Abstract
Objectives. This study examined the association between health insurance coverage, medical care use, limitations in activities of daily living, and mortality among older Mexican-origin individuals.
Methods. We analyzed longitudinal data from the Hispanic Established Populations for Epidemiologic Study of the Elderly (H-EPESE).
Results. The uninsured tend to be younger, female, poor, and foreign born. They report fewer health care visits, are less likely to have a usual source of care, and more often receive care in Mexico. Conversely, those with private health insurance are economically better off and use more health care services. Over time, the data reveal substantial changes in type of insurance coverage.
Conclusions. The data reveal serious vulnerabilities among older Mexican Americans that result from a lack of private Medigap supplemental coverage. (Am J Public Health. 2002;92:1264–1271)
In the United States, employer-sponsored health insurance forms the basis of the health care financing system for the majority of working-age adults. For the middle class, private insurance is carried over into retirement, either as a direct retirement benefit or as a purchase made possible by an adequate retirement income. In contrast, for individuals employed in the service sector or in other low-productivity occupations that do not offer retirement or health benefits, or do so only at a cost that is prohibitive to the employee, health coverage in old age is often inadequate.1 The same factors that influence one’s tie to the labor force and one’s place in it, then, also affect one’s health care coverage in old age.2 These factors include Hispanic ethnicity, nativity, sex, and age at migration.3–6 At every age, a large fraction of Hispanics have either no or only inadequate health care coverage.5 The situation is particularly serious for individuals of Mexican descent, and especially for the foreign born within this group, who lag far behind non-Hispanics in health insurance coverage.7–9
For Hispanics, migration remains a significant component of population growth, and for many immigrants, especially those who come to the United States in mature or late adulthood with little formal education, employment is often restricted to those sectors and occupations that do not offer retirement or health care coverage packages. Even within broad occupational groupings, Hispanics are less likely to have health insurance coverage than are non-Hispanics. In the construction industry, for example, 64% of non-Hispanics hold jobs that offer health insurance, compared with only 46% of Hispanics.10 These working-age disadvantages place an individual at risk of inadequate health care coverage in old age.11 Jobs that do not offer benefits to working-age employees are unlikely to provide supplemental Medigap coverage in retirement or to provide enough of a retirement income to make purchasing private coverage possible. Foreign birth and older age at migration only increase this risk. In addition to the educational and occupational handicaps that the foreign born face, individuals who migrate later in life have less time to accumulate assets that might provide economic security and the ability to purchase supplemental health insurance in retirement.12,13
Social Security has clearly improved the situation of the elderly relative to younger age groups, but substantial racial and ethnic disparities in income and assets among the elderly persist. Hispanics are disproportionately represented among the poor in the United States across the life cycle,14–16 and older Hispanics are far less likely than older nonHispanic Whites to have private pensions or significant assets.3 Medicare has similarly reduced, but not eliminated, racial and ethnic disparities in access to health care.17,18 The relative health care disadvantage that Hispanics face in earlier years persists into old age and has potentially serious health consequences. Medicare premiums and copayments, as well as the portion of hospitalization costs that must be paid by the patient and the cost of uncovered services including prescription drugs, can be substantial.19 If such costs are too high, individuals may simply do without needed health care.
In this report, we examine health insurance coverage and changes in health insurance coverage among older Mexican Americans and identify factors placing individuals at risk of having no or insufficient health insurance within this population. Because people with private Medigap policies have more complete health care coverage than people without it, we were interested in identifying those factors that distinguish older individuals with and without supplemental private insurance. For the impoverished elderly without such private coverage, Medicaid represents the health care safety net. We therefore examine the extent of Medicare and Medicaid coverage among older individuals of Mexican origin. A common assumption is that one’s income and insurance status is fairly well fixed at age 65. By employing longitudinal data, we can examine that assumption as well as changes in the nature and amount of coverage during the latter phases of the life course. Although some health insurance transitions such as acquiring Medicaid as the result of poor health and impoverishment, are predictable, others are not.
METHODS
Data
We analyzed longitudinal data from the Hispanic Established Populations for Epidemiologic Study of the Elderly (H-EPESE), a large probability sample of Mexican American elderly who reside in Texas, California, New Mexico, Arizona, and Colorado. The original survey consisted of a sample of 3050 individuals aged 65 and older interviewed in 1993 and 1994 (time 1). Information from a knowledgeable proxy was obtained for individuals who were unable to complete the entire interview themselves because of infirmity or cognitive incapacity. The original panel was recontacted in 1995 to 1996 (time 2) and again in 1998 to 1999 (time 3). Although the modal age category was 65 to 74 years, almost 7.1% of the sample was older than 85. Nearly half of the sample was foreign born and more than half had household incomes below the poverty level. Only 11.8% had household incomes twice that level. Forty-eight percent were unmarried and living with others.
By the second interview, 224 of the original 3050 respondents had died and 109 refused to be reinterviewed; 278 others either could not be located or had moved to Mexico, where they could not be contacted. From baseline to time 3, there had been 628 deaths. Comparisons of those lost to follow-up at time 2 and those recontacted revealed no major differences according to sex, living arrangements, income, or nativity at time 1. On the other hand, those who were lost to follow-up were older than those who were recontacted. By time 3, a larger fraction of the sample had died, entered a nursing home, or been lost to follow-up. A comparison of baseline characteristics indicates that those lost to follow-up by time 3 were older and were more likely to have been living alone and to have had the lowest incomes at time 1. The 2 groups did not differ by sex or nativity.
Variables
In the following analyses, our focus was primarily on the type of health insurance individuals reported that they had. Our insurance classification consisted of the following: (1) Medicare only, (2) Medicaid, (3) private supplemental coverage, and (4) no insurance of any form. By definition, those individuals who received Medicaid and those who had private policies also qualified for Medicare. We also assumed that anyone who reported receiving any Social Security or Supplemental Security Income (SSI) also qualified for Medicare, even if they reported not receiving it. Information on health insurance was collected at baseline and again at time 2. No information on health insurance was collected at time 3, so our analysis of stability and change in health insurance coverage was confined to the first 2 waves of data. We used the third only to examine potential mortality consequences of transitions between the first 2 periods.
We examined certain health outcomes at time 2 as well. These included subjective assessments of health (excellent/good/fair/poor), problems with at least 1 of 6 activities of daily living (ADLs), and problems with at least 1 of 7 instrumental activities of daily living (I-ADLs). The ADL items included limitations in walking across a small room, bathing, personal grooming and dressing, feeding oneself, getting from a bed to a chair, and using the toilet. The I-ADL items included inability to use the telephone, drive a car, travel alone on buses or taxis, prepare meals, shop for groceries, do light housework, take medicine, and handle money. We also examined certain illness behavior items, including use of prescription drugs, having seen a doctor in the past year, having a usual source of care, having been hospitalized in the past year, and ever having received care in Mexico.
Several demographic characteristics that potentially influence health and health insurance coverage were used descriptively in the multivariate analyses. Age was coded as (1) 65 to 74 years, (2) 75 to 84 years, and (3) 85 years and older. Marital status/living arrangements was coded as (1) married, living with spouse only; (2) living alone; and (3) unmarried, living with others. Nativity and age at immigration were combined and coded as (1) native born, (2) immigrated between ages 1 and 19, (3) immigrated between ages 20 and 49, and (4) immigrated at age 50 or older. Annual household income was coded as (1) under $5000, (2) $5000 to $9999, (3) $10 000 to $14 999, and (4) $15 000 or more. Usual occupation before retirement was based on 1990 US Census Bureau occupational codes, which were combined into (1) managerial, professional, and specialty (white-collar); (2) manufacturing, production, craft, and repair (blue-collar); (3) service; (4) farming, forestry, and fishing; and (5) unemployed (including housewife).
RESULTS
Table 1 ▶ presents changes in health care coverage over the 2-year period between the baseline survey and the 1995/96 follow-up (time 2). Although we present all possible transitions for completeness, we focus only on those changes that are most significant from a policy perspective. Of those individuals who reported that they had only Medicare at time 1, more than 27% were receiving Medicaid by time 2. Nearly 18% had purchased private Medigap coverage, 7% had died, and 5% had entered a nursing home. These transitions are consistent with declining health and the subsequent increasing need for supplemental insurance. For older individuals with adequate income, private coverage is a possibility. For those without sufficient income, Medicaid is the only alternative.
TABLE 1.
—Changes in Health Insurance Coverage Among Older Mexican Americans, 1993/94 to 1995/96: Hispanic Established Populations for Epidemiologic Study of the Elderly
1995/96, % | |||||||||
Type of Health Insurance, 1993/94 | Medicare | Medicaid | Private | None | Don’t Know | Dead | Refused | Lost to Follow-Up | Weighted Marginals |
Medicare | 29.7 | 27.4 | 17.9 | 1.4 | 5.1 | 7.3 | 4.6 | 6.7 | 40.7 (1267) |
Medicaid | 21.4 | 45.7 | 5.5 | 0.5 | 4.5 | 8.0 | 3.4 | 11.0 | 33.0 (1069) |
Private | 12.3 | 5.2 | 59.0 | 0.1 | 1.9 | 7.0 | 8.0 | 6.7 | 19.4 (572) |
None | 5.6 | 16.7 | 3.9 | 33.5 | 4.5 | 5.5 | 3.2 | 27.2 | 6.8 (135) |
Don’t know | 11.2 | 29.9 | 13.8 | 0.0 | 8.7 | 12.7 | 12.8 | 10.9 | 0.2 (7) |
Weighted marginals | 24.7 (749) | 28.5 (1007) | 21.6 (616) | 2.9 (59) | 0.3 (8) | 7.5 (224) | 5.0 (109) | 9.6 (278) | 100.0 (3050) |
Note. Percentages are weighted. Unweighted numbers are given in parentheses.
Of those respondents with private coverage at time 1, 12% had only Medicare by time 2 and 5% were receiving Medicaid. Maintaining private coverage can be expensive on a fixed income. Many of those who had private coverage at time 1 were clearly quite ill, given that 7% had died and 8% had entered a nursing home by time 2. A total of 135 respondents, or nearly 7% of the original sample, reported that they had no insurance at time 1.
By the 1995/96 follow-up, only 59 individuals reported that they had no insurance. Of the remaining 76, 37 (27%) had been lost to follow-up, 7 (5.5%) had died, 7 were receiving Medicare, and 23 (16.7%) were receiving Medicaid. Five (3.9%) reported that they had private coverage.
Even in this economically disadvantaged sample, therefore, changes in income amounts and sources potentially affect Medicaid eligibility. Of course, given that Medicaid is associated with the use of medical services, some portion of the decrease in Medicaid coverage reflects improvements in health or the cessation of a medical care episode. Ancillary analyses (not shown) reveal that global ratings of general health improved between time 1 and time 2. More than 39% of the survivors reported excellent or good health at time 2, compared with 31% at time 1.
Table 2 ▶ presents demographic and health information for these same 6 insurance transition groups. It shows that respondents who reported no insurance at either interview (column 3) tended to be much poorer, to be more likely to have been unemployed throughout life, and to have come to the United States later in life than those in the other groups. By contrast, they tended to have somewhat better general health, fewer ADL or I-ADL limitations, and lower mortality by time 3.
TABLE 2.
—Demographic and Health Profiles of Health Insurance Transition Types, 1993/94 to 1998/99: Hispanic Established Populations for Epidemiologic Study of the Elderly
Changes (%) in Health Insurance Coverage, T2 | ||||||
Independent Variable, T2 | Lost Medicaid | Enrolled in Private Plan | No Insurance, T1 or T2 | Medicaid, Both T1 and T2 | Medicare, Both T1 and T2 | Private, Both T1 and T2 |
Hospital use | 22.9 | 21.7 | 13.6 | 25.1 | 19.1 | 21.2 |
Demographic characteristics | ||||||
Age, y | ||||||
65–74 | 65.8 | 65.0 | 71.7 | 62.6 | 67.7 | 69.0 |
75–84 | 28.6 | 33.2 | 21.8 | 29.7 | 25.0 | 28.8 |
≥ 85 | 5.7 | 1.8 | 6.6 | 7.7 | 7.2 | 2.2 |
Female | 61.0 | 50.4 | 61.0 | 64.6 | 54.4 | 49.7 |
Household income, $ | ||||||
0–4999 | 16.5 | 19.0 | 40.3 | 24.3 | 20.6 | 16.0 |
5000–9999 | 45.2 | 20.8 | 25.0 | 44.2 | 38.5 | 16.4 |
10 000–14 999 | 19.2 | 32.1 | 17.9 | 20.5 | 22.4 | 28.3 |
15 000 or more | 19.2 | 28.2 | 16.9 | 11.1 | 18.6 | 39.4 |
Not reported | 5.4 | 11.9 | 27.9 | 9.5 | 8.0 | 11.2 |
Primary occupation | ||||||
White-collar | 7.0 | 9.4 | 6.7 | 4.6 | 7.0 | 18.4 |
Service | 20.4 | 12.8 | 10.9 | 16.8 | 17.0 | 17.7 |
Farmer | 24.2 | 21.3 | 13.4 | 26.2 | 24.9 | 11.6 |
Blue-collar | 23.6 | 34.9 | 26.8 | 21.6 | 30.5 | 35.0 |
Unemployed | 24.9 | 21.6 | 42.2 | 30.7 | 20.6 | 17.4 |
Migrationa | ||||||
Native of US | 47.8 | 58.3 | 20.4 | 48.1 | 51.2 | 72.9 |
Early | 13.1 | 13.1 | 6.3 | 12.3 | 10.2 | 9.2 |
Midlife | 31.4 | 20.6 | 48.8 | 26.7 | 30.1 | 14.6 |
Old age | 7.7 | 6.4 | 20.4 | 13.0 | 8.5 | 3.4 |
Health and mortality | ||||||
Global health | ||||||
Excellent | 11.7 | 14.4 | 18.2 | 8.0 | 13.9 | 14.6 |
Good | 27.5 | 32.9 | 31.9 | 27.6 | 28.2 | 32.7 |
Fair | 46.3 | 39.1 | 37.8 | 41.2 | 42.9 | 39.9 |
Poor | 14.6 | 13.6 | 12.1 | 23.1 | 14.9 | 12.7 |
ADL limitation | 14.2 | 13.2 | 9.7 | 18.8 | 14.1 | 10.1 |
I-ADL limitation | 58.0 | 48.8 | 44.0 | 60.5 | 54.5 | 44.5 |
Death rate, T3 | 15.8 | 13.1 | 12.5 | 17.7 | 13.6 | 15.2 |
Unweighted | (235) | (227) | (117) | (1276) | (749) | (771) |
Note. T1 = 1993/94; T2 = 1994/95; T3 = 1998/99; ADL = activity of daily living; I-ADL = instrumental activity of daily living.
aImmigration strata are ages 1 to 19 (early), 20 to 49 (midlife), and 50 or older (old age).
Respondents who lost their Medicaid coverage (column 1), as well as those who maintained Medicaid (column 4), tended to have low incomes. Those who maintained Medicaid tended to be older, and both groups, like those who reported no insurance at either interview, tended to be female. In terms of health, both those who lost Medicaid and those who were covered by Medicaid at both interviews were more likely than those in the other groups to have fair or poor global healthand to have ADL and I-ADL limitations, and they had the highest mortality rates of all transition groups by time 3. One receives Medicaid only when one is ill enough to need it, and one receives it only as long as one is ill. For many individuals who enter a nursing home, Medicaid receipt is permanent; for others, it is not. Older individuals who need Medicaid at any time, however, even if they subsequently lose it, are clearly in poorer health than those who do not use it.
Respondents who acquired private insurance between time 1 and time 2 and those who had it at both interviews were far less likely to be female than those in the other groups, and they were far more likely to be in the highest income category, to have been blue-collar workers, and to be native born. Those with private insurance at both interviews were less likely than those in other groups to be farm workers. In terms of health, they were more likely than those with Medicaid to be in excellent or good health, less likely to have ADL or I-ADL limitations, and less likely to have died by time 3.
It is clear from these data that those individuals who reported no insurance were more recent immigrants who came later in their lives and were consequently poorer, more likely to be female, and more likely to have been unemployed or housewives throughout life. Such a combination of vulnerabilities can place one at a further disadvantage. Unemployed and unmarried Hispanic females, for example, are at the highest risk of having no insurance.1,4,20 Those with any private insurance, by contrast, are more likely to have been native born, to have higher incomes, and to have been employed in blue-collar occupations and less likely to have been farm workers.
Table 3 ▶ presents information on the health care use at time 2 of those who changed coverage or maintained the same coverage between time 1 and time 2. The first column refers to those who lost Medicaid, the second to those who enrolled in a private plan, and the third to those who reported no insurance at either time 1 or time 2. The final 3 columns refer to those whose coverage was the same at both interviews. The table reveals a clear health care disadvantage for those who at any point reported no insurance (column 3). These individuals were less likely to have seen a doctor in the previous year, less likely to have had a regular source of care or to have taken prescription drugs, and less likely to have been hospitalized than those in any other group. They were also more likely to have received care in Mexico. Among the other groups, those who had Medicaid at time 1 but not at time 2 (column 1), those who enrolled in a private plan (column 2), and those who had Medicaid at both times (column 4) were more likely than others to have had a large number of doctor’s visits. Those who had Medicare only (column 5) and those with continuous private coverage (column 6) had slightly fewer physician visits. All groups were similar in other health care characteristics, with the exception that those who lost Medicaid (column 1) were somewhat more likely to take a prescription drug, and those with any private coverage (columns 2 and 6) were far less likely to have reported that they had received care in Mexico at time 1. These data make it clear that individuals who reported no insurance at either interview were highly disadvantaged in terms of health care coverage and use. By contrast, those who either maintained or gained private coverage were not noticeably different from those with Medicare only.
TABLE 3.
—Health Service Use Patterns by Health Insurance Transitions, 1993/94 to 1995/96: Hispanic Established Populations for Epidemiologic Study of the Elderly
Changes (%) in Health Insurance Coverage, T2 | ||||||
Independent Variable | Lost Medicaid | Enrolled in Private Plan | No Insurance, T1 or T2 | Medicaid, Both T1 and T2 | Medicare, Both T1 and T2 | Private, Both T1 and T2 |
Hospital use | 22.9 | 21.7 | 13.6 | 25.1 | 19.1 | 21.2 |
No. of doctor visits | ||||||
0 | 11.8 | 9.3 | 46.6 | 13.1 | 18.0 | 17.3 |
1–2 | 17.2 | 17.8 | 20.3 | 14.9 | 21.3 | 23.8 |
≥ 3 | 70.9 | 72.9 | 33.1 | 72.0 | 60.8 | 59.0 |
No usual source of care, T1 | 7.1 | 4.3 | 21.9 | 4.1 | 6.5 | 3.6 |
Takes no prescription drugs | 24.8 | 29.3 | 65.1 | 32.0 | 34.7 | 30.4 |
Obtained care in Mexico | 3.9 | 2.9 | 7.0 | 4.3 | 6.2 | 2.1 |
Unweighted | (235) | (227) | (155) | (1276) | (749) | (771) |
Note. T1 = 1993/94; T2 = 1994/95.
Table 4 ▶ presents the results of multinomial logistic regressions examining the multivariate predictors of the 2 insurance categories into which most respondents fell at time 2: receiving Medicaid and having private health insurance. The comparison category is composed of those having Medicare only. There were too few individuals who reported no insurance or who entered a nursing home for us to analyze those outcomes, so those individuals are not included in this analysis. We present 3 models. The first controls for baseline coverage, age, sex, living arrangements, and household income. The second introduces usual occupation and the third, the age of migration for those who were not native born, to determine whether these 2 variables affect the associations. We control for baseline insurance coverage to isolate the net effects of the independent variables on transitions into Medicaid or private coverage. The coefficients in the models thus reflect the increased or decreased probability associated with each independent variable of having Medicaid or private insurance rather than Medicare only at time 2.
TABLE 4.
—Multinomial Regressions of Changes in Health Insurance and Covariates (n = 2323): Hispanic Established Populations for Epidemiologic Study of the Elderly
Time 2 (1995/96) | ||||||
Model 1 | Model 2 | Model 3 | ||||
Independent Variable, Time 1 (1993/94)a | Medicaid | Private | Medicaid | Private | Medicaid | Private |
Insurance | ||||||
Medicaid | 0.63*** | –0.65*** | 0.60*** | –0.63*** | 0.60*** | –0.64*** |
Private (Medicare only) | –0.36*** | 1.10*** | –0.34*** | 1.07*** | –0.34*** | 1.04*** |
Age, y | ||||||
75–84 | –0.02 | 0.03 | –0.03 | 0.04 | –0.05 | 0.02 |
≥ 85 (65–74) | 0.05 | –0.37 | 0.05 | –0.38* | 0.04 | –0.41* |
Female | 0.16** | 0.01 | 0.07 | –0.04 | 0.08 | –0.04 |
Living arrangements | ||||||
Alone | 0.09 | –0.17 | 0.11 | –0.17 | 0.12 | –0.18 |
Coresidence (couple only) | 0.29*** | –0.23** | 0.29*** | –0.22** | 0.28*** | –0.23** |
Household income, $ | ||||||
0–4999 | 0.43** | –0.70*** | 0.37*** | –0.66*** | 0.37*** | –0.65*** |
5000–9999 | 0.42*** | –0.55*** | 0.41*** | –0.51*** | 0.43*** | –0.51*** |
10 000–14 999 | 0.27** | –0.14 | 0.25* | –0.11 | 0.25* | –0.12 |
Missing ($10 000 or higher) | 0.34** | –0.13 | 0.30* | –0.10 | 0.29* | 0.08 |
Primary occupation | ||||||
Service | 0.04 | –0.21 | 0.05 | –0.19 | ||
Farmer | 0.22 | –0.31* | 0.23 | –0.28* | ||
Blue-collar | –0.06 | –0.19 | –0.04 | –0.17 | ||
Unemployed (white-collar) | 0.27* | –0.09 | 0.27* | –0.06 | ||
Age at migrationb | ||||||
Early | –0.02 | 0.04 | ||||
Midlife | –0.09 | –0.24** | ||||
Old age (native born) | –0.24* | 0.09 |
Note. Proportion in category 1 for Medicaid at time 2 = 0.425. Proportion in category 1 for private plan at time 2 = 0.265.
aComparison category is given in parentheses.
bImmigration strata are ages 1 to 19 (early), 20 to 49 (midlife), and 50 or older (old age).
*P ≤ .05; **P ≤ .01; ***P ≤ .001.
In Model 1, receiving Medicaid or having private coverage at time 1 is, not surprisingly, associated with a greatly increased probability of having the same kind of coverage at time 2 as opposed to having only Medicare. Advanced age reduces the probability of purchasing private coverage (odds ratio [OR] = 0.66), but it has no net effect on Medicaid. Living with someone other than a spouse is associated with increased Medicaid use (OR = 1.32) but a lower probability of having private coverage (OR = 0.80). Living with others is consistent with greater dependency, a traditional family orientation, low personal income, and poor health. Individuals with these characteristics are unlikely to have private coverage. Household income has a large net effect on insurance transitions. Having an annual household income of less than $15 000 increases the likelihood of receiving Medicaid, and having an extremely low household income (less than $10 000) reduces the chances of having private insurance coverage at the follow-up.
Model 2 introduces usual occupation. Having spent one’s life in farm work decreases the probability of gaining private coverage (OR = 0.76). Not having worked or having been a housewife is associated with a greater probability of moving on to Medicaid (OR = 1.31). These associations clearly demonstrate the later-life health insurance disadvantage of farm work.
In Model 3, we introduce nativity and age at migration. Native-born individuals are the reference, so the model contrasts those who came at different ages with those who were born in the United States. We examine the effect of age at migration because of clear evidence that individuals who immigrate later in life have a more difficult time adjusting both culturally and economically and remain more dependent on their families. The introduction of age at migration has no effect on the associations revealed in Model 2. Model 3, however, reveals a significant negative effect of very late immigration on the probability of receiving Medicaid (OR = 0.79). To receive Medicaid, one must provide proof of citizenship, whereas such proof is not necessary for Medicare. States impose strict requirements on qualification for Medicaid in addition to the income test. Late-life immigrants, therefore, may not qualify for Medicaid because of their short residence or their citizenship status.
Model 3 also reveals a strong negative association between midlife immigration and acquiring private supplemental coverage (OR = 0.70). Midlife migration reduces the likelihood of acquiring private insurance at time 2 (OR = 0.79). The effect is direct and suggests that, independent of occupational category, immigrating in mature adulthood increases the likelihood of employment in jobs that do not provide private health insurance coverage. Adult immigrants who find employment in farm or blue-collar work, therefore, face a cumulative disadvantage in terms of private supplemental health coverage. We also controlled for Spanish interview, an indicator of degree of acculturation. Doing so did not change the age-at-migration effect, although it did eliminate the effect of farming on Medicaid coverage at time 2. Additional control for home ownership, an indicator of personal assets, likewise did not modify the impact of migration on type of health coverage. As expected, individuals who owned a home were both less likely to receive Medicaid and more likely to purchase private insurance by time 2.
DISCUSSION
Although Medicare has clearly succeeded in providing health care coverage to the majority of elderly Americans, substantial differences, largely associated with previous employment and income, result in very different levels of coverage. Given the historical role of race and ethnicity in our system of occupational stratification, a disproportionate number of those with the lowest levels of coverage are minority Americans. Historically, Mexican Americans have been the most seriously underinsured members of our society at all ages.8 Our data reveal, however, that even among older Americans of Mexican origin, complete lack of insurance is relatively rare, and those who report not having any coverage are a very different population in all respects. Nevertheless, these findings must be qualified, given that some respondents are no doubt reporting incorrectly and perhaps not realizing that they qualify for Medicare.
Nonetheless, the possible existence of even a small number of older individuals who might fall through the health care safety net completely is cause for concern. Our data suggest that those older individuals who report that they have no insurance are the most economically disadvantaged and have the least medical care contact. Yet they are in relatively good health and have lower mortality than those with coverage. The fact that these individuals are more likely to have received care in Mexico suggests both closer ties to that country and the need to find less expensive alternatives than those available in the United States. Although we do not have information on citizenship status, it seems probable that noncitizens are less likely to use the full range of health care services in the United States and to return to Mexico for routine care if they are able to do so. Current debates over the possibility of regularizing the situations of illegal immigrants have potentially important implications for the health care coverage of this segment of the population.21
Mexican Americans have less complete health care coverage than do non-Hispanics throughout life; however, as is the case for other groups, those with private health insurance tend to be better off materially and to use more medical services. They have higher household incomes, are less likely to have been farm workers, and are far more likely to be native born compared with those without private coverage. A comprehensive old-age safety net clearly requires private supplemental insurance as well as a private pension. The greater reliance among older Mexican Americans on Medicare alone means that they may not receive the full range of needed services. The premiums, the cost sharing for hospitalizations, and the costs of excluded services that one must pay out of pocket if one does not have private insurance and does not qualify for Medicaid may cause some individuals with serious health problems to do without needed services.19
High income and private insurance are clearly more common among the middle class than among the lower classes. In our study, among those with health care coverage, farm workers were more likely than those who had spent their lives in other occupations to have only Medicare. They were also more likely to report no insurance at all. Farm work, in which a disproportionate number of older individuals of Mexican origin have spent their lives, provides one with very few assets and little income during the retirement years. Addressing the health care needs of agricultural workers presents formidable health care policy challenges.
Many older Mexican Americans were born in Mexico and came to the United States in adulthood. In our sample, those who came to the United States in middle age or late in life rely more heavily on Medicare only than do the native born. They are also more likely to report that they have no insurance at all. Many of these individuals were agricultural laborers, so their situation reflects that of farm workers generally. Compounding the challenges of providing adequate health care to older Mexican Americans are issues related to acculturation and the limited opportunities for accumulating assets faced by those who arrive in this country in mature adulthood.13
These results clearly show that although our current system of financing health care for the elderly is far more equitable than a system relying solely on private or employment-based coverage, important differences in the extent of coverage remain. It may be unrealistic to imagine that we could develop a system in which everyone would receive all of the services available to the most affluent retirees. On the other hand, it is imperative to ensure that our universal system of care for the elderly does not leave those at the bottom with serious deficiencies in coverage, especially if those deficiencies fall most heavily on certain disadvantaged groups, such as older Mexican Americans.22 Americans today, including minority Americans, enjoy the longest and healthiest lives in human history. Many of the recent improvements in longevity and health are the result of improvements in the control of chronic debilitating diseases. The control of diabetes, high blood pressure, and other chronic conditions depends on the availability of effective and often extremely expensive drugs. The fact that Medicare does not cover prescription medications is a matter of ongoing concern, and it is likely that we will see some extension of coverage in the near future.19
Mechanisms for improving coverage and for reducing the financial burden on lowincome older persons would be of particular use to the Hispanic and minority elderly. Legislation to achieve this goal has been passed and could conceivably be expanded. For example, increased enrollment of low-income Medicare beneficiaries in Qualified Medicare Beneficiary (QMB) and Specified Low-Income Medicare Beneficiary (SLMB) programs (Medicaid program amendments in the Medicare Catastrophic Coverage Act of 1988) could ease the serious burdens of Medicare’s cost-sharing requirements and premium liabilities.23
Whatever plan is adopted, however, must be designed so that it does not inadvertently leave the most vulnerable citizens without adequate coverage.24 For those with low incomes and no private supplemental insurance, what may seem like moderate individual contributions to the middle class may simply be more than an individual or a couple can afford.
Acknowledgments
This research was supported by National Institute on Aging grant R01-AG10939-03.
This report was presented at the Annual Meeting of the Gerontological Society of America, Chicago, Ill, November 17, 2001.
Human Participant Protection Study protocols were approved by the institutional review board of the University of Texas at Austin.
All 3 authors were involved in the collection and analysis of the Hispanic EPESE data for 8 years. Each contributed to the study in terms of substance, statistical analysis, and critique.
Peer Reviewed
References
- 1.Santos R, Seitz P. Benefit coverage for Latino and Latina workers. In: Perez SM, ed. Moving Up the Economic Ladder: Latino Workers and the Nation’s Future Prosperity. Washington, DC: National Council of La Raza; 2000:162–185.
- 2.Loprest PJ, Zedlewski SR. Health insurance coverage transitions of older Americans. Washington, DC: The Urban Institute; 1998.
- 3.Angel RJ, Angel JL. Who Will Care for Us? Aging and Long-Term Care in Multicultural America. New York, NY: New York University Press; 1997.
- 4.De la Torre A, Friis R, Hunter HR, Garcia L. The health insurance status of US Latino women: a profile from the 1982–1984 HHANES. Am J Public Health. 1996;86:533–537. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Giachello AL. Hispanics and health care. In: PS Cafferty, McCready WC, eds. Hispanics in the United States. New Brunswick, NJ: Transaction Publishers; 1992:159–194.
- 6.Thamer M, Richard C, Casebeer A, Ray N. Health insurance coverage among foreign-born US residents: the impact of race, ethnicity, and length of residence. Am J Public Health. 1997;87:96–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Valdez RB, Morgenstern H, Brown ER, Wyn R, Wang C, Cumberland W. Insuring Latinos against the costs of illness. JAMA. 1993;269:889–894. [PubMed] [Google Scholar]
- 8.Trevino FM, Moyer ME, Valdez RB, Stroup-Benham CA. Health insurance coverage and utilization of health services by Mexican Americans, mainland Puerto Ricans, and Cuban Americans. JAMA. 1991;265:233–237. [DOI] [PubMed] [Google Scholar]
- 9.Angel RJ, Angel JL. The extent of private and public health insurance coverage among adult Hispanics. Gerontologist. 1996;36:332–340. [DOI] [PubMed] [Google Scholar]
- 10.Schur CL, Feldman J. Running in Place: How Job Characteristics, Immigrant Status, and Family Structure Keep Hispanics Uninsured. New York, NY: Commonwealth Fund; 2001.
- 11.Estrada AL, Trevino FM, Ray LA. Health care utilization barriers among Mexican Americans: evidence from HHANES 1982–84. Am J Public Health. 1990;80:27–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Angel JL, Buckley CJ, Sakamoto A. Duration or disadvantage? Exploring nativity, ethnicity, and health in midlife. J Gerontol B Psychol Sci Soc Sci. 2001;56B:S275–S284. [DOI] [PubMed] [Google Scholar]
- 13.Angel RJ, Angel JL, Lee GY, Markides KS. Age at migration and family dependency among older Mexican immigrants: recent evidence from the Mexican American EPESE. Gerontologist. 1999;39:59–65. [DOI] [PubMed] [Google Scholar]
- 14.Zambrana RE, Logie LA. Latino child health: need for inclusion in the US national discourse. Am J Public Health. 2000;90:1827–1833. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Hayes-Baustista DE, Schink WO, Chapa J. The Burden of Support: Young Latinos in an Aging Society. Stanford, Calif: Stanford University Press; 1988.
- 16.Sotomayor M, Garcia A. Elderly Latinos: Issues and Solutions for the 21st Century. Washington, DC: National Hispanic Council on Aging; 1993.
- 17.Healthy People 2010. Washington, DC: US Dept of Health and Human Services; 2000.
- 18.Lillie-Blanton MD, Leigh WA, Alfaro-Corre IA. Achieving Equitable Access: Studies of Health Care Issues Affecting Hispanics and African Americans. Washington, DC: Joint Center for Political and Economic Studies; 1996.
- 19.Moon M. Will the care be there? Vulnerable beneficiaries and Medicare reform. Health Aff. 1999;18:107–117. [DOI] [PubMed] [Google Scholar]
- 20.Lillie-Blanton M, Martinez RM, Taylor K, Robinson BG. Latina and African American women: continuing disparities in health. Int J Health Serv. 1993;23:555–584. [DOI] [PubMed] [Google Scholar]
- 21.Samuelson RJ. It’s more than a drug problem. Newsweek. September 25, 2000:37. [PubMed]
- 22.Schmitt E. Other immigrants, envying Mexicans, demand a break, too. New York Times. July 26, 2001:A1, A21.
- 23.Fox SA, Stein JA, Sockloskie RJ, Ory MG. Targeted mailed materials and the Medicare beneficiary: increasing mammogram screening among the elderly. Am J Public Health. 2001;91:55–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Lamphere J, Rosenbach ML. Promises unfulfilled: implementation of expanded coverage for the elderly poor. Health Serv Res. 2000;35:207–217. [PMC free article] [PubMed] [Google Scholar]