8
Sm. SC;. Med. Vol. 34, No. 2, pp. 191-198. 1992 F’rinted in Great Britain 0277-9536/92 S5.00 + 0.00 Pcrgamon Press plc PREVENTIVE HEALTH BEHAVIOR AMONG BLACK AND WHITE WOMEN IN URBAN AND RURAL AREAS SONJA I. DUELBERG Department of Geography, University of Illinois at Urbana-Champaign, Urbana, IL 61801, U.S.A. Abstract--The relationship of race to preventive health behavior among women is examined using data from the 1985 National Health Interview Survey. We find that black women are less likely to engage in primary prevention behaviors such as exercising, non-smoking and maintaining a favorable weight. However, black women are more likely to engage in secondary prevention behaviors such as receiving a Pap test or a breast exam. These findings are surprising as they indicate a change in secondary prevention behavior among black women. The racial differences in exercising, maintaining a favorable weight and receiving a Pap test or a breast exam cannot fully be explained by the differing levels of socioxonomic status, measured by education and income. However, the higher percentage of smoking among black women is due to their lower levels of education. Urban/rural residence modifies the effect of race on smoking and receiving a Pap test. Black women in urban areas are most likely to be smokers. Almost no difference exists between white women in urban and rural areas concerning their likelihood of receiving a Pap test, we find that black women in urban areas are much more likely to be screened for cervical cancer than black women in rural areas. Key work-health behavior, primary prevention, secondary prevention INTRODUCTION The purpose of this paper is to examine race differ- ences in preventive health behavior among women. If black and white women differ in their likelihood of smoking, exercising, being overweight, getting Pap tests, and breast exams, we then ask two secondary questions. First, are race differences in preventive health behaviors due to education and income, or does race have an independent effect, with behaviors differing between blacks and whites even at the same level of socio-economic status? Second, does urban vs rural residence modify the effect of race on preventive health behaviors? Black women have worse health than white women While the life expectancy for black women has risen during the last decade, it is still about 5 years less than that of white women. In 1987 life expectancy at birth was 78.9 for white women and 73.6 for black women [l]. It is estimated that between 1979-1981 23,261 excess deaths occurred in black females up to age 69 [2]. The leading causes of deaths for white and black females are heart disease, stroke, and cancer. However, differences in these rates exist as well. For all these causes the age-adjusted death-rates are higher for black than for white females. The black/white death rate ratio of females is 1.5 for heart disease, 1.8 for stroke and 1.2 for cancer [2]. Cancer mortality rates differ according to the site of the cancer. While white women have a slightly higher mortality from breast cancer than black women, the breast cancer mortality of black women under the age of 40 is 50% higher than that of white women [3]. Between 1980 and 1987 cervical cancer mortality rates decreased for both black and white women. However, the mortality rates of black females are still more than twice those of white females [4]. The difference in breast and cervical cancer between black and white women varies between geographical areas [5]. Preventive health behaviors affect health Results of previous studies indicate that preventive health behaviors affect health and mortality. For this study we select three health behaviors representing primary prevention and two health behaviors repre- senting secondary prevention. A. Primary prevention. Primary prevention can be defined as behavior by which individuals actively improve or maintain their health status. Physiological changes in the body have not yet occurred. The health behaviors analysed in this study are exercising, main- taining a favorable weight and non-smoking. Lack of exercise, obesity and smoking have a negative effect on health and have been linked to increased mortality [6-lo]. Exercise is positively related to good health status. Women who report never to participate in leisure physical activities have a worse than average health status [ll]. Exercise appears to con- tribute especially to the prevention of coronary heart disease [12, 131. A recent study showed that exercise reduces cardiovascular risk factors of middle- aged women [14]. Exercise reduces the risk of colon cancer [15]. Obesity increases the risk of hyper- tension, heart disease and diabetes mellitus [16]. Even mild- to moderate-overweight in women is positively associated with coronary heart disease [17]. 191

Preventive health behavior among black and white women in urban and rural areas

Embed Size (px)

Citation preview

Page 1: Preventive health behavior among black and white women in urban and rural areas

Sm. SC;. Med. Vol. 34, No. 2, pp. 191-198. 1992 F’rinted in Great Britain

0277-9536/92 S5.00 + 0.00 Pcrgamon Press plc

PREVENTIVE HEALTH BEHAVIOR AMONG BLACK AND WHITE WOMEN IN URBAN AND RURAL AREAS

SONJA I. DUELBERG

Department of Geography, University of Illinois at Urbana-Champaign, Urbana, IL 61801, U.S.A.

Abstract--The relationship of race to preventive health behavior among women is examined using data from the 1985 National Health Interview Survey. We find that black women are less likely to engage in primary prevention behaviors such as exercising, non-smoking and maintaining a favorable weight. However, black women are more likely to engage in secondary prevention behaviors such as receiving a Pap test or a breast exam. These findings are surprising as they indicate a change in secondary prevention behavior among black women. The racial differences in exercising, maintaining a favorable weight and receiving a Pap test or a breast exam cannot fully be explained by the differing levels of socioxonomic status, measured by education and income. However, the higher percentage of smoking among black women is due to their lower levels of education. Urban/rural residence modifies the effect of race on smoking and receiving a Pap test. Black women in urban areas are most likely to be smokers. Almost no difference exists between white women in urban and rural areas concerning their likelihood of receiving a Pap test, we find that black women in urban areas are much more likely to be screened for cervical cancer than black women in rural areas.

Key work-health behavior, primary prevention, secondary prevention

INTRODUCTION

The purpose of this paper is to examine race differ- ences in preventive health behavior among women. If black and white women differ in their likelihood of smoking, exercising, being overweight, getting Pap tests, and breast exams, we then ask two secondary questions. First, are race differences in preventive health behaviors due to education and income, or does race have an independent effect, with behaviors differing between blacks and whites even at the same level of socio-economic status? Second, does urban vs rural residence modify the effect of race on preventive health behaviors?

Black women have worse health than white women

While the life expectancy for black women has risen during the last decade, it is still about 5 years less than that of white women. In 1987 life expectancy at birth was 78.9 for white women and 73.6 for black women [l]. It is estimated that between 1979-1981 23,261 excess deaths occurred in black females up to age 69 [2]. The leading causes of deaths for white and black females are heart disease, stroke, and cancer. However, differences in these rates exist as well. For all these causes the age-adjusted death-rates are higher for black than for white females. The black/white death rate ratio of females is 1.5 for heart disease, 1.8 for stroke and 1.2 for cancer [2]. Cancer mortality rates differ according to the site of the cancer. While white women have a slightly higher mortality from breast cancer than black women, the breast cancer mortality of black women under the age of 40 is 50% higher than that of white women [3].

Between 1980 and 1987 cervical cancer mortality rates decreased for both black and white women. However, the mortality rates of black females are still more than twice those of white females [4]. The difference in breast and cervical cancer between black and white women varies between geographical areas [5].

Preventive health behaviors affect health

Results of previous studies indicate that preventive health behaviors affect health and mortality. For this study we select three health behaviors representing primary prevention and two health behaviors repre- senting secondary prevention.

A. Primary prevention. Primary prevention can be defined as behavior by which individuals actively improve or maintain their health status. Physiological changes in the body have not yet occurred. The health behaviors analysed in this study are exercising, main- taining a favorable weight and non-smoking. Lack of exercise, obesity and smoking have a negative effect on health and have been linked to increased mortality [6-lo]. Exercise is positively related to good health status. Women who report never to participate in leisure physical activities have a worse than average health status [ll]. Exercise appears to con- tribute especially to the prevention of coronary heart disease [12, 131. A recent study showed that exercise reduces cardiovascular risk factors of middle- aged women [14]. Exercise reduces the risk of colon cancer [15]. Obesity increases the risk of hyper- tension, heart disease and diabetes mellitus [16]. Even mild- to moderate-overweight in women is positively associated with coronary heart disease [17].

191

Page 2: Preventive health behavior among black and white women in urban and rural areas

193 SONJA I. DUELBERG

Smoking is mainly associated with increased risk of lung cancer, heart disease, stroke and cervical cancer [ 18-201.

B. Secondary preaention. Secondary prevention in- cludes health behaviors like an examination or test to determine physiological changes in the body, while a person does not display symptoms. Screening for the early detection of disease often requires access to

medical care facilities. For this study we select receiv- ing a Pap test and receiving a breast exam as sec-

ondary prevention behaviors. The Pap test is a device to detect precancerous lesions of the cervix. There is consensus that screening using Pap tests can reduce cervical cancer incidence and mortality [21-241. It is estimated that if women were screened every 3 years the cervical cancer mortality could be reduced by 70-95% [25]. A physical breast examination in women over 50 years is an effective tool to reduce mortality from breast cancer. Evidence about the effect of yearly physical breast exams in younger women is not conclusive [25].

Black women are less likely to practice preaentice

health behaviors than white

In the Alameda county study [7], health practices were measured using an index, which included smok- ing, obesity and exercise. The findings showed that blacks had lower scores on the health practices index than whites. Only a few studies have analysed gender and race differences in preventive health behavior. One study showed that black women are less likely than white women to be physically active [26]. More evidence exists about differences in obesity and smok- ing behavior. Black women in all age-groups are more obese than white women [l, 26, 271. In 1985 black women were more likely to smoke than white women [I]. Data from 1987 shows that the smoking rates of black and white women are almost identical [28, I]. Gottlieb and Green [26] suggest that the differences between black and white women in health behaviors like non-smoking, relative weight and physical exercise can largely be explained by the

differences in education, income, social networks, and life events.

Several studies have investigated secondary preven- tion behavior, especially secondary cancer prevention (29, 301. Studies show that blacks are less likely to be screened for cervical cancer or to receive a breast exam. This is explained by their differing levels of knowledge and attitudes about cancer and their poorer access to medical care [31-331. A recent analysis of a regional survey of women over 50 yr of age distinguished between receiving a physical breast exam and receiving a mammogram. The results show that while white women are more likely to have obtained at least one mammogram, non-white women are more likely to receive a physical breast exam. Reasons for the surprisingly higher likelihood of non-white females to receive a physical breast exam could not be given [34].

Education and income affect health behaciors

Studies show that education is related to primary prevention behaviors. Persons with lower levels of education are more likely to smoke, to be overweight and lack exercise [35-381. Particularly strong is the relationship between low education and smoking

[39,40]. One study shows that the association of education with obesity differs for race/sex groups, Education has no association with obesity in black women but has a negative association with obesity among white women. Low income is linked to smok- ing, obesity and low levels of activity [38,41, 71. Low education and low income are also related to low levels of screening for cervical and breast cancer [30,42,43]. Because education and income affect health behavior in this way differences between black and white women in their health behavior could be caused by differences in education and income.

Urban L’S rural residence affects health behaciors

Concerning differences in health behavior between urban and rural residents, Greenberg [44] found that in 1950 rural residents were more likely to practice primary prevention behaviors, like non-smoking, than urban residents. Between 1950 and 1980 this difference between the two groups diminished. Con- trary to this, Kleinman and Kopstein’s [30] analysis of secondary cancer prevention using the 1973 National Health Survey found that living in a non-metropolitan area was negatively related to having a Pap test. Poor black women living in non-metropolitan areas were least likely to have had a Pap test [30].

Study objectives

This study has two aims. First, we examine the relationship of race and health behavior for black and white women in the United States. Are there differences in the health behavior between blacks and whites? Is the relationship the same for primary and secondary prevention behavior? Our hypothesis is that blacks will display a more negative health behav- ior than whites. It is expected that socio-economic and demographic variables are related to health behavior. Specifically, we expect that education and income will have a positive effect on health behavior. If this is confirmed, the higher percentage of blacks with lower education levels and lower income could explain their poorer health behavior. This study will analyse if this is the case or if race has an independent effect on health behavior.

Secondly, we examine if there are possible differ- ences between rural and urban residents concerning their health behavior. Specifically we are interested in any differences that might exist between primary and secondary prevention. Primary prevention was found to be slightly better in rural areas [44]. We would expect that urban residents have equal or worse primary prevention behavior than rural residents, but

Page 3: Preventive health behavior among black and white women in urban and rural areas

Preventive health behavior among black and white women 193

that rural residents have poorer secondary preven- tion. Secondary prevention is in general linked to use of special services. Availability and access to health services are worse in rural areas than in urban areas [45,46]. Finally, we analyse if the effect of urban/ rural residence differs significantly for black and white women.

METHODS

Data

The data for this analysis come from the 1985 Health Interview Survey representing civilian, non- institutionalized residents of the U.S. and carried out by the U.S. Department of Health and Human Services. The sample was obtained using a multistage sampling technique. Black persons were oversampled. From each household, one adult respondent was selected and interviewed in person for the Health Promotion and Disease Prevention Supplement. The response rate was over 96%. The analysis reported here is based on 19,027 female respondents, 16.2% of whom were black and 83.8% of whom were white. Other racial groups were excluded from the analysis.

Measurement of variables

The independent variables of interest are race, education, income and urban/rural residence. We.also control for age and marital status. Age is measured in number of years. The average age in this sample is 45.71 yr. Marital status is coded not married = 0, married = I. Race is coded white = 0, black = 1. Edu- cation is measured in years of formal schooling completed. Income is total family income during the previous year, coded on a 27-point scale. Urban/rural residence is measured according to the person’s place of residence. If the residence is in a metropolitan area the variable urban is coded I, if the residence is in a non-metropolitan area the variable urban is coded 0.

The dependent variables cover primary and sec- ondary prevention behavior. All are self-reports. Because different socioeconomic variables affect health behavior in opposite directions [IO, 471 no index will be used in this study, but the different health behaviors will be evaluated separately. All

health behaviors are coded so that a high score indicates positive behavior. The primary prevention behavior variables are exercise, favorable weight, and non-smoking. Nonsmokers are coded 1 and current smokers 0. An index is used to measure exercise. Respondents were asked if they had participated in any of 19 physical activities or sports excercises in the two weeks before the interview. These exercises were walking, jogging, hiking, gardening, aerobics, danc- ing, calisthenics, golf, tennis, bowling, biking, swim- ming, yoga, weight lifting, basketball, baseball, football, soccer, volleyball, handball or squash, skat- ing and skiing. In addition the respondents could name up to two activities not on the list. The highest possible value on the variable exercise is 21. The average score on the exercise variable in this sample is 1.708. Degree of overweight is measured by the Quetelet index, which is a ratio of weight to height squared. Of the various weight to height measures, the Quetelet index is the most adequate, because it is the least correlated with height and because it is highly correlated with skinfold measures and thus a good indicator of total body fat [59,60]. Favorable weight is a continuous variable, coded from high to low levels of overweight. This is achieved by sub- tracting the score on the Quetelet index from 70 (a number just above the maximum score). The average score on the variable favorable weight is 45.580.

Secondary prevention behavior is measured by two variables. The first one is Pap test. Respondents were asked “About how long has it been since you had a Pap smear test?” The second one is breast exam. Respondents were asked “About how long has it been since you had a breast examination by a doctor or other health professional?” The responses to both variables are coded on a ten point scale with never = 1, more than 20 yr ago = 2, between 10 and 19yr ago = 3, between 6 and 9yr ago =4, 5yr ago=5,4yrago=6,3yrago=7,2yrago=8, lyr ago = 9 and within the last yr = 10. In sum, women who have never received a pap test or breast exam receive a score of zero on the respective test; those who had one within the last year, a score of IO.

Table 1 shows the correlation matrix, means, and standard deviations of the variables. Data are

Table I. Bivariate correlations, means. and standard deviations (N = 19.027)

I 2 3 4 5 6 7 8 9 IO 11

I Age 2 Married -0.151 3 Urban - 0.064 -0.050 4 Black -0.087 -0.177 0.111 5 Education -0.315 0.101 0.087 -0.1 IO 6 Income -0.185 0.498 0.129 --O-,243 0.414 7 Exercise -0.348 0.004 0.029 -0.042 0.256 0.126 8 Fav. Weight -0.167 0.026 0.038 -0.162 0.204 0. I37 0.145 9 Nonsmoker 0.151 0.024 -0.037 -0.041 0.046 0.033 0.017 -0.087

IO Paptest -0.320 0.160 0.063 0.085 0.199 0.195 0.130 0.068 -0.047 I I Breastexam -0.230 0. I28 0.078 0.057 0.198 0.185 0.119 0.066 -0.012 0.734

Mean 45.706 0.505 0.749 0.161 12.061 15.333 I .708 45.580 0.707 7.995 8.267 SD 19.099 0.500 0.433 0.367 2.936 7.765 I .770 5.066 0.455 2.794 2.684

Correlations greater than 0.026 are significant at P -c 0.001.

Page 4: Preventive health behavior among black and white women in urban and rural areas

193 SONJA 1. DUELBERG

analysed by multiple regression using progressive adjustment. Although most preventive health behav- iors are measured on an interval scale, non-smoking is dichotomous. According to Cleary and Angel [48] the use of dichotomous variables as dependent variables in a regression model using ordinary least squares will not affect the validity of the results if the probability of the dichotomous variable lies between 0.25 and 0.75. In this case, the results of linear regression and logistic regression are not significantly different. The one dichotomous dependent variable in this analysis fits the criteria for the use of OLS regression: 70.7% of the respondents are non-smokers.

RESULTS

In the following, the results of a series of multiple

regressions using progressive adjustment on the 5 dependent variables of health behavior will be discussed. In a first step, each dependent variable is

regressed on race controlling for age, marital status, and urban residence. We expect that age and marital status will be significantly related to health behavior. Since we know that differences in age and marital status exist between black and white women, we control for them. (The unadjusted race differences in health behavior are shown in Table 1.) In step 2, education and income are added. This allows us to evaluate if differences in education and income can explain the differences in health behavior between black and white women. In step 3 the interaction term of race by urban residence is added.*

E?cercise

When we control for age, marital status, and urban

residence in step 1 of the regression analysis with exercise as the dependent variable, we find that black women are less likely to exercise than white women. Urban residence has no significant? effect on exercise. The results of step 2 indicate that education and income both have a significant effect on exercise. Women with high levels of education and income are more likely to exercise than those with low levels of education and income. A comparison of steps 1 and 2 reveals that education and income explain a small part of the effect of race on exercise. Race, however, still has a large direct effect on exercise. This suggests that most of the effect of race on exercise is not due solely to education or income. Urban residence was not significant in either step 1 or step 2. Urban and rural residents do not differ significantly in their exercise behavior. When, in step 3, the interaction of race and urban residence is added it is not significant. The effect of race on exercise is not modified by urban/rural residence.

*Tables containing the detailed results of the regression analyses can be obtained from the author.

tin this discussion ‘significant’ refers to a P-value smaller than 0.01 (a-tailed test).

Favorable weight

In the regression analysis with favorable weight as the dependent variable we find that, controlling for age, marital status, and urban residence, race has a significant effect on favourable weight. Black women are less likely to have favorable weight. In addition, urban residents are more likely to have favorable weight than rural residents. In step 2, education and income are added. Both are statistically significant. Women with higher levels of education and income are more likely than poor persons and those with low levels of education to have favorable weights. Neither education nor income explain the effect of race on favorable weight. When education and income are added in step 2 black women still are significantly less likely than white women to be normal rather than overweight. Controlling for education and income has an effect on urban residence. Urban residence has still a significant positive effect on favorable weight, but the effect is much smaller now. Therefore, a big part of the positive effect of urban residence on favorable weight is due to the higher levels of edu- cation and income for those women living in urban areas. The interaction term of race and urban resi- dence, which is added in step 3, is not significant. The effect of race on favorable weight is not conditioned by urban residence. Both urban and rural black women are less likely than white counterparts to be normal weight rather than overweight.

Non -smoking

By controlling for age, marital status, and urban residence in step 1, we find that race has a significant negative effect on non-smoking. Black women are less likely to be non-smokers than white women. In addition, urban residents are less likely to be non-smokers than rural residents. When education and income are added in step 2, only education is statistically significant. Women with higher education are more likely to be non-smokers than women with lower levels of education. While urban residence has still a significant effect, race is no longer significant once education is added. Education is therefore an important link between race and non-smoking. In step 1 we find that black women are less likely to be non-smokers than white women. When education is added to the equation in step 2, the effect of race becomes nonsignificant. This suggests that levels of education largely explain the effects of race on non- smoking. Thus black women smoke more due to their lower levels of education. However, there is also a significant interaction between race and urban residence. The effect of race on non-smoking is conditioned by urban residence. The interaction is graphed in Fig. 1 at the mean level of all other variables. It shows that black women living in urban areas are more likely to smoke than white women living in urban areas. Rural black women are more likely to be non-smoking than rural white women.

Page 5: Preventive health behavior among black and white women in urban and rural areas

Preventive health behavior among black and white women

El Urban Whites

Urban Blacks

lz Rural Whites

I Rural blacks

El Urban Whites

195

0.6

0.6

0.4

0.2

c

Fig. I. Interaction effect of race and urban residence on non-smoking. The equation at the mean level of all other independent variables are as follows: Urban whites: 0.703 - 0.026 (1) + 0.071 (0) - 0.09 (I x 0) = 0.677. Urban blacks: 0.703 - 0.026 (I) + 0.071 (1) - 0.09 (1 x I) = 0.658. Rural whites: 0.703 - 0.026 (0) + 0.071 (0) - 0.09 (0 x 0) = 0.703. Rural blacks: 0.703 - 0.026 (0) + 0.071 (1) - 0.09 (0 x 1) = 0.774. Note: Non-smoking is coded I for non-

smokers and 0 for current smokers.

Pap test

In step 1, when age, marital status and urban residence are controlled, race has a significant posi- tive effect on receiving a Pap test. Black women are more likely to have had a Pap test than white women. In addition, urban residence has a positive effect when age and marital status are controlled. Urban residents are more likely to have a Pap test than rural residents. Education and income which are added in step 2 both have a significant positive effect on receiving a Pap test. While race is still significant in step 2, urban residence no longer has a significant effect. This suggests that the effect of urban residence can be explained by education and income. Urban residents have more Pap tests because they have higher levels of education and income. When edu- cation and income are controlled for, race still has a large direct effect on receiving a Pap test. It actually increases. The interaction term of race by urban residence, which is added in step 3, is significant. The interaction is graphed at the mean level of all other variables in Fig. 2. White women in urban areas are only slightly more likely to have a Pap test than tihite women living in rural areas. For black women, as well, those in urban areas report receiving a Pap test more often than those in rural areas, but the difference between blacks in urban areas and blacks in rural areas is considerably larger. Black women living in urban areas are the most likely to receive a Pap test.

10

II Urban Blacks

EB Rural Whites

I Rural blacks

6

6

4

2

0 1

Fig. 2. Interaction effect of race and urban residence on Pap test. The equation at the mean level of all other independent variables are as follows: Urban whites: 7.795 + 0.058 (1) +0.396 (0) + 0.510 (1 x 0) = 7.853. Urban blacks: 7.795 + 0.058 (1) +0.396(l) + 0.510 (I x 1) = 8.759. Rural whites: 7.795 + 0.058 (0) +0.396(O) + O.SlO(O x 0) = 7.795. Rural blacks: 7.795 + 0.058 (0) + 0.396 (1) + 0.510 (0 x 1) = 8.191. Note: Pap test is coded on a ten point scale from 0 = never received a Pap test to 10 = received a Pap

test within the last year.

Breast exam

In the regression analysis with receiving a breast exam as the dependent variable, race and urban residence are significant when age and marital status are controlled. Black women are more likely to have a breast exam than white women. Residents in urban areas are more likely to have a breast exam than residents of rural areas. In step 2 we find that education and income have also a positive effect on receiving a breast exam. In contrast to the results for Pap test, urban residence still has a significant effect on having a breast exam. This suggests that education and income do not explain the effect of urban residence on receiving a breast exam. As with Pap tests, race still has an effect when education and income are controlled for. It has an even larger effect. In step 3, the interaction of race and urban residence is added. In contrast to Pap tests, the interaction is not significant.

DISCUSSION

Black women are less likely to exercise, to be normal weight, and to be non-smokers than white women. Black women are more likely to have a Pap test and breast exam than white women. Thus, white women engage in more primary prevention; blacks in more secondary prevention. Furthermore, race has

Page 6: Preventive health behavior among black and white women in urban and rural areas

196 SONJA I. DLZLBERG

an independent effect on certain health behaviors, like exercising and being overweight. Black women are less likely to exercise or to be normal weight, even at the same level of education and income. The high percent- age of smoking among black women, however, can be attributed to their lower educational level. The differ- ences between black and white women in their primary prevention behavior could be explained by a lower level of personal control. Personal control affects behavior patterns, including health behaviors [49]. Personal control can be defined as “a person’s beliefs about how well he or she can bring about good events and avoid bad events” (491. The encounter of overt and covert racism could influence Blacks’ outlook on life and would make them feel in less control. It could negatively influence their attitude about the extent they can control their life, and their future health in particular [5@-521. Studies have found that minorities have a lower sense of control than non-minorities, possibly because of blocked opportunities in a white-

dominated economic system [53]. The results concerning the secondary prevention of

cervical and breast cancer by having a Pap test or a breast exam contrast previous findings. Through the early 198Os, studies found that black women were less likely to have a Pap test or a breast exam. In this study, we find that black women are more likely to have received a Pap test or a breast exam than white women. The findings in this analysis using the most recent data available are supported by unpublished data of the Centers for Disease Control [4] and by a recent study on screening for breast cancer [34]. This change in behavior seems to be fairly recent which could be one reason why mortality data does not reflect this more positive secondary preventive behav- ior by black women [4]. Race has an independent

effect on secondary prevention. Black women are more likely to be screened for cervical cancer and to

receive a breast exam than white women at the same level of socio-economic status.

Thus far dominant theories that have attempted to explain why Blacks receive less cancer screening tests than Whites focused on a low awareness level by Blacks about the benefits of these screenings and/or barriers in the medical care system [3l, 541. This analysis cannot evaluate whether changes in the medical care system or special educational projects have significantly improved the secondary cancer prevention behavior of black women. Possibly certain programs initiated to increase screening among blacks had an effect. One category are those programs focusing on the concept of ‘in-reach’, which promote the increasing use of already existing facilities [55]. This often involves the screening of persons who enter a health clinic for any other reason [31]. This would mean that higher screening levels are not part of conscious decision-making by black women to partici- pate in secondary prevention.

Urban residence has a positive effect on secondary prevention behaviors like having a Pap test or a breast

exam. This could be explained by the higher avail- ability of medical service in urban areas. For receiving a Pap test the positive effect of urban residence however, is largely due to higher socio-economic status of urban residents.

Black women in urban areas are especially likely to have received a recent Pap test. The big difference between urban Blacks and rural Blacks in having a Pap test, compared to the difference between urban Whites and rural Whites, could be explained by the effect of special programs aimed at increasing Pap tests for Blacks which were primarily concentrated on urban areas. Further analysis has to show exactly why the differences between urban and rural Blacks exists.

The impact of urban vs rural residence on health behavior is varied for those behaviors representing primary prevention. Urban residence had no effect on exercise, a negative effect on non-smoking and a positive effect on favorable weight. This does, however, confirm Greenberg’s findings [44] that the

impact of urban residence on health behavior is no longer only negative. The effect of race on smoking differed for urban and rural residents. Black women in urban areas were the group most likely to smoke, whereas black women in rural areas were least likely to smoke. The high levels of black women in urban areas who smoke could be attributed, to a certain extent, to external pressures caused by cigarette advertising which is especially targeted at minorities in cities [56, 571.

These findings have important consequences for the health of black women and for public health policy. The positive secondary prevention behavior among black women could lead to a mortality rate which is more like that of white women. The negative primary prevention which obviously persists means that black women will still develop higher rates of morbidity and mortality. Higher rates of morbidity will have a strong negative effect on the lives of black women, including the burden of having to pay for medical treatment. The application of qualitative research methods in addition to quantitative methods might help to ad- vance our understanding of the underlying causes of the differences between Blacks and Whites in their health behavior. This might be best realized by conducting an in-depth community study, which could examine health behaviors in the broader social and political context in which they take place [58].

In summary, we find that racial and geographical differences in preventive health behavior of women exist. Black women are less likely to engage in primary prevention. However, we find that black women are more likely to engage in secondary prevention. This clearly indicates a reverse of previous behavior pat- terns in secondary prevention. Furthermore, race has an independent effect on health behavior for women. The observed racial differences in preventive health behavior-with the exception of non-smoking-can- not be fully explained by the different levels of education and income between black and white

Page 7: Preventive health behavior among black and white women in urban and rural areas

Preventive health behavior among black and white women 197

women. Lastly, we find that the effect of race on some 22. Boyes D. A. The value of a Pap smear program and

health behaviors, like smoking and receiving a Pap suggestions for its implementation. Cancer 48, 6 13-62 I,

test, is modified by urban/rural residence. 1981.

23. Clarke E. A. and Anderson T. W. Does screening by

Acknowledgemenf--I am grateful to Catherine Ross for her valuable comments on earlier drafts of this paper. 24.

“Pap” smear help prevent cervical cancer? A case- control study. Cancer 2, 14, 1979. Guzick D. S. Efficacy of screening of cervical cancer: A review. Am. J. Publ. Hlfh 68. 125-134, 1978. Greenwald P. and Sondik i. J. Cancer Control Objecfives for the Nation: 1985-2000. National Cancer Institute, Bethesda. MD. 1986. Gottlieb N. H. and Gre& L. W. Ethnicity and lifestyle health risk: Some possible mechanisms. Am. J. Hlfh Promotion 2, 37-51, 1987. Burke G. L. et al. Obesity and overweight in young adults: The CARDIA study. Prevent. Med. 19,476-488, 1990.

1.

2.

3.

4.

5.

6

7.

8.

9.

10.

II.

12.

13.

14.

15.

16.

17.

REFERENCES 25.

26.

27.

28.

29.

30.

31.

32.

National Center for Health Statistics. Healfh, United Sfafes, 1989. Public Health Service, Hyattsville, MD, 1990. Manton K. G., Patrick C. H. and Johnson K. W. Health differentials between blacks and whites: Recent trends in mortality and morbidity. In Health Policies and Black Americans (Edited by Willis D. P.), pp. 129-99. Transactions, New Brunswick, 1989. National Cancer Institute. SEER Program: Cancer Inci- dence and Morfalify in fhe Vnifed Sfafes, 1973-81. NIH Pub. No. 85-1837. National Cancer Institute, Bethesda, MD, 1984. Centers for Disease Control. DHHS. Black-white differ- ences in cervical cancer mortality-United States, 198& 1987. Morbid. Morfal. Weekly Rep. 39, 245-248, 1990. Hayes A. et al. Cancer rate differentials between blacks and whites of three metropolitan areas. J. nafn. Med. Assoc. 81, 237-241, 1981. Breslow L. and Enstrom J. E. Persistence of health habits and their relationship to mortality. Prmenr. Med. 9, 469-483, 1980. Berkman L. F. and Breslow L. Healfh and Ways of Living. Oxford University Press, New York, 1983. Wiley J. A. and Camacho T. C. Life-style and future health: Evidence from the Alameda Country study. Prevent. Med. 9, l-21, 1980. Metzner H. C. er al. Health practices, risk factors, and chronic disease in Tecumseh. Prevent. Med. 12, 491-507, 1983. Pope C. R. Life-styles, health status, and medical care utilization. Med. Care 10. 402-413. 1982. Brock B. M. et al. Alameda county redux: Replication 35 in Michigan. Prpvenf. Med. 17, 483-495, 1988. Haskell W. L. Overview: Health benefits of exercise. In Behauiorul Healfh (Edited by Matarazzo J. D, ef al.), 36. pp. 409-423. Wiley, Chichester, 1985. Berlin J. A. and Colditz G. A. A meta-analysis of physical activity in the prevention of coronary heart disease. Am. J. Epidemiol. 132, 612-628, 1990. Owens J. F. ef al. Physical activity and cardiovascular 37. risk: A cross-sectional study of middle-aged pre- menopausal women. Prer;enr. Med. 19, 147-157, 1990. Severson R. K. ef al. A prospective analysis of physical 38. activity and cancer. Am. J. ,Epidemiol. 130, 522-529, 1989.

33.

34

Van Itallie T. B. Obesity: Adverse effects on health and 39. longevity. Am. J. Clin. Nufr. 32, 273-2733, 1979. Manson J. E. et al. A prospective study of obesity and risk of coronary heart disease in women. N. Engl. J. 40. Med. 322, 887-889, 1990.

18. Centers for Disease Control. DHHS. The Health Conse- quences of Smoking. Cancer. A Report of the Surgeon 41. General. DHHS Publ. No. (PHS) 85-50179. Office on Smoking and Health, Rock;ille, MD, 1982.

19. Fielding J. E. Smoking: Health effects and control_ 1; 42. Public Health and Prevenfive Medicine (Edited by Last J. M.), 12th edn, pp. 999-1038. Appleton-Century- 43. Crofts, Norwalk, 1986.

20. Winkelstein W. Smoking and cervical cancer-current status: A review. Am. J. Epidemiol. 131, 945-957, 1990.

21. Hakama M. and Vistanen U. R. Effect of mass screen- 44 ing programs on the risk of cervical cancer. Am. J. Epidemiol. 103, 512-517, 1976.

Marcus A. C. et al. Prevalence of cigarette smoking in the United States: Estimates from the 1985 Current Population Survey. J. nafn. Cancer Inst. 81, 409-414, 1989. Warnecke R. B. and Graham S. Characteristics of blacks obtaining Papanicolaou smears. Cancer 37, 2015-2025, 1976. Kleinman J. C. and Klopstein A. Who is being screened for cervical cancer? Am. J. Publ. Hlfh 71, 13-16, 1981. Warnecke R. 8. Intervention in black populations. In Cancer Among Black Popularions (Edited by Mettlin C. and Murphy G. P.). pp. 167-183. Liss, New York, 1981. Baquet C. and Ringen K. Health policy gaps in access, delivery, and utilization of the Pap smear in the United States. In Health Policies and Black Americans (Edited by Willis D. P.), pp. 322-347. Tansaction, New Brunswick, 1989. Woolhandler S. et al. Medical care and mortality: Racial differences in preventable deaths. Inc. J. Hlfh Services 15, l-22, 1985. Lerman C. ef al. Factors associated with repeat adher- ence to breast cancer screening. Preuenf. Med. 19, 279-290, 1990. Mechanic D. and Cleary P. D. Factors associated with the maintenance of positive health behavior. Prevenf. Med. 9, 805-814, 1980. Gottlieb N. H. and Green L. W. Life events, social network, life-style, and health: An analysis of the 1979 National Survev of Personal Health Prac- tices and Consequences. Hjfh Educaf. Q. 11. 91-105, 1984. Winkleby M. A. ef al. Social class disparities in risk factors for disease: Eight-year prevalence patterns by level of education. Prevent. Med. 19, I-12, 1990. Gold M. R. and Franks P. The social origin of cardio- vascular risk: An investigation in a rural community. Inf. J. Hlfh Services 20, 40-16, 1990. Wagenknecht L. E. et al. Cigarette smoking is strongly related to educational status: The CARDIA study. Prevent. Med. 19, 158-169, 1990. Novotny T. E. ef al. Smoking by blacks and whites: Socioeconomic and demographic differences. Am. 1. Publ. Hlfh 78, 1187-1189, 1988. Harris D. M. and Guten S. Health-protective behavior: An exploratory study. J. Hlfh Sot. Behav. 20. 17-29. 1979. - _ _ Peters R. K. et al. Barriers to screening for cancer of the cervix. Preuenf. Med. 18, 133-146, 1989. Devesa S. S. and Diamond E. L. Association of breast cancer and cervical cancer incidences with income and education among whites and blacks. J. nam. Cuncer Insl. 65, 515-528, 1980. Greenberg M. Urban/rural differences in behavioral risk factors for chronic diseases. Urban Geog. 8, 146151, 1987.

Page 8: Preventive health behavior among black and white women in urban and rural areas

198 SONJA I. DUELBERG

45. Fiedler J. L. A review of literature on access 53. Mirowsky J. and Ross C. E. Social Causes of Psycho- and utilization of medical care with special emphasis logical Distress. Aldine de Gruyer, New York, 1989. on rural primary care. Sot. Sci. Med. 1X, 129-142, 54. Dutton D. Explaining the low use of health services 1981. by the poor: Costs, attitudes, or delivery system. Am.

46. Joseph A. E. and Phillips D. R. Accessibility and Social. Rev. 43, 348-368, 1979. Utili:ation: Geographical Perspectives on Health Care 55. Howard J. In-reach: An approach to secondary preven- Delivery. Harper and Row, London, 1984. tion of cancer. In Behaaior, Health Risks and Social

47. Hayes S. D. and Ross C. E. Concern with appearance, Disadvantage (Edited by Parron D. L., Solomaon F. health beliefs, and eating habits. J. Hfth Sot. Behav. 28, and Jenkins C. D.). pp. 5 l-61. National Academy Press, 12t&l30. 1987. Washington. DC, 1982.

48. Cleary P. and Angel R. The analysis of relationships involving dichotomous dependent variables. J. Hlth Sot. Behav. 25, 33&348, 1984.

49. Peterson C. and Stunkard A. J. Personal control and health promotion. Sot. Sci. Med. 28, 819-828, 1989.

50. New York Times. Doctors See Gap in Blacks’ Health Hatting a Link to Low Self-esteem, July 17, 1989.

51. Riggs R. S. and Noland M. P. Factors related to the heafih knowledge and health behavior of disadvantaged black vouth. J. Sot. Hlth 54, 431434, 1984.

52. Farley- R. and Allen W. R.. The Color Line and the Quality of Li/e in America. Russell Sage Foundation, New York, 1987.

56. Davis R. M. Current trends in cigarette advertising and marketing. N. Engl. J. Med. 316, 725-732, 1987.

57. Tuckson R. V. Race, sex, economics, and tobacco advertising. J. natn. Med. Assoc. 81, I 119-l 124, 1989.

58. Wallack L. and Winkelby M. Primary prevention: A new look at basic concepts. Sot. Sci. Med. 25,923-930, 1987.

59. Kholsla T. and Lowe C. R. Indices of obesity derived from body weight and height. Br. J. Prevent. Sot. Med. 21, 122-128, 1967.

60. Roche A. F. et al. Grading body fatness from limited anthropometric data. Am. J. C/in. Nutrit. 34, 2831-2838, 1981.