Since 2009, the United States Preventive Services Task Force (USPSTF) has recommended biennial breast cancer screening with mammography for average-risk women aged 50–74 years [1]. Based on the lack of inclusion of women aged ≥ 75 years in key breast cancer screening trials and unanswered questions about reductions in advanced breast cancer and treatment morbidity from mammography screening, the USPSTF concluded that there was insufficient evidence to balance the benefits and harms of screening for women aged ≥ 75 years [1]. Uncertainty about the benefits of routine mammography screening at older ages is reflected in guidelines created by specialty societies with differing recommendations for older women [2]. For example, the American Cancer Society recommended women in good health and with life expectancy of at least 10 years continue screening [3], and the American College of Radiology recommends that screening can be individualized considering life expectancy, comorbidities, and intention to undergo treatment if cancer is diagnosed [4], and the American College of Physicians stated that average-risk women aged ≥ 75 years and those with life expectancy ≤ 10 years should discontinue screening [5]. These differences in recommendations could confuse patients and providers alike [6]. The complexity of decisions around breast cancer screening in older women is compounded by the fact that breast cancer incidence is high in this age group and the highest breast cancer mortality rates occur among women aged ≥ 75 years [7].

In light of the increasingly nuanced guidance for breast cancer screening among older women, the goals of this analysis are first, to examine recent incidence data for breast cancer among older women using a national database, and second, to describe mammography screening use among women near and beyond the upper USPSTF screening age range. These data can be used to inform future research and policy discussions regarding breast cancer screening among older women.

Table 1 displays the incidence and 95% confidence interval of breast cancer across demographic categories and tumor characteristics in 2016. Women aged 70–74 years had the highest incidence of breast cancer (458.3 cases per 100,000 women) and women aged ≥ 85 years the lowest (295.2 per 100,000 women). Within age categories, most cancers were diagnosed at a localized stage; the proportion from highest to lowest by age group was 71.9% for ages 70–74, 70.4% for ages 75–79, 70.1% for 65–69, 67.7% for 80–84, and 62.0% for ≥ 85 (data not shown). The incidence rates for distant stage cancer and cancer of unknown stage also varied by age group; the highest incidence rates for both occurred in age groups above 74 years. The proportion of breast cancer diagnosed at distant stage increased with age and was 2.2% for 65–69, 2.3% for 70–74, 2.7% for 75–79, 3.4% for 80–84, and 6.6% for ≥ 85 years (data not shown). The pattern of rates of ER + and/or PR + cancers by age group was similar to that of breast cancer overall. As for stage at diagnosis, rates of unknown ER/PR status increased with age.

Characteristics of women in the analysis of mammography use overall and by age group are shown in Table 2. Of these women, 31.2% (n = 1085) were aged 65–69 years, 25.8% (n = 898) were 70–74 years old, 18.2% (n = 635) were 75–79 years old, 12.5% (n = 434) were 80–84 years old, and 12.3% (n = 429) were 85 years old or older. Greater percentages of women aged 65–74 years reported higher educational attainment and at least good health status.

Regarding mammography use, 64.4% (n = 2,226) reported having a mammogram within 2 years, 14.7% (n = 506) more than 2 but less than 5 years, 15.8% (n = 558) over 5 years ago, and 5.2% (n = 191) reported never having a mammogram. The percentage of women reporting a mammogram within the prior 2 years decreased with increasing age (Fig. 1).

Table 3 displays demographic and clinical variables by time since mammography among women aged ≥ 77 years. Overall, 46.8% of women aged ≥ 77 years had a mammogram within the last 2 years. The proportion who reported their last mammogram occurred less than 2 years ago decreased with age, from over half of women aged 77–84 years to 26.0% of women aged ≥ 85 years. There was no difference in time since last mammography by race, ethnicity, or region. Women who reported Medicare coverage versus private insurance, not completing high school, fair or poor health, or dependency in one or more ADLs or IADLs had lower use of recent mammography than other women.

This analysis of national data on breast cancer incidence and screening mammography use in older women reveals several points. This analysis found that more than 70% of women aged 65–74 years–for whom the USPSTF recommends routine screening—have received it. Although breast cancer incidence decreased with age from its peak among women aged 70–74 years, incidence remained high among women aged 75–79 years, beyond the upper USPSTF screening age range for recommended routine screening. Many women received screening mammograms at ages ≥ 75 years. However, approximately one quarter of women aged 65–74 years, for whom routine screening is recommended by multiple expert organizations, had not received a mammogram in the prior 2 years. Thus, many women may not be up-to-date with recommended screening as they approach the age at which screening may not be encouraged.

Mammography use declined with increasing age, as did breast cancer incidence after age 74 years. In contrast, the incidence rate of distant stage disease at diagnosis was highest in the 75–79 and 80–84-year-old age groups. Findings of high breast cancer incidence rates and increased risk of advanced disease at diagnosis for some after age 74 years may suggest a potential benefit from continued screening for some older women. Hartman et al. examined screening-detected breast cancer among women aged ≥ 75 years at one institution and reported a favorable cancer detection rate and a high proportion of early stage cancers [6]. A recent review also concluded that screening performance metrics improved with age [7]. However, a study of Medicare beneficiaries concluded that annual mammography screening after age 75 years did not reduce breast cancer mortality over an 8-year time period [19]. Whether newer screening technologies will improve the balance of benefits and harms for older women is uncertain [20].

Our findings of recent screening among women aged ≥ 77 years are consistent with previous work reporting continued screening at older ages after the USPSTF had set an upper age limit [21]. Torke and colleagues reported that some older adults felt stopping cancer screening would “require a major decision,” whereas continuing screening would not. They also found that patients felt mistrustful of panel recommendations about screening cessation, and that a recommendation to stop screening from their physician might threaten their trust [22].

Our study suggests that for some women, neither age nor overall health are considered reasons to stop screening. Among women aged ≥ 77 years who self-reported their health as “fair” or “poor,” 39.1% had a mammogram within the last two years. Mammography’s potential benefit to an individual patient has been based largely on her life expectancy, and there is an estimated 10-year lag time for survival benefit from mammography [23, 24]. Several studies have suggested that comorbidities and life expectancy are associated with mammography screening among older women [25–27], but many women in their 70 s and 80 s with shorter life expectancies receive mammograms [26, 28]. In 2004, prior to the USPSTF recommendation of an upper screening age limit, Schonberg et al. reported that over 50% of women aged ≥ 80 years had received a screening mammogram in the previous 2 years, including 43.9% of women with two or more significant diseases [25]. Lee et al. recently summarized the available evidence on screening after age 74 years and concluded that more than age alone should be considered, including overall health status [7]. As Brawley has proposed, breast cancer screening recommendations for older women in the future are likely to be tailored to the individual woman rather than based on age, to identify women who are at risk for breast cancer and in good health and therefore likely to benefit from screening [29].

In our study, women who reported dependency in one or more ADLs or IADLs were more likely to have had a mammogram > 5 years ago or never (59.1% for ADLs, 54.8% for IADLs), which is consistent with previous work by Warner and colleagues who reported that women with complex multimorbidity were less likely to have had a screening mammogram. However, women who thought they might live another 10–15 years were more likely to have a screening mammogram, even if their health status or advanced age suggested otherwise [30].

Information about healthcare provider discussions with women who had a mammogram after age 74 years was not collected on the 2018 NHIS and therefore not examined in this study. Schoenborn and colleagues suggest that communication about the benefits and risks of screening and shifted health priorities could be preferable to discussions about life expectancy for both primary care providers and older adults [31]. A recent study showed that some clinicians hold negative opinions about the value of life expectancy predictions for cancer screening decisions. Others have underscored the importance of considering a woman’s goals and values in individualizing screening decisions [32]. Decision aids may be useful when engaging older women in conversations about screening [32].

We found that the incidence rate of “unknown” stage disease—possibly suggesting no or limited further workup done after diagnosis—increased with increasing age. The highest rates of unknown stage (6.6% of incident cases) and missing both ER and PR status (8.9% of incident cases) occurred among women aged ≥ 85 years. This raises questions as to whether diagnostic workup or data reporting for breast cancer among women at the oldest ages differs from that of women at younger ages. Many women, particularly those in the oldest age groups, may not receive additional workup after their diagnosis of breast cancer, which may be consistent with a decision not to treat.