Sociodemographic Variations in Women’s Reports of Discussions With Clinicians About Breast Density

Key Points Question Do patient-clinician discussions about breast density vary by women’s sociodemographic characteristics? Findings In this telephone survey study of 770 women, most women reported that clinicians asked questions about breast cancer risk (88%), discussed mammography results (94%), and answered patient questions about breast density (81%); fewer women indicated that clinicians asked about worries or concerns about breast density (69%), future breast cancer risk (64%), or other screening options (61%). Non-Hispanic Black women were asked about breast cancer risk more often, and Hispanic and Asian women as well as those with low literacy were less likely to have questions or worries or concerns addressed. Meaning This study suggests that unaddressed worries or concerns and unanswered questions among Hispanic and Asian women as well as those with low literacy are areas for improvement.


Introduction
Dense breast tissue increases the risk of breast cancer and reduces the sensitivity of screening mammography. 1,2Recognizing these risks, dense breast notifications (DBNs) are mandated by legislation in 38 US states and the District of Columbia, with nationwide US government notification starting in 2024.Dense breast notifications accompany mammography results and provide information about breast density aimed at increasing women's understanding of associated risks.
Most notifications recommend that women with dense breasts discuss their personal risk with their physicians to foster more informed decisions about future and supplemental breast cancer screening. 3wever, prior research shows that less than half of women in the general population have conversations with their clinicians about breast density, 4,5 and little is known about the content of the conversations that do occur.By understanding women's experiences of such conversations, including whether they confer relevant information about breast cancer risk and screening options or address patients' questions and concerns-all necessary components for shared decisionmaking-we can identify strengths and address potential deficits.As prior research showed that women who belong to minoritized racial and ethic groups and with lower literacy were often less knowledgeable or aware of breast density's risks, 4 we hypothesized that women's conversations with clinicians might also vary similarly.

Methods
Using a telephone survey, we assessed women's reported experiences of conversations with their clinicians about breast density.Survey content was based on existing literature, prior studies, [6][7][8][9] and the Health Belief Model. 10,11The research was deemed exempt in the category of survey research by the Boston University Medical Campus institutional review board; however, a consent statement read to prospective participants provided options to decline or continue.This study followed the American Association for Public Opinion Research (AAPOR) reporting guideline.

Participant Sampling and Recruitment
We conducted a cross-sectional, national, random-digit dial telephone survey study using a sampling approach detailed previously. 4In brief, structured telephone surveys from the survey firm SSRS included questions within the SSRS Omnibus survey, a national, weekly, dual-frame bilingual randomdigit dial telephone survey using landlines and cell phones, conducted in English and Spanish.Our questions were included in the Omnibus survey from July 1, 2019, to April 30, 2020.We identified a representative sample of US women 40 to 76 years of age meeting the following eligibility criteria: (1)   had undergone mammography within the prior 2 years, (2) had no personal history of breast cancer, and (3) had heard the term dense breasts or breast density.Eligible participants were also invited to participate using a prescreened callback sample from prior Omnibus waves to increase participation of Hispanic, Black, and Asian women; women with less than a high school education; and those living in less populous nonnotification states.A third group of eligible participants were recruited from a sample specifically modeled by SSRS to reach Asian American women.yield and AAPOR cooperation rates (proportion of completed interviews received out of qualified respondents who were selected to complete the interview 12 ) from each sampling approach; the overall cooperation rate was 85% (2306 of 2718).This analysis was conducted among the 770 of 2306 women in the sample (33%) who reported having had a conversation about breast density with their clinician after their last mammographic screening.

Measures
We queried participants about their most recent discussion with their clinician regarding breast density.As detailed in the eAppendix in Supplement 1, we asked whether the clinician had (1) asked questions about breast cancer risk (family history or prior biopsies), (2) asked about worries or concerns about breast density, (3) discussed mammography results, (4) discussed other options for breast cancer screening (eg, magnetic resonance imaging or ultrasonography), or (5) discussed the woman's future risk of breast cancer (all yes or no responses; answers from respondents who were unsure or declined to answer were designated missing).Then, we asked about the extent to which their clinician had answered their questions about breast density (completely, mostly, somewhat, a little, or not at all).Given our focus on the most desirable outcome, this variable was dichotomized for analyses into completely or mostly and somewhat, a little, or not at all.
We assessed whether or not respondents lived in a state mandating DBNs, sociodemographic characteristics including age (40-49, 50-64, and Ն65 years), educational level (high school or less, some college, and college or more), income (<$50 000, $50 000-$99 999, and Ն$100 000), and race and ethnicity (Asian, Hispanic or Latino, non-Hispanic Black, non-Hispanic White, and other [Native American or American Indian or Alaska Native, Native Hawaiian and Other Pacific Islander, mixed, and other (as indicated by the respondent)]).We assessed literacy, defining low literacy as either having less than a high school education or reporting sometimes, often, or always needing assistance to complete forms using the validated Single Item Literacy Screener. 13To assess and  control for breast cancer risk factors as covariates, we asked each woman whether she had a firstdegree relative with a diagnosis of breast cancer or had ever had a breast biopsy.

Statistical Analysis
Statistical analysis was conducted in April and July 2023.We first conducted bivariate χ 2 analyses assessing associations between each dependent variable and each of the sociodemographic variables listed.We then conducted separate multivariable binary logistic regression analyses, evaluating associations among each dependent variable and all of the sociodemographic variables and breast cancer risk factors simultaneously.We excluded educational level from multivariable models due to its inclusion in the health literacy measure and collinearity with income.All analyses were conducted with the statistical analysis software IBM SPSS Statistics, version 27 (IBM Corp).
Statistical significance was defined at 2-tailed α = .05.Post hoc comparisons in bivariate analyses were Bonferroni corrected for multiple comparisons.

Discussion
As one of the first studies, to our knowledge, to examine women's reports of the frequency and content of clinician-patient breast density discussions, we examined whether or how such discussions might vary by women's sociodemographic characteristics.Overall, only approximately one-third of women surveyed reported such a conversation, similar to other studies. 4,5Findings newly revealed that these women were more likely to be members of historically marginalized groups, including lower income or literacy levels or women of minority racial and ethnic groups.
Among women reporting such conversations, most indicated that clinicians asked about breast cancer risk, discussed mammography results, and answered questions about breast density.
However, women less frequently reported that clinicians had asked about worries or concerns about breast density or discussed future risk of breast cancer or other options for breast cancer screening.
Women's reports of these discussions varied significantly by race and ethnicity.Non-Hispanic Black women (a group at higher risk for breast cancer mortality) more often reported being queried about breast cancer risk, but Hispanic and Asian women were less likely to report being asked about worries or concerns or having their questions about breast density answered completely or mostly.
Although our data do not provide an explanation for this observed difference, clinician behavior may be influenced by the evidence of elevated breast cancer mortality among young Black women in particular. 14The single prior study we located regarding the content of women's discussions with physicians about breast density found no differences between the proportion of non-Hispanic White and non-Hispanic Black women who reported talking with their physicians about breast density notifications, 5 but those results did not characterize the discussions' content.
Women's reported breast density discussions with clinicians varied significantly with literacy level.Women with low literacy less frequently reported being asked about worries or concerns about breast density or that their mammography results were discussed with them and less frequently felt their questions were answered completely or mostly.This finding highlights a mismatch between patients' informational needs and material shared by clinicians; one possibility is that clinicians limit the information offered to patients when discussing breast density due to assumptions about the ability of patients with low literacy to understand information about screening or its consequences, 15 but our data do not allow us to examine this.Perhaps the information given was less comprehensible to women with low literacy, as prior research shows that some clinicians struggle with providing understandable information to such patients. 16Regardless, our findings suggest that women with low literacy want more or different information to meet their needs.
Clinician questioning and counseling did not vary by state DBN legislation status, suggesting that the legislation either was associated with broad or no nationwide outcomes or that other clinician education has generally been associated with care.We are unable to determine this with our data given the single time point of measurement.Clinician questioning and counseling almost never varied by women's age or income.As might be expected, women with risk factors (eg, family history or prior benign breast biopsy result) were more likely to report being asked questions about breast cancer risk; the latter group was also more likely to indicate that their clinician discussed other options for breast cancer screening.Women with a family history of breast cancer could have a heightened sensitivity to the topic, which could in turn have increased recall of discussions of breast cancer risk, although we were unable to examine this possibility directly.
Our findings suggest that when discussions about breast density do occur, over half of women's clinicians are addressing relevant related topics.Shortfalls in addressing breast density may be a function of clinicians feeling ill prepared, insurance coverage of supplemental screening and the absence of a widely agreed-on evidence base 21 or guidelines regarding supplemental screening to guide counseling at the time these data were collected may have inhibited such counseling and the incorporation of recommended shared decision-making, which could improve the quality of conversations. 22The limited clarity regarding supplemental screening may undergird these shortcomings, as clinicians may eschew addressing ambiguous topics.However, the unaddressed worries or concerns and unanswered questions, especially among historically marginalized women, highlighted areas where discussions could be improved.In many situations, it is appropriate for clinicians to not discuss breast density or supplemental screening depending on the woman's overall risk profile or other pressing clinical concerns.Despite some reassuring findings regarding appropriate differential counseling by patient race and ethnicity, other results echo prior research documenting racial and ethnic variations in processes of care [23][24][25] and provide new evidence of differential processes by patient literacy.

Limitations
This study had several limitations.We were able to examine the content of discussions only among the women who had had such discussions, and these women differed in several ways from the national sample.Without clinical assessments of breast density, we were unable to determine whether breast density was a clinically significant or appropriate topic for clinicians to address in their discussions with women.Moreover, women's self-reports of their discussions might be influenced by recall bias, and we did not have data on the length of time that elapsed between the mammography and these conversations.Although language barriers may have been an issue, without data on language we could not address this possibility in the analysis.In addition, we report results of a large number of statistical tests among correlated dependent variables, which warrants caution when interpreting significant findings.Future research efforts are needed to test the generalizability of observed associations, particularly those of smaller effect size.Care processes are not structured such that women are prompted to return to their clinician after mammographic screening (most see them before), so women reporting these discussions could have different health care use patterns relative to those who did not report conversations about breast density with their clinicians.They could have also been more concerned and sought out conversations, but our data did not allow us to examine these possibilities.Although the United States Preventive Services Task Force has not issued new guidelines for supplemental screening for women with dense breasts since this study was conducted, the American College of Radiology guidelines 26 may have influenced counseling since then.Despite such limitations, women's reports represent the most efficient method to assess the content of the conversations.

Conclusions
Taken together, the results of this survey study provide some positive outcomes assessments when clinician-patient conversations about breast density do occur in that most patients report their clinicians are discussing potential breast cancer risk factors and mammography results and answering questions about breast density.However, these results also suggest potential areas for improvement as only a minority of our national sample reported such discussions, and when they occurred, there were limited conversations about other options for breast cancer screening, future breast cancer risk, or queries about patient worries or concerns about breast density.Efforts to improve cultural competency, shared decision-making, and counseling skills, especially when guidelines for care for women with dense breasts become more widespread, can potentially help close the remaining gaps in conversations about breast density.

Figure. Cohort Derivation
Figure.Cohort Derivation

Table 1 .
Sample Characteristics 188 (24%) aged 40 to 49 years and 224 (29%) aged 65 years or older.A total of 348 of 721 women (48%) had incomes less than $50 000, 211 of 721 (29%) had incomes between $50 000 and $99 999, and 162 of 721 (23%) had incomes of $100 000 or higher.A total of 174 of 766 women (23%) had a high school education or less, 226 of 766 (30%) had some college education, and 366 of 766 (48%) had a college education or greater; 156 (20%) had low literacy.A total of 47 women (6%) were Asian, 125 (16%) were Hispanic, 204 (27%) were non-Hispanic Black, 317 (41%) were non-Hispanic White, and 77 (10%) were of other race and ethnicity.Women in the analytic sample were more likely than those not in the analytic sample to reside in a DBN state (82% [630 of 770] vs P < .001).They were also more likely than those not in the analytic sample to be of other race (10% [77 of 770] vs 5% [83 of 1535]) and less likely to be non-Hispanic White (41% [317 of 770] vs 48% [741 of 1535]) (χ 2 4 = 26.15;P < .001 a Individual comparisons are significantly different across groups.The P value is for the χ 2 analysis comparing each categorical variable by conversation status.bIncludes Native American or American Indian or Alaska Native, Native Hawaiian and Other Pacific Islander, mixed, and other (as indicated by the respondent).with

Table 2 )
. Similarly, in multivariable analyses, the odds of a Hispanic woman reporting discussing her mammography results with a

Table 2 .
Women's Recall of Experiences Discussing Breast Density With Clinicians Among Women Who Did Have a Discussion About Breast Density After Last Mammographic Screening (N = 770): Bivariate ResultsWhen you last discussed breast density with your physician or other clinician, did they also ask about any worries or concerns you might have about your breast density?
Abbreviations: DBN, dense breast notification; HS, high school; NH, non-Hispanic.a Significantly different from college or more after Bonferroni correction for multiple comparisons.b Significantly different from Hispanic after Bonferroni correction for multiple comparisons.c Significantly different from 65 years or older after Bonferroni correction for multiple comparisons.d Significantly different from $50 000 to $99 999 after Bonferroni correction for multiple comparisons.e Significantly different from Asian after Bonferroni correction for multiple comparisons.f Significantly different from low literacy after Bonferroni correction for multiple comparisons.g Significantly different from NH White after Bonferroni correction for multiple comparisons.JAMA Network Open | Equity, Diversity, and Inclusion Sociodemographic Variations in Women's Discussions With Clinicians About Breast Density JAMA Network Open.2023;6(11):e2344850. doi:10.1001/jamanetworkopen.2023.44850(Reprinted) November 27, 2023 6/11 Downloaded from jamanetwork.comby guest on 01/01/2024