National yearly cost of breast cancer screening in the USA and projected cost of advocated guidelines: a simulation study with life table modelling

Abstract

Objective

The aim of this study was to estimate the total national direct cost of breast cancer screening from 2019 to 2022 and project the total national cost and average lifetime cost of screening per woman for three current guidelines.

Design

We estimated the national cost of screening from 2019 to 2022, and per cancer detected in 2022, using real-world data on the number of mammograms performed per year. We also projected the national cost of screening using life table modelling for three guidelines: 2021/2023 American College of Radiology (ACR), 2023 American Cancer Society (ACS) and 2024 United States Preventative Services Task Force (USPSTF). The average lifetime cost to screen one woman until age 74 years with each guideline was also estimated. The Optum Labs Data Warehouse was used to estimate commercial and Medicare costs and recall rates. Sensitivity analyses were used to estimate uncertainty and determine which inputs had the largest impact on total national costs.

Setting

This study was conducted for the USA.

Participants

Women eligible for breast cancer screening.

Interventions

Digital mammograms (2D) or digital breast tomosynthesis (3D) and/or MRI.

Primary outcome measure

Total national cost of screening calculated as the sum of screening and recall costs. Average lifetime cost of screening per woman until 74 years.

Results

Nationally, screening cost approximately US$11 billion (B) per year from 2019 to 2022 with approximately 37% of eligible women screened each year. In 2022, screening cost US$55 471 per 3D-detected and US$44 000 per 2D-detected invasive or ductal carcinoma in situ case. Using target yearly participation rates of 54%–78% by age of women, the projected cost of screening was US$30B for ACR, US$18B for ACS and US$8B for USPSTF guidelines. The average lifetime cost to screen an average-risk woman was: US$13 416 for ACR, US$7946 for ACS and US$6931 for USPSTF. Participation rates, the proportion of women with a lifetime risk>20% and commercial MRI and 3D costs had the largest impact on total costs.

Conclusion

The cost of screening varies significantly by guideline (US$8B–US$30B) and was most influenced by participation rates, high-risk population proportions and technology costs. Future work can investigate whether risk-based screening strategies being tested in ongoing clinical trials can reduce national screening costs while improving outcomes.Cite Now

STRENGTHS AND LIMITATIONS OF THE STUDY.

• The cost estimates include commercial and Medicare Advantage costs and recall rates for digital mammograms, digital breast tomosynthesis, and MRI, estimated using real-world data from Optum Labs Data Warehouse.

• Annual screening costs from 2019 to 2022 were calculated using real-world data from the Food and Drug Administration’s Mammography Quality Standards Act and Program.

• The analysis excludes costs related to Medicaid and screening ultrasounds.

• The models assume independence between parameters, even though interdependencies exist in practice.

Introduction

Healthcare spending in the USA has increased substantially over time and is projected to exceed the rate of average gross domestic product growth by 2031.¹ According to the National Health Expenditure Accounts, the cost of healthcare in the USA, which includes Medicare, Medicaid, private health insurance and out-of-pocket costs, grew by 2.7% in 2021 costing US$4.3 trillion or US$12 914 per person which accounted for 18.3% of gross domestic product.¹ This cost is more than any developed nation spends on healthcare.¹ Breast cancer screening is a contributor to the cost of healthcare. O’Donoghue et al estimated that nationally, breast cancer screening cost US$7.8 billion (B) in direct costs in 2010 with a participation rate of approximately 70%. This estimate only included Medicare costs and is likely a significant underestimation of costs.² An updated cost estimate is warranted as there have been several changes to screening guidelines and practices over the past decade and to include costs that are more representative of the payer mix in the USA.

In 2024, the United States Preventative Services Task Force (USPSTF) updated its 2009 breast cancer screening guideline to decrease the age to start biennial screening from 50 years to 40 years for all women at average risk and women who are at an increased risk due to family history or having dense breasts. The guideline does not apply to women with pathogenic mutations or syndromes at high risk, women with a history of high-dose radiation therapy to the chest at a young age, or previous breast cancer or a high-risk breast lesion on previous biopsies.³ The American College of Radiology (ACR) recommends high-risk women begin annual screening at 30 years with adjunctive magnetic resonanace imaging (MRI).⁴ In 2023, ACR updated its guidelines for high-risk women to include recommendations that all women have a risk assessment beginning at 25 years and that women with dense breasts who desire supplemental screening have an MRI beginning at 40 years.⁴ The ACR guidelines for average-risk women were last updated in 2021 which recommends annual screening beginning at age 40 years.⁵ Similar to ACR, the American Cancer Society (ACS) 2007 guidelines recommend high-risk women begin screening at age 30 years with adjunctive MRI.⁶ The ACS 2015 guidelines for average-risk women recommend beginning annual screening at age 45 years and switching to biennial screening at age 55 years.⁷ ACR and ACS share some similarities in defining high risk as women with a lifetime breast cancer risk>20% based on risk assessment tools and women or their first-degree relatives with known pathogenic mutations.

In addition to the updates to screening guidelines, over the last decade there has been an almost complete transition from digital mammograms (2D) to digital breast tomosynthesis (3D).⁸ Given these changes, it is expected that the cost of screening has increased substantially from the previous estimate in 2010. This study had three analytical approaches. First, using real-world data from the commercial payer perspective, we estimated the total, national direct expenditure on breast cancer screening in the USA until the point of diagnosis in 2019 through 2022 and per cancer detected in 2022. Second, we projected the yearly total national cost of screening for three guidelines: 2021/2023 ACR, 2007/2015 ACS, and 2024 USPSTF. Third, we estimated the average lifetime cost to screen one woman until 74 years for each guideline.

Methods

Study design

This study had three analytical approaches. First, we estimated the total national cost of screening per year between 2019 and 2022 and the cost per cancer detected in 2022 using data on the number of mammograms reported each year to the Food and Drug Administration (FDA) Mammography Quality Standards Act (MQSA) and Program⁹ and data on MRI utilisation rates from the literature. The period 2019 to 2022 was selected to account for the decrease in screening uptake due to the COVID-19 pandemic, which began in 2020.¹⁰ Second, we used life table modelling to estimate the total national cost of screening per year for three guidelines: 2021/2023 ACR,^{4 5} 2024 USPSTF³ and 2007/2015 ACS.^{6 7} The age to start and stop screening, frequency and recommended modalities varies by guideline (table 1). Total cost was calculated by summing the costs of screening and recall for commercial and Medicare Advantage (MA) insurance lines using the total cost equation presented in the online supplemental methods. One-way sensitivity analyses were used to investigate which inputs were the largest drivers of total costs for each guideline. Third, we used life table modelling and simulated a cohort of 11 million 30-year-old women and estimated the average lifetime screening cost per woman until age 74 years for each guideline. All modelling was completed using R V.4.1.3 and Microsoft Excel Visual Basic V.16.63.

Table 1. Breast cancer screening guidelines modelled.

Guideline	Risk	Start age	Stop age	Frequency and modality
2021/2023 American College of Radiology4 5	High*	30 or 40†	Per woman’s health status‡	Alternating MRI and mammogram every 6 months
	Average	40		Annual mammogram
2024 United States Preventive Services Task Force3	High*	40	74	Biennial mammogram
	Average
2007/2015 American Cancer Society6 7	High*	30	Per woman’s health status and life expectancy>10 years‡	Alternating MRI and mammogram every 6 months
	Average	45		Annual mammogram: age 45–54Biennial mammogram: age 55 and over

^*See Supplementary Methods andonline supplemental methods and Table S1online supplemental table S1 for definitions of high- risk.

^†Women with a lifetime risk of breast cancer of 20% or greater according to risk assessment tools, with pathogenic mutations or first-degree relatives start screening at 30 years. Women with dense breasts who desire supplemental screening start at 40 years.

^‡Screening simulation stopped at 84 years.

Input variables

Model inputs were obtained from the scientific literature, government databases and the Optum Labs Data Warehouse (OLDW) (table 2). OLDW^{11 12} contains deidentified administrative claims, medical and pharmacy claims, laboratory results and enrolment records for commercial and MA enrolees. The database contains longitudinal health information for over 200 million enrolees and patients, representing a mixture of ages and geographical regions across the USA. The primary study cohort included female patients aged 18–74 years with at least one claim for a breast cancer screening mammogram between 1 January 2018 and 31 December 2018; the date of the first claim for screening was the ‘index date’. Continuous enrolment from 1 January 2016 to the index date (‘baseline period’) and through 6 months after the index date in 2018 (‘follow-up period’) was required to ensure that all administrative claims were available during the study period. Other study cohorts were used to estimate 3D utilisation rates in 2019 and 2020 (see online supplemental methods). All screening and diagnostic procedures were identified based on Current Procedural Terminology and Healthcare Common Procedure Coding System codes (online supplemental table 2).

Table 2. Inputs used for yearly total cost analyses 2019–2022.

Costs*	Median cost of service (IQR)	Median cost per recall after service (IQR)
2D MA	US$139(US$135–US$148)US	$206.00(US$133–US$310)
2D COM	US$252(US$196–US$344)US	$443.00(US$235–US$871)
3D MA	US$195(US$189–US$206)US	$226.00(US$128–US$373)
3D COM	US$360(US$280–US$469)US	$524.00(US$303–US$1045)
MRI MA	US$545(US$501–US$577)US	$1,408.00(US$715–US$1587)
MRI COM	US$1,498.00(US$927–US$2200)US	$1,408.00(US$715–US$1587)
Mammography utilisation*	80% 3D; 20% 2D
Number of screening mammograms9	34 M–35M
Insurance utilisation 14	60% commercial; 40% Medicare
MRI screening rate13	12.1 MRIs per 10 000 women
Median recall rates
3D MA*	9%
3D COM*	11%
2D MA*	11%
2D COM*	12%
MRI COM and MA13	12%

^*Estimated using Optum Labs Data Warehouse.

COMcommercial2Ddigital mammograms3Ddigital breast tomosynthesisIQRinterquartile rangeMAMedicare Advantage

Using OLDW we estimated: (1) utilisation rates of 2D and 3D; (2) median and interquartile range (IQR) for 2D, 3D and MRI screening costs for commercial and MA insurance; (3) median and IQR for MA and commercial insurance recall costs after 2D, 3D and MRI; (4) recall rates after MA and commercial 2D, 3D and MRI. All input screening procedure costs and recall costs obtained from OLDW are the sum of payer-paid and patient-paid amounts. MA costs and recall rates were used as a proxy for traditional MA and recall rates. MA, also referred to as part C, is an alternative to traditional Medicare (parts A and B) that private insurance companies offer. These plans provide both part A (hospital insurance) and part B (medical insurance) benefits. A recall was defined as additional screening mammography, diagnostic mammography, diagnostic ultrasound, MRI, fine needle aspiration, biopsy, tissue examination and immunohistochemistry in the 6-month follow-up period after the index mammography date. Recall costs were calculated as the median cost for the additional tests. The recall rate was the percentage of women recalled for an additional procedure after the index mammography data. Recall rates after the first screening MRI could not be calculated because procedure codes do not differentiate between screening and diagnostic MRIs. Therefore, an MRI recall rate of 12% was used as estimated by Warren et al using a UK screening cohort of women at high genetic risk.¹³

We modelled 80% 3D and 20% 2D utilisation to match real-world utilisation, estimated using OLDW data for 2020. Commercial and Medicare utilisation rates were obtained from the US Census Bureau 2021 estimates.¹⁴ We assumed everyone who screened had insurance coverage to simplify modelling.

Total cost per year from 2019 to 2022 and per cancer detected in 2022

Inputs for total cost per year and per cancer detected are listed in table 2. We estimated the total cost of screening and recall per year between 2019 and 2022 using data on the number of mammograms reported each year to the FDA MQSA which was between 39 and 40 million (M).⁹ We assumed 88% of the mammograms were for screening (34–35M) based on data from 4 million screening mammograms reported to the Breast Cancer Surveillance Consortium between 1996 and 2008.¹⁵ We calculated that there were approximately 95M screening-eligible women (see projected total cost calculation below) which is a yearly participation rate of approximately 37% (35M of 95M). This is consistent with data from the Center for Disease Control and Prevention (CDC) Behavioral Risk Factor Surveillance System that in 2022, 77% of women aged 50–74 years had a mammogram in the last 2 years.¹⁶ An uptake rate of 12.1 MRIs per 10 000 women was used to estimate the number of screening MRIs per year, which was based on breast MRI screening rates in a cohort of 10M US women.¹⁷ We estimated the cost per cancer detected in 2022 via 3D/2D screening by dividing the total cost of 3D/2D screening by the number of cancers detected by 3D/2D. The number of cancers detected was estimated using a 3D and 2D cancer detection rate of 5.3 per 1000 screens and 4.7 per 1000 screens, respectively, which includes invasive and ductal carcinoma in situ cases, based on a screening population from the Breast Cancer Surveillance Consortium.¹⁸ Total costs from 2019 to 2022 and per cancer detected in 2022 calculations are included in online supplemental file 2.

Projected total cost per year for each of the three screening guidelines

Inputs for projected total cost calculations for the three modelled guidelines are listed in table 3 and online supplemental table S3. The number of women alive in each age group was determined using the 2020 American Community Survey 5-year estimate.¹⁹ To limit the cohort to screening-eligible women, we subtracted the number of women diagnosed with invasive breast cancer by age, determined using the Surveillance, Epidemiology and End Results (SEER) 2015–2019 female breast cancer incidence rates per 100 000,²⁰ and the number of women living with invasive breast cancer by age determined using SEER breast cancer prevalence numbers on 1 January 2020.²⁰ There were approximately 95M screening-eligible women. The proportion of women who met the criteria for screening according to each guideline was estimated using the literature (see online supplemental methods and online supplemental table S1).²⁰ We used screening participation rates by age from the 2019 National Health Interview Survey (54%–78%), which reported the proportion of women who were screened in the last 2 years, as the target yearly participation rate.²¹ Projected total cost calculations per guideline are included in online supplemental file 2.

Table 3. Inputs used for projected total cost of screening for the three guidelines, sensitivity analyses and average lifetime cost of screening until age 74 years.

Costs*	See table 2
Mammography utilisation*	80% 3D; 20% 2D
	30–39years	40–44years	45–49years	50–64years	65–69years	70–74years	75+years
Target participation rates of screening-eligible women‡21	60.2%†	60.2%	60.2%	75.7%	78.1%	78.1%	54.2%
Insurance utilisation14
Medicare	0%	0%	25%	25%	95%	95%	95%
COM	100%	100%	75%	75%	5%	5%	5%
Proportion of high-risk women per guideline	See online supplemental table 2
Recall rates*
2D MA	n/a§	n/a§	12.5%	11.2%–12.6%	9.6%	9.2%	7.4%†
2D COM	20.9%	17.90%	14.0%	9.8%–11.9%	9.9%	10.3%	9.4%†
3D MA	n/a§	n/a§	11.2%	8.9%–10.6%	8.5%	8.3%	8.1%†
3D COM	19.9%	15.6%	13.2%	9.3%–11.3%	9.4%	10.3%	9.4%†
MRI13	12%	12%	12%	12%	12%	12%	12%

^*Estimated using Optum Labs Data Warehouse. See Table S3online supplemental table S3 for detailed recall rates for 50-–64 years.

^†The proportion of screening-eligible average-risk and high-risk women in each age group varies by guideline. See Supplementary Methods for calculations.Imputed because data were unavailable. Participation rate for 40–44 years was used for 30–39 years. Recall rates were calculated based on the rate of change across age groups.

^‡Imputed because data was unavailable. Participation rate for 40-44y was used for 30-39y. Recall rates were calculated based on the rate of change across age groups.The proportion of screening-eligible average-risk and high-risk women in each age group varies by guideline. See online supplemental methods for calculations.

^§Not estimated due to low or no Medicare Advantage coverage.

COMcommercial2Ddigital mammograms3Ddigital breast tomosynthesisMAMedicare Advantage

Sensitivity analyses

A probabilistic sensitivity analysis framework was used to estimate the uncertainty in the estimated total projected cost per guideline. One-way deterministic sensitivity analyses were used to determine the impact of different inputs on the total projected cost of each guideline. Inputs with the widest ranges of possible values were selected. See online supplemental methods and online supplemental tables S4 and S5 for details on the inputs and methods used.

Average lifetime cost per woman screened until 74 years

We simulated a hypothetical cohort of 11M 30-year-old women, which is approximately the number of 30-year-old women in the USA in 2019 based on data from the American Community Survey.¹⁹ In each year, the number of women who died from any cause by age and the number of women who were diagnosed with invasive breast cancer by age was subtracted to determine the number of screen-eligible women (see online supplemental methods). The cohort was followed until 74 years to simplify comparisons between the strategies. The total cost over the 44-year follow-up period for each guideline was divided by the total number of screen-eligible women at the end of the follow-up period to estimate an average cost per woman.

Patient and public involvement

Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

Results

OLDW cohort summary

In the OLDW primary cohort, 939 410 women met the study criteria. Of these women, 1.0% were <40 years, 22.7% were 40–49 years, 46.1% were 50–64 years and 30.1% were 65–74 years old. 70% were commercially insured and 30% were MA enrolees. Utilisation of 3D in calendar year 2018 was 60.4% which increased to 73.1% in 2019 and 82.6% in 2020.

Costs and recall rates

The median cost of a mammogram ranged from US$139 for 2D under MA to US$360 for 3D under commercial insurance (table 2). Overall, 3D was US$56–US$108 more than 2D and commercial costs were US$113–US$165 higher than MA costs for the same screen. The median recall costs after a mammogram ranged from US$206 to US$524 (table 2). Overall, recall costs after 3D mammography were US$20–US$81 more than 2D recall costs and commercial recall costs were US$237–US$298 higher than MA recall costs for the same screen. Commercial MRI was approximately four times the cost of commercial 3D.

Recall rates ranged from 8.1% to 20.9% (table 3). Recall rates on MA were overall lower and had a smaller variation by age (8.1%–12.6%) than commercial insurance (9.3%–20.9%). The highest recall rates (15.6%–20.9%) were for commercially insured women aged 30–44 years. In most age groups, recall rates were 1%–2% lower for 3D than 2D for both commercial and MA (table 3).

Total cost of screening and recall from 2019 to 2022 and cost per cancer detected in 2022

The aggregate cost of screening and recall per year between 2019 and 2022 was approximately US$11B (figure 1) with a yearly participation rate of approximately 37% of eligible women. The direct cost of screening was US$9.4–US$9.7B, and the cost of recall was approximately US$1.5B. The cost of screening high-risk women from 2019 to 2022 with an MRI contributed to 0.5% of the cost at approximately US$60 million (table 4) due to low utilisation rates. Based on 3D and 2D cancer detection rates, we estimated that in 2022, US$55 471 and US$44 000 were spent in aggregate to detect one DCIS or invasive cancer via 3D and 2D, respectively.

Figure 1. Estimated total national cost of breast cancer screening in the USA for 2019–2022 using data from the Food and Drug Administration Mammography Quality Standards Act and Program. *The total includes the cost of MRI screening and recall in 2019 to 2022 which was small at $0.06B (Table 4) and therefore not visualized.

Table 4. Estimated total yearly cost for screening and recall in billions of USD using data from the Food and Drug Amdinistration Mammography Quality Standards Act and Program.

Year	Total number of women screened	Total cost of screening and recall in billions of USD (IQR)	Screening mammograms (n)	Cost of mammography screening and recall in billions of USD (IQR)	Women screened with MRI (n)	Cost of MRI screening and recall in billions of USD (IQR)
2019	34.96M	US$11.23 (10.21–14.08)	34.92M	US$11.17 (10.15–14.03)	0.04M	*US$0.06 (0.04–0.07)
2020	34.49M	US$11.08 (10.08–13.89)	34.45M	US$11.02 (10.09–13.89)	0.04M	US$0.06 (0.04–0.07)
2021	34.19M	US$10.98 (9.99–13.77)	34.15M	US$10.92 (9.93–13.72)	0.04M	US$0.05 (0.04–0.07)
2022	34.94M	US$11.22 (10.18–14.07)	34.9M	US$11.16 (10.13–14.01)	0.04M	US$0.06 (0.04–0.07)

Approximately, 37% of eligible women were screened per year. Values in brackets are the interquartile range (IQR) from the probabilistic sensitivity analysis.

MmillionsUSDUS dollars

Projected annual total cost of screening and recall for three guidelines

Using target yearly participation rates of 54%–78% by age, the projected yearly cost of the 2021/2023 ACR guideline was US$29.93B, 2007/2015 ACS was US$17.78B, 2024 USPSTF was US$7.67B (figure 2). Screening high-risk women contributed to 55% of the cost of ACR, 52% of the cost of ACS and 19% of the cost of USPSTF guidelines (table 5).

Figure 2. Projected yearly cost of screening using the 2021/2023 American College of Radiology (ACR), 2007/2015 American Cancer Society (ACS) and 2024 United States Preventative Services Task Force (USPSTF) guidelines assuming target yearly participation rates of 54%–78% by age.

Table 5. Projected total yearly cost of screening and recall in billions of USD for three guidelines assuming target yearly participation rates of 54%–78% by age.

Guideline	Total number of women screened	Total cost of screening and recall in billions of USD (IQR)	Average-risk women (n)	Cost of screening and recall for average-risk women in billions of USD (IQR)	High-risk women (n)	Cost of screening and recall for high-risk women in billions of USD (IQR)
2021/2023 ACR	52.44M	US$29.93 (26.13–35.63)	42.75M	US$13.42 (12.76–17.83)	9.69M	US$16.51 (11.69–19.27)
2007/2015 ACS	32.52M	US$17.78 (13.42–19.66)	27.25M	US$8.51 (6.26–10.21)	5.27M	US$9.27 (5.92–10.39)
2024 USPSTF	25.53M	US$7.67 (6.94–9.8)	20.81M	US$6.25 (5.53–6.97)	4.72M	US$1.42 (0.69–2.16)

Values in brackets is the interquartile range (IQR) from the probabilistic sensitivity analysis.

.ACR, American College of Radiology; ACS, American Cancer Society; MmillionsUSDUS dollarsUSPSTF, US Preventative Services Task Force

Average lifetime screening cost per woman until age 74 years

We estimated the average cost to screen a woman for each guideline until 74 years. The average cost of screening a high-risk woman with pathogenic mutations, first-degree relatives or a lifetime risk>20% was US$89 725 for 2021/2023 ACR and 2007/2015 ACS (online supplemental figure S1). For ACR, screening high-risk women with dense breasts cost US$65 579. For 2024 USPSTF, screening women at increased risk cost US$6931 per woman. The average cost of screening an average-risk woman was: US$13 416 for ACR, US$7946 for ACS, US$6931 for USPSTF (online supplemental figure S2).

Determinants of cost

For the ACR and ACS guidelines, variation in participation rates, the proportion of women with a lifetime risk greater than 20% followed by commercial MRI cost had the largest impact on costs. For the USPSTF guideline, participation rates, commercial cost of 3D and the cost of recall after commercial 3D had the largest impact on cost (figure 3).

Figure 3. One-way sensitivity analysis showing the top five variables that had the largest impact on total yearly costs for each guideline: (A) 2021/2023 American College of Radiology (ACR); (B) 2007/2015 American Cancer Society (ACS); (C) 2024 United States Preventative Services Task Force (USPSTF). The centre line is the estimated total cost in the base case. 3D, digital breast tomosynthesis.

Discussion

We estimated that breast cancer screening in the USA cost approximately US$11B per year from 2019 to 2022, with approximately 37% of eligible women screened. In 2022, it cost approximately US$55 471 and US$44 000 to detect one case of DCIS or invasive cancer via 3D and 2D, respectively. These estimates were based on real-world data from the FDA MQSA on the number of mammograms performed annually. The cost to detect one breast cancer is in the range of reported costs to treat a commercially insured women with stage 0 ‘breast cancer’ in the first year.²² Using target screening yearly participation rates of 54%–78% by age, we projected that if the 2021/2023 ACR guidelines were used nationally, it would cost approximately US$30B per year, or US$19B more than current practice. The 2007/2015 ACS guidelines would cost US$18B per year or US$7B more than current practice. The 2024 USPSTF guidelines would cost US$8B per year or US$3B less than current practice because it recommends the least frequent screening strategy.

The US$11B cost of screening in 2022 is US$3B more than a previous estimate of US$8B in 2010.² The increase in cost over time can be attributed to at least two factors. First, we include commercial and Medicare costs which include payer-paid and patient-paid amounts, whereas the previous estimate only modelled Medicare costs.² Second, the estimated 80% adoption of 3D would increase cost, since commercial 3D costs approximately US$100 more than commercial 2D. The widespread adoption of 3D has been attributed to reported improved detection rates and decreased recall rates.^{23 24} On the other hand, several experts have cautioned against 3D adoption as the clinical benefits are still unclear and a randomised controlled trial comparing the incidence of advanced cancers between 3D and 2D is still underway; the trial is expected to be completed in 2030.²⁴ Of the women who screen, a majority (~82.7%) screen annually, consistent with ACR guidelines.²⁵ Therefore, as expected, our estimated cost of screening from 2019 to 2022 was closest to the projected cost of screening average-risk women with ACR guidelines. The decline in costs between 2019 and 2021 can be attributed to a decline in screening due to the COVID-19 pandemic.¹⁰

The ACR and ACS guidelines are the most expensive partly because a substantial proportion (8%–17%) of women may meet the criteria for high risk and are recommended supplemental MRI screening, the most expensive screening modality. The ACR and ACS guidelines recommend using risk models that consider family history such as BRCAPRO and Tyrer-Cuzick where the proportion of women determined to have a lifetime risk>20% was shown to range widely from 0.6% to 12%, respectively.²⁶ Improved concordance between risk models on who is determined to be high-risk is needed, since this is a significant driver of costs. Further, the 2023 ACR guidelines for high-risk women also recommend that women with dense breasts who desire supplemental screening have an MRI beginning at 40 years. This recommendation was a significant driver of costs, based on our estimate that approximately 8.6% of women would be screened.

Screening breast MRI utilisation in the USA is low, partly due to low availability. Choudhery et al estimated that 20% (ranging from 16% in the West region to 37% in the South region) of clinics surveyed in 2020 reported having supplemental MRI screening available.²⁷ In 2022 0.4% of women aged 25–64 years had a screening breast MRI while approximately 9% were thought to be eligible.²⁸ Therefore, while MRI is an effective screening tool for high-risk women,²⁹ its availability for screening, impact on adherence and cost must be thoughtfully considered by professional organisations that create guidelines. Abbreviated MRI, contrast-enhanced mammography, and high-resolution diffusion-weighted imaging are all promising imaging tools that can reduce the cost of screening and should be further investigated particularly in high-risk women.

Our costs and recall rates estimated using OLDW are in range with those from other datasets and are therefore reasonable estimates of median national costs and rates. For example, 2D and 3D recall rates were reported to range from 9% to 11%,³⁰ consistent with our estimated recall rates (table 2). Our estimated commercial cost of 2D and 3D mammograms are also consistent with costs estimated using the Blue Cross Blue Shield dataset (2D: US$252 vs US$213; 3D: US$360 vs US$299)³¹ (table 2). Medicare reports that a mammogram costs approximately US$170, which is in the range of our median 2D (US$139) and 3D (US$195) MA costs.

The limitations of screening using age as the primary risk factor have been extensively discussed in the literature. In response, a more personalised risk-based approach to screening is being tested in clinical trials such as Women Informed to Screen Depending On Measures of risk in the USA,³² My Personal Breast Cancer Screening in Europe³³ and BREAst screening Tailored for HEr in Singapore.³⁴ Personalised risk assessment uses more advanced risk models that consider additional risk factors such as body mass index and polygenic risk scores. The goal of these trials is to compare the safety, efficacy and acceptability of risk-based screening. Modelling studies have shown that risk-based screening is more cost-effective than age-based screening.35,37 Therefore, it is plausible that risk-based screening would also reduce total screening expenditure, but this has not been shown. Future work can model the cost of different risk-based screening strategies once trial outcomes are published.

Limitations

This study had some limitations. We could not determine recall rates after the first versus subsequent mammogram, which are known to differ. We did not include Medicaid costs or the cost of risk assessment for ACR and ACS guidelines, as these data were not available in OLDW. Traditional Medicare screening and recall costs may vary from MA costs. We did not include the cost of adjunct screening ultrasound for women with dense breasts in modelling the cost as practised which would further increase the cost. We did not model the cost of clinic visits, treatment or surveillance costs, or the indirect costs associated with a screening programme such as human resource and maintenance costs or costs to the patient for transportation and days lost from work. The FDA MQSA aggregate number of yearly mammograms does not include those from non-MQSA certified and Veterans Hospital Administration. The costs of overdiagnosis and missed diagnoses were also not included. In the sensitivity analysis, we assumed independence among costs and recall rates, which have dependencies in practice. Given all limitations, the costs estimated are likely to be an underestimation. However, this work aimed to provide an updated, reasonable estimate of screening cost rather than provide highly accurate point estimates. Future work can address our limitations and investigate how the cost of screening varies under different personalised risk-based algorithms where the method for risk profiling, age to start and stop screening and screening modality are varied.

Conclusion

The total national direct cost of breast cancer screening in the USA was approximately US$11B per year in 2019–2022 and cost US$55 471 and US$44 000 per DCIS or invasive case detected by 3D and 2D, respectively. The 2021/2023 ACR guideline was the most expensive guideline costing US$30B per year. Participation rates, the proportion of women with a lifetime risk>20% and the commercial cost of MRI were key drivers of cost. These costs should be considered alongside data on outcomes when developing screening guidelines. Future work can model the national cost of the risk-based screening algorithms being tested in clinical trials.

Data availability statement

Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as supplemental information.