We thank Lawrence Zaborski, MS, Department of Health Care Policy, Harvard Medical School, for assistance with statistical programming. Mr. Zaborski received no compensation for his contribution to this work.
In November 2009, the US Preventive Services Task Force (USPSTF) issued new recommendations regarding mammography screening. The Task Force recommended against routine screening for women ages 40 to 49 years and recommended biennial screening for women ages 50 to74 years. The recommendations met great controversy in mass media and medical literature; whether they have had an impact on screening patterns is not known. The objective of this study was to determine whether the 2009 USPSTF recommendations led to changes in screening rates among women ages 40 to 49 years and ages 50 to 74 years.
The authors performed cross-sectional assessments of mammography screening in 2005, 2008, and 2011 using data from the National Health Interview Survey, a nationally representative, in-person, household survey of the civilian, noninstitutionalized US population. In total, 27,829 women ages ≥40 years responded to the 2005, 2008, or 2011 surveys and reported about their mammography use. The primary outcome assessed was self-reported mammography screening in the past year.
When adjusted for race, income, education level, insurance, and immigration status, mammography rates increased slightly from 2008 to 2011 (from 51.9% to 53.6%; P = .07) and did not decline within any age group. Among women ages 40 to 49 years, screening rates were 46.1% in 2008 and 47.5% in 2011 (P = 0.38). For women ages 50 to 74, screening rates were 57.2 in 2008 and 59.1 in 2011 (P = 0.09).
In November 2009, the US Preventive Services Taskforce (USPSTF) issued new recommendations regarding mammography screening. The USPSTF had previously recommended screening mammography every 1 to 2 years for all women aged >40 years. In 2009, the task force recommended biennial screening for women ages 50 to74 years and recommended against routine screening mammography for women ages 40 to 49 years, stating that the “decision to start regular, biennial screening mammography before age 50 years should be an individual one and should take into account patient context, including the patient's values regarding specific benefits and harms.” The rationale of the USPSTF focused on the lower breast cancer incidence in younger women, which results in a higher number of women needed to screen to prevent 1 breast cancer death.[1, 2] The USPSTF also cited evidence that the initiation of screening in younger women leads to higher cumulative rates of false-positive results and associated potential harms, such as unnecessary biopsies, altering the risk:benefit ratio of screening in this group.[2-4]
The recommendations regarding screening in women ages 40 to 49 years met with enormous controversy from the advocacy, primary care, and oncology communities.[5-7] Media coverage was generally critical of the recommendations, noting concerns that delayed screening would increase breast cancer mortality and fear about government rationing of health care. Likely in part because of the controversy surrounding the 2009 recommendations, private and public insurers have generally continued to cover annual mammograms for women ages 40 to 50 years; and, in some cases, coverage has expanded. Indeed, the Affordable Care Act requires private insurers and Medicare to cover annual mammography without cost sharing for women starting at 40 years.[9-11]
The impact of the 2009 USPSTF recommendations and the surrounding controversy on actual mammography screening practices in the United States is unknown. In this study, we used data from the National Health Interview Survey (NHIS) to compare self-reported mammography screening in the past year in 2011 versus 2005 and 2008. We also assessed whether screening among younger women changed differentially among subgroups of women who historically have had lower rates of screening, including low-income, uninsured, racial/ethnic minority, and immigrant women, or among less educated women who might have lower awareness of the new recommendations.
MATERIALS AND METHODS
The NHIS is a nationally representative, in-person household survey of the civilian, noninstitutionalized US population. Annual surveys gather demographic, socioeconomic, and medical and health services use information about participants. The sampling design involves stratification and clustering and over sampling of specific subgroups. Within sampled households, 1 adult per family is randomly selected to complete the “sample adult” questionnaire. In select years (most recently, 2005, 2008, and 2011), this questionnaire inquired about cancer screening. Household response rates for the NHIS are high—86.5%, 84.9%, and 82.0%, respectively, for the 2005, 2008, and 2011 surveys. Among eligible adults in households that responded, sample response rates were 80.1%, 74.2% and 81.6%, respectively. Advance letters are sent to invited participants, and informed consent is obtained before questionnaire administration. The study protocol was considered exempt by the Brigham and Women's Hospital Institutional Review Board.
We identified 30,539 women aged ≥40 years who were asked questions about mammography in 2005, 2008, or 2011. We excluded 1304 women with prior breast cancer and 1406 women who did not answer or who answered they did not know to the mammography questions, leaving a final cohort of 27,829 women.
The 2011 NHIS questionnaire asked women: “Have you had a mammogram during the past 12 months?” The 2008 and 2005 questionnaires asked women: “Have you ever had a mammogram?” Women who had had a mammogram were asked: “When did you have your most recent mammogram?” Women who did not provide a complete date were then asked about the time interval since their last mammogram; we examined whether women had had a mammogram in the past 12 months.[14, 15]
The primary independent variables of interest were year and age (categorized as ages 40-49 years, 50-74 years, and ≥75 years). Other independent variables included self-reported race/ethnicity, immigration status, educational status, family income, and insurance, categorized as indicated in Table 1. For immigration status, respondents who reported not knowing whether they were born in the United States, those who refused to answer, those for whom an answer was not ascertained, and those w ho had no data recorded for years in the United States (n = 12), were recorded as born in the United States; those who reported not knowing how many years they had been in the United States (n = 67) were grouped with immigrants who had been in the United States for >10 years.
Table 1. Sample Characteristics According to Survey Year
Abbreviations: HMO, health maintenance organization; NHIS, National Health Interview Survey.
Population-based weights were provided by the NHIS. Percentages were based on unweighted data. Note that percentages may not sum to 100% because of rounding.
Born in US
Immigrant in US <10 y
Immigrant in US ≥10 y
Less than high school
High school graduate
Some college or associates degree
Family household income
Any private, including Medicare HMO
Other public insurance
We analyzed the data using population-standardized weights provided by the NHIS, which take into account the probability of an individual's selection and adjustment for nonresponse; standard errors were adjusted to account for the survey design. We used chi-square tests and Fisher exact tests to compare rates of screening in the past year for 2011 versus 2008 overall and by age group. We interpreted statistical significance as a 2-sided P value < .05, consistent with a type I error of 5%.
Next, we used logistic regression to assess mammography screening over time by age group. We tested the interaction of age group by year, specifically comparing mammography rates in 2011 with those in 2008 to compare rates after versus before the 2009 USPSTF recommendations. We also used these models to calculate case mix-adjusted rates of mammography by year for each age group and for preplanned subgroups of women who were black, Hispanic, immigrants, less educated, low income, uninsured, or publicly insured. All data were analyzed using SAS version 9.3 (SAS Institute, Inc., Cary, NC).
Table 1 lists the demographic characteristics of the 27,829 women in the cohort. Over half of women were ages 50 to 74 years, approximately 30% were of black or Hispanic race/ethnicity, and nearly 25% were college graduates. The overall unadjusted mammography rates increased over time and were higher in 2011 (54%) than in 2008 (52%; P = .03) (Table 2). Mammography rates were higher in 2011 than in 2008 for all age groups, although the differences were not statistically significant.
Table 2. Unadjusted Proportion of Women Reporting Mammography Screening in the Past Year in 2005, in 2008, and in 2011 Overall and by Age Group
P values were based on the chi-square test for comparison of screening rates in 2011 versus 2008 overall and in each age group. Results are weighted to reflect the US population based on the complex survey design.
Women ages 40-49 y
Women ages 50-74 y
Women aged ≥75 y
After adjustment for race, education, income, immigration status, and insurance level, adjusted rates of self-reported mammography in the last year increased slightly (from 51.9% in 2008 to 53.6% in 2011; P = .07) (Fig. 1). Compared with 2008, adjusted rates did not differ significantly in 2011 for women in any age group (all P ≥ .09) (Fig. 1).
Among all subgroups of women examined, including women who were black, Hispanic, immigrants, less educated, low income, uninsured, or publicly insured, there were no significant decreases in screening for women ages 40 to 49 years or ages 50 to74 years from 2008 to 2011 (all P ≥ .08, data not shown).
Our current study demonstrates that mammography screening rates did not decrease in any age group after publication of the USPSTF 2009 recommendations. These findings extend the early results of a study using Medical Expenditures Panel Survey data, which indicated that, from 2009 to 2010, self-reported rates of mammography in the past 2 years did not change significantly among women ages 40 to 49 years or ages 50 to 59 years. In our study, we observed no decreases in screening rates for women ages 40 to 49 years, despite the USPSTF recommendation against routine screening, or for women ages 50 to74 years, for whom the USPSTF recommended biennial (rather than annual) screening. These data suggest that neither of these recommendations has been widely adopted.
Several possible factors may explain the lack of a decrease in screening rates. First, uptake of the recommendations may be delayed, and our study may be too early to detect a decline. The adoption of new clinical guidelines can be delayed by lack of provider awareness[17, 18]; however, the widespread publicity around these recommendations makes this unlikely to account for our findings. Second, many providers who refer patients for mammography may disagree with the USPSTF recommendations. In a survey undertaken from June to December 2009, 84% of family physicians, 81% of internists, and 94% of obstetrician/gynecologists who responded reported recommending mammography to women ages 40 to 49 years. Since 2009, debate about the benefits and harms of mammography has continued in the medical literature, and mammography screening for younger women continues to be recommended by several prominent professional and advocacy organizations.[21-23] Third, patients may continue to request mammography starting at age 40 years or annually, because they either disagree with the recommendations, do not understand them, or are willing to accept the possible harms of mammography after reviewing the risks and benefits. In 1 survey of women in their 40s after publication of the 2009 USPSTF recommendations, 89% believed that women still should receive mammograms between ages 40 and 50 years; in another survey, only 20% of women could correctly identify the new recommendations. Patient preference may have a particularly profound impact on mammography use, because patients can self-refer for mammography screening independent of a physician's recommendation.
The clinical implications of widespread mammography screening, particularly in younger women, are controversial. Recent evidence suggests that the implementation of routine mammography screening over the past 30 years has incurred less significant mortality benefit and more frequent overdiagnoses than previously believed. It seems clear that, for many patients, mammography screening can lead to unnecessary, invasive diagnostic procedures and treatment for tumors that may not have become clinically significant. Furthermore, as noted by the USPSTF, the ratio of possible harms to benefits is greater in younger women. However, experts disagree about both the magnitude of benefits and harms of screening (particularly in younger women) and whether the potential harms should impact screening practices for low-risk women. Furthermore, the importance of false-positive results and overdiagnosis for an individual patient may depend on that individual's personal characteristics and values.
Strengths of this study include the nationally representative, large sample. Limitations include the possibility of nonresponse bias and the reliance of NHIS data on self-report, which may result in overestimation of mammography rates.[28, 29] These potential sources of bias would be unlikely to explain our findings, however, because response rates were similar across years, and presumably bias because of self-report would not change dramatically across years.
In conclusion, we observed no decrease in mammography rates for women aged >40 years after publication of the USPSTF recommendations in 2009. Whether through their impact on patients, providers, or both, it seems that the vigorous policy debates and coverage in the media and medical literature have limited adoption of these recommendations. Continued analysis of mammography trends with assessment of their clinical and cost effects will be essential to further understand the impact of the USPSTF recommendations and the benefits and harms of mammography screening in general.
This research was supported in part by the Global Women's Health Fellowship, Connors Center for Women's Health and Gender Biology, Brigham and Women's Hospital, Boston, Massachusetts (Dr. Pace). The funder had no role in the design or conduct of this study; in data collection, management, analysis, or interpretation; or in preparation, review, or approval of the article.