The Impact of Mammography Screening Guideline Changes in a Universally Insured Population

Abstract

Background:

The U.S. Preventive Services Task Force (USPSTF) modified breast cancer screening guidelines in November 2009. The impact has been studied among privately and Medicare insured populations, but not among universally insured women.

Materials and Methods:

This study compared the proportion of TRICARE beneficiaries aged 40–64 receiving mammograms from fiscal years 2006 to 2015 using an interrupted time series analysis to determine the impact of the 2009 USPSTF guideline changes. Stratified analyses evaluated differences by age (ages 40–49, 50–64), race, care setting, beneficiary type, and military status.

Results:

The proportion of women receiving mammograms increased from October 2005 through September 2009. A small, but significant decrease of 65–66 fewer women screened per 10,000 occurred in the first quarter of 2010 (October 1 to December 31) following the screening guideline update publication. The proportion screened then remained unchanged through 2015. Comparative analysis revealed no differences in impact between age groups, blacks and whites, or military dependents and active-duty/retirees.

Conclusions:

This study determined that the USPSTF guideline updates had a small, but immediate and lasting impact that was not different across age groups, beneficiary type, or race. No racial disparities in the proportion screened or in the impact of the guideline change were noted in our universally insured population.

Introduction

One in eight women in the United States will receive a diagnosis of breast cancer in her lifetime.¹ Breast cancer is the leading incident cancer in women in the United States, after nonmelanoma skin cancer, and the second leading cause of cancer death for women behind lung cancer.² Routine screening via mammography is credited with partially increasing breast cancer survival by identifying the disease process at early, more treatable stages.^3–5

Recommendations vary between medical organizations regarding how often women of average risk should receive screening mammography. In November 2009 following an in-depth review of available evidence on the harms and benefits of mammography screening, the U.S. Preventive Services Task Force (USPSTF), an independent panel of medical experts using evidence-based medicine, changed its screening recommendations from either annual or biennial screening beginning at age 40 to the current recommendation of routine biennial screening beginning at age 50 until age 75.^6,7 For women under 50, the new guideline stressed shared decision making between women and their health care providers to determine the appropriate onset and frequency of mammography. In 2011, the American College of Obstetricians and Gynecologists strengthened their 2003 recommendation for either annual or biennial screenings from age 40 to annual screenings after reviewing the available evidence.^8,9 The American College of Radiology and the Society for Breast Imaging also recommend annual screening beginning at age 40.¹⁰ The American Cancer Society recommended in 2003 annual screening beginning at age 40,¹¹ with an updated recommendation in late 2015 for annual screening from ages 45–54 and biennial screening thereafter.¹²

While mammography has traditionally been accepted as an essential public health success in the reduction of deaths from breast cancer, the risks versus benefits of specific screening schedules have been hotly debated, highlighting the need for more research to satisfy the question of the most appropriate breast cancer screening guidelines.^3,13–16 Although studies conducted shortly after the published USPSTF guideline change showed little change in mammography utilization, the majority of studies comparing rates before and after the November 2009 recommendation change demonstrated a decrease in screenings.^17–25 This impact of the USPTF guideline change has not been addressed in a universally insured population.

Military service members, retirees, and their dependent family members have universal access to health care through the Department of Defense (DoD) health insurance plan, TRICARE. Coverage is provided to all members regardless of socioeconomic status, job title, benefit status (military service member, retiree, dependent), or health status. Research has shown that even among individuals with private health care coverage, there still exists racial differences in access to care,²⁶ which may be related to substantial out of pocket cost not seen in the Military Health System (MHS). Beginning at age 40, and age 30 for women of increased risk, TRICARE allows full coverage for screening mammograms.²⁷ Because the MHS includes the medical departments of the Army, Navy, and Air Force, it currently does not have a shared set of screening guidelines that providers must follow.²⁸ The population of TRICARE recipients is demographically representative of the overall U.S. population under age 65 years.^29–32 This large population offers the unique opportunity to determine the influence of the new screening recommendation on the care received by a nationally representative sample of American women without the effects of differential access. This study examined changes in mammogram screening rates and trends, overall and by demographic subgroups, following the introduction of the new screening recommendations. We expected that screening rates would have declined after the new recommendation and that the magnitude of the change would differ by age but not by race, benefit type, or care setting.

Materials and Methods

Study design

This retrospective cohort study spanned a surveillance period from October 2005 through September 2015 (fiscal years [FY] 2006 through 2015). We queried the Military Health System Data Repository (MDR) to identify all TRICARE Prime/Prime Plus female beneficiaries aged 40–64 years during the study period. The population included both active-duty and retired military members, activated members of the National Guard and Reserves, and civilian dependents of active-duty, retired military members, and activated members of the National Guard and Reserves.

Using the MDR, we obtained all medical encounters and administrative claims for diagnostic and screening mammograms for our study population in both the direct and purchased care setting. Direct care refers to care received at military treatment facilities (MTFs) within the MHS whereas care received at civilian fee-for-service facilities through TRICARE benefits are known as purchased care. The MHS is separate from and does not include care delivered at Veteran's Affairs facilities, nor care delivered in combat zones. This research was considered exempt by the Uniformed Services University of the Health Sciences (USUHS) Institutional Review Board.

Outcomes and measures

Our outcome measure was mammography screening rate, defined as the proportion of women screened quarterly, based on nationally recognized standards used by the Centers for Disease Control and Prevention.³³ We compared the proportion of eligible women who received screening mammograms every fiscal quarter, as calculated by the number of screening mammograms performed divided by the total number of TRICARE Prime/Prime Plus eligible women. Screening was defined by International Classification of Diseases, Ninth Revision (ICD9), Current Procedural Terminology (CPT), and Healthcare Common Procedure Coding System (HCPCS) codes (V76.10, V79.11, V76.12, 87.36, 87.37, 76083, 76085, 76092, 77052, 77057, 77063, G0202, G0203). We excluded diagnostic mammograms defined by CPT/HCPCS (76082, 76090, 76091, 77051, 77055, 77056, G0204, G0205, G0206, G0207, G0279) from the primary analysis.

We also analyzed the influence of the 2009 USPSTF recommendation change through stratification of the data by age group (40–49, 50–64), race (white, black), TRICARE beneficiary status, (sponsors, dependents), and care setting (direct, purchased care). The age groups were chosen to correspond with those addressed in the USPSTF guideline, and limiting to age 64 since TRICARE becomes the secondary payer when patients are eligible and covered by Medicare. Race information is not typically collected for TRICARE military dependents and was missing for a sizable portion of this group (∼75%). Since it is not hypothesized that there is a genetic component to screenings, the decision was made to use the race associated with the dependent's military sponsor as a proxy for the dependent's race. This procedure brought the number of participants missing racial classification to ∼20%. Results were only presented for black and white women due to small numbers of women in other racial categories for some sub-analyses.

Data analysis

We conducted an interrupted time series analysis (ITS) to evaluate changes in screening by taking into account baseline and trends before and after the recommendation change.^34–36 The model used to evaluate trends: Y = β₀ + (β₁) (time_t) + (β₂) (Intervention_t) + (β₃) (time after intervention) + ɛ_t. We further evaluated differences between groups using: Y = β₀ + (β₁) (time_t reference group) + (β₂) (Intervention_t reference group) + (β₃) (time after intervention in reference group) + (β₄) (time_t) + (β₅) (Intervention_t in other group) + (β₆) (time after intervention) (β₇) (time after intervention) + ɛ_t. The change in the level immediately after the guideline change represents the immediate effects whereas the change in slope following the guideline change represents the immediate effect. We addressed seasonal trends in mammography utilization by fitting the model using Newey-West standard errors to address autocorrelation.^37,38

We conducted several sensitivity analyses. An implementation period of one quarter was tested to evaluate the possibility that the first quarter following the 2010 USPSTF guideline change took an additional quarter to fully take effect.³⁴ Additional analyses were performed to include screening and diagnostic mammograms to explore how potential misclassification might affect the results.

Because the USPSTF screening guidelines are based on age, we considered the age structure of the population during the study period. We found that both the mean and median age of women in the 40–49 and the 50–64 years age group remained constant throughout the study period. Since ITS methods are not generally sensitive to changes in population structure over time, we did not perform any age adjustments.³⁴

All p-values are two-sided with a significance level of 0.05. All rates are out of 10,000 women screened quarterly. We used SAS version 9.4 (SAS Institute, Inc., Cary, NC) and Stata, version 15 (StataCorp, College Station, TX) for all analyses.

Results

Study population

The study participants contributed a total of 7,218,162 person-years over the 10-year study period (Table 1). Fifty-seven percent of the women in the study were 40–49 years old. The percent of women aged 40–49 years decreased over time, but the median of each age category remained constant. Approximately 78% in the race-sub analysis were white. Eighty-six percent were military dependents. Sixty-seven percent of screening mammograms were performed in direct care. The percent of white women increased about 2% over the study period, while the percent of military dependents remained steady as did the mammograms performed in direct care.

Table 1.

Population Characteristics

	FY 2006, n = 669,307 (%)	FY 2010, n = 742,164 (%)	FY 2015, n = 700,410 (%)
Age
40–49	51.0	47.9	43.4
50–64	49.0	52.1	56.6
Race
White	55.3	57.9	57.7
Black	14.0	15.7	17.7
Other	6.1	7.4	8.9
Missing	24.6	19.0	15.7
Beneficiary type
Dependent	87.5	86.3	84.7
Sponsor	12.5	13.7	15.3
Care setting
Direct care	70.1	64.7	69.0
Purchased care	29.9	35.3	31.0

Demographic characteristics of the study population by age group, race, beneficiary type, and care setting.

FY, fiscal year.

Overall results

For both age groups, the mammography rate increased between fiscal years 2006 and 2009 (Fig. 1 and Table 2). There was a statistically significant decrease in the rate of mammograms during the first quarter of 2010 (October 1 to December 31) by 65 fewer screened per 10,000 among those aged 40–49 and 66 fewer screened per 10,000 aged 50–64. After 2010, the screening rate in both age groups decreased significantly from that before the guideline change (ages 40–49: 7 fewer screened per 10,000; age 50–64: 6 fewer screened per 10,000) and remained constant through the end of the study period. A comparative analysis shows that the changes in rate and slope were similar across both age groups (Table 3).

FIG. 1.

Quarterly screening rates from FY 2006 through 2015. FY, fiscal year.

Table 2.

Single Group Interrupted Time Series Analysis

Age group	Baseline rate	Slope before guideline change	Slope after guideline change	Change in rate (immediate effect)	Change in slope (effect over time)
Overall
40–49	482 (458 to 506)	7 (5 to 9)	−1 (−2 to 0)	−65 (−83 to −46)	−7 (−10 to −5)
50–64	667 (650 to 684)	6 (4 to 8)	0 (−2 to 1)	−66 (−88 to −44)	−6 (−9 to −3)
Race
White
40–49	481 (455 to 507)	7 (5 to 9)	0 (−1 to 1)	−61 (−79 to −43)	−7 (−10 to −5)
50–64	676 (640 to 712)	10 (7 to 14)	2 (1 to 3)	−76 (−108 to −43)	−8 (−12 to −4)
Black
40–49	498 (476 to 519)	8 (6 to 10)	1 (0 to 3)	−67 (−92 to −41)	−7 (−10 to −4)
50–64	706 (672 to 740)	9 (6 to 13)	5 (2 to 8)	−115 (−161 to −69)	−4 (−9 to 1)
Beneficiary type
Dependent
40–49	482 (469 to 494)	7 (6 to 8)	−1 (−2 to 0)	−65 (−78 to −51)	−8 (−9 to −6)
50–64	672 (646 to 697)	6 (4 to 9)	0 (−2 to 1)	−62 (−87 to −38)	−6 (−9 to −3)
Sponsor
40–49	483 (450 to 515)	7 (4 to 10)	1 (−1 to 3)	−65 (−101 to −29)	−6 (−10 to −3)
50–64	619 (582 to 656)	5 (1 to 9)	1 (0 to 3)	−90 (−138 to −42)	−3 (−8 to 1)
Care setting
Direct care
40–49	297 (279 to 315)	2 (1 to 4)	1 (−1 to 2)	−53 (−72 to −34)	−1 (−3 to 0)
50–64	368 (346 to 389)	1 (−1 to 3)	2 (0 to 3)	−60 (−87 to −34)	1 (−1 to 4)
Purchased care
40–49	185 (178 to 192)	5 (4 to 5)	−1 (−2 to −1)	−11 (−22 to −1)	−6 (−7 to −5)
50–64	299 (291 to 308)	5 (4 to 6)	−2 (−3 to −1)	−5 (−20 to 9)	−7 (−8 to −6)

Results of a single group interrupted time series analysis comparing mammography rates of women screened per 10,000 (95% confidence intervals) per quarter before and after the U.S. Preventive Services Task Force screening recommendation first quarter of fiscal year 2010, by age, race, beneficiary type, and care setting.

Table 3.

Multiple Group Interrupted Time Series Analysis

Age group	Difference in baseline rates	Difference in slopes before guideline change	Difference in slopes after guideline change	Change in rate (immediate effect)	Change in slope (effect over time)
Overall
40–49 to 50–64	−185 (−222 to −148)	1 (−2 to 4)	0 (−2 to 1)	1 (−36 to 38)	−1 (−5 to 2)
Race
White–Black
40–49	−17 (−50 to 17)	0 (−3 to 2)	−1 (−3 to 1)	5 (−26 to 37)	−1 (−4 to 3)
50–64	−30 (−98 to 38)	1 (−6 to 8)	−3 (−6 to 0)	39 (−34 to 113)	−4 (−11 to 4)
Beneficiary type
Dependent—sponsor
40–49	−1 (−39 to 38)	0 (−4 to 3)	−2 (−4 to 0)	0 (−37 to 37)	−1 (−5 to 2)
50–64	53 (7 to 98)	1 (−4 to 6)	−2 (−4 to 0)	28 (−28 to 84)	−3 (−8 to 2)
Care source
Direct—purchased
40–49	112 (93 to 131)	−3 (−4 to −1)	2 (1 to 3)	−42 (−62 to −22)	5 (3 to 7)
50–64	68 (46 to 90)	−4 (−7 to −2)	4 (3 to 6)	−55 (−84 to −26)	9 (6 to 11)
Military status
Dependent—military
40–49	−30 (−82 to 22)	−15 (−20 to −10)	−6 (−9 to −3)	19 (−37 to 76)	9 (4 to 15)
50–64	133 (66 to 201)	−22 (−29 to −14)	−76 (−85 to −67)	177 (26 to 327)	−55 (−67 to −43)

Results of a multiple group interrupted time series analysis comparing the differences between groups (age, race, beneficiary type, care source, and military status) of mammography rates of women screened per 10,000 (95% confidence intervals) per quarter before and after the U.S. Preventive Services Task Force screening recommendation first quarter of fiscal year 2010.

Results by race

Mammography rates among black and white women were similar to that observed in the overall population (Table 2 and Fig. 2). There was an immediate statistically significant decrease in the mammography screening rates among black (aged 40–49: 67 fewer screened per 10,000; aged 50–64: 115 fewer screened per 10,000) and white women (aged 40–49: 61 fewer screened per 10,000; aged 50–64: 76 fewer screened per 10,000) the quarter of the guideline publication. While the screening rate increased through 2015, the rate of increase was lower than before the guideline change for both races (whites aged 40–49: 7 fewer screened per 10,000; whites aged 50–64: 8 fewer screened per 10,000; blacks aged 40–49: 7 fewer screened per 10,000; blacks aged 50–64: 75 fewer screened per 10,000). Direct comparison showed that black and white women received mammograms at equal rates before and experienced similar changes after the USPSTF guideline publication and during the period following the guideline (Table 3).

FIG. 2.

Quarterly screening rates from FY 2006 through 2015 for black and white women aged 40–49 and 50–64.

Results by beneficiary type

The pattern observed in the overall population did not change when stratified by beneficiary status (Table 2). Sponsors and dependents aged 40–49 had an immediate statistically significant decrease in rate after the guideline change of 65 fewer screened per 10,000. Women aged 50–64 also had a statistically significant immediate decrease (dependents: 62 fewer screened per 10,000; sponsors: 90 fewer screened per 10,000). Both dependents (aged 40–49: 8 fewer screened; aged 50–64: 6 fewer screened per 10,000) and sponsors (aged 40–49: 6 fewer screened per 10,000; aged 50–64: 3 fewer screened per 10,000) had lower screening rates after the guideline change through the end of the study period. All changes were similar across age groups and beneficiary types (Table 3).

When limiting the analysis to active-duty compared to military dependents, the patterns did change. Military dependents aged 40–49 had a statistically significant larger increase (9 more screened per 10,000) in the rate than did active-duty women after the guideline change through the study end. Active-duty women aged 50–64 had a statistically significant larger decrease in screening rates immediately after the guideline change (177 fewer screened per 10,000) and a significantly larger increase in quarterly screening rates (55 more screened per 10,000) through the end of the study period compared to dependents of the same age (Table 3).

Results by system of care

The mammography screening rate among women aged 40–49 years who received their screening mammograms at MTFs (direct care setting) increased from 2006 through 2009 (Table 2). After a statistically significant drop in the first quarter of 2010 (aged 40–49: 53 fewer screened per 10,000; aged 50–64: 60 fewer screened per 10,000), screening remained relatively constant. Of those receiving mammography screening in purchased care, only younger women aged 40–49 years of age exhibited a statistically significant rate decrease immediately following the recommendation change (11 fewer screened per 10,000), and a decreased rate from the guideline change through the study end (aged 40–49: 6 fewer screened per 10,000; aged 50–64: 7 fewer screened per 10,000). Note that baseline screening rates for direct and purchased care add to the total baseline screening rate since women could be screened in either system.

Comparing the two care systems, direct care rates decreased significantly more than purchased care rates (aged 40–49: 42 fewer screened; aged 50–64: 55 fewer screened per 10,000; Table 3) immediately after the guideline change. Over time, however, rates decreased after the guideline change in purchased care for both age groups (aged 40–49: 5 fewer screened per 10,000; aged 50–64: 9 fewer screened per 10,000).

Sensitivity analyses

When we repeated the analysis under the assumption that the USPSTF recommendation change would take an additional quarter for full implementation, the results were consistent with the original outcomes. We also conducted separate analyses that included diagnostic mammograms. Overall, the trends were very similar to the original analysis. There was a significant decrease in mammogram rates in almost all the described analyses in the same quarter as the new recommendations issued by the USPSTF.

Discussion

This study was able to evaluate the effects of the Task Force recommendation change on a large nationally representative population of individuals in a universal health care setting, thereby minimizing differential access as a potential confounder. Because the population was so large, this study was able to look at differences by age, race, beneficiary status, and by system of care. Overall, there was an increased rate of TRICARE beneficiaries receiving mammograms from fiscal year 2006 through the end of 2009. There was a small, but statistically significant decrease in this rate in the first quarter of fiscal year 2010 subsequent to the change in USPSTF recommendations for women aged 40–64 regarding mammography screening frequency. Following the drop in the screening rate, the rate in both age groups remained essentially unchanged through the end of the study period. Subgroup analyses did not substantially change the observed patterns. Only older women who received their breast cancer screening through the purchased care system did not see a decrease in screening rates. There were also no significant changes over time for women in the direct care system or among older black women and older military dependents following the guideline publication.

Comparative analysis revealed that the impact of the recommendation change was equivalent among both black and white women and between military dependents and sponsors. The short-term impact was more pronounced in direct care but over time, the impact was greater in purchased care. Older active-duty women displayed a larger impact when compared to their dependent counterparts. It was surprising that there were no statistically significant differences noted between the age groups overall since the recommendations in the updated guideline were different for the two groups. The overall changes resulting from the revised guidelines, while statistically significant, were generally very small in magnitude and comparable to the reductions in cervical cancer and prostate cancer screening following similar changes in USPSTF guideline changes.^39,40

An important finding is the lack of difference in the way the USPSTF recommendation change impacted black and white women. Although most studies that have previously examined the influences of the guideline changes have aggregated racial groups, two have demonstrated that the mammography screening rate of black women did not change as a result of the recommendation change, unlike rates among white and Hispanic women.^24,41 Some studies have observed that black women received breast cancer screenings at a lower rate than do white women.^19,42,43 This study reveals that not only do TRICARE beneficiaries of both races receive preventive breast cancer screening mammograms at the same rate but that the USPSTF recommendation change equally impacted both white and black women. This study contributes further evidence that policy changes in a system with universal access to health care were applied uniformly without regard to race.

There is precedent to have anticipated a change in the proportion of women receiving mammograms following the release of the updated screening guidelines. Such immediate effects were noted for other changed screening recommendations such as those for prostate and cervical cancers, and older recommendation changes for mammography screening.^39,44–46 Soon after the 2009 USPSTF guidelines, several studies evaluated the effects of the updated screening recommendation among women. A few early studies utilizing self-reported survey data of U.S. women failed to find any significant differences in screening rates before or after the recommendation change in any age groups.^17,18 Later studies using health insurance claims data, which included more statistical power and robust techniques, have found declined rates of U.S. women in all age groups receiving mammograms, at least initially following the change in screening guidelines.^{19,21–25,47} Two of these studies, however, found either no decline or an increase in screening among black women following the recommendation change.^24,47 Our study found a significant decrease in the rate immediately and over time following the guideline change of the first quarter of 2010 for both age groups among both whites and blacks compared to before 2010, but the rate through the end of the study period only remained significantly positive for white and black women aged 50–64 years.

It is somewhat unclear why the guideline change had an immediate effect on women receiving mammograms in the direct care setting and longer term decreases in women receiving screening in the purchased care setting. Given that women can be screened in either system, it is difficult to interpret this trend. The immediate decrease noted in the direct care system observed for both age groups may mean that care providers in this system were more responsive to the USPSTF guidelines in the short term. The decrease proportion of women receiving mammography screenings in the purchased care setting may represent differences in out of pocket costs associated not with the breast cancer screening, but potentially with any associated clinical office visits.

While the effects of the USPSTF breast cancer screening guidelines on active-duty women in the U.S. military has been studied elsewhere,²⁸ we have found differences in the effect of the guideline changes noted between active-duty women and military dependents. We believe that the higher rates of breast cancer screening among active-duty military women are a result of mandated medical exams and minimum health and fitness standards that must be maintained. The increased rate observed over time among active-duty women may also correspond to an increased awareness of elevated breast cancer incidence in the military.⁴⁸

Factors other than the specified screening guidelines could explain the changes observed in mammography utilization. Screening guidelines issued by other medical organizations could potentially have affected our results. However, no other medical organizations issued recommendations increasing the recommended age to begin screening or reducing the frequency of breast cancer screening during the surveillance period. The American College of Radiology and the Society for Breast Imaging republished their annual screening recommendations for women of average risk aged 40 or above in January 2010.¹⁰ The American College of Obstetricians and Gynecologists published updated guidelines in July 2011 also recommending the annual screening recommendations for women of average risk beginning at age 40.⁹ Neither of these more aggressive screening guidelines would account for the decline in screening as observed in this study but may have contributed to diminishing the influence of the USPSTF statement among providers and their patients. The medical departments of the Army, Navy, and Air Force did not modify their recommendations during the study period.

Another potential source of influence on screening patterns among women receiving health care benefits through TRICARE was the change in cost sharing during the study period. Beginning in the Fiscal Year before the USPSTF, cost sharing was eliminated for all cancer preventive screenings in the MHS.⁴⁹ This is unlikely to have impacted screening practices after the publication of the updated screening guidelines. If the changed cost sharing policy had an influence, an increase in the proportion of screening mammograms should have been observed instead of the decrease observed in this study.

This study is unable to determine whether the changed screening rates of women receiving mammograms in this military beneficiary population is a result of patient or provider preference. A recent study suggested that while health care providers are in agreement as to the importance and influence of the USPSTF screening recommendations for breast and cervical cancer they are more likely to screen their patients at levels that exceed the Task Force recommendations.⁵⁰ This previous study was limited to women's health care providers in the Northeastern United States, but it may point to the influence of provider preference. We also did not attempt to evaluate adherence to the new screening guidelines. We instead examined whether the USPSTF guideline change affected the number of women receiving mammography screening as the first indicator of potential changes in breast cancer morbidity or mortality.

The major strengths of this analysis were the use of the ITS model, and the size of this nationally representative population. The ITS model is considered the strongest quasi-experimental study design as it takes into consideration the baseline and trends before and after the intervention of interest.^34,36,51,52 This model reduces the influences of potential 1333varying confounders such as changing age distributions, which in this study minimally changed over the surveillance period, and seasonality by adjusting for autocorrelation.^34,37

There are several limitations to be addressed. Despite the best efforts to identify potential sources of the mammography utilization changes observed in this study, there remains the possibility that another undescribed source may be influencing these trends. This study assumed that the trends occurring before FY 2010 would have continued uninterrupted in the absence of the Task Force recommendation. The results, however, consistently show a statistically significant drop in the proportion of women receiving screening the quarter following the recommendation change, not merely a change in the trend. Additionally, we were limited to only studying black and white women because of the small numbers of women after stratification. When necessary, for TRICARE military dependents, we assigned an unknown race to match that of the spouse when that information was available.

This study is a first step in fully evaluating screening for breast cancer. This study has not reported on how the updated USPSTF screening recommendation has affected the rates of breast cancer diagnoses, treatments, or survival. The monetary costs to the health care system have also not been evaluated. All of these are important next steps to being able to completely interpret the results presented here.

Conclusion

This study determined that the change in the Task Force mammography guidelines impacted the mammography screening rates of women both immediately and throughout the years following the change. The guideline changes did not differentially impact women of different age groups, dependent compared to sponsor, or black compared to white women. In absolute terms, the impacts were small but significant. Importantly, no racial disparities in either the rate of women screened or the impact of the guideline change were noted in this population with equal access to health care. Different patterns of mammography utilization were observed between direct and purchased care, and between those of active-duty status and female military dependents. Further studies to evaluate the impact of the guideline changes on active-duty military are warranted. To completely understand the impact of the updated screening guidelines, these results must be incorporated into research focusing on changes in breast cancer morbidity and mortality as well as updated cost benefits analyses.

Disclaimer

The contents of this publication are the sole responsibility of the authors and do not necessarily reflect the views, opinions, or policies of the USUHS, The Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., the DoD, or the Departments of the Army, Navy, or Air Force. Mention of trade names, commercial products, or organizations does not imply endorsement by the U.S. Government.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

All phases of this study were supported by a grant from the U.S. Department of Defense, Defense Health Agency (award No. HU0001-11-1-0023).