Trends in breast cancer screening during the COVID‐19 pandemic within a universally insured health system in the United States, 2017–2022

Abstract Background In the United States, breast cancer is the most commonly diagnosed cancer and second leading cause of cancer death in women. Early detection through mammogram screening is instrumental in reducing mortality and incidence of disease. The COVID‐19 pandemic posed unprecedented challenges to the provision of care, including delays in preventive screenings. We examined trends in breast cancer screening during the COVID‐19 pandemic in a universally insured national population and evaluated rates across racial groups and socioeconomic strata. Methods In this retrospective open cohort study, we used the Military Health System Data Repository to identify female TRICARE beneficiaries ages 40–64 at average risk for breast cancer between FY2018 and FY2022, broken down into prepandemic (September 1, 2018–February 28, 2020), early pandemic (March 1, 2020–September 30, 2020), and late pandemic periods (October 1, 2020–September 30, 2022). The primary outcome was receipt of breast cancer screening. Results Screening dropped 74% in the early pandemic period and 22% in the late pandemic period, compared with the prepandemic period. Compared with White women, Asian/Pacific Islander women were less likely to receive mammograms during the late pandemic period (0.92RR; 0.90–0.93 95%CI). American Indian/Alaska Native women remained less likely to receive screenings compared with White women during the early (0.87RR; 0.80–0.94 95% CI) and late pandemic (0.94RR, 0.91–0.98 95% CI). Black women had a higher likelihood of screenings during both the early pandemic (1.10RR; 1.08–1.12 95% CI) and late pandemic (1.12RR, 1.11–1.13 95% CI) periods compared with White women. During the early and late pandemic periods, disparities by rank persisted from prepandemic levels, with a decrease in likelihood of screenings across all sponsor ranks. Conclusion Our results indicate the influence of race and socioeconomics on mammography screening during COVID‐19. Targeted outreach and further evaluation of factors underpinning lower utilization in these populations are necessary to improve access to preventative services across the population.


| BACKGROUND
In the United States, breast cancer is the most commonly diagnosed cancer and second leading cause of cancer death in women. 1,2It was predicted that, in 2022 alone, almost 290,000 new diagnoses and 43,000 deaths attributable to breast cancer would occur. 3Early detection through mammogram screening has been instrumental in reducing the incidence of advanced stage disease and mortality.Several medical organizations have published guidelines and recommendations on age and frequency of screening mammograms.The US Preventive Services Task Force, since 2016, has recommended biennial breast cancer screening for women aged 50-64 at average risk of disease, while the American Cancer Society recommends annual screening for women 45-54 years of age who are at average risk, with the option to switch to biennial mammograms at age 55. 4,5 There are, however, differences in screening rates across the country based on demographic factors. 6,7Women in lower socioeconomic brackets are less likely to get screened for breast cancer, and differences in the quality of care received during and after screening visits have been reported for racial and ethnic minorities. 7,8hese pre-existing challenges were likely further exacerbated by the COVID-19 pandemic due to restrictions on in-person evaluations and diversion of health care resources. 9This redirection of resources led to delays and cancelations of nonurgent procedures and services.Across the board, there was a significant decrease in preventive screenings and early diagnosis of chronic conditions. 9,10Studies of claims data in April 2020 showed over 80% reduction in mammograms compared with February 2020. 11Much of this reduction was attributed to the onset of the pandemic, and it is estimated that this decrease alone could have resulted in 36,000 delayed breast cancer diagnoses. 11The challenges faced by the health care system during this period may have worsened existing disparities in access to health services, including many people who lost their health insurance as a result of job loss. 12,13he Military Health System (MHS) is unique among other health systems in the United States in that it provides health care to a universally insured beneficiary population. 14,15Through the TRICARE plan, beneficiaries can receive either direct care at military treatment facilities (MTFs) or in private civilian facilities where the plan serves as an insurance benefit.This system minimizes several barriers often encountered in civilian health care settings that prevent individuals from accessing and utilizing medical services, and also mitigate racial disparities in cancer screenings. 16,17In this context, we sought to examine changes in breast cancer screening rates among TRICARE beneficiaries that occurred during the COVID-19 pandemic.Given the nature of the MHS, studying the influence of the pandemic on care in the MHS may provide insight into the current state of preventive care across the nation.Based on previous work, we hypothesized that screening rates decreased during the pandemic, and that there were differences in screening uptake across different demographic categories.The MDR houses all health care encounter and claims data for MHS beneficiaries who received care at a military treatment facility (MTF) (direct care) or at a civilian fee-for-service treatment facility through their TRICARE benefit (private sector care (PSC)). 18Women ages 65 and older were excluded due to Medicare becoming the primary payer, resulting in a loss of transparency of care.Additionally, women who were beneficiaries of the National Guard or Reserves were excluded due to differences in access to care within the MHS.
Utilizing Current Procedural Terminology/Healthcare Common Procedure Coding System (CPT/HCPCS) codes we identified all screening mammography claims for the study population between FY 2018 and 2022 (Table S1).Any women with a history of diagnostic mammography, history of malignancy, or mastectomy within 2 years prior to the screening period were excluded from the analysis (Figure 1; Table S1).Patient demographics such as age (categorized into 40-44, 45-49, 50-54, 55-59, and 60-64), race (White, Black, Asian/ Pacific Islander, American Indian/Alaska Native, Other, and Unknown/Missing; self-reported), beneficiary status (active duty, dependent, retiree, and other), sponsor rank as a proxy for socioeconomic status (Junior Enlisted, Senior Enlisted, Junior Officer, Senior Officer, Warrant Officer, and Other), and associated branch of service (Army, Navy, Marine Corps, Air Force) were obtained from the beneficiary's Defense Enrollment Eligibility Reporting System (DEERS) record during each FY and used in analysis.Data from the MDR does not specify ethnicity for the majority of beneficiaries, therefore we were unable to identify beneficiaries as Hispanic or non-Hispanic.As such, White Hispanic women are grouped in the White cohort and Hispanic women of Afro-Caribbean ancestry are categorized with Black women.

| Study analyses
The primary outcome was receipt of screening mammography and the sociodemographic factors race and sponsor rank were considered the primary predictors.In our study race was missing for 42% of the study population.To resolve this issue, we utilized relational imputation by substituting patient's unknown or missing race with the known race of the sponsor, a method previously published in an analysis of mammography screening rates. 19,20In this setting, and based on previously published work that supports use of sponsor rank as a proxy for socioeconomic status, 14,15,21 enlisted personnel were considered proxies for lower socioeconomic strata.
Study analyses included descriptive statistics on patient demographics and trend analysis of quarterly mammography screening rates overall and by race and rank.Binomial Poisson regressions were performed to calculate unadjusted and adjusted rate ratios and were used to assess for racial and socioeconomic disparities in screening mammography over the total study period and stratified by three summative periods, and the interactions between time periods and race and rank.Based on previously accepted cutoffs for the COVID-19 pandemic, 22 prepandemic (October 1, 2017-February 28, 2020), early pandemic (March 1, 2020-September 30, 2020), and late pandemic (October 1, 2020-September 30, 2022) were used to evaluate the effect of the COVID-19 pandemic on screening mammography rates.Categorical age and beneficiary status were used as adjustment factors in the adjusted regression models.Additionally, adjusted regression models were used in sensitivity analyses of imputed race by comparing screening rates in those with a known race prior to imputation to those with an imputed race; these results can be found in Supplemental Materials (Table S2).Statistical significance was set at α < 0.05, and all analyses were performed using SAS, version 9.4.This study was reviewed and determined to be exempt by the Institutional Review Board at the Uniformed Services University of the Health Sciences.

| RESULTS
We identified 819,833 women eligible for breast cancer screening during the study period (October 1, 2017-September 30, 2022): 702,207 in the prepandemic period, 548,598 during the early pandemic period, and 555,066 during the late pandemic period (Figure 1; Table 1).Among these, 386,336 (55.02%) were screened in the prepandemic period, 77,106 (14.06%) were screened in the early pandemic period, and 237,087 (42.71%) were screened in the late pandemic period; resulting in a -74% rate change in the early pandemic period and a -22% rate change in the late pandemic period compared to the proportion of women screened in the prepandemic period.The total study population consisted primarily of women who are ages 55-64, White, and in or associated with a Senior Enlisted rank (Table 1).
Figure 2 shows a graphical representation of the change in quarterly (per 3-month period) mammography screening rates across the study period.Prior to COVID-19 pandemic, the screening rate for breast cancer was approximately 9 per 100 eligible women.At the onset of the pandemic in March 2020, the rate noticeably drops down to about four per 100 eligible women, after which rates begin to rise again through September 30, 2020.This pattern is also seen across the study period when broken down by race (Figure 3) and the patient's rank or sponsor's rank (Figure 4).
Table 2 displays the unadjusted and adjusted rate ratio (RR) results of race, SES, and time period as predictors for mammography screenings over the full study period.Adjusted analyses of screening rates showed that, compared to White women, Black women (1.09 RR; 1.09-1.1095% CI) and women of 'other' race (1.02 RR; 1.02-1.0395% CI) were more likely to receive mammography screenings, while Asian/Pacific Islander (0.98 RR; 0.97-0.9895% CI) and American Indian/Alaskan Native (0.92 RR; 0.90-0.9595% CI) women were less likely to receive mammography screenings (Table 2).In the evaluation of SES on mammography screening rates, adjusted analyses showed all ranks below Senior Officer were less likely to receive mammography screenings over the full study period (Table 2).In the evaluation of the effects of the COVID-19 pandemic, mammography screenings rates were significantly less likely during both the early (0.25 RR, 0.24-0.25 95% CI) and late pandemic (0.78 RR, 0.78-0.7895% CI) periods compared to the prepandemic period (Table 2).In the evaluation of the interaction between time and race and rank, both interactions were statistically significant in the regression models (type 3 chi-square p-values <0.05); and, significant differences in rates were observed across all races and ranks in both the early and late pandemic periods compared to White and Senior Officer women in the prepandemic period (chi-square p-values <0.05).
Table 3 displays the stratified adjusted analyses evaluating patient demographic associations of screening rates within each time period.Adjusted analyses of screening rates showed that, prior to the onset of the COVID-19 pandemic, Black (1.07 RR; 1.07-1.0895% CI), Asian/Pacific Islander (1.02 RR; 1.01-1.0395% CI), and 'other' race (1.06 RR, 1.06-1.0795% CI) women were more likely to receive breast cancer screening compared to White women (Table 3).In contrast, American Indian/Alaska Native women were less likely to get screened (0.91 RR; 0.89-0.9495% CI).Compared to White women, Black women had a higher likelihood of mammography screenings during both the early pandemic (1.10 RR; 1.08-1.1295% CI) and late pandemic (1.12 RR, 1.11-1.1395% CI) periods; representing a 3% and 5% increase in likelihood from the prepandemic period, respectively.Asian/Pacific Islander women did not experience a significant difference in screenings during the early pandemic period, however they were less likely to receive mammography screenings during the late pandemic period (0.92 RR; 95% CI 0.90-0.9395% CI) compared to White women; representing a 10% decrease in likelihood from the prepandemic period.American Indian/Alaska Native women remained less likely to receive screenings compared to White women during the early pandemic (0.87 RR; 0.80-0.9495% CI) and the late pandemic (0.94 RR, 0.91-0.9895% CI) periods; a 4% decrease and 3% increase in likelihood, respectively, from the prepandemic period.During all periods all sponsor ranks (Junior Enlisted, Senior Enlisted, Junior Officer, and Warrant Officer) had lower rates of screening compared to Senior Officers (Table 3).During both the early and late pandemic periods, a decrease in the likelihood of screenings was observed across all ranks from the prepandemic period (Table 3).

| DISCUSSION
Our analysis of TRICARE beneficiary data showed an overall reduction in breast cancer screening rates in both the early pandemic and late pandemic periods when compared to prepandemic rates.The sharp decline in screening around March and April 2020 coincides with the onset of COVID-19 restrictions and is consistent with previous investigations regarding the impact of the pandemic on cancer screening rates.Although screening rates appear to trend upward immediately after the initial decline, our findings indicate worrisome and persistent reductions for several specific subgroups in the US demographic.For example, Asian/Pacific Islanders experienced a significant decrease in screening during the pandemic period.While screening rates among American Indian/ Alaskan Native women increased slightly in the late pandemic period compared to prepandemic levels, overall they were still less likely to get screened both prior to and during the pandemic.Disparities by sponsor rank, used as a proxy for socioeconomic status, persisted between prepandemic and early and late pandemic periods, with women from lower socioeconomic strata less likely to be screened across all time periods.Although screening rates are trending back upward after the initial onset of COVID-19, important disparities persist for Asian/Pacific Islanders, American Indian/Alaskan Natives, and those in lower socioeconomic strata.
Our findings are consistent with previous studies that assessed the impact of the pandemic on breast cancer screening services.For example, a study of data from the National Breast and Cervical Cancer Early Detection Program showed a sharp decrease in screenings in March-April 2020 when compared to the 5-year average prior to COVID-19, dropping 87% in April. 13Similarly, a study by the National Cancer Institute's Population-based Research to Optimize the Screening Process consortium found a 96% decrease in screening in April and May 2020 compared to rates in 2019. 235][26] Unlike other racial minority groups in the United States, low screening rates among Asian/Pacific Islanders cannot be fully explained by the usual barriers such as income and health care access, and that cultural and social factors may be playing more of a role, particularly during the early days of COVID-19. 27At the start of the pandemic there was a surge in anti-Asian violence, and some literature cites this rise in racial violence and fear among the Asian American population as an additional barrier in deciding to seek preventive services and other medical care. 27,28hile our findings show that American Indian/Alaskan Natives continued to be screened less frequently compared to other groups both prior to the pandemic and during both early and late pandemic, we saw a slight increase in screening uptake during the pandemic.Other literature on the topic has generally shown a substantial decrease in screening mammograms among this population. 13,26,29One study comparing screening rates among a cohort of women in Washington state, for example, found a 60% reduction in screening mammograms among American Indian/Alaska Native women. 301][32] Our study population may not show the same large decline during the COVID-19 period due to the nature of universal insurance in the setting of TRICARE and associated reductions in barriers to accessing care.However, the low rates of screening uptake even in this setting is indicative of other factors that likely contribute to differences in the use of screening mammograms in this minority group.
It is important to note that our results show that the breast cancers screening rate for Black women rose during the pandemic and exceeded those for White women both prior to and during the pandemic.This is in contrast to data from previously published studies which indicated that Black women are less likely to be screened for breast cancer than White women. 33,34One possible explanation for the higher screening mammography rates seen in our study population may have to do with insurance coverage.In the United States, insurance coverage among Black people and other racial minorities is much lower compared with White people, posing a significant barrier to medical care and preventive services.Several studies, including one conducted by Agrawal et al in Texas, have demonstrated positive associations between health insurance coverage and adherence to breast cancer screening among Black women. 35,36Therefore, it would stand to reason that Black TRICARE beneficiaries would be more likely to seek out and receive screening services, since the barrier of insurance coverage is eliminated in the MHS.
While we recognize further studies are needed to fully validate these findings, the disparities in mammography screenings during the COVID-19 period in specific racial and socioeconomic groups identified here represent best available evidence at this time.Particularly among Asian/ Pacific Islanders, American Indian/Alaskan Natives and those of lower socioeconomic strata, the need for both targeted outreach as well as further evaluation of the drivers of low screening uptake during COVID-19 may prove beneficial for improving access to preventative services in these populations.
We recognize several limitations to this study.Given the study design and a reliance on health care claims data, we were restricted to the information that was available in the MHS Data Repository without being able to collect additional information in the event of missing data, and could not account for changing trends in screening that may have been occurring prior to the study period.Additionally, the time range used for the pandemic period was restricted to March 2020-September 2022 as this was the most recent data available, so we were unable to evaluate trends in screening rates beyond this time window.Due to the nature of data reporting in the MDR, ethnicity of beneficiaries was unavailable for study analysis and therefore we were unable to parse out screening rates for Hispanic patients separate from non-Hispanic White and non-Hispanic Black women.Our study population was also limited to women who had been screened via mammography only.However, it is unlikely that this would have a significant impact on our determinations given that mammograms are the most common modality for breast cancer screening.We also restricted analysis to women at average risk of breast cancer diagnosis since the inclusion of individuals at higher risk would have likely artificially increased screening rates and impaired generalizability.It is possible that since we only used a 2-year period prior to screening to assess for exclusionary criteria, some people with a recorded history of malignancy or mastectomy prior to that 2-year period may have been included in our dataset; however, we do not anticipate this to have a significant impact on the findings.And lastly, we recognize the utilization of relational imputation for patient's missing race is imperfect and does not account for interracial couples or marriages.

| CONCLUSION
We found significant reductions in breast cancer screening within the MHS in the time period 2020-2022, which coincides with the onset of the COVID-19 pandemic.We additionally identified new and persistent racial and socioeconomic disparities in the uptake of breast cancer screening among Asians/Pacific Islanders, American Indians/Alaskan Natives, and individuals from low socioeconomic background.Targeted outreach, as well as further evaluation of the factors underpinning lower utilization in these populations are necessary to improve access to preventative services within the MHS and nationwide.

DISCLAIMER
The contents, views, or opinions expressed in this manuscript are those of the author(s) and do not necessarily reflect official policy or position of Uniformed Services University of the Health Sciences, the Department of Defense, or Departments of the Army, Navy, or Air Force, or the Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc. Mention of trade names, commercial products, or organizations does not imply endorsement by the U.S. Government.

2 | METHODS 2 . 1 |
Study design, data source, and study population This retrospective open cohort study analyzed administrative and health care claims data from the MHS Data Repository (MDR) for female TRICARE Prime beneficiaries ages 40-64 during fiscal years (FY) 2018-2022 (October 1, 2017-September 30, 2022).

F I G U R E 2
Total quarterly mammography screening rates in the MHS, FY 2018-2022.

F I G U R E 3
Quarterly mammography screening rates by Race, FY 2018-2022.

F I G U R E 4
Quarterly mammography screening rates by Sponsor Rank, FY 2018-2022.
T A B L E 1 Pandemic period stratified breast cancer screening rate ratio results for MHS beneficiaries ages 40-64, FY 2018-2022.Pre-, early, and late pandemic period multivariate binomial Poisson regression models were adjusted by categorical age and beneficiary status.
Note: Multivariate binomial Poisson regression models were adjusted by categorical age and beneficiary status.Note: