2
The Burden of Cancer Among Ethnic Minorities and Medically Underserved Populations

This chapter focuses on the burden of cancer among ethnic minorities and medically underserved populations. These deceptively simple terms conceal the degree of complexity involved in defining the populations and the disease. The chapter therefore discusses the problems associated with defining ''minorities" and "underserved individuals," the complexity of the disease commonly known as cancer, the findings that describe the burden of cancer in these groups, and the conclusions derived from those data. The limitations of current federal cancer surveillance data are then discussed, followed by recommendations for the improvement of cancer surveillance among ethnic minority and medically underserved populations. Finally, a discussion of surveillance needs for cancer risk factor and control research is presented, with recommendations for National Institutes of Health (NIH) programs.

Defining the Populations

Federal Definitions of Race and Ethnicity

The simplest way of defining ethnic minority groups would be to use the census counts for the distribution of the total U.S. population, but it would first be necessary to have an appropriate method of classifying the various groups in the population. Herein lies the first problem. The existing federal classification by race and ethnicity was established by the U.S. Office of Management and Budget (OMB) in its Statistical Policy Directive No. 15, "Race and Ethnic Standards for Federal Statistics and Administrative



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 33
2 The Burden of Cancer Among Ethnic Minorities and Medically Underserved Populations This chapter focuses on the burden of cancer among ethnic minorities and medically underserved populations. These deceptively simple terms conceal the degree of complexity involved in defining the populations and the disease. The chapter therefore discusses the problems associated with defining ''minorities" and "underserved individuals," the complexity of the disease commonly known as cancer, the findings that describe the burden of cancer in these groups, and the conclusions derived from those data. The limitations of current federal cancer surveillance data are then discussed, followed by recommendations for the improvement of cancer surveillance among ethnic minority and medically underserved populations. Finally, a discussion of surveillance needs for cancer risk factor and control research is presented, with recommendations for National Institutes of Health (NIH) programs. Defining the Populations Federal Definitions of Race and Ethnicity The simplest way of defining ethnic minority groups would be to use the census counts for the distribution of the total U.S. population, but it would first be necessary to have an appropriate method of classifying the various groups in the population. Herein lies the first problem. The existing federal classification by race and ethnicity was established by the U.S. Office of Management and Budget (OMB) in its Statistical Policy Directive No. 15, "Race and Ethnic Standards for Federal Statistics and Administrative

OCR for page 33
Reporting" (U.S. Office of Management and Budget, 1997; see Box 2-1). Directive No. 15 requires that all federal agencies collect and report demographic data according to a limited number of racial and ethnic categories, despite criticism that such a classification scheme fails to recognize the increasing ethnic diversity of the U.S. population (U.S. Bureau of the Census, 1998a). Definitions of race, ethnicity, or origin are not included within the Directive, although the categories and inclusion criteria for the categories are given. OMB specifies four minimum categories for race, which include American Indian or Alaska Native, black or African American, white, and Asian or Pacific Islander. The two categories for data on ethnicity are Hispanic or Latino and Not Hispanic or Not Latino. The latest revisions to the Directive separated the Asian or Pacific Islander populations into two categories, and the term "Hispanic" was changed to "Hispanic or Latino." OMB recommends that these new revisions of the categories be adopted by other federal programs as soon as possible, but the official deadline for adoption of the requirement is January 1, 2003. Modifications to Directive No. 15 were made at the recommendation of the Interagency Committee for the Review of the Racial and Ethnic Standards, which suggested that changes are needed to reflect the nation's population (U.S. Bureau of the Census, 1998a). A number of scientific and professional organizations have criticized the Directive, stating that the standards lack scientific utility and do not encompass the diverse range of peoples in the United States. For example, the American Anthropological Association (AAA) has commented that OMB has noted the absence of "scientific or anthropological" foundations in developing its categories (American Anthropological Association, 1997). However, the race and ethnicity categories of the Directive are regularly used in federally sponsored scientific research and the interpretations of the findings are often based on the variables of race and ethnicity. (The scientific utility of the concept of race and its use in epidemiologic and health research will be discussed later in this chapter.) AAA also criticized the fact that Directive No. 15 did not explain what is meant by "race" or "origin" or the difference between the two concepts. In addition, it is not clear whether the race or origins of people were to be chosen by the individual respondent (self-identification) or by others such as interviewers. Many discrepancies in these two methods may arise in the process of sampling populations, although the Bureau of the Census department disagrees (Sink, 1997). In the U.S. census, respondents are allowed to self-identify their racial and ethnic background. In other health research, however, both methods may be used. Further, it is unclear what the impact will be for the reporting of health statistics for population groups of the policy change in the year 2000 census allowing

OCR for page 33
respondents to report more than one category of racial and ethnic background. Many federal agencies, including the National Cancer Institute (NCI), are attempting to address these challenges to identify populations accurately and trace any possible risk factors affecting their health. Robert Hahn of the Centers for Disease Control and Prevention (CDC) points out the varying and conflicting definitions of race and ethnicity used by federal agencies, and noted the difficulty this situation creates for comparing populations. He notes four assumptions that must be confronted and redressed to operationalize such an effort (Hahn, 1992). These four assumptions have allowed for an ambiguity about terms that define or characterize racial and ethnic groups. Too often, he argues, it is assumed that: 1.   Government agencies have no conflicting classifications of denominator for disease rates, that classifications do not vary by region, sub-group, or individual, and do not change over time; 2.   Racial and ethnic categories are understood by the populations questioned and that understanding does not vary by region, sub-group, or individual; 3.   Survey enumeration, participation, and response rates are high and similar for all racial and ethnic populations; 4.   Individual responses to questions of racial and ethnic identity are consistent in different surveys and at different times. These assumptions lead to inaccurate data by miscounting or misunderstanding both the base population denominator and the numerator(Scrimshaw and McMiller, 1996). Defining Ethnicity and Race Anthropologists and other social scientists view "race" and "ethnicity" differently from the federal definitions of population groups noted above. In fact, the federal definitions as currently applied mix and confuse the concepts of race and ethnicity, thereby reinforcing the outmoded concept of race, failing to allow the important concept of ethnicity to be fully functional, and failing to acknowledge recent advances in our understanding of the role that genetics plays in illness and susceptibility to illness. Ethnicity is a recent term that involves how one sees oneself and how one is "seen by others as part of a group on the basis of presumed ancestry and sharing a common destiny with others on the basis of this background" (Zenner, 1996, p. 393). Common threads that may tie one to an ethnic group include skin color, religion, language, customs, ancestry, and occupational or regional features. In addition, persons belonging to the same

OCR for page 33
BOX 2-1 OMB Directive No. 15 (adopted May 12, 1977): Excerpts from Race and Ethnic Standards for Federal Statistics and Administrative Reporting This directive provides standard classifications for recordkeeping, collection, and presentation of data on race and ethnicity in Federal program administrative reporting and statistical activities. These classifications should not be interpreted as being scientific or anthropological in nature, nor should they be viewed as determinants of eligibility for participation in any Federal program. They have been developed in response to needs expressed by both the executive branch and the Congress to provide for the collection and use of compatible, nonduplicated, exchangeable racial and ethnic data by Federal agencies. 1. 1. Definitions The basic racial and ethnic categories for Federal statistics and program administrative reporting are defined as follows:   A. American Indian or Alaskan [sic] Native.* A person having origins in any of the original peoples of North America, and who maintains cultural identification through tribal affiliations or community recognition.   B. Asian or Pacific Islander. A person having origins in any of the original peoples of the Far East, Southeast Asia, the Indian subcontinent, or the Pacific Islands. This area includes, for example, China, India, Japan, Korea, the Philippine Islands, and Samoa.   C. Black. A person having origins in any of the black racial groups of Africa.   D. Hispanic. A person of Mexican, Puerto Rican, Cuban, Central or South American or other Spanish culture or origin, regardless of race.   E. White. A person having origins in any of the original peoples of Europe, North Africa, or the Middle East. 2. Utilization of Recordkeeping and Reporting To provide flexibility, it is preferable to collect data on race and ethnicity separately. If separate race and ethnic categories are used, the minimum designations are: * This population is currently referred to as "Alaska Native."

OCR for page 33
a. Race:   — American Indian or Alaskan [sic] Native   — Asian or Pacific Islander   — Black   — White b. Ethnicity:   — Hispanic origin   — Not of Hispanic origin When race and ethnicity are collected separately, the number of White and Black persons who are Hispanic must be identifiable, and capable of being reported in that category. If combined format is used to collect racial and ethnic data, the minimum acceptable categories are: American Indian or Alaskan [sic] Native Asian or Pacific Islander Black, not of Hispanic origin Hispanic White, not of Hispanic origin. The category which most closely reflects the individual's recognition in his community should be used for purposes of reporting on persons who are of mixed racial and/or ethnic origins. In no case should the provision of this Directive be construed to limit the collection of data to the categories described above. However, any reporting required which uses more detail shall be organized in such a way that the additional categories can be aggregated into these basic racial/ethnic categories. SOURCE: U.S. Office of Management and Budget (1997).

OCR for page 33
ethnic group share a unique history different from that of other ethnic groups (American Anthropological Association, 1997; Zenner, 1996). Usually, a combination of these features identifies as ethnic group. For example, physical appearance alone does not consistently identify one as belonging to a particular ethnic group; individuals belonging to certain ethnic groups may vary widely in physical appearance (e.g., skin color and hair texture), but they share a common ethnic identity. In the U.S., there are "macro" ethnic groups, such as Latinos or Hispanics, which have many sub-groups, such as Mexicans, Mexican Americans, Cubans, Puerto Ricans, and so on. "Race," in contrast, represents a social or cultural construct of human variability based on perceived differences in biology, physical appearance, and behavior (American Anthropological Association, 1997; Smedley, 1999). The problems with the concept of race for purposes of health research will be discussed later in this chapter. It is useful in medical and epidemiologic research to distinguish ethnic groups from one another, provided that researchers are clear on the nature and source of human variation (e.g., cultural and behavioral patterns, environmental influences) and their relationship to health outcomes. Membership in an ethnic group may be associated with behavioral and environmental factors which may increase or decrease the likelihood of illness. Thus, the availability of pertinent information for a diversity of ethnic groups would assist both those involved in health research and the population as a whole by indicating if any ethnic differences need to be further explored. Such research, however, can be accomplished only by clearly identifying population groups and understanding that human identity is not static or mutually exclusive. Ethnicity can be a product of interaction between people of different origins and identities. The "boundaries" of ethnic identity, however, are still unclear. Some may identify themselves as belonging to a particular group in one context, and to another group in a different context. "Identities thus become circumstantial" (Zenner, 1996, p. 394). Ethnicity is more flexible, fluid, or perceived, than rigid or fixed. This has been compounded by the increased number of mixed ethnicity families in this country, where individuals may claim two or more ethnicities, or give different ethnic identifications at different times or for different purposes. This makes the proposed revisions for the year 2000 census all the more important, as federally sponsored health research will need to account for individuals who identify themselves as from diverse backgrounds. Medically Underserved Populations In the study of cancer, considerable attention has been given to "minorities" as a group, but medically underserved individuals make up a

OCR for page 33
separate group that cuts across all ethnic groups. The term "medically underserved" is sometimes used to include underinsured or uninsured people; those with low levels of education; rural and inner-city populations; unemployed people; or those with low socioeconomic status (SES). As noted below, no consistent definition of "medically underserved" populations appears in the NIH descriptions of this construct. Socioeconomic Status Social class is traditionally measured through socioeconomic indicators such as education, income, and occupation (Adler et al., 1994). Many problems have been linked to low SES such as crime (Reiman, 1997), ill health (Pincus and Callahan, 1995), poor education (Levine and Nidiffer, 1996) and inadequate access to health care. Criticism and controversy have existed over which indexes have appropriately or inappropriately assessed SES to characterize social determinants of health. Schneider (1986) criticizes the use of common measures of SES such as Hollingshead's Four Factor Index of Social Status and Duncan's Socioeconomic Index for all occupations, noting the paucity of information on household patterns, social relationships, and the historical context of people's lives, all of which are factors that may mediate social status. Schneider advocates the use of an index that includes race or ethnicity, a detailed work history, autonomy, and decision-making authority. Lillie-Blanton and LaVeist (1996) agree that these factors in one's social environment (including power arrangements such as political empowerment and individual and community control and influence) are related to one's physical and mental health. A person may be working or living in an environment that exposes him or her to toxins that directly affect health, yet social relations and the conditions of one's neighborhood and work environment can also affect health. The latter experiences may shape the sense of "self worth, influence linkages to social structure and develop adaptive skills" (Lillie-Blanton and La Viest, 1996, p. 85). Many researchers have made the mistake of attributing the health disparities between groups to race or ethnicity without paying close attention to socioeconomic variability. Lillie-Blanton and LaVeist (1996) explain the social context of health by describing an individual's SES as an "expression of the educational and economic opportunities available in one's social environment" (p. 84). They suggest that SES may be more relevant than race or ethnicity in assessing one's socioeconomic context or social environment. This approach not only incorporates the individual, but also incorporates the social forces that also affect individuals and their health.

OCR for page 33
Socioeconomic Status and Cancer A number of factors have been implicated in the relatively poor cancer outcomes for populations of lower SES. Disparities in cancer incidence by SES can be explained by differential levels of exposure to environmental carcinogens, differences in personal health habits (e.g., cigarette smoking, poor diet, and lack of education regarding health risks), and other sociocultural factors. Some negative health behaviors are more prevalent among people of lower SES, but these behaviors should not be considered only as matters of purely personal choice. One's social environment can pose barriers to awareness and to making changes in behavior. Individuals living in poverty, for example, have poorer access to information regarding cancer risks and health behavior, to nutritious foods that may diminish cancer risk, and to role models in the form of cancer survivors who can relay information regarding help-seeking. On the other hand, individuals living in poverty "tend to concentrate on day-to-day survival and often develop a sense of hopelessness and powerlessness and become socially isolated" (Freeman, 1990, p. 18), all of which may contribute to negative health behaviors and decreased survival. Poverty or low SES has not been implicated as a direct cause for a higher incidence of cancer, although research has found that it has an adverse relationship to survival rates. The American Cancer Society estimates that cancer survival rates of poor individuals are 10 to 15 percent lower than those of other Americans (American Cancer Society, 1990). For example, more affluent women have a higher incidence of breast cancer, yet their survival rates surpass those for women living under disadvantaged socioeconomic conditions. Some of the proposed reasons for the disparities in survival rates between the groups are that women of lower SES are diagnosed or seek treatment at later stages of the disease, compared to women in higher income brackets. The results of the 1990 U.S. National Health Interview Survey show that "poorer" women (i.e., those whose annual income was less than $15,200) were significantly less likely to receive Pap smears and mammograms within the previous year than women whose income was at least $46,500. The survey reported that the upper-income women were twice as likely to have received a Pap smear and almost three times as likely to have had a mammogram within the previous year compared with women of low SES. It is suggested that poorer people have fewer opportunities to seek preventive care because of life obstacles, such as difficulties in finding a job or having job security, arranging child care, and arranging transportation (Harvard Center for Cancer Prevention, 1996). Friedell, Linville, and Hullet (1998), in discussing cancer control efforts among low-income women, note that

OCR for page 33
some barriers to increasing participation in cancer control exist at all socioeconomic levels, e.g., lack of information about cancer and about both the availability and the benefit of cancer screening. Fear of what might be found during such an examination mitigates against women either gaining information about cancer or doing something with the information once it is obtained. Other barriers … are more prevalent in low-income, medically underserved populations (Friedell, Linville, and Hullet, 1998, p. 1869). These barriers are noted in Table 2-1. Even after the diagnosis of cancer, not all groups of people are given the opportunity to receive the same quality of treatment. In one study reviewing lung cancer treatment patterns for more than 1,800 patients in New England hospitals, researchers found that more educated and privately insured individuals received more aggressive therapy than those without these attributes. In addition, people of lower SES seek health care more often in emergency room settings because they are underinsured or uninsured, which diminishes the continuity of care (Harvard Center for Cancer Prevention, 1996). This lack of proper and continuous care for individuals of lower SES subjects many to higher cancer mortality rates. Tracing cancer disparities between groups of different SES continues TABLE 2-1 Barriers to Optimal Cancer Screening, Diagnosis, and Treatment Poverty Isolation Geographic Transportation Literacy Age Cultural "Fatalism" Male dominance Putting the family's needs above one's own Fear Fear that cancer will be diagnosed Fear of the exam (e.g., mammogram and proctoscopy) Acceptability of the service provided Physical arrangements, visiting time Hours of service Staff attitudes Language barriers Cultural understanding by staff Lack of continuity of care   SOURCE: Friedell et al., (1998), with additions by the study committee.

OCR for page 33
to prove to be complex when integrating other factors such as psychological responses to illness. Exposure to psychological stresses and a lack of coping skills or resources also affect survival from cancer. Individuals of lower SES are less likely to have social networks and support than those of higher SES, because the former may lack employment or other social resources in their communities. Such social supports contribute to a sense of connection to others, aiding in one's ability to deal with an illness such as cancer (Harvard Center for Cancer Prevention, 1996). Increased research support in this area may demonstrate that strengthening social support for those with cancer in disadvantaged communities can help to lower mortality rates when the many components of lower SES prove less amenable to intervention. In summary, poverty "is a proxy for other elements of living, including lack of education, unemployment, substandard housing, poor nutrition, risk-promoting lifestyle and behavior, and diminished access to health care" (Freeman, 1990, p. 18), all of which affect an individual's chances of developing cancer and surviving it. NCI does not define SES in its report to this committee (National Cancer Institute, 1998b), although it does consider individuals of low income as "special populations," without further defining that group. Nevertheless, studies are underway at NCI to evaluate the extent to which racial and ethnic disparities in cancer incidence may be related to SES factors (see Chapter 3). As noted above, however, the term "low SES" could imply the existence of a number of cancer risk factors, yet for the low-income group within the NCI's "special populations" portfolio the focus seems to be only on the lack of monetary resources. It is therefore unclear what populations are referred to, or what needs should be addressed first (i.e., education, income, insurance status, or social environment) to reduce the risk of cancer. Ethnicity, Socioeconomic Status and Health Unfortunately, differences in SES among U.S. ethnic groups exist. When examined collectively, African Americans and Hispanics are three times as likely as whites to be poor, according to U.S. Census Bureau (as reported by the National Center for Health Statistics, 1998; see Table 2-2). Such gaps have been related to inequalities in the levels of educational and economic attainment associated with racial barriers (Gimenez, 1989; Lillie-Blanton and LaVeist, 1996; Harvard Center for Cancer Prevention, 1996). Low SES and ethnic minority status are not synonymous, but many members of ethnic minority groups who also have low income comprise an important proportion of underserved populations in the United States. Low-income ethnic minorities are least likely to have a regular source of

OCR for page 33
TABLE 2-2 Percentage of People Below Poverty Level, 1996 Race and Hispanic Origin Percent Below Poverty Level All races 12.8 White 10.0 Black of African American 30.7 Asian or Pacific Islander 14.1 Hispanic origin 26.2 American Indian 31.2   SOURCE: U.S. Bureau of the Census, as reported by the National Center for Health Statistics (1998). TABLE 2-3 Percent and Number of Individuals Below Poverty Level, by Ethnicity and Metropolitan Region, 1987   Metropolitan Non-Metropolitan Ethnicity Percent Number Percent Number White 9.6 13.3 million 13.7 6.6 million African American 30.7 7.3 million 44.1 2.2 million Hispanic 27.6 3.9 million 35.6 0.6 million   SOURCE: U.S. Congress, Office of Technology Assessment (1990) primary care (i.e., family doctor or clinic; Giachello, 1994). Many of these ethnic minority group members do not know where to go when they get sick (Robert Wood Johnson Foundation, 1997). When a regular source of care is reported, it tends to be a public health care facility, a hospital outpatient clinic, or an emergency room (Giachello, 1994; Robert Wood Johnson Foundation, 1997). Members of ethnic minority groups are also least likely to have insurance coverage. Approximately one-third of the Hispanic population does not have health insurance coverage, and the problem is most severe among Mexican Americans and Central Americans (Giachello, 1993; Naranjo, 1992; Valdez et al., 1993). Conversely, the largest numbers of low-SES persons are white, and many of them have the same health care access problems as do members of minority groups (Friedell, Linville, and Hullet, 1998). Nearly 20 million white Americans lived below the poverty line in 1987, as depicted in Table 2-3, with many of these living in non-metropolitan, rural areas.

OCR for page 33
recently published survival data for minorities other than African Americans is from the 1978 to 1981 SEER program registry and is therefore nearly 20 years old. These data suggest that the rate of survival from cancer is much poorer among minorities and that additional follow-up is needed. The most recent SEER program monograph to address racial and ethnic patterns of cancer (1988 to 1992) evaluates survival rates only among African Americans and whites. In addition, the reported SEER program data do not adequately address patterns of cancer survival among medically underserved white populations. Recommendation 2-3: Annual reporting of cancer surveillance data and population-based research needs to be expanded to include survival data for all ethnic groups, as well as for medically underserved populations. Human Diversity, Population Groups, and Cancer As noted earlier in this chapter, NCI's population-based data collection efforts are shaped by Directive No. 15 of the OMB, which stipulates that the U.S. population be classified according to one of four basic "racial" categories (American Indian or Alaska Native, Asian or Pacific Islander, black or African American, or white) and one of two ethnic groups (Hispanic or non-Hispanic). Although these classifications carry important historical, social and political significance in the United States, they are of limited utility for purposes of health research because the concept of race rests upon unfounded assumptions that there are fundamental biological and behavioral differences among racial groups (Cooper, 1984; Williams, Lavizzo-Mourey, and Warren, 1994; President's Cancer Panel, 1997; American Anthropological Association, 1998). In reality, human diversity cannot be adequately summarized according to the broad, presumably discrete categories assumed by a racial taxonomy. Furthermore, "racial" groups as defined by OMB are not discernible on the basis of genetic information (President's Cancer Panel, 1997; American Anthropological Association, 1998). Although the four racial groups defined by OMB are broad and imprecise, and greater genetic heterogeneity exists within groups than between groups, health researchers may nonetheless benefit from understanding differences in health status between these groups. Health differences between "racial" groups may be due to many factors, including discrimination in the health care system, limited access to prevention and treatment services, poverty and socioeconomic factors, exposure to environmental toxins, and cultural factors, such as attitudes about health, beliefs, diet, and lifestyle patterns. Health is therefore a biological response

OCR for page 33
to all of these conditions—conditions that are more accurately emphasized when groups are defined on the basis of ethnic background. Ethnic groups include individuals who share a unique history different from that of other groups, in addition to other attributes, such as language, customs, ancestry, and religion. Usually, a combination of these features identifies an ethnic group. In the U.S., many groups commonly referred to as "racial groups" may be more accurately referred to as "macro-ethnic" groups. These include "white" Americans of European descent, African Americans, Asian Americans, Hispanics, and Native Americans. It is important to recognize, however, that there is considerable cultural and biological heterogeneity within these groups, and therefore the precision of population-based research can be enhanced by referring to specific subgroups. For example, within the Asian American population, there are many ethnic subgroups, including individuals of Southeast Asian, Korean, Japanese, Chinese, and Indian descent. Distinguishing many ethnic groups from one another is therefore useful in medical and epidemiologic research, provided that researchers are clear on the nature and source of human variation (e.g., cultural and behavioral patterns, environmental influences, and genetic variation) and their relationship to health outcomes. Researchers must therefore use caution in interpreting the sources of observed differences between these groups. Ultimately, greater precision in understanding and describing human diversity is needed to distinguish genetic and environmental contributions to cancer risk and the complex effects of the gene-environment interaction. This precision can be improved with greater clarity in the conceptualization and definition of population groups. The views of AAA are pertinent to this approach. AAA recognizes that classical racial terms may be useful for many people who prefer to use such terms about themselves with pride, but it recommends phasing out the term "race" and recommends that it be replaced with more correct terms related to ethnicity, such as "ethnic origin," which would be less prone to misunderstanding (American Anthropological Association, 1997). Recommendation 2-4: The committee recommends an emphasis on ethnic groups rather than on race in NIH's cancer surveillance and other population research. This implies a conceptual shift away from the emphasis on fundamental biological differences among "racial" groups to an appreciation of the range of cultural and behavioral attitudes, beliefs, lifestyle patterns, diet, environmental living conditions, and other factors that may affect cancer risk. This change should not be difficult because, under the present arrangements,

OCR for page 33
the aggregations which are called "races" are really macro-ethnic groups. Scientifically speaking, there is only one race, Homo sapiens , but many ethnic groups and the entire population can be described within the five macro-ethnic groups we have indicated. This arrangement recognizes both the unity of the human race and the diversity of the ethnic groups, without any major disturbance in the data collection. Macro-ethnic groups can then be subdivided as indicated by the needs of cancer research to permit studies within such groups. The committee feels that it is important to study cancer within as well as across macro-ethnic groups. In this respect the diversity of the U.S. population offers an excellent opportunity to clarify issues relating to prevention and control. The study of several ethnic groups permit a better assessment of the factors contributing to cancer than studies based on "race," especially when these studies are limited to black-white differences. The racial emphasis is often associated with supposed genetic differences, but these assumptions are inconsistent with our current knowledge of the genetic diversity of the human race. Cancer Risk Factor Research and Cancer Control Cancer surveillance data are often used to measure the progress toward reducing the incidence of cancer that has already occurred, but the potential of those data is far greater than that. The differences in incidence, survival, and mortality rates for various ethnic groups raise critical questions about the causes of cancer and how it can best be prevented and controlled. The data suggest that many lives could be saved if more were understood about the role of behavior, the environment, socioeconomic factors, and genetic factors related to cancer and their interplay. To date, the research effort has failed to take adequate advantage of the increasing diversity of the U.S. population as a tool in understanding the interplay of cancer risk factors. Such an effort would benefit not only ethnic minorities and medically underserved individuals, but also the entire U.S. population. It requires, however, that the appropriate data be collected. This is the challenge facing the newly organized Division of Cancer Control and Population Science (DCCPS). NCI's Cancer Control Review Group report (National Cancer Institute, 1997a) made recommendations regarding the pursuit of research opportunities most likely to accelerate reductions in the nation's cancer burden (see Appendix B for the Review Group's recommendations). One of the items highlighted in that report was the need for basic behavioral and social science research in NCI to enhance the focus on primary prevention efforts. Included in the scope of this recommendation was the

OCR for page 33
need to measure risk factors and screening behaviors, especially among medically underserved individuals. The Review Group defined cancer control research as the conduct of basic and applied research in the behavioral, social, and population sciences that, independently or in combination with biomedical approaches, reduces cancer risk, incidence, morbidity, and mortality. Thus, optimum cancer prevention and control strategies are those that combine biomedical and public health research to address the process of carcinogenesis across the life span, from prevention to screening and treatment. The surveillance data reviewed earlier in this chapter reveal considerable gaps in the understanding of cancer risk factors among ethnic minority and medically underserved populations. Greater research is needed to illuminate risk factors both within and across population groups. Such research should address the full range of cancer risk factors, as noted above, including cultural factors affecting health attitudes, behaviors, diet, and other factors, as discussed below. The Role of Genetics As in the case with many other diseases such as hypertension or diabetes, large disparities in cancer incidence, mortality, and survival rates are sometimes observed between "racial" groups. These differences are sometimes assumed to be due to genetically determined differences between "races." However, it is important to understand the true nature of genetic variability both within and among "racial" groups and how evolutionary and sociocultural forces have shaped human genetic diversity to understand the meaning of the observed differences. As noted above, assumptions that differences are due to "race" or genetics may not be justified by the evidence. Behavioral factors (e.g., smoking), environmental factors (e.g., chemical and viral exposure), and socioeconomic factors (e.g., availability, affordability, and accessibility of diagnostic, therapeutic and preventive services) are likely to be the major links responsible for a higher (or lower) incidence or prevalence of cancer in ethnic minority and medically underserved populations (or in any population). The distribution of particular genetic polymorphisms in a population may be a significant factor, however, and must also be considered in comprehensive evaluations of cancer causation. In all comprehensive population-based research on carcinogenesis, the genetic constitutions of the study subjects must be taken into account. In a small proportion of patients with cancers of various types, mutations in a single gene can be identified as a predominant cause. In such instances, cancers of specific types show a strong tendency to run in families. Examples of genes causing hereditary breast cancer are the BRCA1

OCR for page 33
gene on chromosome 17 and the BRCA2 gene on chromosome 13; another example is the p53 gene, some mutant forms of which lead to the Li-Fraumeni syndrome, which has breast cancer as one feature. These genes, however, account for no more than 3 to 5 percent of cases of breast cancer. As to the genetic factors involved in causation, most cancers are the consequence of an interplay between more than one gene collaborating with one or more environmental factors. The environmental factor may be of overwhelming importance, such as cigarette smoking in lung cancer and papillomavirus infection in cervical cancer; however, even in these instances the existence of genes that make some individuals more or less vulnerable to the environmental factors are suspected or such genes can be identified. In general, genetic factors responsible for common disorders with multifactorial causes, such as cancers, are usually "susceptibility genes," and multiple genes are often responsible for the disorder. They represent a variable form (or allele) of a particular gene and have common variations, called polymorphic alleles. Alone, each variation has little effect, but in combination with other specific polymorphic alleles or nongenetic influences, these variations give rise to a particular disorder. Figure 2-1 diagrams the contrast between disorders caused by a mutation in a single gene, such as cystic fibrosis, Huntington disease, and Li-Fraumeni syndrome on the one hand and common disorders with multifactorial causation such as hypertension, manic-depressive illness, and lung cancer on the other. Each polymorphic allele occurs relatively frequently in the general "normal" population; it is present at a higher frequency (and usually in combination with other specific alleles) in people with a particular common disorder. The variable gene is a susceptibility factor, not a causative factor. Genetic factors in common disorders with multifactorial etiology are usually investigated by allele association studies with DNA markers. These markers are sometimes variations in the sequences of specific genes that are investigated as candidate genes; because of the known functions of the genes, implication of their association in the pathogenesis of the given disorder is plausible. Alternatively, the DNA marker may be located in an "anonymous" section of the genome, that is, in a gene with an unknown function or in a noncoding region of the genome. A common variation, or "polymorphism," of a gene may alter the function of the gene product, thereby contributing to susceptibility to a disease. In such instances, the marker itself is the disease-related change. In the case of a common variation not involving the coding or gene-regulating sequence of DNA, linkage (in which genes are located close to each other on the same chromosome) is presumed to exist between the DNA variation and a specific allelic form of a susceptibility gene. Studies of associations between allelic variants

OCR for page 33
FIGURE 2-1 The contrast between disorders caused by a mutation in a single gene and common disorders with multifactorial etiology. SOURCE: Victor A. McKusick. and common disorders are being undertaken with increasing frequency, in part under the stimulus of the Human Genome Project, which is identifying many genes and many polymorphisms. Caution in the interpretation of results in such studies is indicated for at least two reasons. First, population admixture, a conspicuous feature of the United States and other ethnically heterogeneous parts of the world, can cause an artifactual association if the study includes genetically distinct subpopulations, one of which coincidentally displays a higher frequency of both the disorder and the allelic variant. In such a situation, an observed association may be explained by differences in the genetic contributions of subpopulations to the case patient and control groups rather than by a physiologic effect of the genetic variant. Consideration of the ethnic backgrounds of subjects and the study of multiple, independent populations can help avoid the subpopulation stratification problem. Also, family-based studies can provide a test of the validity of an association between an allelic variant and a disorder. In the transmission

OCR for page 33
disequilibrium test, for example, if a given allele contributes to a given disease, then the probability that an affected person has inherited the allele from a heterozygous parent should vary from the expected Mendelian ratio of 50:50; association with a neutral polymorphism due to population admixture displays no such deviation in family data (Spielman et al., 1993). A second pitfall is that testing of multiple hypotheses, aggravated by publication bias, can lead to fallacious conclusions concerning associations. Researchers who test a single genetic change (mutation) for its association with a single disorder base their statistical conclusions on a single hypothesis. Many researchers, however, seek associations using multiple genetic variants. Each test represents an independent hypothesis, but there is a tendency to publish only positive results. In fact, these may merely represent those cases expected to fall outside the 95 percent confidence limits in tests with multiple genetic variants. Statistical correction for multiple testing is possible, but such corrections result, in a loss of statistical power (Altshuler et al., 1998). The types of DNA markers used in association studies include restriction fragment length polymorphisms, variable-number tandem repeats, microsatellites (short tandem repeats), and single nucleotide polymorphisms (Collins et al., 1997). These markers are listed roughly in the order in which they have been used historically. NCI is initiating a research program to identify all variations in the human genome (Richard Klausner, Director, National Cancer Institute, personal communication, 1998). The Human Genome Diversity Project is a follow-up to the Human Genome Project. When the goals of the Human Genome Project—mapping of all human genes and determination of the sequences of all DNA in those genes—are achieved, it will still be unknown how the genes and the DNA vary among the 6 billion or 7 billion or more people who will inhabit the Earth at that time. The functions of all those genes will also still be unknown, as will how variations in the structure of the DNA and genes relate to variations in function, including variations in susceptibility to common disorders, including cancers. The Human Genome Diversity Project proposes to identify the range of DNA variation that can be the basis for studies of genetic factors in common disorders. In the pursuit of the Human Genome Diversity Project or any DNA-based studies with special populations, including ethnic minorities and medically underserved individuals, concerns about the risks of discrimination and stigmatization have been raised. The ethical, legal, and societal implications of the project have been addressed by Knoppers et al., (1996, 1998) and by the report of a committee of the National Research Council. The newly organized DCCPS recognizes that epidemiology and genetics

OCR for page 33
represent the foundations of cancer control research. Its Epidemiology and Genetics Program has studied the following: Interactions of genetic and metabolic factors with lifestyle, social and behavioral factors, diet and nutrition, hormones, and medications. Gene prevalence-associated metabolic markers and predictive value of identified genes and markers. Definitions of ethnicity using molecular genetics and; application of these definitions, as well as data on behavior and lifestyle choices, to studies of the effects of migration on the incidence of cancers. The genetic determinants, tumor markers, and cancer risk from immune function. Special populations with different patterns of cancer risk. Means of improving estimate of exposures by direct and indirect means. Studies of Human Behavior NCI research in the area of human behavior has not been strong in the past, but an expanded behavioral research program is planned in DCCPS, according to NCI officials (Barbara Rimer, National Cancer Institute, communication with the study committee, June 12, 1998). This program will place greater emphasis on the development of balanced behavioral research portfolios that include a range of research in all areas, spanning basic research, dissemination research, and policy development. The literature from anthropology, psychology, sociology, behavioral medicine, and public health has shown that different ethnic groups vary in their attitudes, perceptions, and behaviors toward health. As noted earlier, macro-ethnic groups, such as Asian Americans, include many sub-groups, such as Chinese Americans, Japanese Americans, and Vietnamese Americans, to name only a few. For each ethnic group, culture influences their health, their attitudes toward health, and their health practices, although individual beliefs, attitudes, and behaviors may vary. Another factor is acculturation, the degree to which immigrants leave behind a culture of origin and assimilate aspects of their new culture. Some health behaviors may change within a few years of taking up residence in the U.S., while others may persist for generations. For example, newly arrived Mexican women are less likely to smoke than women of Mexican origin raised in the U.S. In addition, a study on Pap smear and mammogram screening in Mexican-American women found that the prevalence of these screenings increased with acculturation (Suarez, 1994). Given the diversity of health practices even within one ethnic group, it is essential to obtain accurate

OCR for page 33
and detailed ethnic histories of people to obtain a better understanding and knowledge of differences in health practices and possible risk. Research on the different health care practices of various ethnic minority groups has described the existence of subcultures with different values, actions, and perceptions about health and illness (e.g., Saunders, 1954; Kleinman et al., 1978; Leslie, 1976; Good and DelVecchio, 1981; Hahn, 1995). These also include behaviors such as diet and exercise patterns, and interaction with the health care system (Salazar, 1996; Crane et al., 1996). For many of these groups the family and other social networks are important in providing support during illness in and health-seeking behaviors. During the testimony heard by this committee from representatives of several ethnic minority and community groups (see Chapter 4 for a description of this testimony), two conflicting concerns about ethnicity were raised. One was that by identifying cultural habits that increase risk or reduce the likelihood of carrying out treatment, one is ''blaming the victim" by "blaming culture." In contrast, the concern was also expressed that not enough attention is paid to cultural norms, values, and concerns. Since all human beings are influenced by culture, these influences and their effects must be taken into account, but in ways which are sensitive to community concerns. It is understandable that many individuals in ethnic groups share common fears and distrust of both the medical care system and the research community, based on a history of disrespect and mistreatment of minorities by the research community. The Tuskegee syphilis experiment (summarized in greater detail in Chapter 5) is cited by African Americans and by Hispanics as a reason not to trust the government or researchers (Robinson et al., 1996). It is therefore important to understand the factual basis for this mistrust, and assess its impact on health care delivery and health behaviors among ethnic minority populations. Summary of Cancer Risk Factor Research and Surveillance Needs The committee finds that the newly reorganized DCCPS at NCI is well poised to address a wide range of challenges to improving our understanding of cancer risk among ethnic minority and medically underserved populations. Greater precision is needed in the definition and conceptualization of high-risk populations, and research must examine the diverse range of cancer risk both within and between ethnic groups. The committee is confident, however, that such research conducted under DCCPS's auspices will prove fruitful for the national cancer effort.

OCR for page 33
Recommendation 2-5: The committee commends the proposed NCI program of expanded behavioral and epidemiological research examining the relationship between cancer and cancer risk factors associated with various ethnic minority and medically underserved groups, and recommends that these studies be conducted both across and within these groups. Summary In this chapter, the committee has reviewed some of the difficulties associated with understanding and defining the problem of cancer among ethnic minorities and medically underserved individuals. In a large and diverse nation such as the United States, with its many overlapping population groups that may experience differences in cancer risk as a result of a complex interplay of environmental, cultural, socioeconomic, behavioral, and other factors, the use of clear and consistent definitions of populations is imperative to understanding how cancer differentially burdens various groups among the overall population. Depending on how populations are defined, cancer incidence, mortality, and survival rates vary considerably. Some ethnic minority groups, for example, may experience lower rates of some forms of cancer relative to the white majority; others, such as African Americans, experience higher rates of cancer incidence and mortality, and lower cancer survival rates. Groups of lower SES experience cancer incidence and mortality rates that are, in many instances, as high as the highest of any ethnic minority group. Understanding why cancer differentially affects these groups offers important clues to cancer etiology and control research. Because cancer surveillance is critical to this effort, the committee finds that greater resources must be committed to expanding upon the existing, high-quality SEER program data. Additional data must be collected to understand cancer among medically underserved groups, and SEER program coverage of other important population groups should be improved. Because the problem of defining populations is complex, the committee offers recommendations on how population groups may be conceptualized, but recognizes that the NIH has been constrained in its data collection efforts by current federal guidelines. The following recommendations were offered: Recommendation 2-1: NIH should develop and implement across all institutes a uniform definition of "special populations" with cancer. This definition should be flexible but should be based on disproportionate or insufficiently studied burdens of cancer, as measured by cancer incidence, morbidity, mortality, and survival statistics.

OCR for page 33
Recommendation 2-2a: To further enhance the excellent data provided in the SEER program database, adequate resources should be provided to expand SEER program coverage beyond the existing sites to include high-risk populations for which SEER program coverage is lacking. This expansion should address a wider range of demographic and social characteristics by using consistent nomenclature and a uniform data set and by reflecting the diverse characteristics of the current U.S. population. Recommendation 2-2b: NCI should continue to work with the North American Association of Central Cancer Registries and other organizations to expand the coverage and enhance the quality of the 45 non-SEER program state cancer registries with the intent of ultimately achieving—together with the SEER program state registries—two goals: (1) a truly national data set through a system of longitudinal population-based cancer registries covering the entire country, and (2) a reliable database for each state to serve as the basis for both the development and evaluation of cancer control efforts in that state. Recommendation 2-3: Annual reporting of cancer surveillance data and population-based research needs to be expanded to include survival data for all ethnic groups, as well as for medically underserved populations. Recommendation 2-4: The committee recommends an emphasis on ethnic groups rather than on race in NIH's cancer surveillance and other population research. This implies a conceptual shift away from the emphasis on fundamental biological differences among "racial" groups to an appreciation of the range of cultural and behavioral attitudes, beliefs, lifestyle patterns, diet, environmental living conditions, and other factors that may affect cancer risk. Recommendation 2-5: The committee commends the proposed NCI program of expanded behavioral and epidemiological research examining the relationship between cancer and cancer risk factors associated with various ethnic minority and medically underserved groups, and recommends that these studies be conducted both across and within these groups.