% 32% 60% 0% Entire year Income-to-Poverty Ratio 80% Black 23% 60% Children 69% Other, 6% 100% 20% <100% 52% <100% 39% <100% 27% 0% Entire year Part of year Never Portion of year with Medicaid/SCHIP as sole source of coverage Entire year Part of year Never Portion of year with Medicaid/SCHIP as sole source of coverage Source: Congressional Research Service (CRS) analysis of the Household Component of the 2001 Medical Expenditure Panel Survey (MEPS). Note: These figures are based on the estimated 40.8 million noninstitutionalized individuals ever enrolled in Medicaid/SCHIP in 2001. An estimated 18.8 million individuals relied on Medicaid/SCHIP as their sole source of health insurance for the entire year. Medicaid/SCHIP was the sole source of coverage for part of the year for 12.3 million individuals. Another 9.7 million were enrolled in Medicaid during the year but never relied on it as their sole source of coverage. Totals may not sum to 100% due to rounding. Hispanics may be of any race. In the race/ethnicity chart, whites and blacks are those who are non-Hispanic. Among non-Hispanics, individuals who report any other race (e.g., Asian, American Indian) are categorized as “other.” CRS-5 Among the enrollees who relied on Medicaid/SCHIP as their sole source of health insurance for the entire year, more than two-thirds (69%) were children (under age 19) and none were aged, since virtually all the aged had Medicare. On the other hand, among the Medicaid enrollees who never relied on it as their sole source of health insurance, less than a quarter (23%) were children but nearly half (46%) were aged. Figure 2 also shows differences by race. For example, among the enrollees who relied on Medicaid/SCHIP as their sole source of health insurance for the entire year, 40% were non-Hispanic white. However, this racial/ethnic group made up 57% of Medicaid enrollees who never relied on it as their sole source of health insurance. More than half (52%) of those who relied on Medicaid/SCHIP as their sole source of health insurance for the entire year had income below the poverty level, as shown in Figure 2. On the other hand, 27% of Medicaid enrollees who never relied on it as their sole source of health insurance were poor. In fact, 39% of Medicaid enrollees who never relied on it as their sole source of health insurance had incomes at least twice the poverty level, compared to 17% of those who relied on Medicaid/SCHIP as their sole source of health insurance for the entire year.3 Many of these are elderly individuals with recurring medical expenses that deplete their income to levels that allow them to qualify for Medicaid. Figure 3. Source of Health Insurance Prior to Medicaid/SCHIP Enrollment, Among Those Who Gained and Ever Relied on Medicaid/SCHIP as Sole Source of Health Insurance in 2001 Uninsured 85% Job-based 14% Other 1% Source: Congressional Research Service (CRS) analysis of the Household Component of the 2001 Medical Expenditure Panel Survey (MEPS). Note: These figures are based on the estimated 8.0 million noninstitutionalized individuals who gained Medicaid/SCHIP in 2001 and ever relied on that coverage as their sole source of coverage in 2001. 3 The 2001 poverty threshold for a single nonaged adult was $9,214 and increases with family size. For a single aged adult, the poverty threshold was $8,494. CRS-6 Enrollees with Part-Year Reliance on Medicaid/SCHIP In 2001, 12.3 million individuals relied on Medicaid/SCHIP as their sole source of coverage but for less than the entire year, as previously mentioned.4 This section focuses on this population, of which more than three-quarters (77%) were uninsured at some point in 2001. Of those with part-year reliance on Medicaid/SCHIP, 8 million gained Medicaid/SCHIP in 2001. Of those, 85% (6.8 million) were uninsured just before obtaining that coverage, as shown in Figure 3. Nearly all of the remainder (1.1 million) were covered by job-based coverage before obtaining Medicaid/SCHIP. Of the 5.1 million individuals who relied on Medicaid/SCHIP but then lost that coverage, 87% (4.5 million) became uninsured immediately afterward. Eleven percent (less than 600,000) had job-based health insurance in the month following their final month in Medicaid/SCHIP. Loss of Job-Based Coverage. Since 2000, the percentage of people with job-based health insurance has steadily fallen while the percentage with public coverage has risen. The decreases in job-based coverage since 2000 could have occurred for a number of reasons, including the following: ! ! ! Economic conditions. The softening economy could have led to falling rates of job-based coverage as well as the loss of jobs that provided health insurance. Many of those who lost job-based coverage may have enrolled in Medicaid or SCHIP rather than becoming uninsured. Increases in premiums. In 2001, job-based health insurance premiums increased by 10.9% while workers’ earnings and overall inflation increased by less than half that amount.5 The cost of health insurance may have compelled many workers to forgo their employers’ coverage. Some of these people may have been able to obtain Medicaid/SCHIP for themselves or their dependents. Availability of Medicaid/SCHIP. Awareness of Medicaid/SCHIP eligibility could have led some employers or workers to drop their job-based coverage and caused increased enrollment in publicly available programs. In general terms, some call the substitution of public coverage for private insurance “crowd-out.” However, experts on crowd-out define the term more 4 Even though these individuals relied on Medicaid/SCHIP for only part of calendar year 2001, that enrollment could have marked the beginning of a full year or even multiple years of enrollment. Likewise, these individuals could have lost their Medicaid/SCHIP coverage after years of coverage but be considered to have relied on Medicaid/SCHIP for part of the year simply because of the months included in this analysis. 5 “Employer Health Benefits: 2003 Annual Survey,” The Kaiser Family Foundation and Health Research and Educational Trust, 2003, p. 2, at [http://www.kff.org/insurance/ ehbs2003-1-set.cfm]. CRS-7 narrowly: the percentage of enrollees who obtain coverage from a public program expansion but would have been covered by some other form of health insurance in the absence of the expansion. Using the narrower definition, measuring crowd-out is not straightforward. It requires determining the portion of enrollment growth attributable to the program expansion and how that growth would differ in the hypothetical world in which the expansion did not take place. The results depend on the data used and the assumptions made, which causes the estimates to vary wildly.6 Thus, the analysis that follows is not called a crowd-out analysis. Approximately 1.8 million individuals who relied on Medicaid/SCHIP as their sole source of coverage at some point during 2001 lost their job-based health insurance during the year. Looking at the characteristics of the policyholders through whom individuals had coverage, one finds substantial disruption in their employment and insurance status. Nearly half of these Medicaid/SCHIP enrollees (47%, 860,000) were linked to policyholders who had some period during the year in which they did not work, as shown in Figure 4. An additional 14% (260,000) were linked to policyholders who never went without work but experienced a drop in their wages, either by having a wage reduction in their current job or by taking a different job with a lower wage. Only 39% (720,000) were linked to policyholders who worked for the entire year and did not experience a wage drop. Figure 4. Changes in Policyholder’s Work Status and Wages, Among Individuals Who Ever Relied on Medicaid/SCHIP as Sole Source of Health Insurance and Who Lost Job-Based Coverage in 2001 No wage drop; always worked 39% Wage dropped but always worked 14% Went without work 47% Source: Congressional Research Service (CRS) analysis of the Household Component of the 2001 Medical Expenditure Panel Survey (MEPS). Note: These figures are based on the estimated 1.8 million individuals who ever relied on Medicaid/SCHIP as their sole source of coverage in 2001 and who lost job-based coverage in 2001. 6 The issue of crowd-out is not explored here. Secondary literature exists that discusses crowd-out and the complicated issues around it, such as how to measure it and the wide range of estimates produced by different researchers. For example, see “Public program crowd-out of private coverage: What are the issues?” by Gestur Davidson et al., Research Synthesis Report No. 5, Robert Wood Johnson Foundation, June 2004, at [http://www.rwjf.org/publications/synthesis/reports_and_briefs/pdf/no5_researchreport.pdf]. CRS-8 Data Sources and Limitations The source of data for this report is the Household Component of the 2001 Medical Expenditure Panel Survey (MEPS), which contains estimates of monthly Medicaid/SCHIP enrollment along with other important information on the noninstitutionalized U.S. population. MEPS is administered by the federal Agency for Healthcare Research and Quality (AHRQ). Most surveys are cross-sectional, meaning that they obtain information from a set of respondents a single time. Cross-sectional surveys may occur regularly, but not for the purpose of obtaining follow-up information from the same individuals. However, MEPS is a longitudinal survey, meaning that it obtains information from the same respondents over a period of time in order to observe changes in certain characteristics. MEPS interviews the same individuals every three to five months for two years. Because of its longitudinal design and the detailed health insurance questions it asks, MEPS is particularly useful for examining changes in individuals’ health insurance status.7 Methodology and Variable Definition. It is not possible to use MEPS to obtain estimates of SCHIP enrollment separately from Medicaid enrollment. The MEPS questionnaire obtains information about individuals’ Medicaid and SCHIP enrollment in the same question. However, even when using surveys that have separate variables for Medicaid and SCHIP, analysts usually combine the results because of respondents’ tendency to confuse the two types of coverage. In addition, the survey asks whether anyone in the household has coverage “through any state or local agency which provided hospital and physician benefits.” Similar questions appear in other health insurance surveys as well, and analysts typically subsume them under Medicaid/SCHIP. Experts at AHRQ recommended this approach for MEPS. In order to accurately estimate the prevalence of changes among Medicaid/SCHIP enrollees, 12 full months of data were needed, except for newborns. However, nearly 2% of individuals did not have 12 months of information. They were excluded, and the weights of the remaining individuals were increased proportionally to make up the difference. The enrollees who relied on Medicaid/SCHIP for part of 2001 and who lost jobbased coverage during 2001 were linked to the person most likely to have been the policyholder of that coverage.8 The first attempt to link individuals to their 7 For additional information, see CRS Report RL31275, Health Insurance: Federal Data Source for Analyses of the Uninsured, by Chris L. Peterson and Christine Devere. 8 These 1.8 million individuals consist of those who had employment-based health insurance in Jan. 2001 or afterward and had lost that coverage by Dec. 2001. This estimate does not include those who had job-based coverage in Dec. 2000 and lost that coverage by Jan. 2001, in part because only a third of the respondents in the survey in Jan. 2001 were in the survey in Dec. 2000. However, an analysis of those individuals found that the percentage of (continued...) CRS-9 policyholder was with the MEPS variable that groups individuals into Health Insurance Eligibility Units (HIEUs). HIEUs are sub-family relationship units constructed to include adults plus those family members who would typically be eligible for coverage under the adults’ private health insurance family plans. Of the 248 individuals in this sample, all but 24 matched using the HIEU. Those 24 were linked instead to the sole policyholder in the family or household. Limitations. Statistics from surveys such as the MEPS are subject to sampling and non-sampling error. Since survey estimates come from only a sample of the population, the estimates could differ from the results of a complete census using the same questions. All other sources of error in survey estimates are collectively called nonsampling error. Sources of nonsampling error include the following: ! ! ! differences in individual respondents’ interpretation of questions; respondents’ inability to recall information; and errors made in data collection, such as recording and coding data. Although it is not possible to measure the extent of nonsampling error, it is possible to calculate the sampling error, based on measurements like the survey’s sample size. This allows analysts to calculate a confidence interval around each estimate. For example, based on responses for 32,122 individuals in the 2001 MEPS, the percentage of individuals uninsured at any time during the year is estimated at 22.7%. This single value is called the point estimate. The 90% confidence interval for this estimate ranges from 21.8% to 23.5%. In other words, if all possible samples were surveyed under essentially the same general conditions and the same sample design, approximately 90% of those samples would have an estimate between 21.8% and 23.5%. In this case, each of these numbers varies by less than 4% (approximately 0.85 percentage points) from the point estimate. As sample size decreases, the magnitude of the sampling error increases; performing the same analysis on a smaller group increases the size of the confidence interval. For example, consider the individuals who relied on Medicaid/SCHIP as their sole source of coverage for part of 2001 and also lost job-based health insurance during the year (an estimated total of 1.8 million people versus the national total of 284.2 million people). Of these individuals, 23.3% were estimated to have been uninsured at some point during the year, based on responses for 248 individuals. The confidence interval for this estimate ranges from 19.7% to 27.0%. These numbers vary from the point estimate by nearly 16% (approximately 3.65 percentage points), rather than the 4% above, demonstrating that smaller sample sizes yield wider confidence intervals. Additional examples of confidence intervals from this report are shown in Table 1 and Table 2. 8 (...continued) individuals who lost coverage at the end of the year was the average monthly amount for all of 2001. In other words, it appears that those who relied on Medicaid/SCHIP for part of 2001 were no more likely to have lost job-based coverage at the end of the year 2000 than for any given month in 2001. CRS-10 Table 1. Selected Characteristics of U.S. Noninstitutionalized Population and Medicaid/SCHIP Enrollees, with 90% Confidence Intervals, 2001 Population (est. 284.2 million) 22.7% Ever uninsured during year Uninsured entire year 11.7% Enrolled in Medicaid/SCHIP 14.4% Age Children 27.1% Non-aged adults 60.3% Aged 12.6% Race White 69.7% Black 12.3% Hispanic 13.3% Other 4.7% Income-to Less than 100% 11.9% Poverty Ratio 100% to 199% 18.5% 200% and above 69.7% Ever uninsured during year Uninsured entire year Enrolled in Medicaid/SCHIP Age Children Non-aged adults Aged Race White Black Hispanic Other Income-to Less than 100% Poverty Ratio 100% to 199% 200% and above 21.8%-23.5% 11.1%-12.2% 13.6%-15.2% 26.5%-27.7% 59.7%-60.9% 12.1%-13.1% 68.4%-71.0% 11.5%-13.1% 12.3%-14.2% 4.1%-5.3% 11.2%-12.5% 17.7%-19.3% 68.5%-70.9% Among enrollees, Medicaid/SCHIP was sole source of coverage for: Enrolled in Part of the year Medicaid/SCHIP Entire year (est. (est. 12.3 Never (est. 9.7 (est. 40.8 million) 18.8 million) million) million) 23.9% 0.0% 76.8% 2.9% 0.0% 0.0% 0.0% 0.0% 100.0% 100.0% 100.0% 100.0% 54.5% 68.7% 57.5% 23.1% 34.5% 31.3% 42.0% 31.1% 11.0% 0.0% 0.0% 45.9% 47.3% 39.9% 50.6% 57.3% 24.6% 29.4% 18.8% 22.7% 22.3% 24.6% 25.7% 13.6% 5.8% 6.2% 4.9% 6.4% 42.2% 51.6% 39.4% 27.4% 34.2% 31.8% 37.8% 34.2% 23.7% 16.7% 22.7% 38.5% 90% Confidence intervals 22.3%-25.5% NA 74.0%-79.7% 1.8%-4.0% NA NA NA NA NA NA NA NA 53.1%-55.9% 67.0%-70.3% 55.0%-60.1% 20.0%-26.2% 33.2%-35.8% 29.7%-33% 39.4%-44.6% 28.6%-33.6% 10.0%-12.1% NA NA 42.3%-49.4% 44.6%-49.9% 36.4%-43.3% 47.2%-54.0% 53.9%-60.7% 22.2%-27.0% 25.4%-33.3% 16.0%-21.7% 19.8%-25.7% 20.2%-24.5% 21.7%-27.6% 23.2%-28.3% 11.4%-15.8% 4.3%-7.3% 3.9%-8.4% 3.3%-6.4% 4.5%-8.3% 40.2%-44.1% 48.8%-54.3% 36.2%-42.7% 24.4%-30.3% 32.4%-36.0% 29.2%-34.3% 34.7%-41.0% 30.9%-37.5% 22.0%-25.3% 14.7%-18.6% 20.1%-25.4% 35.4%-41.6% Source: Congressional Research Service (CRS) analysis of the 2001 Medical Expenditure Panel Survey (MEPS). CRS-11 Table 2. Selected Characteristics of Those Who Relied on Medicaid/SCHIP as Their Sole Source of Health Insurance for Part of the Year and Who Lost Job-Based Coverage in 2001, with 90% Confidence Intervals Ever uninsured during year Changes in policyholder’s job employment Policyholder went without work during year Policyholder experience wage drop but always worked Policyholder always worked; no wage drop Point estimate 23.3% 90% confidence interval 19.7%-27.0% 46.7% 14.3% 40.7%-52.6% 9.8%-18.7% 39.1% 33.5%-44.6% Source: Congressional Research Service (CRS) analysis of the 2001 Medical Expenditure Panel Survey (MEPS). Note: The percentages are out of an estimated total of 1.8 million individuals.