Public Health Nursing Vol. 32 No. 4, pp. 287–297 0737-1209/© 2014 Wiley Periodicals, Inc. doi: 10.1111/phn.12162
POPULATIONS
AT
RISK ACROSS
THE
LIFESPAN: POPULATION STUDIES
Mammogram Use and Self-Efficacy in an Urban Minority Population Bonnie Jerome-D’Emilia, PhD, MPH, RN, and Patricia Dunphy Suplee, PhD, RNC-OB School of Nursing – Camden, Rutgers, The State University of New Jersey, Camden, New Jersey Correspondence to: Patricia Dunphy Suplee, School of Nursing – Camden, Rutgers, The State University of New Jersey, 311 N. 5th St., Camden, NJ 08102. E-mail: [email protected]
ABSTRACT Objectives: (1) To compare a sample of low-income African American and Hispanic women in general and mammogram specific self-efficacy and other factors potentially associated with screening to identify any differences related to ethnicity and in the use of mammogram screening; and (2) to examine the association of general self-efficacy and mammography specific self-efficacy and mammogram screening in these two ethnically different groups of women. Design and Sample: Cross-sectional. A convenience sample of 139 women. Measures: General and mammogram specific self-efficacy, having ever had a mammogram, acculturation, and demographics. Results: Mammogram specific self-efficacy was significantly associated with having had a mammogram (p < .001), as was insurance status (p = .027). Using logistic regression, older women (OR: 1.3) and those with insurance (OR: 4.8) were more likely to have been screened. When mammogram specific self-efficacy was added to the model, overlap between this construct and insurance prevented insurance from reaching significance. Conclusions: An association between insurance status and mammogram specific self-efficacy was found. It is likely that mammogram specific self-efficacy will vary with mammogram adherence and insurance status, rather than predict screening. General self-efficacy, higher in screened women, may be an effective mediator through which to develop interventions to increase preventive health-seeking behaviors. Key words: African Americans, Hispanics, mammography, self-efficacy.
Background People who regard themselves as highly efficacious act, think, and feel differently from those who perceive themselves as inefficacious. They produce their own future, rather than simply foretell it. (Bandura,1986)
Breast cancer is the second greatest cause of cancer mortality for women, exceeded only by lung cancer (American Cancer Society [ACS], 2014). An estimated 232,670 new cases of invasive breast cancer in women are expected in the United States in 2014 (ACS, 2014). Although non-Hispanic White women have a higher incidence of breast cancer than African American women, African American women have a higher death rate from breast cancer across all age ranges (ACS, 2013a). Hispanic
women are less likely to be diagnosed with breast cancer than White or African American women, yet breast cancer is the most frequently diagnosed cancer and the greatest cause of cancer death in Hispanic women (ACS, 2013b). Hispanic and African American women are more often diagnosed at later stages and with negative prognostic features leading to the increased mortality rate noted in these populations (DeSantis, Ma, Bryan, & Jemal, 2014). Lower income women have a significantly greater risk of diagnosis at a later stage and dying of breast cancer than do higher income women (Sprague et al., 2011; Vona-Davis & Rose, 2009). Despite widespread availability of mammography services, women with less education and lower income, the uninsured and recent immigrants all
287
288
Public Health Nursing
Volume 32
Number 4
have lower rates of mammography screening than found in the middle-class non-Hispanic White or African American populations (Kaiser Family Foundation, 2014; Sprague et al., 2011). According to the 2012 National Health Interview Survey, 66.5% of women 40 and older of all races and ethnicities reported having had a mammogram in the last 2 years. When race and ethnicity were considered, rates were fairly consistent with 67.4% of non-Hispanic White women, 67.9% of African American women and 64.2% of Hispanic women 40 and older reporting having had a mammogram in the past 2 years (National Center for Health Statistics, 2014). However, only 36% of uninsured women 40 and older and 53% of women with less than a high school degree reported having had a mammogram in the past 2 years (National Center for Health Statistics, 2014). In addition to various demographic variables, such as lack of education, studies have found insurance status to be the primary predictor of cancer screening across ethnicities (Henry et al., 2011; Nuno, Castle, Harris, Estrada, & Garcia, 2011). Selvin and Brett (2003) found that regardless of race/ ethnicity, maintaining a usual source of care was the strongest predictor of mammography screening. Jerome-D’Emilia (2014), in a meta-analysis of barriers and facilitators to screening in the Latina population, found that for low-income Hispanic women socioeconomic barriers were the most significant factors determining a woman’s likelihood of being screened. This study concluded that, excluding language barriers, poor women, whether Hispanic, African American, or White, seem to experience the same barriers and facilitators to mammogram screening. Removing financial and access barriers to screening has not been effective in increasing screening rates in low-income women (Teran, Baezconde-Garbanati, Marquez, Castellanos, & Belkic, 2007). As the country heads toward full implementation of the Affordable Care Act (ACA), an estimated 6.8 million low-income women will gain access to insurance and nearly all insured women will be entitled to screening mammograms without cost sharing (Levy, Bruen, & Ku, 2012). According to a recent Centers for Disease Control update of breast cancer screening services (considering mammogram utilization following full implementation of the ACA), barriers such as a women’s low self-effi-
July/August 2015
cacy toward breast cancer screening will likely continue to impede appropriate utilization of screening services (Centers for Disease Control and Prevention [CDC], 2014). To improve rates of mammography and cancer screening in general, it is imperative to examine factors such as self-efficacy to understand how they may influence a women’s utilization of screening services. Self-efficacy refers to the confidence a person feels about taking a certain action, including the confidence to overcome barriers that may stand in the way of taking that action. A higher level of selfefficacy may be what allows some women to access preventive care, while many others with the same socioeconomic conditions and facing the same potential barriers do not. Derived from Social Cognitive Theory, a theory that explains how people develop certain behavioral patterns (Bandura, 1997), general self-efficacy has been widely applied to predict and explain health-seeking behavior and has been found to be associated with positive health outcomes such as: more frequent use of active coping (Luszczynska, Scholz, & Schwarzer, 2005); self-management of chronic illness and perception of wellness (Joekes, Van, & Schreurs, 2007); and health-related quality of life (Kvarme, Haraldstad, Helseth, Sørum, & Natvig, 2009). The construct of self-efficacy has been measured in many and varied ways when used to examine the influence of self-efficacy on utilization of cancer screening services. Schwarzer and Jerusalem (1995) developed the Generalized Self-Efficacy Scale (GSES) to measure the level of perceived general self-efficacy. Although this tool has been used effectively to measure confidence in goal setting, effort, and persistence, other studies have opted to construct different measures of self-efficacy more specific to the behavior in question. Cronan and colleagues, in a survey of African American, Mexican American and White women, measured self-efficacy with three 10-point response scaled questions that assessed a women’s confidence that mammography can detect cancer, prevent cancer, and allow early diagnosed cancer to be cured (Cronan et al., 2008). This measure of self-efficacy was a statistically significant predictor of mammogram screening in the White and Mexican American samples, but not for the African American women. Another study, Stewart, Rakowski, and Pasick (2009) measured self-efficacy with one yes or no
Jerome-D’Emilia and Suplee: Mammogram Use and Self-Efficacy survey question, “Do you think that you could get a mammogram every year?” This operationalization of self-efficacy was able to predict mammogram use significantly only in White women. The association between self-efficacy and mammogram utilization in non-White samples of women is less clear. Champion, Skinner, and Menon (2005) developed and tested the validity of a self-efficacy scale specific to mammography. This scale, which includes items such as: “I can arrange transportation to get a mammogram,” measures a woman’s perceived confidence in her ability to obtain mammogram screening (Champion et al., 2005). Using this mammogram specific scale, these researchers found that individuals with higher mammogram specific self-efficacy were more likely to express the intention and follow through by obtaining a mammogram. In the initial testing of this tool, the total selfefficacy scores increased over time in the adherent group of women. This could reflect the fact that the more adherent a woman is to screening, the higher her level of mammogram specific self-efficacy. It could also be construed as evidence that these questions are too specific to the act of getting a mammogram, so that the self-efficacy score goes up due to familiarity with the screening facility and process. In that case, the self-efficacy score can be considered more of a proxy for adherence than a predictor for mammogram screening. The GSES may provide a more reliable and stable measure of mammography use. Champion et al. (2008) state that the need to identify valid and reliable measures of women’s health-seeking behavior that are specific to race and ethnicity is essential to eliminate disparities. It is not clear if the construct measures of self-efficacy are valid and reliable across ethnicities (Burke et al., 2009).
Purpose The overall aims of this study were to: (1) compare low-income African American and Hispanic women in terms of socioeconomic status, general and mammogram specific self-efficacy, and other factors that are potentially associated with screening to see if there are any differences related to ethnicity which may affect the likelihood of using mammogram screening; and (2) examine the association of general self-efficacy and mammography specific
289
self-efficacy and mammogram screening in this sample of women to see which measure may be the better predictor of mammogram screening in this population.
Methods Design and sample This study used a descriptive cross-sectional research design. Originally, the researchers identified a Catholic church in a primarily Hispanic neighborhood and a Baptist church with a primarily African American congregation, both in a lowincome, underserved urban area of New Jersey, as the recruitment sites. These churches were chosen as study sites due to the race or ethnicity of their populations and their clergy members’ willingness to participate. The eligibility criteria for participation was literacy in English or Spanish and age 40 and older (the ACS recommends that screening mammography begin at 40 years of age [ACS, 2013c]). In early 2012, all self-reported Hispanic and African American women attending Sunday services at both churches in a low-income urban area of NJ were approached by the researchers (with the assistance of one African American and one bilingual Hispanic nursing student) and recruited to participate in the study. The sample size was not predetermined, and all eligible women were asked to participate. However, the sample of 139 was adequate to determine whether those women with high mammogram specific self-efficacy (above the median) would be more likely to have ever had a mammogram than participants with low mammogram specific self-efficacy, with 80% power to detect a difference as low as 21% between the groups to be statistically significant at the 0.05 level. Women who were eligible and agreed to participate signed informed consents (available in Spanish and English), completed the surveys, and received five dollar gift cards for their effort. Data collection took place on church grounds. This study was approved by the Institutional Review Board of Rutgers University. The population of Hispanic women at the Catholic Church was a younger population of women than found at the Baptist church. This is consistent with the finding by Lugo, Smith, Cox, and Pond (2008) that Hispanics comprise almost half of all
290
Public Health Nursing
Volume 32
Number 4
Catholics in the United States under age 40. Of the 41 Hispanic women who attended the services on the day of data collection, only 15 (37.5%) of the women were 40 and over. Therefore, given the researchers’ goal of recruiting relatively equal samples of African American and Hispanic women, the decision was made to recruit women attending a primarily Hispanic community center, following the same procedure for consent and completion of the self-reported survey.
Measures Women were asked to complete the General SelfEfficacy Scale (GSES; Schwarzer & Jerusalem, 1995), the Mammography Specific Self-Efficacy Scale (MSSES; Champion et al., 2005), and a 10-item researcher created demographic survey. All of the surveys were available in Spanish and English. Acculturation was measured by the proxy “How many years have you lived in the United States,” which has been found to be highly correlated with other, more detailed measurements of acculturation (Brown, Consedine, & Magai, 2006). The GSES consists of ten items with a 4-point Likert type scale (1 = Not at all true, 4 = Exactly true) which take the general form: “I have the capacity to do X even if Y (barrier) occurs” and yields a final composite score with a range of 10–40. This scale has been translated and used in many and varied studies in 23 countries, with Cronbach’s alpha ranging from 0.76 to 0.90, with the majority in the high 0.80s (Finlayson, Edwards, & Courtney, 2011; Kvarme et al., 2009). Criterion-related validity has been documented in numerous correlational studies (Carlsson, Bjorvatn, Engebretsen, Berglund, & Natvig, 2004; Joekes et al., 2007). The MSSES was developed to provide a reliable and valid measure of self-efficacy that could predict or explain a women’s decision to have a screening mammogram (Champion et al., 2005). This scale, which includes items such as: “I can arrange transportation to get a mammogram,” measures a woman’s confidence in her ability to obtain mammogram screening. This scale includes ten 5-point scale Likert items (1 = Strongly Disagree, 5 = Strongly Agree); the composite score has a range of 10–50. The Cronbach’s alpha for the English version of this scale in its initial testing was 0.87, indicating a good internal consistency (Champion et al., 2005). The Spanish translation of the
July/August 2015
MSSES was validated by the authors in a previous study, with a Cronbach’s alpha of 0.94 (JeromeD’Emilia, Suplee, & Akincigil, unpublished data). This scale has been validated in various and diverse populations (Champion et al., 2005; Secginli, 2012). The demographics portion of the survey was translated from English to Spanish by a professional translator and translated back into English by a Spanish speaking research assistant to ensure readability, appropriateness, and conceptual equivalence with the English version. The three short surveys combined took the women approximately 15 min to complete.
Analytical strategy Three outcome variables were identified. The first outcome variable was the continuous measure of general self-efficacy, the second was the continuous measure of mammography specific self-efficacy, and the third was the binary ‘having ever had’ or ‘not having ever had a mammogram.’ Data were analyzed with the STATA-MP 12 program (StataCorp, 2011). Univariate statistics were generated to describe the sample demographics. Bivariate analyses were performed using t tests for continuous variables and chi-squares for categorical variables to identify differences between the two groups of women in variables such as marital status, education, country of origin, acculturation and insurance, factors that may potentially be associated with mammography screening. Total general and mammogram specific self-efficacy were compared to identify any differences that may be associated with ethnicity. T tests and chi-square were used to test the associations of demographic variables, acculturation, ethnicity, and total general self-efficacy and mammogram specific self-efficacy on the binary, ‘having ever had a mammogram’. Logistic regression was used to identify a possible association between the self-efficacy measures and the binary “ever had a mammogram”. Covariates found to have a significant association with mammogram screening from bivariate analyses were entered into a logistic regression model to test for the effect of general and mammogram specific self-efficacy on the dependent variable having ever had a mammogram. An estimate of the area under the receiver operating characteristic (ROC) curve was derived to assess the predictive value of the self-efficacy model.
Jerome-D’Emilia and Suplee: Mammogram Use and Self-Efficacy
Results Of the 139 women who completed the survey, 70 were African American and 69 Hispanic. None of the African American women self-identified as having Hispanic ethnicity, leaving the two groups divided by ethnicity. The two samples were significantly different in age (p = 00014). The majority (62%) of the Hispanic women were born in Puerto Rico, 17% were born in the Dominican Republic, 14% in the continental United States, and 4% in Mexico. All but one African American woman was born in the United States. African American women were more likely to be insured; the difference was statistically significant (p = .003). While 40% of the
291
sample reported having private insurance, the majority of women were either uninsured (18%), or relied on public programs: 21% received Medicaid and 20% (primarily the older African American women) reported receiving Medicare (See Table 1 for sample characteristics). Self-efficacy as measured by the GSES varied just slightly by ethnicity, and Hispanic women had a marginally higher composite mean score. The only item that was significantly different by ethnicity was “I can solve most problems if I invest the necessary effort,” for which the African American women had a higher mean score (p = .01). MSSES varied between these two racial/ethnic groups, but not significantly. African American women had
TABLE 1. Sample Characteristics of Survey Respondents by Race/Ethnicity (N = 139) Hispanic women (N = 69) %a
n
%a
p-value
37 20 9 3
53.6 28.9 13.0 4.3
28 15 12 15
40.0 21.4 17.1 21.4
.0014*
7 62
1.0 89.8
16 52
5.4 94.6
.036*
21 27 5 14
31.3 40.3 7.4 20.9
23 22 8 17
32.8 31.4 11.4 24.2
.678
31 15 6 14
46.9 22.7 9.0 21.2
22 13 21 10
33.3 19.6 31.8 15.1
.003*
10 3 43 12 1
14.4 4.3 62.3 17.3 1.4
69 0 0 0 1
98.5 – – – 1.4
.000*
1 8 37 23 41.3 31.3
1.4 11.5 53.6 33.3
1 0 0 69
1.4 – – 98.5 42.9 31.2
.000*
n Age 40–50 51–60 61–70 71+ Education 5 years 5–10 years 11–20 years All my life Mean total MSSES Mean total GSES
African American women (N = 70)
Note. aPercentages are based on nonmissing values. *Significant p < .05.
.125 .537
292
Public Health Nursing
Volume 32
Number 4
higher mean values for all but one of the items on the MSSES (“I can arrange other things in my life to have a mammogram”). Three of the items had significantly higher means for the African American women: “I can talk to people at the mammogram center about my concerns” (p = .02); “I can find a way to pay for a mammogram” (p = .04); and “I know for sure I can get a mammogram if I really want to” (p = .009). Eighty-nine percent of women in this sample reported ever having had a mammogram; 91% of African American women and 88% of Hispanic women. There was no significant difference in screening based on ethnicity. When comparing the women who reported ever having had a mammo-
July/August 2015
gram with those who had not, screened women were older (55.7 as compared to 45.9) than the women who had never been screened, but this difference was not statistically significant. Screened women were significantly more likely to be insured (p = .027). There was a significant difference in MSSES scores with both screened women (p < .001) and insured women (p < .001) scoring higher. No significant differences were noted in screening behavior based on insurance type (private or public) (See Table 2). Variables were added in a stepwise fashion to model the probability that a woman had ever had a mammogram. Ethnicity was not entered into the final model, as only a marginal difference in screen-
TABLE 2. Demographic Characteristics of Survey Respondents Screened and Nonscreened (N = 139) Screened women (N = 125) Age 40–50 51–60 61–70 71+ Education
POPULATIONS
AT
RISK ACROSS
THE
LIFESPAN: POPULATION STUDIES
Mammogram Use and Self-Efficacy in an Urban Minority Population Bonnie Jerome-D’Emilia, PhD, MPH, RN, and Patricia Dunphy Suplee, PhD, RNC-OB School of Nursing – Camden, Rutgers, The State University of New Jersey, Camden, New Jersey Correspondence to: Patricia Dunphy Suplee, School of Nursing – Camden, Rutgers, The State University of New Jersey, 311 N. 5th St., Camden, NJ 08102. E-mail: [email protected]
ABSTRACT Objectives: (1) To compare a sample of low-income African American and Hispanic women in general and mammogram specific self-efficacy and other factors potentially associated with screening to identify any differences related to ethnicity and in the use of mammogram screening; and (2) to examine the association of general self-efficacy and mammography specific self-efficacy and mammogram screening in these two ethnically different groups of women. Design and Sample: Cross-sectional. A convenience sample of 139 women. Measures: General and mammogram specific self-efficacy, having ever had a mammogram, acculturation, and demographics. Results: Mammogram specific self-efficacy was significantly associated with having had a mammogram (p < .001), as was insurance status (p = .027). Using logistic regression, older women (OR: 1.3) and those with insurance (OR: 4.8) were more likely to have been screened. When mammogram specific self-efficacy was added to the model, overlap between this construct and insurance prevented insurance from reaching significance. Conclusions: An association between insurance status and mammogram specific self-efficacy was found. It is likely that mammogram specific self-efficacy will vary with mammogram adherence and insurance status, rather than predict screening. General self-efficacy, higher in screened women, may be an effective mediator through which to develop interventions to increase preventive health-seeking behaviors. Key words: African Americans, Hispanics, mammography, self-efficacy.
Background People who regard themselves as highly efficacious act, think, and feel differently from those who perceive themselves as inefficacious. They produce their own future, rather than simply foretell it. (Bandura,1986)
Breast cancer is the second greatest cause of cancer mortality for women, exceeded only by lung cancer (American Cancer Society [ACS], 2014). An estimated 232,670 new cases of invasive breast cancer in women are expected in the United States in 2014 (ACS, 2014). Although non-Hispanic White women have a higher incidence of breast cancer than African American women, African American women have a higher death rate from breast cancer across all age ranges (ACS, 2013a). Hispanic
women are less likely to be diagnosed with breast cancer than White or African American women, yet breast cancer is the most frequently diagnosed cancer and the greatest cause of cancer death in Hispanic women (ACS, 2013b). Hispanic and African American women are more often diagnosed at later stages and with negative prognostic features leading to the increased mortality rate noted in these populations (DeSantis, Ma, Bryan, & Jemal, 2014). Lower income women have a significantly greater risk of diagnosis at a later stage and dying of breast cancer than do higher income women (Sprague et al., 2011; Vona-Davis & Rose, 2009). Despite widespread availability of mammography services, women with less education and lower income, the uninsured and recent immigrants all
287
288
Public Health Nursing
Volume 32
Number 4
have lower rates of mammography screening than found in the middle-class non-Hispanic White or African American populations (Kaiser Family Foundation, 2014; Sprague et al., 2011). According to the 2012 National Health Interview Survey, 66.5% of women 40 and older of all races and ethnicities reported having had a mammogram in the last 2 years. When race and ethnicity were considered, rates were fairly consistent with 67.4% of non-Hispanic White women, 67.9% of African American women and 64.2% of Hispanic women 40 and older reporting having had a mammogram in the past 2 years (National Center for Health Statistics, 2014). However, only 36% of uninsured women 40 and older and 53% of women with less than a high school degree reported having had a mammogram in the past 2 years (National Center for Health Statistics, 2014). In addition to various demographic variables, such as lack of education, studies have found insurance status to be the primary predictor of cancer screening across ethnicities (Henry et al., 2011; Nuno, Castle, Harris, Estrada, & Garcia, 2011). Selvin and Brett (2003) found that regardless of race/ ethnicity, maintaining a usual source of care was the strongest predictor of mammography screening. Jerome-D’Emilia (2014), in a meta-analysis of barriers and facilitators to screening in the Latina population, found that for low-income Hispanic women socioeconomic barriers were the most significant factors determining a woman’s likelihood of being screened. This study concluded that, excluding language barriers, poor women, whether Hispanic, African American, or White, seem to experience the same barriers and facilitators to mammogram screening. Removing financial and access barriers to screening has not been effective in increasing screening rates in low-income women (Teran, Baezconde-Garbanati, Marquez, Castellanos, & Belkic, 2007). As the country heads toward full implementation of the Affordable Care Act (ACA), an estimated 6.8 million low-income women will gain access to insurance and nearly all insured women will be entitled to screening mammograms without cost sharing (Levy, Bruen, & Ku, 2012). According to a recent Centers for Disease Control update of breast cancer screening services (considering mammogram utilization following full implementation of the ACA), barriers such as a women’s low self-effi-
July/August 2015
cacy toward breast cancer screening will likely continue to impede appropriate utilization of screening services (Centers for Disease Control and Prevention [CDC], 2014). To improve rates of mammography and cancer screening in general, it is imperative to examine factors such as self-efficacy to understand how they may influence a women’s utilization of screening services. Self-efficacy refers to the confidence a person feels about taking a certain action, including the confidence to overcome barriers that may stand in the way of taking that action. A higher level of selfefficacy may be what allows some women to access preventive care, while many others with the same socioeconomic conditions and facing the same potential barriers do not. Derived from Social Cognitive Theory, a theory that explains how people develop certain behavioral patterns (Bandura, 1997), general self-efficacy has been widely applied to predict and explain health-seeking behavior and has been found to be associated with positive health outcomes such as: more frequent use of active coping (Luszczynska, Scholz, & Schwarzer, 2005); self-management of chronic illness and perception of wellness (Joekes, Van, & Schreurs, 2007); and health-related quality of life (Kvarme, Haraldstad, Helseth, Sørum, & Natvig, 2009). The construct of self-efficacy has been measured in many and varied ways when used to examine the influence of self-efficacy on utilization of cancer screening services. Schwarzer and Jerusalem (1995) developed the Generalized Self-Efficacy Scale (GSES) to measure the level of perceived general self-efficacy. Although this tool has been used effectively to measure confidence in goal setting, effort, and persistence, other studies have opted to construct different measures of self-efficacy more specific to the behavior in question. Cronan and colleagues, in a survey of African American, Mexican American and White women, measured self-efficacy with three 10-point response scaled questions that assessed a women’s confidence that mammography can detect cancer, prevent cancer, and allow early diagnosed cancer to be cured (Cronan et al., 2008). This measure of self-efficacy was a statistically significant predictor of mammogram screening in the White and Mexican American samples, but not for the African American women. Another study, Stewart, Rakowski, and Pasick (2009) measured self-efficacy with one yes or no
Jerome-D’Emilia and Suplee: Mammogram Use and Self-Efficacy survey question, “Do you think that you could get a mammogram every year?” This operationalization of self-efficacy was able to predict mammogram use significantly only in White women. The association between self-efficacy and mammogram utilization in non-White samples of women is less clear. Champion, Skinner, and Menon (2005) developed and tested the validity of a self-efficacy scale specific to mammography. This scale, which includes items such as: “I can arrange transportation to get a mammogram,” measures a woman’s perceived confidence in her ability to obtain mammogram screening (Champion et al., 2005). Using this mammogram specific scale, these researchers found that individuals with higher mammogram specific self-efficacy were more likely to express the intention and follow through by obtaining a mammogram. In the initial testing of this tool, the total selfefficacy scores increased over time in the adherent group of women. This could reflect the fact that the more adherent a woman is to screening, the higher her level of mammogram specific self-efficacy. It could also be construed as evidence that these questions are too specific to the act of getting a mammogram, so that the self-efficacy score goes up due to familiarity with the screening facility and process. In that case, the self-efficacy score can be considered more of a proxy for adherence than a predictor for mammogram screening. The GSES may provide a more reliable and stable measure of mammography use. Champion et al. (2008) state that the need to identify valid and reliable measures of women’s health-seeking behavior that are specific to race and ethnicity is essential to eliminate disparities. It is not clear if the construct measures of self-efficacy are valid and reliable across ethnicities (Burke et al., 2009).
Purpose The overall aims of this study were to: (1) compare low-income African American and Hispanic women in terms of socioeconomic status, general and mammogram specific self-efficacy, and other factors that are potentially associated with screening to see if there are any differences related to ethnicity which may affect the likelihood of using mammogram screening; and (2) examine the association of general self-efficacy and mammography specific
289
self-efficacy and mammogram screening in this sample of women to see which measure may be the better predictor of mammogram screening in this population.
Methods Design and sample This study used a descriptive cross-sectional research design. Originally, the researchers identified a Catholic church in a primarily Hispanic neighborhood and a Baptist church with a primarily African American congregation, both in a lowincome, underserved urban area of New Jersey, as the recruitment sites. These churches were chosen as study sites due to the race or ethnicity of their populations and their clergy members’ willingness to participate. The eligibility criteria for participation was literacy in English or Spanish and age 40 and older (the ACS recommends that screening mammography begin at 40 years of age [ACS, 2013c]). In early 2012, all self-reported Hispanic and African American women attending Sunday services at both churches in a low-income urban area of NJ were approached by the researchers (with the assistance of one African American and one bilingual Hispanic nursing student) and recruited to participate in the study. The sample size was not predetermined, and all eligible women were asked to participate. However, the sample of 139 was adequate to determine whether those women with high mammogram specific self-efficacy (above the median) would be more likely to have ever had a mammogram than participants with low mammogram specific self-efficacy, with 80% power to detect a difference as low as 21% between the groups to be statistically significant at the 0.05 level. Women who were eligible and agreed to participate signed informed consents (available in Spanish and English), completed the surveys, and received five dollar gift cards for their effort. Data collection took place on church grounds. This study was approved by the Institutional Review Board of Rutgers University. The population of Hispanic women at the Catholic Church was a younger population of women than found at the Baptist church. This is consistent with the finding by Lugo, Smith, Cox, and Pond (2008) that Hispanics comprise almost half of all
290
Public Health Nursing
Volume 32
Number 4
Catholics in the United States under age 40. Of the 41 Hispanic women who attended the services on the day of data collection, only 15 (37.5%) of the women were 40 and over. Therefore, given the researchers’ goal of recruiting relatively equal samples of African American and Hispanic women, the decision was made to recruit women attending a primarily Hispanic community center, following the same procedure for consent and completion of the self-reported survey.
Measures Women were asked to complete the General SelfEfficacy Scale (GSES; Schwarzer & Jerusalem, 1995), the Mammography Specific Self-Efficacy Scale (MSSES; Champion et al., 2005), and a 10-item researcher created demographic survey. All of the surveys were available in Spanish and English. Acculturation was measured by the proxy “How many years have you lived in the United States,” which has been found to be highly correlated with other, more detailed measurements of acculturation (Brown, Consedine, & Magai, 2006). The GSES consists of ten items with a 4-point Likert type scale (1 = Not at all true, 4 = Exactly true) which take the general form: “I have the capacity to do X even if Y (barrier) occurs” and yields a final composite score with a range of 10–40. This scale has been translated and used in many and varied studies in 23 countries, with Cronbach’s alpha ranging from 0.76 to 0.90, with the majority in the high 0.80s (Finlayson, Edwards, & Courtney, 2011; Kvarme et al., 2009). Criterion-related validity has been documented in numerous correlational studies (Carlsson, Bjorvatn, Engebretsen, Berglund, & Natvig, 2004; Joekes et al., 2007). The MSSES was developed to provide a reliable and valid measure of self-efficacy that could predict or explain a women’s decision to have a screening mammogram (Champion et al., 2005). This scale, which includes items such as: “I can arrange transportation to get a mammogram,” measures a woman’s confidence in her ability to obtain mammogram screening. This scale includes ten 5-point scale Likert items (1 = Strongly Disagree, 5 = Strongly Agree); the composite score has a range of 10–50. The Cronbach’s alpha for the English version of this scale in its initial testing was 0.87, indicating a good internal consistency (Champion et al., 2005). The Spanish translation of the
July/August 2015
MSSES was validated by the authors in a previous study, with a Cronbach’s alpha of 0.94 (JeromeD’Emilia, Suplee, & Akincigil, unpublished data). This scale has been validated in various and diverse populations (Champion et al., 2005; Secginli, 2012). The demographics portion of the survey was translated from English to Spanish by a professional translator and translated back into English by a Spanish speaking research assistant to ensure readability, appropriateness, and conceptual equivalence with the English version. The three short surveys combined took the women approximately 15 min to complete.
Analytical strategy Three outcome variables were identified. The first outcome variable was the continuous measure of general self-efficacy, the second was the continuous measure of mammography specific self-efficacy, and the third was the binary ‘having ever had’ or ‘not having ever had a mammogram.’ Data were analyzed with the STATA-MP 12 program (StataCorp, 2011). Univariate statistics were generated to describe the sample demographics. Bivariate analyses were performed using t tests for continuous variables and chi-squares for categorical variables to identify differences between the two groups of women in variables such as marital status, education, country of origin, acculturation and insurance, factors that may potentially be associated with mammography screening. Total general and mammogram specific self-efficacy were compared to identify any differences that may be associated with ethnicity. T tests and chi-square were used to test the associations of demographic variables, acculturation, ethnicity, and total general self-efficacy and mammogram specific self-efficacy on the binary, ‘having ever had a mammogram’. Logistic regression was used to identify a possible association between the self-efficacy measures and the binary “ever had a mammogram”. Covariates found to have a significant association with mammogram screening from bivariate analyses were entered into a logistic regression model to test for the effect of general and mammogram specific self-efficacy on the dependent variable having ever had a mammogram. An estimate of the area under the receiver operating characteristic (ROC) curve was derived to assess the predictive value of the self-efficacy model.
Jerome-D’Emilia and Suplee: Mammogram Use and Self-Efficacy
Results Of the 139 women who completed the survey, 70 were African American and 69 Hispanic. None of the African American women self-identified as having Hispanic ethnicity, leaving the two groups divided by ethnicity. The two samples were significantly different in age (p = 00014). The majority (62%) of the Hispanic women were born in Puerto Rico, 17% were born in the Dominican Republic, 14% in the continental United States, and 4% in Mexico. All but one African American woman was born in the United States. African American women were more likely to be insured; the difference was statistically significant (p = .003). While 40% of the
291
sample reported having private insurance, the majority of women were either uninsured (18%), or relied on public programs: 21% received Medicaid and 20% (primarily the older African American women) reported receiving Medicare (See Table 1 for sample characteristics). Self-efficacy as measured by the GSES varied just slightly by ethnicity, and Hispanic women had a marginally higher composite mean score. The only item that was significantly different by ethnicity was “I can solve most problems if I invest the necessary effort,” for which the African American women had a higher mean score (p = .01). MSSES varied between these two racial/ethnic groups, but not significantly. African American women had
TABLE 1. Sample Characteristics of Survey Respondents by Race/Ethnicity (N = 139) Hispanic women (N = 69) %a
n
%a
p-value
37 20 9 3
53.6 28.9 13.0 4.3
28 15 12 15
40.0 21.4 17.1 21.4
.0014*
7 62
1.0 89.8
16 52
5.4 94.6
.036*
21 27 5 14
31.3 40.3 7.4 20.9
23 22 8 17
32.8 31.4 11.4 24.2
.678
31 15 6 14
46.9 22.7 9.0 21.2
22 13 21 10
33.3 19.6 31.8 15.1
.003*
10 3 43 12 1
14.4 4.3 62.3 17.3 1.4
69 0 0 0 1
98.5 – – – 1.4
.000*
1 8 37 23 41.3 31.3
1.4 11.5 53.6 33.3
1 0 0 69
1.4 – – 98.5 42.9 31.2
.000*
n Age 40–50 51–60 61–70 71+ Education 5 years 5–10 years 11–20 years All my life Mean total MSSES Mean total GSES
African American women (N = 70)
Note. aPercentages are based on nonmissing values. *Significant p < .05.
.125 .537
292
Public Health Nursing
Volume 32
Number 4
higher mean values for all but one of the items on the MSSES (“I can arrange other things in my life to have a mammogram”). Three of the items had significantly higher means for the African American women: “I can talk to people at the mammogram center about my concerns” (p = .02); “I can find a way to pay for a mammogram” (p = .04); and “I know for sure I can get a mammogram if I really want to” (p = .009). Eighty-nine percent of women in this sample reported ever having had a mammogram; 91% of African American women and 88% of Hispanic women. There was no significant difference in screening based on ethnicity. When comparing the women who reported ever having had a mammo-
July/August 2015
gram with those who had not, screened women were older (55.7 as compared to 45.9) than the women who had never been screened, but this difference was not statistically significant. Screened women were significantly more likely to be insured (p = .027). There was a significant difference in MSSES scores with both screened women (p < .001) and insured women (p < .001) scoring higher. No significant differences were noted in screening behavior based on insurance type (private or public) (See Table 2). Variables were added in a stepwise fashion to model the probability that a woman had ever had a mammogram. Ethnicity was not entered into the final model, as only a marginal difference in screen-
TABLE 2. Demographic Characteristics of Survey Respondents Screened and Nonscreened (N = 139) Screened women (N = 125) Age 40–50 51–60 61–70 71+ Education
