AREA-LEVEL VARIABLES AND CANCER
Cancer Incidence
Area-SES
Numerous studies have examined the relationship between area-SES and cancer incidence, and found that results vary by cancer site/type (Singh et al., 2003; Yin et al., 2010). For example, low area-SES is associated with elevated incidence for cancers of the stomach, lung, and cervix among Blacks and Whites alike (Baquet, Horm, Gibbs, & Greenwald, 1991; Gorey & Vena, 1994; Krieger et al., 2002). Alternatively, high area-SES consistently has been found to be related to increased incidence for cancers of the female breast, prostate, and colon (Whites only—Cheng et al., 2009; Reynolds et al., 2005; Robert et al., 2004). All of the above studies adjusted for individual-level characteristics; a few are summarized in Table 6.2.
Rurality and Segregation
The association of rurality with cancer incidence generally is consistent across studies, with higher incidence in urban than in rural communities (e.g., Reynolds et al., 2005; Robert et al., 2004; Sung, Blumenthal, Alema-Mensah, & McGrady, 1997). For example, a study on urban/rural differences in cervical cancer in Georgia Medicaid recipients (N = 111,208 women, 1988-1992) found a higher incidence of cervical cancer in metropolitan Atlanta (vs. rural areas), among Black women in particular (Sung et al., 1997). Data on segregation and cancer incidence are limited. High segregation is related to increased cancer incidence associated with ambient air toxics, with Black-White disparities in cancer incidence widening with increasing segregation (Morello-Frosch & Jesdale, 2006).
Cancer Mortality and Stage at Diagnosis
Area-SES
Although the role of area-SES in cancer incidence varies by cancer type, its association with stage at diagnosis and with cancer mortality is by and large consistent across cancer types. Lower area-SES is associated with late-stage diagnosis and higher mortality for both Blacks and Whites, above and beyond individual-level demographics (Cheng et al., 2009; Echeverria, Borrell, Brown, & Rhoads, 2009; Greenlee & Howe, 2009; MacKinnon et al., 2007; Yabroff & Gordis, 2003). However, because Blacks are overrepresented in low-SES areas, they are more likely to be diagnosed with late-stage disease (e.g., Campbell et al., 2009) and exhibit higher mortality than Whites (e.g., Gerend & Pai, 2008). For example, Yabroff and Gordis (2003) examined county-level SES and breast cancer stage at diagnosis and mortality among Blacks and Whites ages ^ 55 years. They found that lower area-SES was associated with higher late-stage breast cancer incidence and death rates. Similarly, Echeverria et al.(2009) investigated 4,589 New Jersey urban Black and White breast cancer cases, and found that the odds of late-stage disease for women living in low-SES neighborhoods were 1.6 times higher than that of women living in more advantaged neighborhoods, even when controlling for age and race/ethnicity. Likewise, the Cheng et al. (2009) study of 8,997 California prostate cancer deaths found that higher area-SES was associated with lower prostate cancer death rates; moreover, Blacks had a twofold to fivefold increased risk of prostate cancer deaths compared to Whites across all levels of area-SES.
TABLE 6.2 Selected Studies on the Role of Area Variables in Cancer
|
Association of Area-SES and Cancer Incidence, Mortality, Stage at Diagnosis, Treatment, and Survival |
|||
|
Author |
Participants |
Area-SES Measures |
Low Area-SES Associated With |
|
Yin et al. (2010) |
CA statewide national sample of breast, prostate, colorectal, cervical, and lung cancer survivors diagnosed between 1 998 and 2002 (n = 376,158) |
CT and CB SES (occupation, unemployment, median household income, % BPL, median gross rent, median value of owner-occupied houses, etc. |
Decreased cervical, lung, and colorectal cancer incidence (especially White); increased breast and prostate cancer incidence (Blacks and Whites) |
|
Cheng et al. (2009) |
CA statewide sample of prostate survivors diagnosed between 1998 and 2002 (n = 98,484) |
CB SES (education, income, and occupation) |
Decreased prostate cancer incidence; increased mortality |
|
Yabroff and Gordis (2003) |
National sample of breast cancer cases from SEER |
County-level SES (median family income, % BPL, education, unemployment) |
Decreased localized breast cancer incidence; increased distant breast cancer incidence; increased case-fata I ity rate |
|
Zell et al. (2007) |
CA statewide sample of pancreatic cancer cases diagnosed between 1989 and 2003 (n = 24,735) |
CB SES (education, median household income, % BPL, median house value, median rent, occupation, employment) |
Poor survival |
|
Ward et al. (2004) |
National sample from SEER diagnosed between 1975 and 2000 |
CT % BPL |
13% higher death rates from cancer in men and 3% higher rates in women; 10% points lower in survival |
TABLE 6.2 (continued)
|
Association of Urbanicity and Cancer Incidence, Mortality, Stage at Diagnosis, Treatment, and Survival |
|||
|
Author |
Participants |
Rural-Urban Measures |
High Urbanicity Associated With |
|
Robert et al. (2004) |
Population-based Wisconsin breast cancer survivors and control (n = 14,667) |
CT and zip-code Census Bureau-defined rural and urban categories |
Increased breast cancer incidence |
|
McLafferty and Wang (2009) |
IL statewide sample of breast, prostate, colorectal, and lung cancer cases diagnosed between 1998 and 2002 |
Zip-code level RUCA (Rural-Urban Continuum/Commuting codes from the Department of Agriculture) |
Increased cancer risk, following a J-shaped progression (with small upturn in the most isolated rural areas) |
|
Hao et al. (2010) |
GA statewide colorectal survivors diagnosed between 2000 and 2004 (n = 4,748) |
RUCA classification |
Increased odds of receiving chemotherapy, but urban Blacks were less likely to receive the treatment than Whites |
|
Author |
Participants |
Segregation Measures |
High Segregation Associated With |
|
Morello-Frosch and Jesdale (2006) |
45,710 census tracts in 309 U.S. metropolitan areas |
MSA Dissimilarity Index |
Increased estimated cancer risk associated with ambient air toxics |
|
Dai (2010) |
Detroit breast cancer cases in 156 zip codes |
Zip-code level Isolation Index |
Increased risk of late stage breast cancer diagnosis |
|
Haas et al. (2008) |
SEER breast cancer senior survivors diagnosed between 1992 and 2002 (n = 47,866) |
CT Isolation Index |
Decreased odds of receiving adequate breast cancer care |
Rurality and Segregation
Unlike the consistent, inverse relationship between rurality and cancer incidence, studies of rurality and late-stage diagnosis and cancer mortality have yielded mixed results. Some have found late-stage diagnosis to be associated with rural residence (e.g., Jemal et al., 2005) among Blacks in particular (Amey, Miller, & Albrecht, 1997); alternatively, others have found late-stage cancers (of the female breast, colorectal, lung, and prostate) to be highest in the most highly urbanized areas and to decrease as rurality increases, with a small increase in isolated rural areas (McLafferty & Wang, 2009). Such differences may reflect use of different definitions of rural versus urban.
There are only a few studies of residential segregation and stage at diagnosis and cancer mortality. Dai (2010) found a significant association between Black segregation and late-stage diagnosis among breast cancer patients in Detroit, but Haas et al. (2008) reported no such association between segregation and breast cancer mortality, even though increased segregation is strongly associated with increased all-cause mortality (Jackson, Anderson, Johnson, & Sorlie, 2000). Again, these studies controlled for individual-level demographics.
Cancer Treatment and Survival
Area-SES
Lower area-SES is associated with a lower probability of early detection, preferred treatment, and survival of cancers for Blacks and Whites (Campbell et al., 2002; Singh et al., 2003; Tewari et al., 2005; Ward et al., 2004; Zell, Rhee, Ziogas, Lipkin, & Anton-Culver, 2007). However, the Black-White disparity in receipt of adequate cancer treatment and in survival remains irrespective of area-SES (Singh et al., 2003; Tewari et al., 2005; Ward et al., 2004). For example, Singh et al. (2003) found that the percentage of Black patients, diagnosed with localized or regional prostate cancer, receiving radical prostatectomy was lower than the percentage of Whites within each area-poverty group (CT % BPL = >10%, 10% to < 20%, 20% +). Similarly, Tewari et al. (2005) found that surgical treatment rate for clinically localized prostate cancer (diagnosed between 1980 and 1997) was 17% for Blacks versus 28% for Whites; survival rates were lower for Blacks than for Whites among patients treated conservatively or by radiation therapy, while survival rates did not differ for Blacks and Whites among those undergoing surgery. Likewise, Ward et al. (2004) found that even when CT poverty was accounted for, Blacks still had a lower 5-year survival than Whites. This pattern has been found for many other cancer sites as well (e.g., Singh et al., 2003; Zell et al., 2007).
Rurality and Segregation
Studies of the association between rurality and cancer treatment are generally consistent, with rural residence associated with underuse of surgical, radiation, and chemotherapy (Esnaola, Knott, Finney, Gebregziabher, & Ford, 2008; Hao et al., 2010; Sankaranarayanan et al., 2010). In urban areas however, a Black-White disparity in receipt of cancer treatment consistently has been found (Esnaola et al., 2008; Hao et al., 2010). For example, Hao et al.(2010) reported a rural disadvantage in receipt of chemotherapy for stage III colon and stages II/III rectum cancer patients in Georgia (n = 4,748 diagnosed during 2000-2004), after controlling for area-SES and other relevant factors (e.g., comorbidities); a Black-White disparity in receipt of adjuvant chemotherapy was observed only for patients residing in urban (vs. suburban and rural) areas. Similarly, being Black is independently associated with underuse of surgery among urban patients with nonmetastatic breast cancer in South Carolina, whereas rural residence is associated with underuse of surgery irrespective of patient race (Esnaola et al., 2008).
Studies of residential segregation and cancer treatment are less consistent. Haas et al. (2008) found that high Black segregation was associated with inadequate cancer care among 47,866 senior Black and White breast cancer survivors diagnosed during 1992-2002. However, no significant association was found between segregation and the receipt of adjuvant chemotherapy among patients with colorectal cancer in Georgia (Hao et al., 2010).
POSSIBLE MECHANISMS OF AREA-LEVEL INFLUENCES
Researchers have identified a variety of neighborhood-level resources, hazards, and barriers that play a role in the higher prevalence of cancer and cancer risk factors in urban, low-SES (i.e., inner-city), and urban Black-segregated (i.e., inner-city Black) neighborhoods, and in rural communities as well. These specific neighborhood disparities have been highlighted as contextual contributors to and social-contextual determinants of cancer disparities.
Smoking
Inner-city and inner-city Black neighborhoods contribute to high smoking among their residents through at least four pathways. The first is the significantly higher exposure to tobacco advertising—outdoors, on store windows, and at the point of purchase—in inner-city areas, with such exposure known to play a role in increased tobacco use (Barbeau, Wolin, Naumova, & Balbach, 2005; Diez-Roux, Merkin, Hannan, Jacobs, & Kiefe, 2003; Hackbarth et al., 2001; Luke, Esmundo, & Bloom, 2000). Moreover, the tobacco industry sponsors music and sporting events in inner-city Black neighborhoods, with this including displays of tobacco advertisements and logos at the events, and on the free t-shirts, hats, bags, and other items distributed (Lee, Cutler, & Burns, 2004; Yerger, Przewoznik, & Malone, 2007). Such exposure contributes to higher smoking rates among inner-city Blacks (59% of men, 41% of women) relative to their suburban cohorts and to Whites (19-25%; Delva et al., 2005).
A second mechanism is the significantly higher availability of tobacco in inner-city neighborhoods, that is, the greater number and density of convenience stores and tobacco shops, both strongly associated with increased tobacco use (Chuang, Cubbin, Ahn, & Winkleby, 2005; Siahpush, Jones, Singh, Timsina, & Martin, 2010). In addition, sales of single cigarettes— removed from the pack and sold individually for pennies—are higher in inner-city areas irrespective of area ethnicity, and in inner-city Black neighborhoods in particular; access to this cheap form of tobacco facilitates smoking among adults and youth who may be unable to afford a pack due to rising prices secondary to increasing excise taxes.
A third mechanism is the high levels of stress associated with life in inner-city neighborhoods (Elliot, 2000), with smoking used as a (stress-reducing) coping mechanism for many (Bennett, Wolin, Robinson, Fowler, & Edwards, 2005; Janzon et al., 2005). Finally, the fourth mechanism is the well-known failure of inner-city physicians to advise patients—Blacks in particular—to quit smoking (Ashford et al., 2000; Bach, Pham, Schrag, Tate, & Hargraves, 2004; Gemson, Elinson, & Messeri, 1988; van Ryn, 2002). Alternatively, in rural communities, lack of access to smoking prevention and cessation play a role in excess tobacco use, along with community norms (e.g., Cox et al., 2008).
Obesity
At least six neighborhood disparities have been demonstrated to contribute to (i.e., mediate) the high rates of obesity in inner-city and inner-city Black-segregated neighborhoods. These are: (1) the failure of inner-city physicians to advise dietary modification and weight loss, especially to Blacks (Ashford et al., 2000; Bach et al., 2004; van Ryn, 2002); (2) the significantly lower access to healthy foods, supermarkets, and fresh fruits and vegetables in inner-city areas (Morland & Filomena, 2007; Morland, Wing, Diez-Roux, & Poole, 2002; Powell, Chaloupka, & Bao, 2007; Powell, Slater, Mirtcheva, Bao, & Chaloupka,2007); (3) the significantly higher access to high-calorie, high-fat, fast-food outlets (Block, Scribner, & DeSalvo, 2004; Kwate, 2008; Powell et al., 2007); (4) the significantly lower access to exercise and other recreational facilities (Mobley et al., 2006; Moore, Diez-Roux, Evenson, McGinn, & Brines, 2008; Powell, Slater, Chaloupka, & Harper, 2006); (5) the significantly lower levels of outdoor activity and walking due to fear of crime and to absence of sidewalks, lighting, and green space (Boardman et al., 2005; Chang et al., 2009; Massey, 2001); and (6) the high stress of such neighborhoods (Elliot, 2000; Glass et al., 2006).
Specifically, inner-city and inner-city Black neighborhoods contain 2-4 times more fast-food outlets and convenience stores (Block et al., 2004; Kwate, 2008), 3 times fewer large supermarkets (Morland & Filomena, 2007; Morland et al., 2002), and are 3-8 times more likely to lack recreational facilities than other neighborhoods (Moore et al., 2008; Powell et al., 2006). Indeed, more than 30 studies have documented the paucity of healthy food choices in inner-city areas, with such neighborhoods now referred to as "food deserts," known to partially mediate poor diet and high BMI among residents (Black & Macinko, 2008; Mehta & Chang, 2008; Walker, Keane, & Burke, 2010; Wang et al., 2007). Similarly, the absence of recreational facilities has been shown to partially mediate low levels of physical activity and high BMI in inner-city neighborhoods; 70% of Black versus 38% of White neighborhoods have no recreational facilities (Gordon-Larsen, Nelson, Page, & Popkin, 2006; Mobley et al., 2006; Moore et al., 2008).
Equally important is the high level of ongoing, chronic stress associated with residing in inner-city neighborhoods. Stressful neighborhood features include garbage-filled vacant lots, abandoned and boarded buildings, graffiti, trash, constant noise, stray dogs, and other signs of neighborhood danger and deterioration (Accordino & Gary, 2000; Cohen et al., 2000; Glass et al., 2006; Massey, 2001). These features are strongly associated with chronic stress, vigilance, and poor physical and mental health (Cohen et al., 2003; Glass et al., 2006). Indeed, neighborhood danger has been shown to cause increased cortisol production, with ensuing high BMI and central adiposity irrespective of diet and physical activity (Glass et al., 2006). Likewise, chronic stress causes increased production of pro-inflammatory factors (e.g., tumor necrosis factor alpha, C-reactive protein), and other forms of physiological "wear and tear" known to be predictive of obesity and central adiposity (Black, 2002; Geronimus, Hicken, Keene, & Bound, 2006; Glass et al., 2006; Miller, Cohen, & Ritchey, 2002; Segerstrom & Miller, 2004).
Cancer Screening
Similarly, low rates of cancer screening in inner-city and inner-city Black neighborhoods in part reflect low access to health care and to cancer screening in those communities—and play a role in subsequent diagnosis of late-stage cancers as well (Coughlin et al., 2006, 2008; Dai, 2010; Datta et al., 2006; Onega et al., 2008; Schootman et al., 2006; Zenk, Tarlov, & Sun, 2006). For example, Zenk et al. (2006) found that the distance from Black neighborhoods to low- or no-cost cancer screening facilities in Chicago was significantly further than from White neighborhoods—by more than a mile. For rural populations, low access to cancer screening facilities and high barriers to transportation to such facilities are significant contributors to low screening rates (Coughlin et al., 2006, 2008; Larson & Fleishman, 2003; Probst et al., 2004). For example, Huang et al. (2009) found that the farther a woman had to travel to receive a mammogram, the greater her chances were of being diagnosed with a late-stage breast cancer. Rurality interacts with segregation to yield extremely low access to screening and to cancer care among rural Blacks (Probst, Laditka, Wang, & Johnson, 2007).
Cancer Incidence
Where cancer incidence is concerned, the higher rates of smoking in rural and in inner-city areas no doubt contribute, along with the higher rates of obesity in inner-city and inner-city Black neighborhoods, detailed above. Differences between segregated Black versus White neighborhoods in environmental exposures likewise appear to contribute to Blacks’ higher incidence of certain cancers (Olden & White, 2005). National and regional studies consistently have found that environmental exposures in minority neighborhoods are 5-20 times higher than those in White neighborhoods (even after controlling for area-SES) as a result of the deliberate placement of air-polluting factories and toxic waste dumps in minority neighborhoods (Centers for Disease Control, 2005; Morello-Frosch & Jesdale, 2006; Morello-Frosch & Lopez, 2006; Olden & White, 2005). Segregated Black neighborhoods are characterized by significantly higher exposure to air toxins and persistent organic pollutants (e.g., dioxins, pesticides), as well as by significantly higher exposure to mercury, arsenic, lead, sulfur dioxide, and a myriad of known carcinogens (Centers for Disease Control, 2005; Morello-Frosch & Jesdale, 2006; Morello-Frosch & Lopez, 2006). Such higher environmental exposures have been implicated in Blacks’ higher incidence of some cancers (Morello-Frosch & Jesdale, 2006).
Cancer Treatment, Mortality, and Survival
Racial and rural disparities in cancer treatment, mortality, and survival at least in part reflect the low access to high-quality, cancer treatment facilities and to oncologists in segregated Black and rural communities—resulting in underuse of surgical, radiation, and chemotherapy in those communities even when cancer is diagnosed at an early or treatable stage (Esnaola et al., 2008; Haas et al., 2008; Hao et al., 2010; Sankaranarayanan et al., 2010). Such disadvantages in the receipt of cancer treatment (particularly when coupled with late-stage diagnosis) in part explain poorer cancer survival and higher cancer death rates in those segregated Black and rural communities (Ward et al., 2004).
CONCLUSIONS
The numerous studies reviewed here reveal that residence in rural—and in inner-city and inner-city Black—neighborhoods is associated with higher rates of smoking and obesity, lower rates of cancer screening, higher cancer incidence and mortality, and lower-quality cancer treatment—irrespective of individual-level SES (personal education and income), and in some cases, irrespective of race as well. Such data imply the need to focus perhaps not so much on Black or poor people, but instead on these problematic geographic areas—in other words, on neighborhood disparities—in the effort to reduce cancer disparities.
