Cancer and disparities in health: Perspectives on health statistics and research questions

The finding that black Americans, compared with white Americans, have inferior 5-year survival rates and death rates from cancer has been noted for more than 40 years. Disparities in other causes of death were noted even earlier. Race and differences in health outcomes have been topics for politicians and scientists ever since.1 In many respects, the academic field of health outcomes research owes its creation to concerns about disparate outcomes among the races. In America, race is a difficult, volatile topic. Medical science and the interpretation of data have been affected greatly by America's obsession with race. The finding that there are disparities in cancer incidence and mortality is incontrovertible.2 The reasons for the disparities and even how to define the populations with disparities are not as clear and, indeed, are quite controversial. However populations are categorized, there are disparate influences on the cause of the disease. the course of the disease, and the outcome of the disease. Defining these influences is the most scientifically appropriate way to develop and test the interventions that may overcome these disparities. Five-year survival rate Incidence rate Mortality rate Stage distribution at diagnosis Adequacy of treatment received Quality of life. Incidence rates, 5-year survival rates, and stage distribution are affected most by screening and early detection activities. Populations that have a higher prevalence of screening for a particular cancer frequently have a higher incidence of that cancer. A noted exception is colorectal screening, in which finding and removing benign polyps can decrease colorectal cancer incidence (prevent cancer). Table 1 lists the incidence and mortality rates for the 15 most common cancers in the United States.3 The racial/ethnic categories are those mandated by the US Office of Management and Budget in 1998 and have been used by the US Census Bureau since 2000. With few exceptions, Americans of black or African heritage have the highest mortality rates and the worst survival of any population.4 Some have suggested that genetic differences among the races are the cause of differences in mortality for breast, prostate, and colon cancer.5 The National Institutes of Health (NIH) Revitalization Act of 1993 mandates that all NIH-funded phase 3 clinical trials be designed to assess the differences in interventions between the races.6 The wording of the legislation actually attributes some if not most racial disparities on treatments, especially drugs, that have different affects on 1 race versus another. Less emphasis has been placed on the possibility that environmental influences associated with being black affect the biology of cancer and cause worse prognoses. Even less emphasis is put on the finding that a lower proportion of blacks or African Americans receive high-quality medical care compared with white Americans. Racial disparities in quality of care received have been documented in the screening or prediagnostic setting as well as during treatment. Most interest in disparities has centered on black-white differences. It is of concern that the incidence and mortality rates of some minority groups actually may be higher than reported because of coding discrepancies or lack of access to care, leading to a lack of diagnosis.7 Some populations, such as Native Americans, have such low cancer rates that it must asked whether the system is failing to count some Native-American cancer patients. The low cancer rates among Asians, Native Americans, and Hispanics, if they are accurate and are studied appropriately, have the potential to identify strategies that may be useful for blacks and whites. If these rates truly are low, then study of these populations may lead to preventive interventions for blacks and whites. Although our labels and categories do predict for groups of individuals that are less likely to do well with cancer and other diseases, we must realize the sociopolitical nature of these categories. The concept of race is a remnant of the justification for slavery. The original categories date back several hundred years.8, 9 Negroid, Caucasoid, Mongoloid, and later derivations all are constructs rejected by anthropologists and based in societal politics. These categories are not based in science (taxonomy or anthropology). It is noteworthy that the racial categories used in our country have changed over time. A man born in India in 1948 who immigrated to the United States before the 1950 Census would have been categorized into 3 different racial/ethnic groups by his 50th birthday. He would have been counted as Caucasian, Indian, and Asian. It is truly unfortunate that data concerning health are collected using these sociopolitical categories and that scientific conclusions are drawn from them. The concept that phenotypic differences translate into biologic differences was the basis of 'Race Medicine.' This concept commonly was accepted in the 19th century and early 20th century America. It was talked about in numerous medical publications from the 1860s to the 1940s, and it was one of the reasons that the study commonly known as the 'The Tuskegee Syphilis Study' was deemed reasonable and ethical. It began in 1932 as a cohort study in which the affects of untreated syphilis were observed in Negro men, and it was deemed reasonable because it was believed that syphilis was a very different disease in blacks versus whites. Many actually believed that syphilis rarely killed Negroes but that it did kill whites.10 It is unfortunate today to read that breast cancer is a different disease in blacks versus whites or that prostate cancer is a different disease in blacks versus whites. This view can obscure the truth and actually can impede science from benefiting disparate populations. Race is a sociopolitical categorization, and it can be a surrogate for other extrinsic causes of disease, such as socioeconomic status (SES) or culture. If we are truly to attack disparities in health outcomes, we must be open minded, question the standard paradigms, and carefully define the scientific and medical questions that need addressing. Other more causally related and, thus, more scientific and more relevant ways of categorizing populations also can be used. These include ethnicity, SES, and area of geographic origin. All 3 can be correlated loosely with the sociopolitical concept of race within the United States. For example, the proportion of black Americans in poverty or near poverty in the United States is higher than the proportion of white Americans. For another example, ethnic Chinese (Asian Americans) are more likely to have a diet higher in vegetables compared with white Americans. Ethnicity relates to a group of behaviors and habits, such as diet and other culturally related or environmental influences. Indeed, migration studies indicate that Asians who move to the United States from China and Japan and acculturate increase their risk of breast, prostate, and colon cancers, presumably through diet. Similar data have been reported regarding Eastern Europeans migrating from their homeland to Western Europe or Australia.11 In 2003, the US National Cancer Institute (NCI) published a monograph reporting that SES, which is influenced by education, income, and poverty level, is correlated with cancer outcomes. Americans living in poorer census tracts of the United States have inferior 5-year survival rates from most cancers. Individuals living in poorer areas also have higher incidence and higher mortality for most major cancers.12 American Cancer Society intramural epidemiologists have reported that lower education, lower SES, and lack of insurance are correlated with the greater likelihood of a poor outcome.13 Those with lower SES are more likely to present with higher stage disease; and, even within individuals who have the same disease stage, low SES is a predictor of worse outcome. SES influences the quality and types of foods we eat, the environment in which we live, and the work that we do.14 It correlates with several environmental influences that can cause disease. The National Health Interview Survey suggests that poor Americans are more likely to have harmful health behaviors and are less likely to practice healthy behaviors. Compared with higher SES Americans, they consume higher calorie diets and have diets high in carbohydrates and fats. Poor Americans also are more likely to be overweight or obese and are less likely to consume diets high in fruits and vegetables compared with middle class and upper middle class Americans.15 Poor Americans may be more likely to live in neighborhoods with environmental pollutants, which may influence cancer, or in crime ridden neighborhoods, where exercise and recreation are not safe options. Many of the factors that determine SES in the United States are related intimately to race. A higher proportion of Americans of African heritage are poor compared with whites. It is appropriate to ask about the affect poverty has on our race-based health statistics? Indeed, SES and its incumbent environmental influences may be the cause of many of our cancer disparities. It is possible that socioeconomic factors that act largely through and are associated with race are responsible for much of the disparity between blacks and whites. The influence of social deprivation or SES and overall health also is observed in studies done in Scotland. Scotland is overwhelmingly Caucasian or white and has a very good health registry. At each decennial census, all Scots are categorized into 1 of 7 deprivation index categories. This index uses an extensive questionnaire to determine those who are the wealthiest (Category 1) and those who are the poorest or most deprived (Category 7). Figure 1 shows that the most deprived have a greater risk of death compared with the wealthiest or least deprived.16 Table 2 shows that the most affluent have better 5-year survival rates compared with the most deprived.16 This bar chart illustrates the death rate in Scotland for all cause mortality, 1991. Adapted from: McLaren G, Bain M. Deprivation and Health in Scotland: Insights from NHS Data. Edinburgh, Scotland: ISD Scotland; 1998.16 Scottish and American data reveal the influences that SES and social deprivation can have not only on the risk of malignant disease but also on pathologic factors.17, 18 The Europeans use the phrase 'social deprivation' in lieu of poverty. In population studies, they have demonstrated a correlation between social deprivation and the increase risk of having poor prognostic breast cancer biologies. Thomson and colleagues18 suggest that poor women with breast cancer are more likely to be diagnosed with estrogen receptor-negative tumors. In the middle 1990s, Gordon demonstrated similar findings when studying poor white women in Ohio. It is unknown how poverty influences the pathology of disease. This is likely through lifelong diet and birthing habits. Diets high in calories during childhood can affect menarche, which affects breast cancer risk later in life. Another way of categorizing populations is the area of geographic origin, which can correlate with ethnicity and race. The concept of area of geographic origin can be more scientific than race. It recognizes that someone may be of mixed origins, eg, both European and African origins or of indigenous American or Asian origins. In the United States, we often categorize an individual as black when their heritage is of several races or when they have various areas of geographic origin. There are small genetic differences among populations as defined by area of geographic origin. Some of these differences have been associated with an environmental influence and/or a survival advantage. Sickle cell disease, for example, is associated with individuals from Spain, Italy, Greece, the Middle East, and northern and sub-Saharan Africa.19 It is noteworthy that, in the United States, sickle cell disease often thought of as a marker for Africans. Although the prevalence of sickle cell trait and sickle cell anemia is greater in sub-Saharan Africa, there is a prevalence among individuals who originate from southern Europe and who are considered white. Sickle cell is not observed among black Africans originating in southern Africa. It is believed that this genetic mutation is an example of genetic selection. Those who have sickle cell trait had some advantage during a massive malaria epidemic several thousand years ago. It is an example of environmental influences on genetics. There are genetic markers that correspond with areas of geographic origin far better than with skin color or race. Glucose 6 phosphate dehydrogenase deficiency is common but not monopolized among individuals who originate in the Middle East and the Mediterranean.20 Alcohol dehydrogenase deficiency is another example: It is common in individuals from certain areas of Asia.21 There are populations often defined by other than racial criteria for whom certain drugs are less likely to work or are more likely to cause side affects. The epithelial growth factor inhibitor gefitinib (trade name, Iressa) clearly has greater activity in Japanese patients with nonsmall cell lung cancer as a group versus Americans of European ancestry.22 The drug irinotecan (trade name, Camptosar) may cause diarrhea in twice as many individuals who identify themselves as African in origin versus those who identify themselves as European in origin.23 Indeed, the scientific data suggest that pharmacogenomic differences among populations, however they are defined, are not the major reasons for disparities in health. These pharmacogenetic differences rarely mean that a particular drug should not be administered for a specific disease. The literature is filled with comparison studies demonstrating that there are racial differences in the amount and quality of care received. The major factor in health disparities is that a drug does not work because the poor, disenfranchised, or underserved patient does not get the drug. Numerous patterns-of-care studies have been published that examine the black-white differences in healthcare. Unfortunately, few studies have studied disparities by SES, and very few have studied other racial/ethnic groups, such as Hispanics, Asians, or Native Americans. Irrespective of race, lung cancer mortality is strikingly higher in less educated men compared with men who have higher levels of education. Over the past 30 years, smoking rates have been higher among the less educated. Data suggest that equal treatment yields equal outcomes among black and white patients with nonsmall cell lung cancer and small cell lung cancer.24 Surgical and radiation therapy outcomes all are similar.25, 26 Patterns-of-care studies indicate that black Americans less frequently receive surgical resection,24, 27-29 radiation therapy,24 chemotherapy,30 or any definitive treatment29 compared with whites. In an article that received much press attention, Bach and colleagues noted that blacks with early-stage nonsmall cell lung cancer were less likely to receive curative resection compared with whites. It also was noted that blacks were more likely to die sooner than whites. In a population-based study of nearly 11,000 Medicare recipients with resectable nonsmall cell lung cancer (stage I and II) who were diagnosed between 1985 and 1993, 64% of blacks and 77% of whites underwent a resection. Blacks as a group have a 5-year survival rate of 26.4% versus 34.1% for whites. Among patients undergoing surgery, no racial difference was reported in survival.25 The mortality rates for black and white Americans from 1975 to 2005 for breast cancer and colon cancer are provided in Figure 2. The major changes in the mortality rate over 3 decades probably are not attributable to differences in inherent genetic makeup of the populations. The growing disparity after a period of relative equivalence a more likely are because of differences in disease behavior brought on by differing changes in diet or, even more likely, differences in early detection and treatment received once patients are diagnosed. These charts illustrate cancer mortality for colon cancer (Top) and breast cancer (Bottom). Adapted from: Ries LAG, Melbert D, Krapcho M, et al, eds. SEER Cancer Statistics Review, 1975-2005 (based on November 2007 SEER data submission, posted to the SEER website, 2008). Bethesda, Md: National Cancer Institute; 2007.3 In Scotland and in the United States, the poorest Americans tend to present with higher stage disease at diagnosis. Studies of Scots and Americans demonstrate that wealthier women with breast cancer are more likely to present with estrogen receptor-positive disease compared with poorer women, even within the same disease stage. A higher portion of poorer women are overweight or obese.17, 18 In a large study of Americans, breast cancer mortality rates were higher among women with less education than among women with more education. This finding was true regardless of race.13 Important questions that deserve more scientific resources include: 'What is it about poverty that influences breast cancer tumor biology and the molecular markers of breast tumors?' What are the environmental influences on tumor biology? The word 'environmental' has broad meaning and includes dietary influences and other chemical pollutants as well as reproductive habits. We in medicine often do a sort of medical racial profiling to identify individuals who are at higher risk for a disease. A black American with breast cancer is more likely to have a triple-negative tumor, which means that the tumor is less likely to express estrogen receptors, progesterone receptors, and HER-2/neu.31 Most studies suggest that equal treatment yields equal outcomes among equal patients, and there is not equal treatment for many in the United States; this means that a black woman is less likely to receive optimal care in the United States compared with a white woman who has the same disease. The study by Lund et al32 is eye-opening. Those authors assessed the care received by black and white women in a population-based Surveillance, Epidemiology, and Ends Results registry. Black women experienced longer treatment delays after diagnosis, regardless of their disease stage at diagnosis. Black women were 4- to 5-fold more likely to experience delays in treatment >60 days after diagnosis. For locoregional disease, significantly more black women did not undergo cancer-directed surgery (7.5% vs 1.5% of white women). Among patients who were eligible for breast-conserving surgery, only 61% of blacks versus 72% of whites received radiation. Among women who should have received hormone therapy, black women were less likely to receive it.32 Griggs et al33 found evidence that some chemotherapy drugs are given in lower than recommended doses because of obesity. In the United States, blacks have a higher prevalence of obesity than whites. American men of African heritage have a higher incidence and mortality from prostate cancer compared with white men. Many have speculated there may be a genetic mutation, series of or common in men of African heritage that cause the have racial differences in and in of the and the growth factor 1 All data are and reported that of men of European and of black American men at least 1 of of the on which an of for prostate cancer. The various prostate cancer incidence rates in several populations of African origin suggest that some environmental influences do influence prostate cancer incidence rates among men of African It should be noted that the black population in the United States is among the only of these groups in which screening is for prostate cancer began in the United States early in the 1990s, and the screening rates to have among black Americans by Even for screening, the incidence rates of prostate cancer in black Americans are higher than those reported in other African populations. Disparities in index and obesity may be the in prostate cancer and in many black-white cancer Black Americans tend to have higher than white Americans. A higher proportion of Americans of African heritage have the and This is a disease in which the levels of and growth factors are The that these levels of growth factors have on prostate cancer and other cancer is unknown but of for is believed to an in the differing prevalence of the and breast cancer The American Cancer Society Cancer is a study of men who an questionnaire in and were through et reported that, among black men, and was associated with a higher risk of prostate cancer with a increase in relative The influence of was especially with a relative risk of the highest versus the studies also have observed that the of may to prostate cancer in black men in the United States. This is an environmental influence associated with ethnicity and culture. There to be a correlation between tumor and treatment in prostate cancer. et studied patients black men and white and demonstrated that obese men to have worse and tumor and survival compared with Black American men tend to have higher than white Americans. The affect on pathology that is associated with obesity is It is likely a factor associated with diet and calorie The poor and those with lower tend to present with more disease. In several studies of American men with prostate cancer, was correlated with disease, and race was et observed that SES was a major in outcomes for men with prostate cancer. is one of the few population-based studies that a large of patients in to blacks and whites. those authors observed that the influence of race and ethnicity was not nearly as as the influence of cancer is another disease in which there to be some racial differences in of care received. In the Cancer black men with disease to receive less therapy compared with white men with the same disease. There were no racial differences in the care of lower have demonstrated black-white differences in treatment. population-based studies and have demonstrated differences in of colorectal cancer treatment by There is a of blacks to get less treatment than whites. This includes studies that assessed and studies that assessed In a of studies on black-white disparities in survival that on colon cancer, differences were only after for socioeconomic factors and The mortality rate from cancer of the is but most of these are through early detection and appropriate treatment. The incidence and mortality rate is highest among the less educated and most There is a difference in quality of screening and quality of of findings among poor Americans compared with the middle There are several on population-based of care in the Studies also have reported racial/ethnic disparities in the of treatment for cancer. There are differences in the quality of clinical and differences in the definitive treatment or Blacks are more likely than whites to receive no treatment after A major of this is that disparities by race, area of geographic origin, and SES do Many of the disparities are related to disparities or in of care among the poor and especially among black Americans. There are differences in preventive as well as differences in treatment once patients are diagnosed. It is an truth that, as a we have this Although disparities are a is a Many and the biology of race have our to with all of the disparities our in with racial our to the A large of literature to demonstrate that the poor, and individuals of other are less likely to receive optimal In studies, Bach and have the disease by disease, to demonstrate that the literature the concept that equal treatment equal outcomes in the treatment of cancer, and and data demonstrating that a proportion of minority and poor patients receive less than optimal care compared with patients. In there is not equal treatment. using Medicare data for treatment of breast, or prostate cancers during the period from to and observed that black patients were significantly less likely than white patients to receive therapy for cancers of the lung resection for early-stage 64% vs for blacks and breast after vs colon therapy for stage vs and prostate therapy for early-stage vs There was or no over the years in the proportion of patients therapy for the cancers and there was no decrease in the of any of the racial disparities. These patients all had Medicare The from that study suggest that SES and insurance status are not the only reasons for the disparities in quality of to care is a in the United States. It is only 1 in the of disparities in health. Americans, are poor, and have no health The of Americans who have health insurance is even greater than the of Americans living health they include of all of all Hispanics, and of all whites. The of whites health insurance is greater than the of blacks and We be to more Americans to to health disparities if those we defined those disparities in socioeconomic than racial disparities research should be at major and biologic these cancer disparities. The most question in health disparities research can we high-quality care include preventive to a population that often does not receive We must to define that The of poor whites less than optimal care is greater than the of blacks and who receive less than optimal and also must ask whether they are the scientific questions and whether questions about race are to What is the influence of poverty or social deprivation on cancer biology? are the poor at higher do the poor have the more What is the of diet and cancer among these What environmental influences There are factors that can cause lack of optimal Some care is not given because the patient reasonable treatment or because social such as lack of not Some care may be because of the of diseases, such as health and of is if we are to disparities. race, we as must realize that poverty and unfortunate many in the race to receive less than optimal care and to habits. scientists have about biologic differences among racial groups and have not on the do we care to include preventive care and education to those who often do not receive These are academic for and as well as social and