Racial and ethnic differences in late-life health may reflect the disadvantaged status of some groups and reinforce their disadvantage. Some health differences may also reflect unequal treatment. It is reasonable for society to assign some priority to interventions to reduce health differences, and this chapter considers the evidence on the effects of intervention.
The goal of reducing differences needs to be carefully defined. What is of interest is improving group health, reducing differences without impairing the health of other groups. Interventions that improve health overall while possibly exacerbating differences should be considered, especially if the effect on differences is temporary, but they need to be assessed against other, neutral or differential-reducing interventions. It is important to recall that the health differences we consider involve groups, not individuals, though our focus is on interventions that improve individual health. Researchers need to be aware of events that could affect group differences simply by altering group membership—by facilitating return migration of ill elderly immigrants, for instance, which would have an effect on group differences.
We consider interventions in two categories: specific, directed interventions to change behavior and broader policy changes that may also have behavioral effects.
CHANGING INDIVIDUAL BEHAVIOR
Interventions to promote behavior change can play a role in improving health at older ages. Whether they can also contribute to reducing racial
and ethnic differences in health is less certain, and it may depend not just on the particular interventions, but also on their specific design, targeting, and implementation. We consider the role of health promotion at older ages and then touch briefly on its possible effect on differences, about which little research has been done.
Health Promotion at Older Ages
Health promotion can make a difference in the care of older persons. Some health care providers are trained to counsel older persons to stop smoking, moderate their alcohol use, exercise, and lose weight. Older persons on their own sometimes attempt to improve their health by modifying behavior, making choices about medical care, or using complementary or alternative therapies.
Health promotion interventions take place in a complex environment that includes family and social relationships, economic and geographic factors, and physical barriers and opportunities, all of which influence older persons’ ability to process health information and translate it into new behavior. Furthermore, outcomes are also influenced by how individuals choose to incorporate health promotion ideas that may reach them, with various levels of accuracy, from various sources. Cultural factors may also modify individual responses.
Health promotion may be more effective with older than younger people, who may be more willing to adopt healthy life-styles (Stewart et al., 2001). To reduce the incidence of diabetes, Knowler et al. (2002) found that a life-style intervention was superior to Metformin (a drug which keeps the liver from making too much sugar), with the life-style intervention being particularly effective in an older age group. Likewise, Appel et al. (2001) showed that hypertension could be reduced in a cohort of older persons through a 9-month sodium intervention. There may be important differences between younger and older persons in how they respond to health promotion activities.
The focus of health promotion in later life is different than at younger ages. For those over age 65, the focus is primarily to reduce age-associated morbidity and disability and the effects of cumulative disease comorbidities. For instance, the Fitness Arthritis in Seniors Trial achieved a significant reduction in disability and an improvement in physical performance in an older cohort despite having no effect on the actual disease process (Ettinger et al., 1997). In the over-85 age group, health promotion focuses almost exclusively on maintaining function and enhancing the quality of life.
Possible Effects on Differences
Health promotion could have different effects for various racial and ethnic groups. We illustrate possible reasons for this, though the research on each is not conclusive.
First, racial and ethnic groups may differ initially on the behavior that one seeks to influence. As discussed above, some unhealthy behaviors are more characteristic of older whites—smoking, less frequent Pap tests and mammogram screening—and others are more characteristic of older blacks and Hispanics—less physical activity and obesity (Winkleby and Cubbin, 2004). The relative physical inactivity of blacks and Mexican Americans cannot be explained by the socioeconomic status of individuals (Crespo et al., 2000). Also, even at similar socioeconomic levels, Mexican American women have more body fat than white women of similar body mass (Casas et al., 2001). (There is more limited information on these behaviors for Asians and American Indians and Alaska Natives.) It is reasonable to expect that health promotion interventions targeted at specific behaviors should have differential effects across groups, but whether those with more or less healthy behavioral profiles would be more affected is undetermined.
Second, some interventions may affect socioeconomic groups differently, with some being more responsive to them because of differences in education or income. Some interventions, such as smoking cessation, have been more successful among more educated people, but others, such as cholesterol screening, appear to affect socioeconomic groups equally (Cutler, 2004). Since racial and ethnic groups differ in socioeconomic status, interventions with disproportionate influence on higher status groups could affect differences, increasing the black disadvantage but reducing the Asian advantage (and the Hispanic advantage for mortality) relative to whites. Particularly if such effects are temporary, this difference is not necessarily a reason for avoiding such interventions.
Third, the social situation for members of some racial or ethnic groups may favor a stronger or weaker response to a specific intervention. The social environment plays a large role in the health of older persons, with younger family members involved in caregiving to maintain the health of older persons and keep them in their home environment. Extended caregiving networks are reported to be more active among some racial and ethnic minorities than others (Navaie-Waliser et al., 2001; Roth et al., 2001; Williams and Dilworth-Anderson, 2002). Eyler et al. (1999) found that, in comparison with black and American Indian women, Hispanic women were more likely to have high levels of social support, which promoted physical activity. The amount of social support, including also support from the wider community, may determine whether health promotion
succeeds or fails. Support, however, is double-edged: it can also reinforce opposition to behavioral change.
Fourth, some groups may react more positively, for cultural reasons, to particular types of interventions, as can be seen for the use of alternative therapies (Jones, 2001; Loera et al., 2001). Hispanic and Asian older adults use both folk medicine and Western medicine, depending on the illness and symptoms, as well as their ability to obtain traditional health care (AppleWhite, 1995; Torsch and Ma, 2000). Despite the heavy promotion of alternative therapies and their common use by older adults, such therapies are inadequately studied. If they have any effect, they could produce health differences. The effect could also be negative, in deterring or delaying the search for more effective therapies. Not only healthy behaviors are promoted. Nutritional supplements that may or may not be useful are also promoted, as more broadly, are such unhealthful behaviors as smoking and the consumption of fast food. Racial and ethnic groups differ in such behaviors, but what role promotion plays and how best to communicate opposing messages to particular groups is unknown.
Given how little is known with any certainty about the effects of health promotion interventions on racial and ethnic differences, fairly basic research issues need to be addressed. What health care seeking behavior is typical of different groups is probably the place to start.
Research Need 17: Measure the use of complementary and alternative therapies by racial and ethnic groups.
How are these therapies incorporated into self treatment? Do these therapies have any substantial health effect, and how do they affect quality of life? The answers to such questions may provide guides for the design of health promotion interventions.
Subsequently, research could proceed to identify the appropriate health care seeking behavior and risk behaviors to target across different ethnic groups. Information is limited on the prevalence of health risk behaviors in older cohorts, especially Asian and American Indian and Alaska Native populations. Reliable information on obesity, diet, and sedentary life-style patterns would allow for a better understanding of how these risks affect morbidity and functional disability in these populations. Better information would help guide the selection of interventions and the selection of target groups in the population.
Some research could also be directed at understanding differential responses to health promotion by age, socioeconomic status, and racial and
ethnic group. Health promotion strategies that succeed in one population may not be successful in another. The evidence that older cohorts respond better than younger cohorts is intriguing, but it requires confirmation and explanation. The evidence of differential response by education also needs study: How should interventions be designed to more effectively reach less educated, lower income groups? Racial and ethnic differences in response to interventions need to be identified and explained: Are they due to socioeconomic factors, to differences in the support environment, to particular cultural traits, or to some combination of these factors?
Direct promotion of individual behavior change is only one of the tools available to reduce health disadvantages among particular racial and ethnic groups. The range of possible interventions to achieve such public health objectives includes
economic incentives and disincentives linked to health-relevant behaviors;
changes in the informational environment—education, product labeling, and regulation of commercial speech;
direct regulation through penalties for behavior risky to oneself or others or for organizations that fail to deliver contracted health benefits;
indirect regulation through the legal tort system; and
deregulatory actions that dismantle legal barriers to desired public health behaviors (Gostin, 2001).
Such measures can be designed to affect personal behavior (through education, incentives, or penalties) or to work indirectly: through modifying the general environment for behavior or through affecting health care providers or other producers of relevant products and services (Sampson and Morenoff, 2001).
Effectiveness of Interventions
Relatively little is known about how interventions affect health differences. In fact, their effectiveness at improving population health, regardless of their effect on differences, is still a matter for discussion. Large-scale community experiments aimed at changing behavior in the population at large (not just among older adults) have been disappointing. Multifaceted community intervention experiments that promoted healthy behaviors have shown some effects, such as those in the 12,866-subject experiment labeled
MRFIT (Multiple Risk Factor Intervention Trial Research Group, 1982, 1990, 1996). However, changes were modest, and control groups showed similar outcomes (Cutler, 2004; Syme, 2002).
In contrast, some health risk behaviors have changed over time, and public interventions seem to have played an important role: an example is smoking. National cigarette consumption per capita reached its peak in the early 1960s. Since the 1964 Surgeon General’s report on the dangers of smoking (U.S. Department of Health, Education, and Welfare, 1964), consumption has steadily dropped to less than half the peak level, declining roughly in parallel among whites and blacks. Standing by itself, the Surgeon General’s report would probably not have generated such a wide-ranging and sustained response. But the report did not stand by itself. It mobilized public opinion and eventually generated a variety of legislative responses, including taxes on cigarettes, restrictions on broadcast advertising, antitobacco advertising campaigns, limits on access to cigarettes, and bans on smoking in various establishments and public places.
Antismoking policies have been extensively studied (Chaloupka and Warner, 2000; U.S. Department of Health and Human Services, 2000); as summarized by Cutler (2004), many of them are effective. Generally, a 10 percent increase in prices (through taxation, for instance) reduces consumption of cigarettes by 4 percent, with young men affected more than young women (Chaloupka and Pacula, 1998) and the poor affected substantially more than the rich (Gruber and Kosygi, 2002). Antitobacco advertising can have large effects: California’s $26 million media campaign of the early 1990s reduced smoking by an average (per year) of eight packs per person. Workplace smoking bans reduce the number of workers smoking by 5 percent and also reduce consumption among workers who continue to smoke (Cutler, 2004). Such policy changes, while effective, are nevertheless insufficient to account for the overall historical reduction in smoking, which must reflect, in addition, a broad-based movement away from smoking among millions of people. Somehow, in the late 20th century, the political timing was right for antitobacco measures, and growing public approval helped make them effective.
Blacks have benefited as much or slightly more than whites from smoking declines over the last four decades. The same cannot be said across education groups. Smokers among high school dropouts were 6 percentage points more numerous than among college graduates in 1966, and 19 percentage points more numerous in 1995.
When promotion of health behavior becomes part of a broad social movement, involving and mobilizing a range of social actors—national and local legislators, the media, private business, volunteer groups, etc.—wide-ranging effects such as those from antismoking campaigns are possible. Cutler (2004) refers to this phenomenon as permeation, the saturation of
the environment with informational messages, including and perhaps especially messages from one’s peers. The same phenomenon arguably has been at work in the reduction in drunk driving since the early 1980s, spurred by the activities of Mothers Against Drunk Driving and other grassroots programs.
One factor that has helped keep campaigns against smoking and drunk driving going has been the presence of externalities, the argument that others are hurt when people smoke or drive drunk (Cutler, 2004; Warner, 2001). In contrast, this argument has not been prominent so far in attempts to reduce obesity and has not been sufficient in the public debate about guns. Campaigns also have costs, of course. In addition to the resources needed for intervention, there are the burdens, difficult to quantify, that may be imposed on individual behavior, as when smokers are required to limit where they smoke (Gostin, 2001; Warner, 2001).
If whites and minorities benefit equally from society-wide interventions, there would appear to be little reason to target interventions to particular groups. However, interventions do not necessarily have the same effect across racial and ethnic groups. For instance, long-term declines in smoking, roughly similar among whites and blacks, were paralleled among Hispanics and Asian and Pacific Islanders, but at somewhat lower levels, while declines were much more limited among American Indians and Alaska Natives. Particular subgroups among these minorities—reproductive-age American Indian women and Southeast Asian men, for instance, show strikingly high tobacco use (U.S. Department of Health and Human Services, 1998). Aside from smoking policy, quite different types of health policies, such as changes in health care regulation and funding, could also have disparate effects.
Are targeted interventions ever worth consideration? They raise various concerns relating to equity (Hudson, 2002). To the extent that federal social policy has increasingly benefited the aged, what effects have there been on equity across age groups (Meyers and Darity, 2000; Preston, 1984)? What resources are appropriate to address the special needs of particular immigrant groups among the aged, in comparison with groups that have spent their entire lives in the United States (Jasso et al., 2004)? How can one think about the fairness of the distribution of scarce national resources that do not provide broad assistance across the entire population but may be effective in addressing the needs of particular groups? These are broad philosophical and legal questions that are central in the history of the country’s public policy making and its democratic ideals (Gostin, 2001).
There are two approaches that largely avoid these difficult equity issues. A simple approach is to build on the fact that racial and ethnic groups differ in their health problems, as already observed. Focusing on some health problems rather than others, even if no particular groups are targeted, would have a differential effect. Devoting more resources to diabetes, in particular, would probably benefit blacks, Hispanics, and American Indians more than whites or Asians, if the interventions are effective. This approach requires that all groups be reached equally and that the intervention effects do not differ by economic status, important assumptions that cannot be taken for granted. Similarly, programs to improve hospital quality are likely to reduce black-white differences in care to the extent blacks go to poorer quality hospitals, as could efforts to direct patients to better quality hospitals (Skinner et al., 2003).
A second approach comes from noting other differences between racial and ethnic groups. They may react differently to interventions, and some degree of specificity may be appropriate in order to reach members of particular groups. Health motivations, and resulting health-seeking behavior, could vary. While long life is presumably an important personal goal for everyone, the emphasis on it may differ for groups who suffer a multitude of environmental and life stressors (Baum, 2001). Cultural and socioeconomic considerations may influence the perceptions, experiences, and expression of health and illness (Dilworth-Anderson and Gibson, 2002; Goldman and Smith, 2002). For example, for those groups with relatively low life expectancy, quality of life may be a more important consideration than longevity. Thus, interventions that focus on promoting life extension may not be as effective as those that target quality of life.
Similarly, interventions that focus on the personal consequences of behavior may be less effective, for some groups, than those that focus on significant others who may be affected (Warner, 2001). For instance, behavioral interventions directed at the wives of men who have had a heart attack may be more effective, for some groups, than those directed at the men themselves (Taylor et al., 1985).
Finally, different racial and ethnic groups may show age-related changes at different times and to different degrees (Crimmins et al., 2004). Treatment may therefore be more effective at different ages for different groups; for instance, bone density testing at younger ages for white and Asian women in comparison with Hispanic and black women, mammograms at younger ages for black women, and blood pressure screening at younger ages for black men than other groups.
Indirect Effects of Policy
The policy possibilities we have considered are limited in an important respect. A broader conception of the possibilities would include any government measure that results in some transfer of resources, as well as other government regulations that, intentionally or not, affect health differences or the factors underlying them. Though such policies may be adopted for reasons unrelated to health, they may have important indirect effects on health differences.
Since socioeconomic factors have an important relationship to health differences, any program that affects resource distribution in the society could have health implications. For instance, changes in Medicare funding could affect health differences. As noted above, exclusive dependence on Medicare is higher among Hispanics and blacks than among whites, so that changes in the program could have more effects on them. Given the known and hypothesized effects on health in late life of childhood conditions, one may consider racial and ethnic differences in the socioeconomic conditions of childhood and government programs that exacerbate those effects or do not attempt to remedy them (Warner and Hayward, 2002).
Changes in immigration regulations could also affect health differences, by affecting the number, selectivity, or experience of the recent immigrants who constitute an important proportion of some racial and ethnic groups. Tighter controls on immigration, for instance, could increase the risks that illegal immigrants take, reduce the propensity for circular migration and therefore modify salmon bias (the return to their native countries of ill immigrants), and make immigrants a more select group in ways that could have unpredictable implications for population health.
National policy may in fact have had substantial influence on health differences from the 1960s to the 1990s. From 1968 to 1978, blacks showed larger gains than whites in life expectancy and larger declines in mortality rates from a variety of causes (Cooper et al., 1981). But from 1980 to 1991, the black-white gap in health status widened, whether measured by life expectancy, excess deaths, or infant mortality (National Center for Health Statistics, 1994; Williams and Collins, 1995), before narrowing again in the 1990s. The trends in the mortality gap also appeared at older ages, among those aged 65-74 (National Center for Health Statistics, 2003). The narrowing of the gap in the late 1960s and the early 1970s coincided roughly with gains from the civil rights movement and a parallel narrowing in the income gap between blacks and whites (Economic Report of the President, 1998; Williams, 2001a). The reversal of the trend in health differences beginning around 1980 coincided with substantial changes in national social and economic policies, during which the health status of economically vulner-
able populations worsened in several states (Lurie et al., 1984; Mandinger, 1985). What influence on health differences was actually exerted by national policy, by economic gains, or by the empowerment and increasing social acceptance of blacks are obviously research questions of interest.
When considering public health measures targeted at reducing racial or ethnic health disadvantage, therefore, one should keep in mind the ways in which other government actions, taken for a variety of extraneous reasons, can also affect racial and ethnic differences, supporting or possibly undermining measures directed at improving health.
Much is obviously not known about the policy possibilities for dealing with racial and ethnic differences in late-life health. There is a need to examine the effects of broad social policies on a wide variety of health behaviors among older people of different racial and ethnic groups. The effects of specific policies, such as taxation, also require corraboration for various population groups, such as recent immigrants (Jasso et al., 2004). A simple distinction between whites and minority groups is clearly insufficient in assessing differential effects. More broadly, there appears to be little analysis of the differential effects of interventions at the individual, community, or national levels by age and race or ethnicity (Emmons, 2001). There may be large cohort or period influences that would make public policies more or less effective for different groups at different times (Warner, 2001). Analysis should use as much existing data as possible, but it may be hampered by the lack of good, representative data, especially for older adults and recent immigrants.
Research Need 18: Characterize long-term trends (and possible lags) in the effects of changing social policy—federal, state, and local—on health differences and on public health.
For instance, popular expectations may have been too high that the 1964 civil rights laws, which marked a significant public policy change, would have large and immediate effects on the health and well-being of that generation of blacks. Much time was required to prepare regulations to implement legislation—to change the nature of segregated health care facilities, to improve the availability of good prenatal care and well-baby check-ups, to expand the possibilities for better jobs, and to improve environmental and housing conditions. Overcoming long-term neglect and lack of access are long-term projects that are still in process. How lags in the effects of public policy differentially affect racial and ethnic groups at different points in the life course requires sustained scrutiny (Hudson, 2002; Warner, 2001).
This research issue also has a prospective dimension. As policy options and interventions are developed to deal with current major issues in public health, research is needed of the ways in which these policies and interventions may affect racial and ethnic groups differently. The epidemic of obesity, for instance, is commanding increasing public and professional attention. Researchers need to attend to the mechanisms of influence of legal interventions and the ways they may have differential effects (Emmons, 2001; Gostin, 2001; Warner, 2001). For instance, one possible reason for the limited progress that has been made to combat obesity, particularly among the disadvantaged, may have to do with the price structure for various foods. If a premium must be paid for lean meats, reduced fat, and low-calorie foods, availability and consumption of these will be more limited at lower socioeconomic levels, with possible differential effects across racial and ethnic groups. When combined with cultural food patterns, the wide availability of high-starch and high-fat food stuffs, lack of leisure time opportunities, limited leisure time outlets, and the cheap, sedentary entertainment of television, public policies to affect health behaviors face difficult barriers (Warner, 2001).
One option, requiring special research attention, is supplementing universal health intervention policies with policies targeted at specific groups (Syme, 2002). Universal health policies (e.g., Medicare and Medicaid) are reported to have improved the health of blacks and other disproportionately low-income groups (Meyers and Darity, 2000). Whether some targeting of interventions would be useful to reduce group health disadvantages is worth investigation (Hudson, 2002; Syme, 2002).