Breast density knowledge and awareness among Latinas in a low-resource setting

INTRODUCTION

Mammographic breast density (MBD) describes the proportion of dense fibroglandular tissue relative to fat on a mammogram. It is important for women to know about their MBD because: 1) the masking effect of dense tissue limits early detection of breast cancers on screening mammograms [1] and 2) women with dense breast tissue have a higher risk of developing breast cancer [2]. Since 2009, most states have passed MBD legislation, with the goal of informing women about their individual MBD in their mammogram results letter. In doing so, public health policy aims to enhance awareness and knowledge regarding MBD [3; 4] and to promote conversations with healthcare providers about risk-based screening [5]. In 2014, Arizona became the 15th state to enact a breast density notification law [6].

There have been various studies assessing women’s knowledge of their MBD prior to, and following this breast density notification legislation [3]. Knowledge has been studied amongst women of varying socioeconomic statuses, educational backgrounds, and ethnic groups. Findings suggest that MBD awareness increased in the U.S. as more states passed notification laws, but awareness continues to be lower among Latina women than Non-Hispanic White women and in women with lower education and household incomes [7–9]. While notification laws are a broad-based strategy for increasing MBD knowledge and awareness, written letters may not be effective in all populations; educational messages may need to be targeted to local needs. Women in low-resource settings, for example, may face multiple barriers to education through written notification alone [10–12].

The purpose of our study was to compare MBD knowledge and awareness among Latinas in a federally qualified safety-net health clinic in Maricopa County, Arizona (AZ) with a national sample of women participating in an online research panel. Identification of potential differences in MBD awareness and knowledge between Latinas nationally and those in a low resource setting can highlight potential sources of disparities and inform targeted educational strategies, including those tailored to underserved populations with limited resources.

MATERIALS AND METHODS

Study Design

Data Sources and Study Procedures

This study was conducted with approval from two IRBs, one for the data survey data from a nationally representative online panel of Latinas (NS cohort) and another for the collection and evaluation of self-reported questionnaire data from Latina women undergoing screening mammography at a federally qualified health center, FQHC [13], in Maricopa County, AZ who participated in a clinical trial evaluating 3 different approaches on receiving information on breast density.

National survey cohort (NS Cohort)

A cross-sectional national online survey was administered in 2017 to screening-age women (ages 40–74) who were randomly selected from a national panel recruited to Knowledge Networks. Briefly, eligible participants were identified by Knowledge Networks and recruited to participate via email. Written informed consent was completed online in the preferred participant language before survey completion. The survey was available in English or Spanish [14]. All data from the national survey were weighted to adjust for probability of selection into the panel and post-stratified to match known U.S. Latina population distributions between the ages of 40–74 (NS cohort), using the 2016 American Community Survey as benchmarks. Weights were balanced with respect to age, education, household income, Census region, Hispanic origin, and language proficiency. At the time of the survey, 28 states had passed MBD notification laws. The NS cohort in this study includes responses from respondents who identified as Latina.

Arizona clinical trial cohort (AZ Cohort)

Between 2016 and 2019, women ages 40–74 were recruited to enroll in a randomized clinical (RCT) comparing the impact of three notification approaches on behavioral and psychological outcomes in Latina women [15]. Recruitment and enrollment took place at a FQHC in Maricopa County, AZ. Paper surveys were self-administered, or completed with the assistance of a bilingual study coordinator, in the mammography unit at the time of enrollment [15]. Surveys were available in English or Spanish and assessed awareness of MBD, and understanding of impact of MBD on masking and risk. They repeated the MBD awareness and knowledge questions from the NS Cohort study, which also included questions regarding mammography history, education attainment, and language preference. The AZ cohort in this study includes respondents who identified as Latina.

Survey content

Awareness of breast density was assessed with the question “Have you heard of something called breast density before today?” [response options: Yes / No]. Knowledge regarding the impact of MBD on masking of cancers was assessed with the question “If a woman has dense breasts, what impact does this have on the ability of a mammogram to correctly detect cancer?”, where the response options of ‘dense breasts make it easier’; ‘do not impact the ability’; or ‘I don’t know’ are considered incorrectly, and ‘make it more difficult to see cancer on a mammogram’ as correctly identifying the impact of MBD on detecting cancer. Knowledge regarding the association of MBD with cancer risk was assessed with the question “Women who have dense breasts are at greater risk for getting breast cancer” where a response of true is considered knowledgeable and ‘False’ or ‘ I don’t know’ [incorrect]] as not being knowledgeable as to the impact of MBD on risk of BC.

Primary Outcomes

Thus, the primary outcomes of interest are analyzed as binary: MBD awareness (yes/no), knowledge of MBD masking cancers (yes/no), and able to identify that MBD increases the risk of BC (yes/no).

Sample size and Analytic Methods

The study cohort consisted of Latina respondents who completed the survey items regarding breast density awareness and knowledge (N=1327 AZ, N=152 NS). Rao-Scott chi-square tests were used to assess whether demographic characteristics of the AZ cohort differed significantly from the NS cohort. For each of the 3 primary outcomes of interest, logistic regression modeling was used to assess whether they differed significantly between the AZ and NS cohorts (exposures). These analyses were repeated for the subset of women who came from states where breast density notification legislation was in place at the time of the surveys. Analyses were performed with SAS version 9.4 SURVEY procedures (SAS Institute Inc., Cary, NC). P-values less than 0.05 were considered statistically significant.

RESULTS

The AZ cohort (N=1327) was slightly younger (median age: 46.4 vs 48.7), had less education (68.8% vs 34.8% had less than high school), more often preferred the survey written in Spanish language (92.3% vs 53.1%) and less often reported having had a prior mammogram at any time (82.9% vs 90.4%) as compared to the 152 women in the NS cohort (all p≤ 0.03). See Table 1. In the AZ cohort, 92.3% of participants requested a Spanish language version of the survey, and 65% received assistance from the bilingual study coordinator to complete the survey. nd 9

Table 1:

Demographics from Latina women in Arizona and national cohorts

	Arizona N=1327 N (%)	National N=152 Weighted N (%)	P-value (Arizona vs National)	Subset of National Cohort with BD notification Legislation1 N=92 Weighted N (%)
Age (years)
40 to <50	851 (64.1)	79.9 (52.6)	<.0001	49.7 (54.0)
50 to <60	354 (26.7)	31.7 (20.9)		17.7 (19.2)
60+	122 (9.2)	40.4 (26.6)		24.7 (26.8)
Median (range)	46.4	48.7		47.8
Education
Less than HS	908 (68.8)	52.9 (34.8)	<.0001	31.5 (34.2)
HS/GED	264 (20.0)	40.6 (26.7)		26.3 (28.6)
Some college	110 (8.3)	35.0 (23.0)		19.8 (21.5)
4-year college degree +	38 (2.9)	23.5 (15.5)		14.4 (15.7)
Language of Survey			<.0001
Spanish	1225 (92.3)	80.6 (53.1)		46.2 (50.2)
Any prior mammogram history before today	1092 (82.9)	137.4 (90.4)	0.03	84.3 (91.6)

The NS cohort was more likely to be aware of MBD than the AZ cohort (32.6% vs 20.7%, respectively, p=0.005).Among the subset of 322.5 (weighted N) women who indicated MBD awareness (n=49.5 NS cohort n=275 AZ cohort), the NS cohort also had greater understanding of the masking effect of MBD, (67.8% vs 37.0%, p=0.001) and breast cancer risk (72.2% vs 32.6%, p<0.0001). See Figure 1. In the subset of women among the NS cohort respondents from states with notification laws at the time of the survey (n=92), 31.4% were aware of MBD, and 72.0% and 74.5% of those aware of MBD reported correct knowledge about masking and risk respectively.

Figure 1:

Awareness and knowledge regarding breast density. Latina women’s awareness and knowledge regarding mammographic breast density in a local federally-qualified health center (AZ) cohort vs national (NS) cohort

Awareness was positively associated with more education, prior mammogram history, and English language. See Table 2. When adjusted for differences in education, age, language, and screening mammography history, MBD awareness was not found to be significantly different between the two cohorts (unadjusted odds ratio (OR)=1.85 [95% CI: 1.21–2.83], p=0.005; adjusted odds ratio (ORadj)=0.95 [95% CI: 0.60–1.51], p=0.83). Considering the subset of women from states with notification laws, MBD awareness was not found to differ between the AZ cohort (all pts.) to NS cohort (n=92) (ORadj=0.84, 95% CI: 0.49–1.45, p=0.54).

Table 2:

Awareness of Breast Density. All frequencies, percentages, and odds ratios are weighted.

Subgroup [Weighted N]	N (%) Aware	Unadjusted1		Adjusted2
		OR (95% CI)	P-value	OR (95% CI)	P-value
Cohort
AZ [1327.0]	275.0 (20.7%)	Reference		Reference
NS [152.0]	49.5 (32.6%)	1.85 (1.21, 2.83)	0.005	0.95 (0.60, 1.51)	0.83
Age
40–49 [930.9]	191.3 (20.5%)	Reference		Reference
50–59 [385.7]	90.6 (23.5%)	1.19 (0.89, 1.59)	0.25	0.98 (0.72, 1.34)	0.91
60+ [162.4]	42.6 (26.3%)	1.38 (0.93, 2.05)	0.11	0.96 (0.63, 1.48)	0.86
Education
<High school [960.9]	166.7 (17.4%)	Reference		Reference
High school/GED [304.6]	79.0 (25.9%)	1.66 (1.21, 2.26)	0.001	1.48 (1.07, 2.05)	0.02
Some college [145.0]	49.5 (34.1%)	2.40 (1.60, 3.61)	<.0001	1.99 (1.28, 3.09)	0.002
4-year degree or more [61.5]	27.3 (44.3%)	3.77 (2.13, 6.67)	<.0001	2.67 (1.46, 4.87)	0.001
Mammogram history
Never [240.6]	20.4 (8.5%)	Reference		Reference
Ever [1229.0]	300.1 (24.4%)	3.49 (2.18, 5.59)	<.0001	3.43 (2.08, 5.65)	<.0001
Language on survey
Spanish [1306.0]	250.6 (19.2%)	Reference		Reference
English [173.4]	74.0 (42.7%)	3.13 (2.20, 4.47)	<.0001	2.36 (1.57, 3.53)	<.0001

¹From univariate logistic regression models predicting the odds of awareness

²From multivariable logistic regression model including all characteristics listed in the table

Among respondents who were aware of MBD, significant differences in knowledge of MBD masking (OR=3.58 [95% CI: 1.67–7.67], p=0.001; ORadj=2.78 [95% CI: 1.13–6.82], p=0.03) and knowledge of MBD increasing breast cancer risk (OR=5.36 [95% CI: 2.41–11.94], p<0.0001; ORadj=7.22 [95% CI: 2.99–17.45], p<0.0001) were found after adjusting for education, age, language, and screening mammography history. (Tables 3 and 4). Considering the respondents from states with legislation, findings were similar (knowledge of masking for NS vs AZ: ORadj=3.33 [95% CI: 1.04–10.61], p=0.04; knowledge of risk for NS vs AZ: ORadj=7.22 [95% CI: 2.45–21.28], p=0.0003).

Table 3:

Knowledge regarding masking effect of breast density on cancer detection, among those who had awareness of breast density (weighted N=322.5 who answered the masking effect survey question). All frequencies, percentages, and odds ratios are weighted.

Subgroup [Weighted N]	N (%) Correct	Unadjusted1		Adjusted2
		OR (95% CI)	P-value	OR (95% CI)	P-value
Cohort
AZ [273.0]	101.0 (37.0%)	Reference		Reference
NS [49.5]	33.6 (67.8%)	3.58 (1.67, 7.67)	0.001	2.78 (1.13, 6.82)	0.03
Age
40–49 [189.3]	80.9 (42.7%)	Reference		Reference
50–59 [90.6]	35.4 (39.1%)	0.86 (0.50, 1.47)	0.58	0.87 (0.48, 1.56)	0.64
60+ [42.6]	18.3 (42.9%)	1.01 (0.49, 2.05)	0.98	0.77 (0.34, 1.77)	0.54
Education
<High school [164.7]	49.4 (30.0%)	Reference		Reference
High school/GED [79.0]	46.3 (58.5%)	3.29 (1.86, 5.84)	<.0001	3.07 (1.70, 5.56)	0.0002
Some college [49.5]	18.3 (36.9%)	1.41 (0.68, 2.90)	0.35	1.14 (0.51, 2.51)	0.75
4-year degree or more [27.3]	19.6 (72.0%)	6.01 (2.22, 16.32)	0.0004	3.86 (1.26, 11.86)	0.02
Mammogram history
Never [20.4]	4.0 (19.6%)	Reference		Reference
Ever [298.1]	128.6 (43.1%)	3.11 (1.01, 9.57)	0.048	2.72 (0.92, 8.03)	0.07
Language on survey
Spanish [248.6]	95.2 (38.3%)	Reference		Reference
English [74.0]	39.4 (53.3%)	1.84 (1.04, 3.25)	0.04	1.02 (0.51, 2.05)	0.96

¹From univariate logistic regression models predicting the odds of correct knowledge with respect to breast density masking effect of cancer detection

²From multivariable logistic regression model including all characteristics listed in the table

Table 4:

Knowledge regarding association of dense breasts with increased risk of breast cancer, among those who had awareness of breast density (weighted N=322.5 who answered the breast cancer risk survey question). All frequencies, percentages, and odds ratios are weighted.

Subgroup [Weighted N]	N (%) Correct	Unadjusted1		Adjusted2
		OR (95% CI)	P-value	OR (95% CI)	P-value
Cohort
AZ [273.0]	89.0 (32.6%)	Reference		Reference
NS [49.5]	35.7 (72.2%)	5.36 (2.41, 11.94)	<.0001	7.22 (2.99, 17.45)	<.0001
Age
40–49 [189.3]	64.7 (34.2%)	Reference		Reference
50–59 [90.6]	35.0 (38.6%)	1.21 (0.70, 2.09)	0.49	1.18 (0.66, 2.13)	0.57
60+ [42.6]	25.0 (58.7%)	2.74 (1.34, 5.59)	0.006	1.88 (0.82, 4.32)	0.14
Education
<High school [164.7]	57.3 (34.8%)	Reference		Reference
High school/GED [79.0]	31.5 (39.8%)	1.24 (0.70, 2.18)	0.45	1.06 (0.57, 1.95)	0.86
Some college [49.5]	23.7 (49.8%)	1.64 (0.81, 3.32)	0.17	1.11 (0.50, 2.47)	0.80
4-year degree or more [27.3]	11.3 (41.3%)	1.32 (0.51, 3.37)	0.56	0.64 (0.19, 2.10)	0.46
Mammogram history
Never [20.4]	2.0 (9.8%)	Reference		Reference
Ever [298.1]	120.7 (40.5%)	6.26 (1.41, 27.73)	0.02	4.55 (0.98, 21.03)	0.05
Language on survey
Spanish [248.6]	90.4 (36.4%)	Reference		Reference
English [74.0]	34.4 (46.5%)	1.52 (0.85, 2.71)	0.154	0.58 (0.27, 1.28)	0.18

¹From univariate logistic regression models predicting the odds of correct knowledge with respect to association of breast density and breast cancer risk

²From multivariable logistic regression model including all characteristics listed in the table

DISCUSSION

In this study we found significant differences in MBD awareness and knowledge that dense breasts makes it more difficult to identify a mass on mammography or that dense breasts increase the risk of breast cancer between Latinas from a nationally-representative sample and Latinas from a low-resource setting. Differences in awareness, but not knowledge could be explained by education, screening history and preferred language. The majority of Latina participants in the AZ cohort had less than a high school degree and spoke primarily Spanish.

Within the AZ cohort, we also found significant differences of MBD awareness by various demographic factors, including literacy and educational training. Although we did not limit this cohort by immigration history, this finding is consistent with prior literature from Latina focus groups that found knowledge about breast cancer and health risks is relatively low in Latina immigrants in the US [16–18]. Reasons for lack of knowledge about breast cancer and health risks are likely multifactorial and may include lack of discussion about cancer screening in countries of origin [19–22], embarrassment or fear of breast cancer diagnosis [23], and language barriers in the United States, which may limit the opportunity for breast cancer risk discussions with English speaking providers. Prior studies have found the use of social support via peer counselors to be an effective form of medical education outreach in Latino communities [24–26]. The larger LLEAD clinical trial is investigating interventions to inform patients about MBD targeted to this specific socioeconomic, predominantly Spanish speaking Latina population [15].

The major limitation for this study is the variability in the breast density notification legislation in the two cohorts. The NS cohort was a sample of the national population with varying MBD state legislations in place when the study was conducted in 2017, whereas, in the AZ cohort state MBD legislation had been in place for two years prior to the initiation of assessment. The notification language is consistent for the AZ cohort but will vary state to state for the NS cohort. We were unable to look at the impact of notification on awareness in the NS cohort as we do not know whether respondents from states with notification laws ever had a mammogram after such legislation was passed. However, analyses subset to those states with MBD notification legislation at time of survey did not find statistically significant differences.. Another limitation of our study is that the AZ cohort was comprised of patients getting their care in a single FQHC facility with providers likely providing similar counseling regarding breast cancer screening and dense breasts. The NS study was constructed to be representative of the Latina population whereby the practicing styles of providers and guidelines for MBD counseling may vary compared to the AZ cohort. This variability in provider counseling within the NS cohort, may be a truer reflection of clinical practice variations. Another characteristic of the AZ cohort that may differ from the NS cohort is whether the respondent required assistance from the bilingual study coordinator to complete the survey. It is unclear what assistance was required by the Latinas in the national cohort but it is known to 65% in the AZ cohort. Thus, it is not clear how generalizable the findings from each of these cohorts reflects the general Latina population or those who receive medical care at a FQHC . Finally, we were unable to assess the impact of immigrant status on MBD awareness and knowledge.

Our findings suggest that a tailored approach to increasing MBD awareness and knowledge will be needed. To ensure equitable dissemination of health care resources to Latina women undergoing screening mammography, future work should be aimed at developing effective tools and resources for educating Spanish-speaking women with lower education about MBD and risk-based screening. These strategies may also address between-race differences in women’s knowledge of dense breast tissue [27; 28]. These educational efforts may also facilitate patient-provider conversations and education or counseling regarding MBD [16; 18].