Screening mammography reduces breast cancer mortality by about 15% in women ages 40–49 [1], somewhat less than the 20–30% mortality reduction in women ages 50–69. Considering the balance of benefits and harms, in 2009 the US Preventive Services Taskforce (USPSTF) recommended against routine screening mammograms for women in their 40s, advising them instead to make individual decisions about screening with their physicians in the context of their personal breast cancer risk, values and preferences [2]. The USPSTF reaffirmed this recommendation in 2016 following an updated review of the evidence [3]. The American Medical Association, American College of Physicians, American Academy of Family Physicians and the American College of Preventive Medicine supported the panel’s 2009 recommendation. The American Congress of Obstetricians and Gynecologists (ACOG) and the American College of Radiology expressed opposition and recommended annual screening mammography starting at age 40. Until recently, the American Cancer Society also recommended annual screening for women 40 and older [4].

Controversy and conflicting guidelines have led to confusion among women and their clinicians about screening mammography [5–7]. We created Breast Screening Decisions (BSD), a web-based tool to help women in their 40s at low-to-average risk of breast cancer decide when to start and how often to have screening mammograms, in the context of a shared decision-making process with their clinicians. The goal of BSD is not necessarily to change screening behavior, but rather to help women make decisions about screening mammography that are informed and consistent with their values and preferences [8], and to serve as a foundation for discussions about screening mammography between women and their health care providers. We conducted a large pilot trial of BSD to assess its feasibility and acceptability in clinical practice, and to evaluate decisional conflict, knowledge and worry about breast cancer and screening mammography in BSD users.

Users of our personalized, web-based decision aid reported low levels of decisional conflict and high levels of knowledge about breast cancer risk and screening. Most discussed breast cancer screening with a physician, and their clinicians perceived them to be well informed about screening mammography.

Many women overestimate the benefits of mammography, some even holding the mistaken belief that screening can prevent breast cancer [21]. Most BSD users in our study (79%) agreed that having a screening mammogram would reduce their chance of dying from breast cancer, a fact stated in the decision aid. Contrary to information presented in BSD, 8% of users said that screening mammography can detect all breast cancers. It is not clear whether these women were unable to process or retain information presented in the decision aid, or whether their beliefs about screening mammography were too strong to counteract conflicting information.

BSD was designed for women at low-to-average risk of developing breast cancer, and eligible users are told that they are at low-to-average risk based on their personalized risk prediction. Despite this, 10% of BSD users perceived that they were “more likely to get breast cancer than the average woman” their age. Thus, for some women there is discordance between perception and information, even when breast cancer risk estimates are personalized and stated explicitly, perhaps due to lack of confidence in those estimates [22].

In the context of disease screening, decision aids have been shown to improve knowledge, reduce decisional conflict and anxiety, allow patients to be active participants in their care and lead to informed, values-based decisions [23–25]. A screening mammography decision aid for Australian women under 50 increased knowledge and reduced indecision without increasing feelings of anxiety [26], and information about “overdetection” of breast cancer further enhanced the likelihood of making an informed decision [27].

In the absence of a comparison or control group in our study, we cannot make causal inferences about BSD and its relationship with decision-making outcomes. BSD users were a self-selected group, and their prior knowledge, attitudes and beliefs about screening may have differed from those of non-users. However, the high levels of knowledge we observed, and relatively low levels of decisional conflict and worry, suggest that BSD did not have a detrimental impact on these important outcomes. Similarly, we did not find a large difference between the proportion of users with evidence of prior screening (74%), and the proportion of users who reported actual or intended use of screening mammography following their post-BSD preventive care visit (77%), suggesting that BSD was not biased for or against any particular screening strategy and did not cause dramatic changes in behavior. BSD did not appear to influence clinician perceptions of patient knowledge, but perhaps more importantly, it did not significantly increase clinician-perceived patient anxiety about screening or the length of discussions about screening. These results suggest that BSD did not add a burden to the clinical encounter or adversely affect the clinician-patient interaction from the provider perspective.

Our study included women of different racial and ethnic backgrounds, who also likely varied in their socioeconomic status. However, 80% of the sample was white, 77% were married, and 96% had commercial health insurance, and our findings may not be generalizable to different populations, particularly those with a greater proportion of African-American women or a greater proportion of uninsured women. Similarly, the numeracy, health literacy and internet comfort of our sample may not be generalizable to all US women in their 40s. Socially disadvantaged groups, including those with limited education or low literacy, may benefit from decision aids specifically tailored to their needs [28].

BSD itself has several limitations, some due to limitations of available evidence. For example, breast cancer risk estimates were personalized, but estimates of screening benefit were age group-specific. Other limitations reflect deliberate decisions to maximize user comprehension and reduce confusion. For example, breast cancer risk and screening benefit are presented as point estimates without ranges or confidence intervals. Effectively communicating uncertainty with lay audiences is challenging, and there is currently no consensus on best practices [29]. Although we reviewed many decision aids for cancer screening as we developed BSD, we cannot say whether BSD is superior to other web-based decision aids for breast cancer screening or even if a web-based tool is appropriate for all women in their 40s at low-to-average risk of breast cancer.

There are approximately 22 million women ages 40–49 in the US, each of whom is advised, by the USPSTF and others, to make an individualized decision regarding screening mammography in consultation with her physician. While some expert groups continue to recommend that women ages 40–49 be offered annual mammography, not all women accept this offer. Use of a decision aid could alleviate medico-legal concerns that some clinicians have in supporting their patients’ individualized screening mammography decisions, especially amid conflicting and often sensationalized reports in the popular press [30]. Even ACOG, which recommends annual screening mammograms, recently advised that decisions about screening start age and frequency “should be made through shared decision making,” and that “health care providers should work with patients to determine the best screening strategy based on individual risk and values” [31]. BSD directly addresses this recommendation, offering women personalized estimates of their breast cancer risk and an opportunity to explore their values and preferences. BSD also provides balanced, evidence-based information about screening options, consistent with recent calls for patient-centered, informed consent that meets a “reasonable patient” standard and advances shared decision making [32].

In a large pilot study, users of Breast Screening Decisions, a web-based decision aid, reported high levels of knowledge about breast cancer risk and screening and low levels of decisional conflict and worry. Most users subsequently discussed breast cancer screening with a clinician, and BSD did not appear to adversely affect the patient-clinician interaction from the clinician’s perspective. Our results suggest that decision aids like BSD may help women in their 40s make informed, personalized decisions about screening mammography.