Mammography Use from 2000 to 2006: State-Level Trends with Corresponding Breast Cancer Incidence Rates

Women s Imaging Original Research Miller et al. -Level Trends in Mammography Use and Breast Cancer Incidence Women s Imaging Original Research FOCUS ON: Jacqueline W. Miller 1 Jessica B. King 2 A. Blythe Ryerson 1 Christie R. Eheman 2 Mary C. White 1 Miller JW, King JB, Ryerson AB, Eheman CR, White MC Keywords: breast cancer incidence, breast cancer screening, mammography, women s imaging DOI:10.2214/AJR.08.1757 Received September 3, 2008; accepted after revision October 1, 2008. The findings and conclusions of this article represent those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. 1 Epidemiology and Applied Research Branch, Division of Cancer Prevention and Control, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, 4770 Buford Hwy., NE, Mailstop K-55, Atlanta, GA 30341. Address correspondence to J. W. Miller (aci8@cdc.gov). 2 Cancer Surveillance Branch, Division of Cancer Prevention and Control, National Center for Chronic Disease Prevention and Health Promotion, Centers for Disease Control and Prevention, Atlanta, GA. AJR 2009; 192:352 360 0361 803X/09/1922 352 American Roentgen Ray Society WOMEN S IMAGING Mammography Use from 2000 to 2006: -Level Trends with Corresponding Breast Cancer Incidence Rates objective. Although breast cancer incidence and mortality rates have declined, a significant number of women are affected by this disease each year. Mammography is currently the most effective way to detect breast cancer at an early stage when it is most treatable, but there have been reports of decreasing or stagnant mammography use in the United s. For this study, we investigated the trend in mammography use for each state in comparison with the corresponding breast cancer incidence trend. materials and methods. We used data from the Behavioral Risk Factor Surveillance System for 2000, 2002, 2004, and 2006 to assess the percentage of women 40 years old who reported undergoing mammography within the past 2 years and data from the National Program of Cancer Registries and the Surveillance, Epidemiology, and End Results Program to assess breast cancer incidence rates from 2000 through 2004, the latest year for which data were available at the time of the study. results. The majority of states had a decreasing tendency in mammography use from 2000 to 2006. Only one state had a statistically significant increase in reported mammography use, whereas two states had significant decreases. There was a correlation between breast cancer incidence rates and mammography use by states (r = 0.6), but no correlation between the time trends in breast cancer incidence rates and mammography use was observed. conclusion. There was little statistically significant change in self-reported mammography use from 2000 to 2006. Continued monitoring of breast cancer screening practices and breast cancer incidence trends is important for targeting at-risk populations with effective interventions to improve breast cancer prevention and early detection. B reast cancer is the most commonly diagnosed cancer and the second leading cause of cancer deaths among women in the United s. Approximately 180,000 women are diagnosed with breast cancer and 40,000 die from the disease each year [1]. Although breast cancer incidence and mortality rates declined in 2002 through 2004 [2 4], a significant number of women are affected by this disease each year. Recent reports about the decline in breast cancer mortality have attributed the decline to both widespread use of mammography screening and improvements in cancer therapy [5 7]. Because systematic reviews of previous studies have shown that mammography use decreases breast cancer mortality, the United s Preventive Services Task Force (USPSTF) recommends screening mammography every 1 2 years for all women 40 years old [8, 9]. Regular screening with mammography is currently the best way to detect breast cancer at an early stage when it is most treatable. Although mammography use increased dramatically in the 1990s [10, 11], there have been more recent reports of declining mammography use from 2000 through 2005 across the United s [12, 13]. This decline has caused concern that delays in breast cancer diagnosis could result in an increase in breast cancer mortality. A more recent report using data from the Behavioral Risk Factor Surveillance System (BRFSS) from 2000 through 2006 found that the downward trend in mammography use was not persistent nationally but that use had significantly declined among women 40 59 years old and non-hispanic white women [14]. Our analysis further examines the BRFSS data by state to determine whether there are any significant changes on a more local level. This report provides detailed information for local stakeholders to develop and target interven- 352 AJR:192, February 2009

-Level Trends in Mammography Use and Breast Cancer Incidence tions that promote breast cancer screening in the appropriate populations. Materials and Methods Supported by the Centers for Disease Control and Prevention (CDC), the BRFSS is a large statebased telephone survey of the civilian, noninstitutionalized adult population 18 years old and older. The survey gathers information about health behaviors, chronic diseases, and preventive health practices from all 50 states, the District of Columbia, and three U.S. territories (Guam, Puerto Rico, and U.S. Virgin Islands). Respondents are selected with a random-digit-dialing tech nique and interviewed using a computer-assisted telephone interview program. Further details about the history and methodology have been published elsewhere [15]. The core survey includes a wo men s health module that is conducted every evennumbered year. Female respondents are asked whether they have ever had a mammogram. Those who respond affirmatively are then asked how long it has been since their last mammogram. For this study, BRFSS data from the 50 states and the District of Columbia were included for the years 2000, 2002, 2004, and 2006, except data from Hawaii for 2004, when they did not have available data. Recent mammography use was defined as having reported obtaining a mammogram within the past 2 years. All data were weighted to the respondents probability of selection and sex, age, and race and ethnicity from each state s census populations. Percentages were age-adjusted to the 2000 U.S. standard population using 5-year age groups. Percentage and standard errors (SEs) were calcu lated using SAS software (version 9.1, SAS Insti tute) and SUDAAN software (version 10.0, Research Triangle Institute), with SEs calculated using Taylor series linearization. Logis tic regression was used to assess for linear time trend. The time trend was considered significant if the beta coefficient for the year was nonzero at p 0.01. The sample sizes for each year were 65,774; 93,657; 125,195; and 156,982, respectively. The corres ponding state Council of American Research Organizations (CASRO) response rates ranged from 28.8% to 82.6%, and the cooperation rates for each state ranged from 35.5% to 99.8% for all years com bined [16]. We also assessed corresponding state breast cancer incidence rates. Because in situ breast cancers are more likely to be identified by mammography [17], we assessed the time trend for all breast cancers and for in situ and invasive breast cancers separately. Data were obtained from cancer registries participating in the CDC s National Program of Cancer Registries (NPCR) and the National Cancer Institute s Surveillance, Epidemi ology, and End Results Program (SEER) from 2000 through 2004, the latest year for which data were available at the time of the study. Registries from 44 states and the District of Columbia met the United s Cancer Statistics (USCS) criteria [18] for all years, covering 92.1% of the U.S. population. We calculated Pearson s correlation coefficient to assess for any correlation between mammography use percentages and breast cancer incidence rates for all years combined and between the time trends in mammography use and breast cancer incidence. Incidence rates were calculated per 100,000 women and age-adjusted to the 2000 U.S. standard population using 5-year age groups. Breast cancer cases were defined in women 40 years old or older at diagnosis with an in situ or invasive morph ology cancer. Incidence rates, SEs, annual percent changes, and percent changes were calculated using SEER*Stat software version 6.3.6 (National Cancer Institute) [19]. SEs were calculated using the Tiwari modification [20]. Results Figure 1 compares the percentage of women 40 years old who reported undergoing a mammogram within the past 2 years in 2000 and 2006. There was a lot of variability among the states. In general, there appears to be an increase in mammography utilization within the recommended interval among only a few states and a decrease in approximately 30% of the states. In 2006, the states with lower reported mammography use were generally in the mid to western portion of the country. Estimates of mammography utilization rates ranged from 65.7% to 83.8% by state for 2000 2006 combined (Table 1). In 2000, the age-adjusted percentage of women reporting mammography use within the past 2 years ranged from 65.9% to 86.5%. The variation among states narrowed in 2006, with the mammography rates ranging from 66.7% to 84.8%; however, this lower variation is due in part to a drop in the highest reported rates. Although mammography use in 17 states had an increasing tendency, ranging from an absolute increase of 0.4 8.6% from 2000 to 2006, this increase was significant in only one state (Tennessee). Thirty-four states and the District of Columbia reported a decreasing tendency in mammography use (range, 0.3% to 5.3% absolute change), with only New Mexico and Delaware reporting statistically significant declines. Of those states where mammography use decreased, most had the largest declines between 2002 and 2004 and slight increases from 2004 to Fig. 1 Age-adjusted percentage of women 40 years old who reported undergoing mammography within the past 2 years in the United s for 2000 and 2006. Data are from the Behavioral Risk Factor Surveillance System. AJR:192, February 2009 353

Miller et al. TABLE 1: Age-Adjusted a Percentages, with Standard Errors (SEs), of Women 40 Years Old Who Reported Undergoing Mammography Within the Past 2 Years by and Year, United s, 2000 2006 2000 (n = 65,774) 2002 (n = 93,657) 2004 (n = 125,195) 2006 (n = 156,982) Years Combined (n = 441,608) % SE % SE % SE % SE % SE Absolute % Change Alabama 74.5 1.7 78.8 1.3 75.2 1.2 76.7 1.3 76.3 0.7 2.2 Alaska 77.3 2.2 73.5 2.2 67.1 2.4 73.1 2.2 72.4 1.2 4.3 Arizona 80.2 2.1 77.4 1.6 74.5 1.6 76.8 1.6 77.1 0.9 3.4 Arkansas 71.1 1.5 68.3 1.3 66.5 1.2 69.0 1.1 68.7 0.6 2.1 California 78.4 1.4 76.2 1.5 76.5 1.2 78.0 1.1 77.3 0.6 0.3 Colorado 73.5 1.8 74.2 1.3 71.6 1.2 71.7 1.0 72.8 0.7 1.9 Connecticut 84.4 1.1 82.2 1.0 81.0 0.9 81.9 0.8 82.3 0.5 2.5 Delaware b 86.5 1.4 83.6 1.3 82.2 1.2 83.2 1.2 83.8 0.6 3.3 District of Columbia 82.5 1.8 81.3 1.8 80.3 1.5 81.2 1.2 81.4 0.8 1.3 Florida 78.0 1.1 77.5 1.1 75.2 1.1 77.0 0.9 76.9 0.5 0.9 Georgia 75.6 1.3 75.1 1.2 74.1 1.3 78.4 0.9 75.8 0.6 2.8 Hawaii 78.5 1.1 69.9 1.3 0.0 76.9 1.2 75.0 0.7 1.6 Idaho 65.9 1.4 66.5 1.2 63.4 1.2 66.7 1.2 65.7 0.6 0.8 Illinois 76.1 1.3 74.1 1.7 75.9 1.2 74.2 1.0 75.2 0.6 1.9 Indiana 72.8 1.5 73.1 1.0 68.9 1.0 71.3 1.0 71.5 0.6 1.5 Iowa 76.5 1.2 76.4 1.3 75.0 1.1 77.3 1.0 76.2 0.6 0.8 Kansas 75.9 1.2 75.7 1.1 75.9 0.8 74.3 0.8 75.4 0.5 1.7 Kentucky 75.0 1.2 75.8 1.1 75.2 1.1 74.7 1.2 75.1 0.6 0.4 Louisiana 74.8 1.1 75.9 1.2 73.5 0.9 75.2 0.9 74.8 0.5 0.4 Maine 79.1 1.6 81.6 1.3 81.6 1.1 81.4 1.0 81.0 0.6 2.3 Maryland 81.2 1.2 81.6 1.2 78.7 1.2 79.6 0.9 80.3 0.6 1.6 Massachusetts 83.9 0.8 83.1 0.9 82.5 0.9 84.8 0.7 83.5 0.4 0.9 Michigan 81.5 1.4 77.9 1.2 78.8 1.0 79.3 0.9 79.4 0.6 2.2 Minnesota 72.5 1.6 80.7 1.0 80.5 1.0 81.1 1.0 78.8 0.6 8.6 Mississippi 68.0 1.8 67.2 1.3 66.0 1.1 67.5 1.0 67.1 0.7 0.5 Missouri 74.2 1.4 74.0 1.3 68.6 1.3 70.8 1.4 71.8 0.7 3.4 Montana 74.1 1.6 72.1 1.5 71.5 1.2 71.1 1.1 72.2 0.7 3.0 Nebraska 75.8 1.4 75.2 1.2 76.0 0.8 72.8 1.0 74.9 0.6 3.0 Nevada 73.7 2.5 72.8 2.0 68.8 2.1 70.1 1.7 71.2 1.0 3.6 New Hampshire 81.5 1.5 80.1 1.0 79.9 1.0 78.8 0.9 80.0 0.6 2.6 New Jersey 76.3 1.3 77.6 1.6 74.6 0.7 78.0 0.7 76.6 0.6 1.6 New Mexico b 74.6 1.5 69.3 1.2 68.9 1.0 69.2 1.1 70.3 0.6 5.3 New York 80.4 1.3 78.9 1.2 75.1 1.1 79.1 0.9 78.3 0.6 1.3 North Carolina 77.7 1.4 80.8 1.1 77.1 0.7 78.7 0.6 78.6 0.5 1.0 North Dakota 75.6 1.8 77.0 1.3 71.8 1.4 76.9 1.1 75.3 0.7 1.3 Ohio 78.0 1.7 75.8 1.3 72.7 1.6 75.9 1.6 75.6 0.8 2.1 Oklahoma 69.3 1.4 69.3 1.0 67.1 1.0 66.8 1.0 68.1 0.6 2.5 Oregon 75.9 1.2 75.4 1.4 71.5 1.1 75.5 1.1 74.5 0.6 0.4 Pennsylvania 77.2 1.3 76.3 0.8 73.3 1.0 75.0 1.0 75.5 0.5 2.2 Note Data are from the Behavioral Risk Factor Surveillance System. Dash ( ) indicates that no data were available. a Age-adjusted to the 2000 U.S. standard population (19 age groups: census P25 1130 [31]) standard. b Statistically significant change. c Includes the 50 states and the District of Columbia but does not include the U.S. territories. (Table 1 continues on next page) 354 AJR:192, February 2009

-Level Trends in Mammography Use and Breast Cancer Incidence TABLE 1: Age-Adjusted a Percentages, with Standard Errors (SEs), of Women 40 Years Old Who Reported Undergoing Mammography Within the Past 2 Years by and Year, United s, 2000 2006 (continued) 2006. However, these increases did not offset the decrease in previous years, and mammography use remained below the 2000 level. Overall, there was no statistically significant change in the percentage of women who reported undergoing mammography in the past 2 years from 2000 to 2006 across the United s. Patterns of overall breast cancer (i.e., in situ and invasive cases combined) incidence rates are shown in Table 2. Among the states that met the USCS criteria, all but one had a total percentage decrease in overall breast cancer incidence rates from 2000 to 2004, ranging in magnitude from 0.2% to 21.1%. Among these states, 20 had a significant decrease in the annual percent change ranging from 2.0% to 6.8%. A majority of the decline in breast cancer incidence occurred between 2002 and 2004. However, there was no clear pattern among these states with regard to region of the country, average population age, average income, or population density. The annual percent change among the remaining 23 states and the District of Columbia had a downward trend, although it was not significant. There were very few similarities between the changes in invasive and in situ rates per state from 2000 to 2004 (Table 3). Invasive breast cancer rates showed a decline in the total percent change and the annual percent 2000 (n = 65,774) 2002 (n = 93,657) 2004 (n = 125,195) 2006 (n = 156,982) change overall, with 24 states having significantly decreasing annual percent change from 2000 to 2004. In contrast, 30 states and the District of Columbia showed an increase in total percent change among in situ cancers from 2000 to 2004, but only four of these states had a significantly increasing annual percent change. Of the states with a decreasing total percent change in the incidence of in situ breast cancer, only three had a significant decreasing annual percent change. With the data for all years combined, a correlation was observed between reported mammography use within the past 2 years (2000 through 2006) and overall breast cancer incidence per state (r = 0.6). This observation was mostly due to in situ breast cancers, with a correlation coefficient of 0.7, whereas invasive breast cancers had a correlation coefficient of 0.4. However, there was no correlation between the total percent change in mammography use and overall breast cancer incidence annual percent change (r = 0.2), in situ annual percent change (r = 0.2), or invasive annual percent change (r = 0.3) per state. To minimize the potential discrepancies between the years of data, we combined the BRFSS data for the years of 2000 through 2004 and found the same correlations between mammography use and overall breast cancer incidence (r = 0.6), in situ breast cancers (r = 0.7), and invasive breast cancers (r = Years Combined (n = 441,608) % SE % SE % SE % SE % SE Absolute % Change Rhode Island 83.4 1.1 85.1 1.0 82.0 1.1 84.6 0.9 83.8 0.5 1.2 South Carolina 78.4 1.4 75.1 1.4 71.3 1.0 73.8 0.9 74.5 0.6 4.6 South Dakota 75.7 1.1 75.6 1.1 75.5 1.0 74.0 1.0 75.1 0.5 1.7 Tennessee b 74.9 1.4 74.9 1.4 77.5 1.2 79.0 1.1 76.7 0.7 4.1 Texas 69.2 1.3 68.9 1.1 67.1 1.2 70.6 1.4 68.9 0.6 1.5 Utah 72.6 1.9 68.8 1.7 66.4 1.4 67.8 1.3 68.9 0.8 4.8 Vermont 77.2 1.2 76.6 1.1 74.8 0.9 78.9 0.8 76.9 0.5 1.7 Virginia 76.4 1.9 74.2 1.5 73.2 1.4 75.6 1.3 74.9 0.8 0.9 Washington 76.1 1.3 74.0 1.2 72.8 0.6 75.5 0.6 74.6 0.5 0.6 West Virginia 73.8 1.6 74.2 1.3 71.8 1.3 73.4 1.2 73.3 0.7 0.4 Wisconsin 74.8 1.5 79.8 1.1 75.3 1.2 77.7 1.1 76.9 0.6 2.9 Wyoming 66.9 1.7 67.8 1.4 67.6 1.2 68.1 1.1 67.7 0.7 1.2 United s c 76.6 0.3 76.0 0.3 74.2 0.2 76.1 0.2 75.7 0.1 0.4 Note Data are from the Behavioral Risk Factor Surveillance System. Dash ( ) indicates that no data were available. a Age-adjusted to the 2000 U.S. standard population (19 age groups: census P25 1130 [31]) standard. b Statistically significant change. c Includes the 50 states and the District of Columbia and does not include the U.S. territories. 0.4). However, there was a slight correlation between the total percent change in mammography use from 2000 to 2004 and in situ annual percent change (r = 0.5). Discussion This report expands the findings of a previous report by assessing the trends in mammography use within the past 2 years from 2000 to 2006 by state and by comparing changes in breast cancer incidence with changes in mammography use. We found that although few states had a statistically significant change in self-reported mammography use from 2000 to 2006, a majority of states had a decreasing tendency in mammography use. Only two states had statistically significant decreases, whereas one state had a statistically significant increase. These findings are a stark contrast to the trend in the 1990s that showed increases in mammography use across all 38 states participating in the BRFSS from 1989 to 1997 [10]. We found a correlation between the combined percentage of women reporting mammography use from 2000 to 2006 (and from 2000 to 2004) to the combined breast cancer incidence rates from 2000 to 2004 within each state. This relationship was mostly because of the correlation between in situ breast cancer incidence and mammography use. This finding is expected because the most AJR:192, February 2009 355

Miller et al. TABLE 2: Age-Adjusted Incidence Rates a, with Standard Errors (SEs), of Breast Cancer b Among Women 40 Years Old by and Year, United s, 2000 2004 2000 2001 2002 2003 2004 2000 2004 Rate SE Rate SE Rate SE Rate SE Rate SE Rate SE Total % Change c Alabama 305.9 5.3 315.1 5.4 315.2 5.3 289.0 5.1 291.9 5.1 303.3 2.3 4.6 1.8 Alaska 391.5 20.9 369.9 20.0 356.8 19.2 370.4 19.2 329.3 17.7 362.5 8.7 15.9 3.4 Arizona 325.8 5.3 329.0 5.3 317.2 5.1 287.1 4.8 272.7 4.6 305.5 2.2 16.3 4.8 e Arkansas 325.7 7.1 318.1 7.0 327.8 7.0 299.4 6.6 277.2 6.4 309.4 3.0 14.9 3.7 California 355.3 2.3 352.3 2.2 350.1 2.2 323.4 2.1 317.9 2.0 339.3 1.0 10.5 3.0 e Colorado 369.7 6.5 362.5 6.4 360.8 6.3 353.9 6.1 330.3 5.9 354.9 2.8 10.6 2.4 e Connecticut 411.3 7.0 419.0 7.0 395.5 6.8 375.7 6.5 374.7 6.5 394.7 3.0 8.9 2.9 e Delaware 364.2 13.9 365.7 13.8 356.4 13.5 349.3 13.2 334.0 12.8 353.6 6.0 8.3 2.2 e District of Columbia 375.1 16.7 400.0 17.2 357.9 16.4 337.5 16.1 363.5 16.9 366.7 7.4 3.1 2.3 Florida 335.4 2.8 341.2 2.8 333.7 2.7 305.5 2.6 295.8 2.5 321.8 1.2 11.8 3.6 e Georgia 334.8 4.5 340.3 4.4 333.9 4.3 329.0 4.2 329.3 4.2 333.3 1.9 1.6 0.7 Hawaii 328.3 10.8 377.3 11.4 374.3 11.2 331.4 10.4 331.2 10.3 347.8 4.8 0.9 1.3 Idaho 348.3 11.3 328.4 10.9 348.9 11.0 310.3 10.3 294.8 9.9 325.6 4.8 15.3 3.8 Illinois 358.6 3.6 362.4 3.6 345.6 3.5 322.1 3.3 325.3 3.3 342.4 1.5 9.3 3.1 e Indiana 331.4 4.9 346.6 4.9 341.9 4.9 308.5 4.6 292.7 4.5 323.8 2.1 11.7 3.6 Iowa 334.5 6.8 342.8 6.9 343.1 6.9 320.4 6.6 322.9 6.6 332.5 3.0 3.5 1.4 Kansas 353.7 7.6 353.4 7.6 346.9 7.4 327.3 7.2 323.6 7.1 340.6 3.3 8.5 2.5 e Kentucky 326.8 5.8 332.4 5.8 322.4 5.7 313.0 5.6 303.6 5.5 319.4 2.5 7.1 2.1 e Louisiana 319.8 5.6 319.9 5.6 312.3 5.5 310.1 5.5 314.2 5.5 315.3 2.5 1.8 0.7 Maine 359.9 10.4 386.5 10.7 360.5 10.2 355.9 10.0 338.8 9.7 359.9 4.6 5.9 2.1 Maryland Massachusetts 428.5 5.3 408.0 5.1 405.1 5.1 392.1 5.0 391.7 5.0 405.0 2.3 8.6 2.2 e Michigan 370.5 4.0 370.1 4.0 369.3 4.0 346.4 3.8 326.2 3.7 356.0 1.7 12.0 3.1 e Minnesota 378.9 5.9 369.0 5.8 357.5 5.6 339.0 5.4 336.6 5.4 355.5 2.5 11.1 3.2 e Mississippi Missouri 339.2 5.0 332.6 4.9 334.4 4.9 320.4 4.8 311.7 4.7 327.3 2.2 8.1 2.0 e Montana 363.4 12.9 366.2 12.8 331.2 12.1 340.8 12.1 314.9 11.6 342.6 5.5 13.4 3.5 e Nebraska 366.0 9.7 370.6 9.6 365.4 9.5 328.7 9.0 332.0 9.0 352.3 4.2 9.3 3.1 Nevada 350.0 9.1 363.5 9.1 339.8 8.6 289.0 7.8 276.2 7.4 321.8 3.7 21.1 6.8 e New Hampshire 397.0 11.8 408.2 11.8 393.3 11.4 362.4 10.8 358.9 10.6 383.1 5.0 9.6 3.2 New Jersey 376.0 4.3 376.1 4.3 362.8 4.1 353.6 4.1 360.9 4.1 365.6 1.9 4.0 1.4 New Mexico 307.9 8.7 298.0 8.5 297.1 8.4 273.4 7.9 287.7 8.0 292.3 3.7 6.6 2.2 New York 361.1 2.8 348.7 2.7 341.0 2.7 345.8 2.7 342.3 2.7 347.5 1.2 5.2 1.2 North Carolina 333.6 4.2 335.7 4.2 316.4 4.0 309.6 4.0 308.0 3.9 320.2 1.8 7.7 2.4 e North Dakota Ohio 343.8 3.5 341.4 3.5 337.1 3.5 320.3 3.4 321.5 3.4 332.4 1.5 6.5 2.0 e Note Data are from the National Program of Cancer Registries and Surveillance, Epidemiology, and End Results Program. Dash ( ) indicates that statistic could not be calculated. APC = annual percent change. a Rates are per 100,000 and age-adjusted to the 2000 U.S. standard population (19 age groups: census P25 1130 [31]) standard; SEs are 95% for rates (Tiwari modification) and trends. b Breast cancer includes both in situ and invasive cancers. c Percent changes were calculated using 1 year for each end point. d APCs were calculated using weighted least-squares method. e The APC is significantly different from zero (p < 0.05). f United s refers to the states meeting United s Cancer Statistics criteria for all years, 2000 2004, and covers 92.1% of the U.S. population. (Table 2 continues on next page) APC d 356 AJR:192, February 2009

-Level Trends in Mammography Use and Breast Cancer Incidence TABLE 2: Age-Adjusted Incidence Rates a, with Standard Errors (SEs), of Breast Cancer b Among Women 40 Years Old by and Year, United s, 2000 2004 (continued) 2000 2001 2002 2003 2004 2000 2004 Rate SE Rate SE Rate SE Rate SE Rate SE Rate SE Total % Change c Oklahoma 335.3 6.4 335.6 6.4 332.8 6.3 316.7 6.1 334.5 6.2 331.0 2.8 0.2 0.6 Oregon 392.4 7.0 398.5 7.0 382.3 6.8 348.5 6.4 354.4 6.4 374.8 3.0 9.7 3.3 e Pennsylvania 353.1 3.3 356.1 3.3 347.1 3.3 330.9 3.2 338.3 3.2 344.9 1.5 4.2 1.6 Rhode Island 372.9 12.0 356.7 11.7 355.4 11.6 321.7 11.0 368.2 11.7 355.0 5.2 1.3 1.2 South Carolina 328.6 5.9 328.8 5.8 327.4 5.8 321.3 5.6 302.2 5.4 321.3 2.5 8.0 1.9 South Dakota Tennessee Texas 317.3 2.8 316.6 2.7 321.6 2.7 302.7 2.6 292.4 2.5 309.7 1.2 7.9 2.1 Utah 322.4 9.3 325.7 9.3 304.5 8.8 291.2 8.5 300.0 8.5 308.3 4.0 7.0 2.5 Vermont 378.8 16.0 377.3 15.7 372.9 15.5 361.8 15.2 341.5 14.5 365.2 6.9 9.8 2.5 e Virginia Washington 400.2 5.6 405.6 5.6 406.3 5.5 375.2 5.2 362.4 5.1 389.4 2.4 9.4 2.7 West Virginia 309.8 8.0 314.1 8.0 300.4 7.8 301.6 7.8 309.1 7.9 306.6 3.5 0.2 0.5 Wisconsin 358.7 5.4 363.5 5.4 340.0 5.1 324.3 5.0 320.0 4.9 340.9 2.3 10.8 3.4 e Wyoming 345.9 17.6 284.9 15.9 334.1 17.0 319.7 16.4 313.1 16.1 319.1 7.4 9.5 1.0 United s f 350.7 0.8 350.6 0.8 344 0.7 325.9 0.7 321.4 0.7 338.2 0.3 8.3 2.4 e Note Data are from the National Program of Cancer Registries and Surveillance, Epidemiology, and End Results Program. Dash ( ) indicates that statistic could not be calculated. APC = annual percent change. a Rates are per 100,000 and age-adjusted to the 2000 U.S. standard population (19 age groups: census P25 1130 [31]) standard; SEs are 95% for rates (Tiwari modification) and trends. b Breast cancer includes both in situ and invasive cancers. c Percent changes were calculated using 1 year for each end point. d APCs were calculated using weighted least-squares method. e The APC is significantly different from zero (p < 0.05). f United s refers to the states meeting United s Cancer Statistics criteria for all years, 2000 2004, and covers 92.1% of the U.S. population. common form of in situ breast cancers is usually detected by mammography [17]. We found significant differences in the direction of change for in situ versus invasive breast cancer rates. Whereas invasive breast cancer tended to decrease throughout all states, in situ breast cancer increased in the majority of states. There was also no correlation between the directions or magnitudes of changes in incidence rates of invasive compared with in situ breast cancer. Overall, an increase in the incidence rates of in situ breast cancer with a corresponding drop in invasive breast cancer rates could be a marker for the consistent use of mammography. Regular mammography use should be associated with fewer invasive breast cancer diagnoses and more in situ diagnoses. However, this analysis is an ecologic comparison at the state level and cannot detect associations that exist at an individual level. Individual differences among women in a given population, such as sociodemographic characteristics, individual risk factors, and environmental or genetic risk factors, may have enough influence on both breast cancer incidence and mammography use that any associations may be difficult to detect. This study is subject to several limitations. First, the BRFSS collects self-reports of mam mography use that are not validated, and respondents may not accurately recall the length of time since their last mammogram. However, this factor should not affect the degree of change when assessing a trend over time. The results of a study using BRFSS questions showed that self-reports of the time interval since most recent mammogram were fairly accurate [21]. Second, although BRFSS is one of the largest surveys conducted, the sample sizes for women 40 years old and older in many states were too small to provide adequate statistical power to detect even modest changes in mammography use as statistically significant. Third, the USPSTF recommends screening every 1 2 years for women 40 years old and older [9], and the American Cancer Society recommends annual screening for breast cancer beginning at the age of 40 years [22]. APC d We examined reports of mammography use within the last 2 years, the upper limit of the USPSTF-recommended screening interval. Overall, the proportion of women who reported having undergone mammography within the past 2 years on BRFSS was substantially higher than the proportion reported for just the past year [23]. Previous research has indicated that some overreporting of mammogram use occurs when women underestimate the length of time since their last mammogram [24]. Therefore, these data may overestimate the proportion of women who were truly up-to-date with breast cancer screening. Fourth, because the survey is administered by landline telephones, the data may not be truly representative of low-income women who are more likely not to have consistent telephone service. Last, the breast cancer incidence data after 2004 were not available. Hence, the true relationship examined between mammography utilization and breast cancer incidence rates is not exact because lack of mammography use resulting AJR:192, February 2009 357

Miller et al. TABLE 3: Age-Adjusted Incidence Rates a, with Standard Errors (SEs), of In Situ and Invasive Breast Cancer Among Women 40 Years Old by and Year, United s, 2000 and 2004 In Situ Cancer Invasive Cancer 2000 2004 Total % b 2000 2004 Total Rate SE Rate SE Change APC c Rate SE Rate SE Total % Change b Alabama 53.1 2.2 56.6 2.2 6.5 1.2 252.8 4.8 235.3 4.5 6.9 2.5 Alaska 62.5 8.0 81.3 8.7 30.0 4.5 328.9 19.3 248.0 15.4 24.6 5.3 Arizona 55.8 2.2 47.7 1.9 14.4 4.8 d 270.0 4.8 225.0 4.2 16.7 4.8 d Arkansas 56.6 3.0 48.2 2.7 14.9 3.0 269.0 6.4 229.0 5.8 14.9 3.8 California 60.3 0.9 60.0 0.9 0.5 0.5 295.0 2.1 257.9 1.8 12.6 3.6 d Colorado 69.8 2.8 71.7 2.7 2.7 1.0 299.8 5.9 258.6 5.2 13.8 3.3 d Connecticut 100.0 3.5 90.2 3.2 9.8 2.1 d 311.3 6.1 284.5 5.7 8.6 3.2 Delaware 68.9 6.1 81.8 6.3 18.8 4.0 d 295.3 12.5 252.2 11.1 14.6 3.8 d District of Columbia 72.2 7.4 80.5 8.0 11.5 1.1 302.9 15.0 283.0 14.8 6.6 2.6 Florida 59.6 1.2 60.3 1.1 1.1 0.5 275.8 2.5 235.5 2.2 14.6 4.3 d Georgia 59.7 1.9 64.5 1.9 8.1 1.8 275.1 4.0 264.7 3.8 3.8 1.2 d Hawaii 70.7 5.0 75.5 5.0 6.7 1.0 257.6 9.5 255.7 9.0 0.7 1.4 Idaho 52.6 4.4 60.1 4.4 14.3 3.9 295.7 10.4 234.7 8.8 20.6 5.3 Illinois 68.0 1.6 68.3 1.5 0.4 0.7 290.6 3.2 257.0 2.9 11.5 3.6 d Indiana 54.9 2.0 57.1 2.0 3.9 0.0 276.5 4.4 235.7 4.0 14.8 4.3 Iowa 57.7 2.9 62.1 2.9 7.6 2.3 276.8 6.2 260.8 5.9 5.8 2.1 Kansas 54.0 3.0 64.5 3.2 19.5 4.9 d 299.7 7.0 259.1 6.3 13.5 3.9 d Kentucky 49.7 2.3 52.3 2.3 5.1 0.6 277.1 5.4 251.4 5.0 9.3 2.6 d Louisiana 42.9 2.1 53.0 2.2 23.4 3.6 276.9 5.2 261.2 5.0 5.6 1.4 d Maine 61.6 4.4 70.0 4.5 13.6 2.1 298.3 9.5 268.8 8.6 9.9 3.1 Maryland Massachusetts 110.8 2.7 99.2 2.5 10.5 1.9 317.7 4.5 292.6 4.3 7.9 2.3 d Michigan 74.7 1.8 67.7 1.7 9.3 2.4 295.9 3.6 258.5 3.3 12.6 3.3 d Minnesota 63.9 2.4 64.9 2.4 1.5 0.8 314.9 5.4 271.7 4.8 13.7 4.0 d Mississippi Missouri 55.3 2.0 55.6 2.0 0.5 0.1 283.9 4.6 256.1 4.2 9.8 2.4 d Montana 65.4 5.5 69.7 5.5 6.6 0.9 298.1 11.7 245.2 10.2 17.7 4.6 d Nebraska 67.7 4.2 61.6 3.9 9.1 2.8 d 298.3 8.7 270.4 8.1 9.4 3.1 Nevada 58.2 3.7 44.7 3.0 23.2 8.3 291.8 8.3 231.6 6.8 20.6 6.5 d New Hampshire 79.1 5.3 79.1 5.0 0.0 0.5 317.9 10.5 279.9 9.4 12.0 3.9 New Jersey 74.8 1.9 79.2 1.9 5.9 1.8 301.2 3.8 281.7 3.6 6.5 2.3 New Mexico 46.5 3.4 54.3 3.5 16.7 4.4 261.3 8.0 233.3 7.2 10.7 3.4 d New York 74.7 1.3 76.3 1.3 2.3 1.0 286.4 2.5 265.9 2.4 7.2 1.7 d North Carolina 57.9 1.8 57.5 1.7 0.6 0.7 275.7 3.8 250.5 3.5 9.1 3.0 d North Dakota Ohio 60.7 1.5 62.8 1.5 3.5 0.0 283.1 3.2 258.6 3.0 8.6 2.4 d Note Data are from National Program of Cancer Registries and Surveillance, Epidemiology, and End Results Program. Dash ( ) indicates that statistic could not be calculated. APC = annual percent change. a Rates are per 100,000 and age-adjusted to the 2000 U.S. standard population (19 age groups: census P25 1130 [31]) standard; SEs are 95% for rates (Tiwari modification) and trends. b Percent changes were calculated using 1 year for each end point. c APCs were calculated using weighted least-squares method. d The APC is significantly different from zero (p < 0.05). e United s refers to the states meeting United s Cancer Statistics criteria for all years, 2000 2004, and covers 92.1% of the U.S. population. (Table 3 continues on next page) Total APC c 358 AJR:192, February 2009

-Level Trends in Mammography Use and Breast Cancer Incidence TABLE 3: Age-Adjusted Incidence Rates a, with Standard Errors (SEs), of In Situ and Invasive Breast Cancer Among Women 40 Years Old by and Year, United s, 2000 and 2004 (continued) in increased cancer rates is likely to be a delayed effect. This study found that mammography use showed only slight variability in each state for the years 2000 2006. Of even more importance, this study identifies a relative stagnation in mammography use among women 40 years old. The lack of increase in women being regularly screened for breast cancer is of concern. Women who are not undergoing regular mammography screening are at increased risk of having breast cancer detected at a later stage when treatment is less effective. Understanding why some women are still not getting the recommended screening is critical in the continued battle against breast cancer mortality. Previous studies reported that women without health insurance and women who did not visit their primary care physician within the past year were more likely to not undergo mammography screening [25]. Access to care and health care provider recommendation have already been documented as being important predictors of women getting a mammogram [26, 27]. In Situ Cancer Continued research to monitor breast cancer screening practices and breast cancer incidence trends is important for targeting at-risk populations with effective interventions to improve prevention and early detection. This research also includes understanding issues about access to mammography, women s knowledge of the importance of mammography screening, and why specific groups of women are more likely not to get regularly screened. Broadening of public health programs that target these at-risk women is a key intervention step. One such program is CDC s National Breast and Cervical Cancer Early Detection Program that provides free or lowcost mammography to low-income, uninsured, and underinsured wo men in all 50 states, the District of Columbia, 12 tribes and tribal organizations, and five territories. Although this program has provided screening services to more than 3 million women since its inception in 1991 [28], it currently reaches only about 15% of the eligible population [29]. Another such program is CDC s Comprehensive Cancer Control Program from which Invasive Cancer 2000 2004 Total % b 2000 2004 Total Rate SE Rate SE Change APC c Rate SE Rate SE Total % Change b Oklahoma 49.4 2.5 60.3 2.7 21.9 5.2 d 285.8 5.9 274.2 5.6 4.1 1.7 Oregon 66.1 2.9 62.6 2.7 5.3 2.6 326.3 6.4 291.8 5.8 10.6 3.5 d Pennsylvania 65.6 1.5 69.8 1.5 6.3 1.2 287.5 3.0 268.5 2.9 6.6 2.2 d Rhode Island 79.6 5.7 84.2 5.7 5.8 0.1 293.3 10.6 284.0 10.3 3.2 1.4 South Carolina 59.3 2.5 56.6 2.4 4.6 1.0 269.3 5.3 245.6 4.9 8.8 2.1 South Dakota Tennessee Texas 53.8 1.1 53.0 1.1 1.5 0.3 263.5 2.5 239.4 2.3 9.2 2.4 Utah 53.7 3.8 53.6 3.6 0.1 0.4 268.7 8.5 246.4 7.7 8.3 3.0 Vermont 84.0 7.5 63.6 6.3 24.3 3.9 294.8 14.1 277.9 13.1 5.7 2.0 Virginia Washington 72.8 2.4 75.8 2.3 4.1 0.6 327.4 5.0 286.6 4.5 12.4 3.5 West Virginia 51.5 3.3 57.5 3.4 11.8 2.6 258.3 7.3 251.5 7.1 2.6 1.1 Wisconsin 48.5 2.0 66.9 2.3 37.8 6.7 d 310.2 5.0 253.1 4.4 18.4 5.3 d Wyoming 51.5 6.8 59.3 7.0 15.1 5.1 294.4 16.3 253.8 14.5 13.8 2.2 United s e 63.6 0.3 64.4 0.3 1.3 0.1 287.1 0.7 257 0.6 10.5 3.0 d Note Data are from National Program of Cancer Registries and Surveillance, Epidemiology, and End Results Program. Dash ( ) indicates that statistic could not be calculated. APC = annual percent change. a Rates are per 100,000 and age-adjusted to the 2000 U.S. standard population (19 age groups: census P25 1130 [31]) standard; SEs are 95% for rates (Tiwari modification) and trends. b Percent changes were calculated using 1 year for each end point. c APCs were calculated using weighted least-squares method. d The APC is significantly different from zero (p < 0.05). e United s refers to the states meeting United s Cancer Statistics criteria for all years, 2000 2004, and covers 92.1% of the U.S. population. Total APC c all states receive funding to develop a coordinated and comprehensive approach to reducing cancer burden [30]. One of the program s priorities is early detection of breast cancer. The variability we observed in mammography use from state to state indicates that efforts to improve mammography use should be focused at individual state or local levels and should target specifically identified populations to ensure that all women 40 years old undergo the recommended breast cancer screening. Acknowledgments We acknowledge the contributions of the BRFSS state coordinators and the state and regional cancer registry staffs for their work in collecting the data used in this study. References 1. U.S. Cancer Statistics Working Group. United s cancer statistics: 2004 incidence and mortality. Atlanta, GA: Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute, 2007 2. Espey DK, Wu XC, Swan J, et al. Annual report to AJR:192, February 2009 359

Miller et al. the nation on the status of cancer, 1975 2004: featuring cancer in American Indians and Alaska Natives. Cancer 2007; 110:2119 2152 3. Jemal A, Ward E, Thun MJ. Recent trends in breast cancer incidence rates by age and tumor characteristics among U.S. women. Breast Cancer Res 2007; 9:R28 4. Centers for Disease Control and Prevention. Decline in breast cancer incidence: United s, 1999 2003. MMWR 2007; 56:549 553 5. Tabar L, Yen MF, Chen HT, Smith RA, Duffy SW. Mammography service screening and mortality in breast cancer patients: 20-year follow-up before and after introduction of screening. Lancet 2003; 361:1405 1410 6. Das B, Feuer EJ, Mariotto A. Geographic association between mammography use and mortality reduction in the US. Cancer Causes Control 2005; 16:691 699 7. Berry DA, Inoue L, Shen Y, et al. Modeling the impact of treatment and screening on U.S. breast cancer mortality: a Bayesian approach. J Natl Cancer Inst Monogr 2006; 36:30 36 8. Humphrey LL, Helfand M, Chan BKS, Woolf SH. Breast cancer screening: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2002; 137:347 360 9. U.S. Preventive Services Task Force. Screening for breast cancer: recommendations and rationale. Ann Intern Med 2002; 137(5 Pt 1):344 346 10. Blackman DK, Bennett EM, Miller DS. Trends in self-reported use of mammograms (1989 1997) and Papanicolaou tests (1991 1997): Behavioral Risk Factor Surveillance System. MMWR CDC Surveill Summ 1999; 48(6):1 22 11. Swan J, Breen N, Coates RJ, Rimer BK, Lee NC. Progress in cancer screening practices in the United s. Cancer 2003; 97:1528 1540 12. Centers for Disease Control and Prevention. Use of mammograms among women aged > or = 40 years: United s, 2000 2005. MMWR 2007; 56:49 51 13. Breen N, Cronin KA, Meissner HI, et al. Reported drop in mammography. Cancer 2007; 109: 2405 2409 14. Ryerson AB, Miller JW, Eheman CR, Leadbetter S, White MC. Recent trends in U.S. mammography use from 2000 2006: a population-based analysis. Prev Med 2008; 47:477 482 15. Hughes E, McCracken M, Roberts H, et al. Surveillance for certain health behaviors among states and selected local areas: Behavioral Risk Factor Surveillance System, United s, 2004. MMWR Surveill Summ 2006; 55(7):1 124 16. Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System summary data quality reports. Atlanta, GA: Department of Health and Human Services. www.cdc. gov/brfss/technical_infodata/quality.htm. Access ed July 31, 2008 17. Ernster VL, Barclay J. Increases in ductal carcinoma in situ (DCIS) of the breast in relation to mammography: a dilemma. J Natl Cancer Inst Monogr 1997; 22:151 156 18. U.S. Cancer Statistics Working Group. United s Cancer Statistics publication criteria. USCS technical notes. Atlanta, GA: Centers for Disease Control and Prevention and National Cancer Institute, 2004. www.cdc.gov/cancer/ npcr/npcrpdfs/uscs_ 2004_technical_notes.pdf. Accessed July 31, 2008 19. SEER*Stat Software. National Cancer Institute: Surveillance, Epidemiology, and End Results. seer.cancer.gov/seerstat/. Published April 24, 2008. Accessed June 10, 2008 20. Tiwari RC, Clegg LX, Zou Z. Efficient interval estimation for age-adjusted cancer rates. Stat Methods Med Res 2006; 15:547 569 21. Caplan LS, McQueen DV, Qualters JR, Leff M, Garrett C, Calonge N. Validity of women s selfreports of cancer screening test utilization in a managed care population. Cancer Epidemiol Biomarkers Prev 2003; 12:1182 1187 22. Smith RA, Saslow D, Sawyer KA, et al.; American Cancer Society High-Risk Work Group; American Cancer Society Screening Older Women Work Group; American Cancer Society Mammography Work Group; American Cancer Society Physical Examination Work Group; American Cancer Society New Technologies Work Group; American Cancer Society Breast Cancer Advisory Group. American Cancer Society guidelines for breast cancer screening: update 2003. CA Cancer J Clin 2003; 53:141 169 23. Smith RA, Cokkinides V, Brawley OW. Cancer screening in the United s, 2008: a review of current American Cancer Society guidelines and cancer screening issues. CA Cancer J Clin 2008; 58:161 179 24. Rauscher GH, Johnson TP, Cho YI, Walk JA. Accuracy of self-reported cancer-screening histories: a meta-analysis. Cancer Epidemiol Biomarkers Prev 2008; 17:748 757 25. Meissner HI, Breen N, Taubman ML, Vernon SW, Graubard BI. Which women aren t getting mammograms and why? (United s). Cancer Causes Control 2007; 18:61 70 26. Zapka JG, Puleo E, Taplin SH, et al. Processes of care in cervical and breast cancer screening and follow-up: the importance of communication. Prev Med 2004; 39:81 90 27. Snell JL, Buck EL. Increasing cancer screening: a meta-analysis. Prev Med 1996; 25:702 707 28. Centers for Disease Control and Prevention. National Breast and Cervical Cancer Early Detection Program. Updated August 29, 2008. www.cdc. gov/cancer/nbccedp/about.htm. Accessed July 31, 2008 29. Tangka FK, Dalaker J, Chattopadhyay SK, et al. Meeting the mammography screening needs of underserved women: the performance of the National Breast and Cervical Cancer Early Detection Program in 2002 2003 (United s). Cancer Causes Control 2006; 17:1145 1154 30. Centers for Disease Control and Prevention. National Comprehensive Cancer Control Program (NCCCP). www.cdc.gov/cancer/ncccp/about.htm. Accessed July 31, 2008 31. Day JC. Population projections of the United s by age, sex, race, and Hispanic origin: 1995 to 2050, U.S. Bureau of the Census, Current Population Reports, P25 1130. U.S. Government Printing Office, Washington, DC, 1996 360 AJR:192, February 2009