Improving State-Mandated Breast Density Notifications

Improving State-Mandated Breast Density Notifications

ORIGINAL ARTICLE Improving State-Mandated Breast Density Notifications Derek L. Nguyen, MD a, Emily B. Ambinder, MD a, Mary Kate Jones, MA a, Lisa A. ...

170KB Sizes 0 Downloads 14 Views

ORIGINAL ARTICLE

Improving State-Mandated Breast Density Notifications Derek L. Nguyen, MD a, Emily B. Ambinder, MD a, Mary Kate Jones, MA a, Lisa A. Mullen, MD a, Susan C. Harvey, MD a Abstract Purpose: Effective written communication directly affects health care outcomes. Since 2016, the complex language of state-mandated breast density notifications (BDNs) has been challenged, because it is perceived to be beyond the comprehension of most patients. The aim of this study was to assess whether a revised BDN written at a lower reading grade level improves understanding compared with the current state-mandated BDN. Methods: A revised notification with similar content to the current state-mandated BDN was developed. Both notifications were presented to patients for direct comparison, using a paper survey asking questions that evaluated patients’ perceptions and convictions associated with breast density. Surveys were distributed at four outpatient imaging centers to screening mammography patients. Results: The current BDN’s mean readability metric was 13.4, and that of the revised BDN was 6.6. Five hundred surveys were analyzed. Survey data demonstrated that 56.6% of all women perceived that dense breast tissue results indicated a “high” associated lifetime breast cancer risk from the current state-mandated BDN compared with only 2.2% with the revised notification (P < .001). Nearly all women were more likely to initiate discussions with their providers regarding their breast tissue density after reading the revised notification (96.0%) as opposed to the current state-mandated BDN (32.8%; P < .001). Conclusions: A significant portion of women misinterpret the intended messages of the current state-mandated BDN. Thus, a revised notification at a lower reading grade level may improve understanding of breast density, leading to improved individualized breast cancer screening for women with dense breasts. Key Words: Breast density notification, screening mammography, readability, state-mandated, breast imaging J Am Coll Radiol 2019;-:---. Copyright  2019 American College of Radiology

INTRODUCTION For the past decade, increased awareness of the relationship between dense breast tissue and breast cancer risk has led to state legislation requiring notification statements alerting women that breast tissue density may affect their health [1]. As of August 2019, breast density notifications (BDNs) are mandated in 38 states [2]. Typically, BDNs attempt to relay to patients the key issues associated with dense breast tissue, such as masking effect and associated increased risk for developing breast cancer [3]. BDNs also encourage a discussion between a patient and her primary a

Johns Hopkins Medicine, Baltimore, Maryland. Corresponding author and reprints: Susan C. Harvey, MD, Johns Hopkins Medicine, 720 Rutland Avenue, Baltimore, MD 21205; e-mail: susan. [email protected]. Dr Harvey is the vice president of global medical affairs in the Breast and Musculoskeletal Division of Hologic. All other authors state that they have no conflict of interest related to the material discussed in this article.

ª 2019 American College of Radiology 1546-1440/19/$36.00 n https://doi.org/10.1016/j.jacr.2019.08.023

provider, including consideration of supplemental screening studies [3]. The underlying goal of the notifications is to educate patients and potentially facilitate the early detection of breast cancer in women with dense breast tissue. Dense breast tissue may warrant supplemental screening modalities for early breast cancer detection. Yet not all women have dense tissue; about half of all women older than 40 years have either heterogeneously or extremely dense breast tissue [4]. It should be noted that for women without personal or genetic histories of breast cancer, dense breast tissue marginally raises the estimated 5- and 10-year risk for developing invasive breast cancer as calculated by the Breast Cancer Surveillance Consortium risk calculator [5,6]. This is true regardless of age or race. Thus, if misinterpreted, BDNs may inaccurately influence a woman’s perception of her risk for developing breast cancer. Generally, the language used in

1

the state-mandated BDNs has been characterized as lacking readability and potentially causing patient confusion [3]. A study by Kressin et al [3] in April 2016 demonstrated that although many states require specific language for their respective BDNs, nearly all notifications exceeded the eighth grade reading level recommended by the Centers for Disease Control and Prevention and the National Institutes of Health for patient-directed written materials [3]. A recent follow-up study by Saraiya et al [7] in January 2019 reported that despite the published findings of Kressin et al, the readability of the majority of BDNs remained above national recommendations, even with 12 additional states having passed BDN legislation and 3 states having amended their notifications in the interim of study publication. This demonstrates a fundamental failure to accomplish the one intent of the notification laws: improving the ability of women to advocate for their individualized health care. To realize the full benefits and intent of BDNs, it is critical that women understand these state-mandated communications. Thus, a study to assess the impact of the language in these notifications is timely and relevant. The purpose of this study is to assess whether a revised notification, written at a lower reading grade level, increases understanding compared with our state-mandated BDN. Specifically, we assessed whether a revised notification improved both the perceived associated lifetime breast cancer risk and the likelihood of patient-initiated breast tissue density discussions with their providers.

METHODS The institutional review board at our institution acknowledged this study as exempt from review under US Department of Health and Human Services regulations. This work was HIPAA compliant, because no protected health information was included in the survey.

Study Design and Participants The readability of our state’s BDN (see Appendix 1) was assessed and calculated by using the multiple readability techniques described by Saraiya et al [7]: automated readability index, Coleman-Liau index, Gunning fog index, Flesch-Kincaid grade level (FKGL), Flesch-Kincaid reading ease, and simple measure of gobbledygook [7]. A revised notification with equivalent medical information was formulated at a lower FKGL (see Appendix 1). 2

Similar to the methodology of our prior study [8], a single three-page paper survey was created that presented both our state-mandated and revised notifications for review. The survey asked questions evaluating patients’ perceived associated lifetime breast cancer risk and likelihood of initiating conversations with their providers regarding their breast tissue density. Half of the surveys showed the current state-mandated BDN first, then the revised notification on the next page. The other half of the surveys presented notifications in the reverse order. The surveys were presented in English only. To assess the perception of associated lifetime breast cancer risk, patients were shown the notification and asked, “What would you think this statement says about your overall lifetime risk of getting breast cancer?” after reading each notification and were provided multiplechoice answers. The appropriate response was defined as either “low risk” or “medium risk,” referring to a low or moderate increase, respectively, in a woman’s baseline risk for developing breast cancer. Other possible responses were “high risk,” “no risk,” and “unsure.” To assess the likelihood of discussion with their providers, patients were then asked, “After reading this statement, what is the likelihood you would want to talk to your doctor about your breast density?” after each notification and were provided multiple-choice answers. The ideal response to this question was either “likely” or “very likely.” Other possible responses were “unlikely,” “very unlikely,” and “in between.” Demographics were recorded at the end of the survey. Between January 2019 and February 2019, surveys were distributed by simple randomization in waiting rooms at our institution’s four outpatient breast imaging centers. As indicated on the top of each survey, patient participation was strictly voluntary, and completion of the survey served as consent to participate. Inclusion criteria were female patients undergoing only screening mammography who were at least 30 years of age. Because women undergoing examinations other than screening mammography may exhibit emotional bias when reading the notifications, responses from these women were excluded. Exclusion criteria were any female patient undergoing an examination other than screening mammography (such as diagnostic mammography or ultrasound), age less than 30 years, and a survey insufficiently completed to allow us to make comparisons between the two BDN versions. Patients who possessed at least a college degree were classified in our study as having higher levels of education. All other patients (including those reporting “some college”) were classified as having lower levels of education. Journal of the American College of Radiology Volume - n Number - n Month 2019

Statistical Analysis A sample size of at least 384 was determined to achieve adequate power to yield results at a confidence level of 95% with a margin of error of 5% for our local population of women older than 30 years. The primary outcomes of interest were selecting the appropriate associated lifetime breast cancer risk and intending to initiate discussions with a provider regarding breast tissue density. These outcomes were stratified by education level and compared using the McNemar test between the two notifications for the entire sample size. The primary outcomes’ different distributions between the two notifications were assessed using the Fisher or c2 test for our selected sociodemographic variables. By designating the patient as the random effect, multivariate mixed-effect logistic regression was used to evaluate the impact of notification version on the result when controlling for independent covariates. Multivariate model covariates are as followed: age, being a health care provider, education level, and race. The computing software program R 2017 (R Foundation for Statistical Computing, Vienna, Austria) was used to calculate all statistical analyses. For all analyses, P values < .05 were considered to indicate statistical significance. RESULTS Demographics and Readability A total of 527 patients participated in the survey at our institution’s four outpatient breast imaging centers between January and February 2019. Of these, 26 were excluded because they underwent examinations other than screening mammography. One additional survey was excluded because the patient’s age was less than 30 years. Table 1 presents the patient demographics, with a total sample size of 500. The readability metric mean was found to be 13.4 for the current BDN and 6.6 for the revised notification (Table 2). Specifically, the FKGL was found to be 12.2 for the current BDN and 4.9 for the revised notification (Table 2).

Table 1. Patient demographics (n ¼ 500) Characteristic Age (y) 31-40 41-50 51-60 61-70 71-80 81-90 Older than 91 Race White Black Asian Other Highest level of education attained Less than high school High school Some college College Postgraduate Native English speaker Yes No Health care provider Yes No Survey version Version A Version B

Journal of the American College of Radiology Nguyen et al n Breast Density Notifications

6 (1.2) 151 (30.2) 130 (26.0) 114 (22.8) 76 (15.2) 23 (4.6) 0 (0.0) 325 (65.0) 154 (30.8) 13 (2.6) 8 (1.6) 8 (1.6) 155 (31.0) 80 (16.0) 150 (30.0) 67 (13.4) 470 (94.0) 30 (6.0) 65 (13.0) 435 (87.0) 252 (50.4) 248 (49.6)

Note: Data are expressed as number (percentage).

perceived a “high” associated lifetime breast cancer risk (74.9% [212 of 283]; Table 3). The majority of women who achieved lower levels of education interpreted the appropriate associated lifetime breast cancer risk after reading the revised notification (95.4% [270 of 283]) compared with the state’s BDN

Table 2. Breast density notifications’ readability metrics Metric

Perceived Associated Lifetime Breast Cancer Risk Our results demonstrated that more women perceived that dense breast tissue results indicate a “high” associated lifetime breast cancer risk with the state’s BDN (56.6% [283 of 500]) compared with the revised notification (2.2% [11 of 500]; P < .001; Table 3). Furthermore, we noted that with the state’s notification, the majority of patients who achieved less than a college degree

Value

Automated readability index Coleman-Liau index Gunning fog index Flesch-Kincaid grade level Flesch-Kincaid reading ease Simple measure of gobbledygook Mean (95% confidence interval)

Current 13.8 12.2 15.2 12.2 49.5 13.6 13.4 (12.3-14.5)

Revised 5.3 7.5 7.3 4.9 81.1 8.1 6.6 (5.4-7.9)

Note: Mean was calculated by summing the scores of all metrics except the Flesh-Kincaid reading ease and dividing by 5.

3

Table 3. Perceived breast cancer risk associated with dense breast tissue between the current and revised notifications Respondent’s Characteristics Overall (n ¼ 500) Revised Current Less than college degree (n ¼ 283) Revised Current At least college degree (n ¼ 217) Revised Current Health care providers (n ¼ 65) Revised Current

Appropriate Risk

High Risk

No Risk

Unsure

481 (96.2) 132 (26.4)

11 (2.2) 283 (56.6)

0 (0.0) 26 (5.2)

8 (1.6) 59 (11.8)

270 (95.4) 24 (8.5)

10 (3.5) 212 (74.9)

0 (0.0) 21 (7.4)

3 (1.1) 26 (9.2)

211 (97.2) 108 (49.8)

1 (0.5) 71 (32.7)

0 (0.0) 5 (2.3)

5 (2.3) 33 (15.2)

65 (100.0) 30 (46.2)

0 (0.0) 24 (36.9)

0 (0.0) 2 (3.1)

0 (0.0) 9 (13.9)

Note: Data are expressed as number (percentage).

(8.5% [24 of 283]; P < .001; Table 3, Appendix Table 1). Yet despite achieving at least a college degree, only 49.8% of these women (108 of 217) perceived that the state’s BDN implied an appropriate associated lifetime breast cancer risk compared with 97.2% (211 of 217) with the revised notification (P < .001; Table 3, Appendix Table 1). Regarding women who self-reported as health care providers, surprisingly only 46.2% (30 of 65) believed that the state’s BDN implied an appropriate associated lifetime breast cancer risk compared with 100.0% (65 of 65) after reading the revised notification (P < .001; Table 3). No woman, irrespective of education level, concluded the revised notification communicated no increased associated lifetime risk for developing breast cancer. This is in contrast to the state’s BDN (n ¼ 26; Table 3). The odds ratio (OR) of the revised notification for selecting the appropriate associated lifetime breast cancer risk with univariate analysis was 70.6 (95% confidence interval [CI], 42.8-116.3). As detailed in Table 4, when controlling for sociodemographic covariates, the adjusted OR of the revised notification for selecting the appropriate associated lifetime breast cancer risk with multivariate analysis was 146.1 (95% CI, 81.5-262.1). Furthermore, the only sociodemographic factor that was an independent predictor of increasing the likelihood of selecting the appropriate associated lifetime risk was achieving at least a college degree (adjusted OR, 7.3; 95% CI, 4.7-11.7).

Likelihood of Patient-Initiated Discussion With Provider The majority of women were more likely to initiate discussions with their providers regarding breast tissue 4

density after reading the revised notification (96.0% [480 of 500]) as opposed to the state’s BDN (32.8% [164 of 500]; P < .001; Table 5). The majority of patients who were unlikely to initiate this discussion after reading the state’s BDN achieved less than a college degree (72.0% [242 of 336]; Table 5, Appendix Table 2). Women who achieved lower levels of education were more likely to initiate discussions after reading the revised notification (97.5% [276 of 283]) compared with the state’s BDN (14.5% [41 of 283]; P < .001; Appendix Table 2). Women who achieved at least a college degree were also more likely to initiate discussions after Table 4. Multivariate analysis of sociodemographic factors on choosing the appropriate associated breast cancer risk between the current and revised notifications Variable Notification version Current Revised Education level Less than college degree College degree Age (y) <60 61 Race White Black Asian Other Health care provider No Yes

Adjusted OR

95% CI

Reference 146.1

Reference 81.5-262.1

Reference 7.3

Reference 4.6-11.7

Reference 0.75

Reference 0.49-1.13

Reference 0.66 0.34 0.86

Reference 0.42-1.03 0.10-1.16 0.18-4.13

Reference 0.73

Reference 0.41-1.28

Note: CI ¼ confidence interval; OR ¼ odds ratio.

Journal of the American College of Radiology Volume - n Number - n Month 2019

Table 5. Likelihood of patient-initiated discussion regarding breast density (n ¼ 500) Response Likely and very likely In between Unlikely and very unlikely

Revised Notification

Current Notification

P

480 (96.0)

164 (32.8)

<.001

18 (3.6) 2 (0.4)

86 (17.2) 250 (50.0)

Note: Data are expressed as number (percentage).

reading the revised notification (94.0% [204 of 217]) compared with the state’s BDN (56.7% [123 of 217]; P < .001; Appendix Table 2). Health care providers were also more likely to initiate discussions after reading the revised notification (93.8% [61 of 65]) compared with the state’s BDN (36.9% [24 of 65]), which was statistically significant (P < .001; Appendix Table 2). The OR of the revised notification for the likelihood of patient-initiated discussion with a provider with univariate analysis was 50.5 (95% CI, 31.1-82.1). As detailed in Table 6, when controlling for the sociodemographic covariates, the adjusted OR of the revised notification for the likelihood of patient-initiated discussion with a provider with multivariate analysis was 72.7 (95% CI, 43.1-122.4). Furthermore, the only sociodemographic factor that was an independent predictor of increasing the Table 6. Multivariate analysis of sociodemographic factors on likelihood of patient-initiated discussions regarding breast tissue density between the current and revised notifications Variable Notification version Current Revised Education level Less than college degree College degree Age (y) <60 61 Race White Black Asian Other Health care provider No Yes

Adjusted OR

95% CI

Reference 72.7

Reference 43.1-122.4

Reference 5.2

Reference 3.4-8.0

Reference 0.75

Reference 0.51-1.11

Reference 1.01 1.31 0.67

Reference 0.66-1.54 0.40-4.27 0.15-3.08

Reference 1.70

Reference 0.96-3.02

Note: CI ¼ confidence interval; OR ¼ odds ratio.

Journal of the American College of Radiology Nguyen et al n Breast Density Notifications

likelihood of patient-initiated discussion with a provider was achieving at least a college degree (adjusted OR, 5.2; 95% CI, 3.4-8.0).

DISCUSSION Revising the breast density language to an appropriate reading grade level significantly increased the likelihood of selecting the appropriate associated lifetime breast cancer risk by 146-fold and the likelihood of patientinitiated discussions with providers regarding breast density by 73-fold. The only important independent predictor of these outcomes was highest level of education achieved. When controlling for the notification version, achieving at least a college education led to a 7-fold increase in choosing the appropriate associated lifetime breast cancer risk. This also led to a 5-fold increase in the likelihood of patient-initiated breast density discussions with providers. Our study determined that our state-mandated BDN caused misperceptions related to perceived associated lifetime risk for developing breast cancer. This validates the findings of Yeh et al [9] showing that women had a significantly greater perceived associated risk for developing breast cancer after reading New York’s notification, which was written with less consideration of readability. Our work also showed that with our state’s BDN, women are less likely to initiate discussions with their providers regarding their breast tissue density. This is consistent with the findings of Cappello et al [10], which demonstrated that in the setting of the current BDNs, breast density discussions were more often initiated by providers than patients. Our study, however, took this one step further to demonstrate that by rewording BDNs to improve readability, these negative outcomes can be significantly mitigated for the majority of women regardless of education level. These prior observations combined with our study strongly suggest that if written communication is too complex for patients to understand, the result is misinterpretation of the message of the BDN. The overall intent of statelegislated BDNs is to inform patients of associated risk and to counsel on next steps for their individualized patient-centered care. However, the consequences reported adversely affect the ability of patients to advocate for their own personalized care when current statemandated communications are used. Surprisingly, at this time, none of the 38 statemandated BDNs defines “dense breast tissue” for 5

patients [2]. Among all the notifications, there are multiple versions that emphasize different concepts related to dense breast tissue [2]. These concepts include epidemiology, masking effect for breast cancer detection, and the fact that dense breast tissue is both normal and common. However, no explanation of “dense breast tissue” is presented [2]. Therefore, the written communication assumes that women should inherently know the meaning of “dense breast tissue” before reading the notification, irrespective of education level and life experience. It is logical that only after women understand the meaning of dense breast tissue will the information in the notification be relevant to their personalized care. A previous study by Gunn et al [11] evaluating women’s perceptions of the Massachusetts BDN found that 11 of their interviewed participants felt “the [breast density] notification language was either too clinical or vague” with some women stating “when you say dense, I don’t understand what it is,” or, more alarming, “dense breasts means that you got cancer or something that’s not supposed to be in ’em.” This work further demonstrates that it is impractical to expect women to already know what “dense breast tissue” means. If women cannot understand the communication because they have no prior context for understanding breast tissue density, then these notifications are much less effective. Therefore, including a statement such as “‘dense breast tissue’ means that there is more tissue than fat in your breasts” could provide context for women to better understand the notification. Importantly, we did not explicitly state the definition of “dense breasts” in our revised notification while conducting the study in order to limit response bias, because our original state-mandated BDN did not define dense tissue. On March 28, 2019, the FDA released a proposed rule to modernize the Mammography Quality Standards Act and implement requirements for a standardized federal BDN [12]. Unfortunately, despite multiple recent articles published detailing the current lack of readability of nearly every state-mandated BDN, readability as a requirement was not included in the federal directive [3,7,12]. Furthermore, the proposed language of the new BDN is still written higher than the recommended eighth grade reading level [3,12]. Our study illustrates that the necessary aspects of breast density information can be incorporated in a notification written at an appropriate reading grade level that allows improved patient comprehension for all education levels. Thus, because our data show that language drastically influences patient understanding 6

and their future intentions, appropriate readability is critical for effective patient communications and must be a vital addition to the proposed rule. There are some limitations of this study. The conclusions of this study are based on the findings derived from direct comparison of the state’s BDN with a revised notification that we formulated with the intention of lowering the FKGL. Generalizability of our results may be limited related to a number of other language permutations that can be used to create a revised notification with a lower mean readability metric. Because participation in our study was completely voluntary, our findings include only women who chose to respond, so our conclusions stratified by sociodemographic factors may be biased in ways we cannot accurately estimate. Being a non-native English speaker could have potentially affected the interpretation of both notifications in ways we cannot fully estimate. Not all survey participants actually had dense breast tissue, so responses may not reflect what would happen in real life. Last, the survey addressed intent to initiate discussion with a provider rather than actual behavior. In summary, our results illustrate that a significant portion of women undergoing screening mammography lack understanding of the intended message of our current state-mandated BDN. Our study emphasizes that clarity in these notifications may encourage women to have an active voice in their own personalized care. Therefore, it is imperative to revise any BDN written higher than an eighth grade level to decrease misperceptions related to breast cancer risk associated with dense breast tissue. In addition, it is important to include a definition of dense breast tissue to provide women with clinical context and to facilitate understanding of the notification. We are optimistic that this study may provide a template to promote change of BDN language as well as influence federal regulations, thus potentially improving the quality of individualized breast cancer screening for women with dense breasts.

TAKE-HOME POINTS -

-

Women are significantly less likely to overestimate their breast cancer risk associated with dense breast tissue when reading a BDN statement written at a lower reading grade level. This is most notable if the woman achieved less than a college degree. Regardless of education level, women report they are more likely to initiate discussions with their health care providers regarding their personal breast Journal of the American College of Radiology Volume - n Number - n Month 2019

-

tissue density after reading a BDN statement written at a lower reading grade level compared with the current state-mandated notification. Revising the language of state-mandated BDNs to an appropriate reading grade level may improve patient comprehension and potentially facilitate more patient-initiated discussions with their providers.

ACKNOWLEDGMENTS We thank all the site managers, mammography technologists, and front desk staff members at the four outpatient imaging centers for their help in handing out and collecting the surveys used in this study. ADDITIONAL RESOURCES Additional resources can be found online at: https://doi. org/10.1016/j.jacr.2019.08.023. REFERENCES 1. Are You Dense. About Are You Dense, Inc. Available at: https://www. areyoudense.org/about/. Accessed February 5, 2019. 2. DenseBreast-info. Legislation and regulations—what is required? Available at: https://densebreast-info.org/legislation.aspx. Accessed June 8, 2019.

Journal of the American College of Radiology Nguyen et al n Breast Density Notifications

3. Kressin NR, Gunn CM, Battaglia TA. Content, readability, and understandability of dense breast notifications by state. JAMA 2016;315: 1786-8. 4. Price ER, Hargreaves J, Lipson JA, et al. The California breast density information group: a collaborative response to the issues of breast density, breast cancer risk, and breast density notification legislation. Radiology 2013;269:887-92. 5. Breast Cancer Surveillance Consortium. Breast Cancer Surveillance Consortium (BCSC) risk calculator. Available at: http://tools.bcsc-scc. org/BC5yearRisk/. Accessed March 8, 2019. 6. Vachon CM, Pankratz VS, Scott CG, et al. The contributions of breast density and common genetic variation to breast cancer risk. J Natl Cancer Inst 2015;107:dju397. 7. Saraiya A, Baird GL, Lourenco AP. Breast density notification letters and websites: are they too “dense”? J Am Coll Radiol 2019;16:717-23. 8. Nguyen DL, Ambinder EB, Jones MK, Mullen LA, Harvey SC. Improving patient comprehension of screening mammography recall lay letters. J Am Coll Radiol. https://doi.org/10.1016/j.jacr.2019.05. 029. 9. Yeh VM, Schnur JB, Margolies L, et al. Dense breast tissue notification: impact on women’s perceived risk, anxiety, and intentions for future breast cancer screening. J Am Coll Radiol 2015;12:261-6. 10. Cappello NM, Richetelli D, Lee CI. The impact of breast density reporting laws on women’s awareness of density-associated risks and conversations regarding supplemental screening with providers. J Am Coll Radiol 2019;16:139-46. 11. Gunn CM, Battaglia TA, Paasche-Orlow MK, et al. Women’s perceptions of dense breast notifications in a Massachusetts safety net hospital: “so what is that supposed to mean?”. Patient Educ Couns 2018;101:1123-9. 12. Mammography Quality Standards Act. Available at: https://www. federalregister.gov/documents/2019/03/28/2019-05803/mammographyquality-standards-act. Accessed June 8, 2019.

7