Abstract
Introduction:
Fear has been described as potentially important in affecting breast cancer screening completion. Limited information is available on the prevalence and determinants of fear among Mexican American women. This study describes perceived breast cancer fear and its association with personal characteristics and screening behavior among Mexican American women.
Methods:
This is a secondary analysis of data collected during the implementation of the Breast Cancer Education, Screening and NavigaTion program among eligible uninsured women in two Texas border counties. Participants completed a 26-item survey to assess eligibility, risk status, prior screening, and breast cancer fear. Descriptive statistics and multivariable analyses were used to determine associations between the fear score, personal characteristics, and mammography screening.
Results:
In all, 1916 of 2012 eligible women completed the study. The mean age was 57.3 years, 99.2% were Hispanic, and 88% were born in Mexico; 15% had a family history of breast cancer and 14% had never had a mammogram. The mean breast cancer fear score was 25.5 (standard deviation: 10.52; range: 8-40); 54.0 % (95% confidence interval: 52.1%-56.7%) had a high fear level. In multivariable analyses, better health status (P < .001), older age (P = .039), birth in the United States (P = .020), and having a regular doctor (P = .056) were associated with lower fear scores. There was no association between breast cancer fear and mammography screening.
Conclusion:
Breast cancer fear is high and varies by personal characteristics and health status among uninsured Mexican American border-residing women due for screening, but is not associated with screening behavior. Further research is needed to clarify the effect of interventions designed to help reduce breast cancer fear among these women, including educational interventions to reduce breast cancer fear.
Introduction
Fear was defined by Witte 1 as “a negatively toned emotion accompanied by a high level of physiologic arousal stimulated by a threat that is perceived to be significant and personally relevant.” During a person’s lifetime, each individual perceives different levels of health threat to falling ill, but the diagnosis of cancer could be considered one of the most frightening diagnoses for anyone. Cancer fear is a negative emotional reaction to the threat of cancer that has implications for both quality and duration of life. 2 Previous reports have shown that around 64% of the adult population report being afraid or worried about getting a cancer diagnosis. 3 The fear of cancer can vary by personal characteristics: higher cancer fear levels have been associated with lower education, poor health, and ethnic minority status2,4 Breast cancer fear has been reported to be higher in women with hereditary risk factors.2,5
In the United States, 1 out of every 8 women will be diagnosed with breast cancer in their lifetime, and approximately 276 480 new cases of breast cancer are expected to be diagnosed in US women in 2020. 6 In 2016, there were an estimated 3 477 866 women living with breast cancer in the United States. Furthermore, breast cancer is the second commonest cause of overall cancer deaths in women and the leading cancer killer among Hispanic women. 7 There is evidence that breast cancer screening reduces mortality by 15% to 54% through early detection when treatment is more successful. 8 However, a significant number of women are not screening according to national recommendations. Mammography screening rates in the United States vary by race and ethnicity; according to the National Health Interview Survey 2018, the screening rate among Hispanic women aged 45 years or older was 60% compared with 64% among non-Hispanic whites. Hispanic women are also less likely to be diagnosed with local-stage disease (58%) compared with non-Hispanic white women (66%). 7
Although cancer fear has been shown to affect screening health behaviors, studies have yielded contradictory results; it is not clear whether negative emotions, such as fear, act as a barrier or facilitator of mammography screening. 9 Various studies have indicated that fear, which is often linked to uncertainty, naturally causes a person not to act or induce themselves to make proper decisions and to avoid huge risks as well as potential threats because of fear that the screening process could reveal that they have cancer, and most of them desired not knowing.5,10-12 Fear has also been associated with other negative effects in women’s lives. Cancer worries have been reported to be higher among women at higher risk of breast cancer because of family history, 5 and the level of worry or fear interfered with their daily life functioning especially among women who required follow-up testing after screening. 13 Although information about how breast cancer fear affects racial/ethnic groups is scarce, several researchers have studied fear and anxiety related to breast cancer screening among Hispanic women. They have reported that fear of cancer and their fatalistic cancer perception are a barrier to screening in addition to linguistic barriers and culturally based embarrassment. 14 Similarly, others note that the main reason why many Hispanic women avoid mammography is because of fear, busy schedules, and also feeling uncomfortable during the screening procedure. 15 In contrast, various studies have indicated that fear can motivate routine screening for cancer and that worry about breast cancer risk appears to be associated with mammography use in an inverted u-shaped pattern in which most of the women who reported moderate levels of worry were more likely to use mammography annually than those who were either mildly or severely worried. 16 Another study revealed that self-regulation and cancer worry were both positively associated with mammography screening and clinical breast examination frequency. 12 In addition, some studies have indicated that fear does not affect the frequency of screening for breast cancer. According to Ghahramanian et al, 17 although fear had a stimulating effect on breast cancer screening performance, the association was not significant. They argued that the frequency of breast cancer screening or the behaviors related to the process were affected by family status and lifestyle. 17 The purpose of this study was to expand our knowledge about breast cancer fear among Hispanic women. The main aim was to estimate the prevalence of fear in a predominantly Hispanic underinsured population who were due for screening. We also aimed to determine sociodemographic factors associated with the fear score. In the exploratory analysis, the role of levels of fear as a predictor of future mammographic screening was also examined.
Methods
Sample
A secondary data analysis of baseline data collected during a breast cancer screening intervention called the Breast Cancer Education, Screening and NavigaTion (BEST) program was conducted. This funded program provided outreach, education, navigation, and no-cost screening and diagnostic testing to uninsured women. The BEST program was implemented in a 2-county area between June 2014 and May 2017 by 2 program community health workers. Recruitment occurred at 61 participating community sites such as community centers, churches, food pantries, and health fairs. Eligibility criteria for the study included women who were aged 50 to 75 years, uninsured, who had not had a mammogram within the past 2 years, and self-reported a Texas address. The exclusion criterion for this study was a previous diagnosis of breast cancer.
Of the total 2115 women approached, 103 were not eligible for screening because of age (91 were aged <50 years and 12 were aged >75 years); of the remaining 2012 women, 96 did not complete all items in the fear score scale and were excluded. Thus, 1916 participants were included in these analyses. The study was approved by the Institutional Review Board (IRB) of Texas Tech University Health Sciences Center El Paso (IRB Protocol No. E17066).
Measures
Survey items were available in English and Spanish and were collected in-person by a bilingual community health worker during BEST recruitment. Survey items covered age, race and ethnicity, years of education, country of birth, marital status, preferred language, work status, whether they have a regular doctor, personal and family history of cancer, and self-reported health status. The perception of health was assessed with a self-reported response to the question, “Would you say that for someone of your age your health is?” (excellent; very good; good; fair; poor). The self-reported health status items have been validated as a measure of chronic health and have been used as an item in the national Behavioral Risk Factor Surveillance System since 198418,19.
Breast cancer fear was assessed with the validated Champion Breast Cancer Fear Scale. These questions included 8 items covering the emotional responses when women thought of breast cancer, such as “When I think about breast cancer, my heart beats faster,” “When I think about breast cancer, I feel nervous,” and “The thought of breast cancer scares me.” Response categories were scored on a 5-item Likert-type scale and ranged from strongly agree (5) to strongly disagree (1). Women with a higher score feel more fear about breast cancer. The scale had a good level of internal reliability reported with a Cronbach coefficient α = 0.91. 20 Postintervention follow-up screening was determined by documented completion of screening mammography in the BEST program database.
Statistical Analysis
Quantitative variables were described using mean and standard deviation (SD). Categorical variables were described using frequencies and proportions. Scoring of the fear variable was done by coding the 8 components that made up the fear score: scared, nervous, upset, depressed, jittery, heartbeat, uneasy, and anxious. Strongly agree = 5, agree = 4, undecided = 3, disagree = 2, and strongly disagree = 1. These components were further summed with a total possible score of 8 to 40. We assessed the reliability of the fear instrument on our data set using a Cronbach α. As the reliability of the fear instrument was very high in our population, we conducted all analyses based on a composite fear score. However, the distributional differences of individual items were also reported using descriptive data analysis. The composite fear score was calculated using 3 categories (strongly agree/agree, strongly disagree/disagree, and undecided). The fear score was also categorized into low, moderate, or high fear categories. Low fear was defined as a total score of 8 to 15; moderate fear, as a score of 16 to 23; and high fear, as a score of 24 to 40. 20 Univariate and multiple linear regression analysis was conducted to examine the relationship between baseline variables and the total fear score as a dependent variable. In the main analysis, a simple ordinary linear regression was used to examine the unadjusted relationship between the baseline cofactors and the total fear score as the dependent variable. A univariate linear regression coefficient (RC) along with the 95% confidence interval (CI) was reported in addition to the P value. Statistical significance was considered for P values less than 5%. Variables that were significant in the unadjusted analysis at a level of significance of 20% were included in the multivariable analysis. In addition, any clinical significant variables regardless of the significance level were also included in the final multivariable model.
To explore the relationship between fear and screening completion, we performed logistic regression analyses. In these analyses, we used the categorical fear levels of low, moderate, or high as the independent predictor for follow-up screening (with low level of fear as a reference category), and follow-up screening was considered a dichotomized variable with mammography screening completed as the dependent variable. Results are summarized by odds ratio, 95% CI, and P value.
Prior to developing regression models, assumptions were assessed, including multicollinearity using variance inflation criteria (VIF). A variable with VIF >5 indicates a potential presence of collinearity of that variable with other variables in the model. We have used multiple model fit criteria such as negative log likelihood, Akaike information criteria (AIC), and Bayesian information criteria (BIC) to evaluate the goodness of fit of the developed multivariable linear model. A higher value of these indices indicates an improvement in the model fit; SPSS version 25.0 was used for all analyses.
Results
Table 1 shows the demographic characteristics of the participants in this study. A total of 1916 women were included in the analysis. The mean age of participants was 57.3 years (SD: 5.42), with an average of 23 years of living in the United States. Overall, 99.2% (n = 1901) of the women were Hispanic and 93.3% (n = 1788) stated a Spanish language preference; 87.9% (n = 1684) of the respondents were born in Mexico. Almost half (53.5%, n = 1026) were either married or living with a partner. The majority were not working (63.3%, n = 1213) and had less than high school education (83.9%, n = 1607). Regarding their health, the majority (94.8%, n = 1816) reported not having a regular doctor, whereas 1113 (58.1%) self-reported their health as poor or fair. Table 1 also shows that 15.5% (n = 297) had a family history of breast cancer, and among 40% (n = 119) of these, the diagnosis was made before the age of 40 years. About half of them (52%, n = 994) reported completing a mammogram more than 3 years previously, and 14.2% (n = 272) had never had a mammogram.
Demographic characteristics (N = 1916).
Abbreviations: BEST, Breast Cancer Education, Screening and Navigation.
The mean fear score was 25.54 (SD: 10.52); the lowest fear score recorded was 8 and the highest was 40. The prevalence of a high fear level was 54% (n = 1042) (95% CI: 52.1%-56.7) and a moderate fear level was 35.7% (n = 684) (95% CI: 33.5%-37.9%). Overall, about half of the women agreed or strongly agreed with each fear statement. The most prevalent item was being scared (60.1%, n = 1152), followed by feeling nervous (58.1%, n = 1113), uneasy (54%, n = 1034), anxious (53.3%, n = 1021), and depressed (52.8%, n = 1012) when they think of breast cancer. More than a third (39%, n = 749) reported not being scared when they thought about breast cancer. The Cronbach α reliability coefficient of the Fear Scale in this population was 0.978 (Table 2).
Breast cancer fear scale response distribution.
Abbreviations: A, agree; CI, confidence interval; D, disagree; SA, strongly agree; SD, strongly disagree.
Table 3 shows the association between baseline factors and the average fear score. In the univariate regression analysis, the following variables were associated with a lower fear score: higher education (RC =−1.59, P = .015), working (RC =−1.26, P = .011), birth in the United States (RC = −2.16, P = .020), self-reported better health (RC = −1.86, P < .001), and having a regular doctor (RC =−2.28, P = .035). Although not statistically significant, subjects with a family history of breast cancer also reported a lower fear score (RC =−1.09, P = .09).
Linear regression of baseline factors and fear score.
Only 5 participants reported a previous abnormal mammogram and it was removed from the analysis. All bold values are statistically significant p values.
Abbreviations: 1.0, reference category; CI, confidence interval; RC, regression coefficient.
Baseline factors that showed a value of P ⩽ .20 or if a plausible association was considered were included in the multivariable analysis; this made no difference to effect sizes associated with cancer fear except for higher education that became nonstatistically significant and older age being associated with a reduced fear score (RC = −0.094, P = .039; Table 4).
Multivariable linear regression of baseline factors and fear score.
All bold values are statistically significant p values.
Abbreviations: 1.0, reference category; CI, confidence interval; RC, regression coefficient.
We did not find any sign of multicollinearity. The developed model showed improved overall fit compared with the model with individual predictors. A significant value of the omnibus test (log likelihood ratio test) also indicated a significant fit of the developed model compared with the null model (Table 4).
For the exploratory analysis, a total of 1583 (82.6%) participants completed follow-up mammography screening as part of the BEST program. The logistic regression analysis did not show a significant association between levels of fear (low, moderate and high) and the follow-up screening completion (Table 5).
Logistic regression of future mammogram completed (dependent) and fear score levels (independent) (N = 1916).
Discussion
The results indicate that despite numerous public health interventions related to cancer prevention and early detection, negative emotions about breast cancer persist in Hispanic women along the border. In this population, more than half of the women indicated that they were fearful of cancer (60%, n = 1152). Furthermore, the mean score for fear of breast cancer (25.5 [SD: 10.5]) was within the higher fear category previously defined by Champion et al 20 and was higher than that reported among non-Hispanic women (21.1 [SD: 8.5]) 20 and similar to that reported by women in Turkey (26.36 [SD: 7.29]) 21 when using the same fear scale. Participants in Champion’s study were mainly African American and Caucasian women, with older average age (66 years vs 57 years) who received health services from a Health Maintenance Organization or were seen at an indigent clinic. The main difference between the women in the Turkish study and our population is the access to health care and socioeconomic status. Participants were younger (47 years vs 57 years in this study) and had a higher socioeconomic status, with 93% reporting their income in the middle or higher level, compared with our population who had >90% working less than full-time and >80% having less than a high school education. Most of the Turkish participants had health insurance (75%) compared with 100% with no insurance, and nearly 95% reported no regular doctor in this sample. In a meta-analysis that included 10 studies and 3342 women, the mean score for breast cancer fear or worry was also in the lower third of the score range, although it was measured using a different tool. 9
Little is known about the distribution of fear as a single component across demographic characteristics. In our population, higher cancer fear scores were associated with lower educational attainment, unemployment, younger age, and poor self-reported health. These findings are consistent with the relationship between lower educational attainment and higher cancer fear reported for prostate cancer 22 and any cancer in general. 4 In our study of the participants aged 50 years, only 31 (17.6%) reported never having a previous mammogram, and of these, 8 reported their doctor had recommended a mammogram but did not seek a mammogram until this program reached them, so only 2.5% (n = 23) had never previously been advised to get a mammogram because of their age. Similar findings about the association of age and health status were reported in a population from Spain that included adults over the age of 50 years. In that study, higher scores of cancer worry were associated with younger age and worse self-perceived health. 2 In a different study that included UK women and men, the prevalence of “cancer as greatest health fear” was 59%. They showed that higher fear of cancer was associated with respondents with lower education and ethnic minority backgrounds. No association was found with age. In addition, respondents with higher general anxiety were more likely to have cancer as the greatest health fear. However, the authors assessed cancer fear using a different method by selecting 3 questions from a Cancer Attitude Inventory. 4 On the contrary, other authors did not observe any significant association between breast cancer fear scores and self-reported education, marital status, health insurance, 23 or health status. 4
Nonetheless, most of the findings suggest that probably focusing public health interventions on young women with lower education, higher anxiety levels, and low self-perceived health status might reduce the fear of breast cancer.
The relationship between breast cancer fear and screening completion is complicated and difficult to clarify. On the one hand, a positive association between breast cancer fear and previous mammography completion has been reported 12 ; on the other hand, studies report breast cancer fear has a nonsignificant effect on mammogram completion. 17 Also, a previous report showed that both low and high levels of fear may be detrimental to screening, whereas a moderate level of fear may encourage it. 20 In our analyses, we did not observe an association; however, although not significant, there is a suggestion that women with a moderate level of fear may have had a higher screening rate than those with a low level of fear. Comparing the effect of cancer fear on screening behavior is complex because terms are operationalized differently; often synonymous terms such as “cancer fear,” “cancer worry,” and “cancer anxiety” are used, and results vary depending on the focus of worry. The focus of fear or worries has been “general” as in the fear of being diagnosed with cancer or can be “specific,” for example, fear or worry about physical or emotional consequences (eg, pain during mammography, financial problems if diagnosed with cancer). Previous research suggests that general worries about breast cancer have been associated with improved screening, but fear of specific consequences may discourage screening. 12 In this Hispanic population, general fear for breast cancer was not related to future screening; this is similar to the findings of a recent study among British women where general cancer worry was not associated with screening, but specific fears about physical or emotional consequences were associated. 3
Our study confirms that screening behavior was not associated with a high level of fear in this study sample, although in this population the level of fear was not associated with follow-up mammography completion. It is plausible that the differences in the breast cancer fear score may be attributed to different personality types and consideration of different individual sentiments regarding perceptions of breast cancer. There are potentially many background variables that may affect the fear score, some of which may not be related to or affect breast cancer screening. Such factors may include general anxiety, 4 cultural stigma, 24 faith or religious beliefs, 25 past encounters with breast cancer with relatives or close friends, or individual perception. 26 Barriers to access due to financial challenges, lack of insurance, and immigration status might be associated with whether or not a women engages in breast screening behavior 27 as well as social cognitive factors and theoretical concepts from the Health Belief Model, such as perceived threat or harm, benefits, self-efficacy, barriers, and fatalism. 20 Champion proposed an integrated model in which the perception of higher threat and fatalism would increase fear, whereas a higher perception of benefits and self-efficacy would decrease fear with more possibilities to participate in screening when women experience a moderate level of fear. 20
The Attitude-Social influence-self-Efficacy (ASE) model is a social psychology model that suggests that the intention to perform a certain behavior predicts this behavior, and “intention” is determined by 3 important psychosocial factors: attitude, social influence, and self-efficacy. This behavioral intent–oriented model has been reported to determine health preventive behaviors such as intention of smoking cessation 28 and intention of cervical cancer screening. 29 The attitude reflects previous knowledge of the advantages and disadvantages of a particular behavior. Also, attitudes toward “health” and “illness” are shaped by cultural beliefs: a fear of “finding something wrong” has been reported as a main obstacle to screening among Hispanic women 14 as has the fear associated with the screening procedure. 12 Albert Bandura 30 defined self-efficacy as one’s ability to succeed in specific situations or accomplish a task. Higher levels of self-efficacy defined as an individual’s confidence to overcome barriers have been associated with previous breast cancer screening among Hispanic women. 31 Theoretically, lower levels of self-efficacy have been associated with greater breast cancer fear. In this study, no association was found between total fear and mammography completion, although it is possible that other social and psychological factors included in the ASE model are predictors of compliance with screening.
Strengths
This is one of the first studies to systematically describe breast cancer fear in a population of predominantly Hispanic women along the US-Mexico border. Strengths include a large sample size with countywide recruitment from predominantly community sites, which is more representative of the population and increases the external validity of the study. Implementation by bilingual staff and outreach workers who belong to the same community reduced barriers to reporting emotions. Using a validated questionnaire to collect the data makes the data more reliable and increases the internal validity.
Limitations
Limitations of the study include that although it provides much needed information about breast cancer fear, we were not able to include other psycho-cognitive mediators of mental or psychological status that can also influence screening behavior. Future studies should include these measures for a more complete understanding of the role of fear in breast cancer screening behavior.
This sample reflects the demographics and culture of the US-Mexico border region, and so generalizations to the larger population of women in the United States should be done cautiously. Further limitations relate to self-report bias and the inclusion of uninsured women who were due for breast cancer screening.
Implications/Conclusions/Further Research
The study describes that a high rate of breast cancer fear is present among Mexican American women due for breast cancer screening. It shows that greater breast cancer fear is associated with personal characteristics like younger age, being unemployed, and self-reported poor health . There is a suggestion that both too little fear and too much fear may reduce screening behavior, but this needs more study. More longitudinal research using standardized measures needs to be done to better understand both the level and type of fear and how that in turn impact subsequent screening behavior. Interventions designed to help women better manage their fear could be very helpful. Those interventions might include providing education about risk factors that can be changed and promoting early detection with regular mammography screenings.
Footnotes
Funding:
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Cancer Prevention & Research Institute of Texas (PP130068).
Declaration of conflicting interests:
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Author Contributions
SFL: Conceptualization, statistical analysis, main writer; SD: Assisted with manuscript development; AKD: Assisted with manuscript editing, analysis plan and interpretation of statistical analysis; GS: Manuscript editing; NKS: Principal investigator, conceptualization, program implementation and manuscript development.
