Implications of the initial Braidwood v. Becerra ruling for colorectal cancer outcomes: a modeling study

Abstract

The Affordable Care Act (ACA) eliminated patient cost-sharing for United States Preventive Service Task Force (USPSTF) recommended services. However, if the US Court of Appeals for the Fifth Circuit fully upheld a US District Court ruling in Braidwood Management v. Becerra, 666 F. Supp. 3d 613 (N.D. Tex 2023), cost-sharing for USPSTF recommendations made after ACA passage would have been reinstated for more than 150 million people. The case would have reinstated cost-sharing for colorectal cancer (CRC) screening for ages 45-49 years and for polyp removal during (diagnostic) colonoscopy across all ages. Using the MISCAN-Colon model, we simulated the potential impact on CRC outcomes, assuming early-onset CRC trends and lower screening participation. An 8-percentage-point decline in screening participation could increase CRC incidence by 5.1% and CRC mortality by 9.1%, with slightly lower costs due to increased cost-sharing. Larger decreases in screening participation can result in higher costs from increased incidence and delayed diagnoses.

The Affordable Care Act (ACA) became law in March 2010 and increased access and affordability of preventive services by requiring nearly all private health insurance plans and Medicaid programs in expansion states to cover a wide range of recommended preventive services without patient deductibles, co-payments, or co-insurance. The ACA provision eliminating patient cost-sharing was informed by research demonstrating higher rates of preventive services use among individuals without financial barriers (1). It includes all recommendations that received an “A” or “B” rating by the United States Preventive Service Task Force (USPSTF) (2-5). Reducing these financial barriers increased the use of preventive services (6,7) and reduced health disparities (8). The federal court case of Braidwood Management v. Becerra, 666 F. Supp. 3d 613 (N.D. Tex 2023) (9), which claims that the preventive services requirements in the ACA are unconstitutional, could reinstate patient cost-sharing for USPSTF-recommended services that were made after the passage of the ACA in 2010 and eliminate coverage for those services that were recommended after the passage of the ACA.

Since the ACA’s implementation, the recent 2021 USPSTF recommendation for colorectal cancer (CRC) screening lowered the recommended screening start age from 50 to 45 (10), based on increases in early-onset CRC (11). Furthermore, cost-sharing for colonoscopy screening with polyp removal (12) or follow-up colonoscopy after a positive noninvasive test across all ages was recommended to be waived, because diagnostic colonoscopies are part of the screening process (13). If the initial Braidwood ruling is ultimately upheld, CRC screening rates will likely decline due to increased financial barriers (14,15). Missed opportunities for prevention and early detection can also result in higher cancer treatment costs and worse survival.

In this study, we estimated the potential impact on CRC outcomes and associated costs if the initial Braidwood ruling is upheld. We used the MISCAN-Colon microsimulation model, which has been calibrated to US population incidence and mortality data (16,17), to simulate the effect of this ruling for the US population aged 40 years and older between 2025 and 2055. We accounted for the latest trends in CRC incidence by increasing population risk over time based on an age-period-cohort model fitted to SEER data (18) (Supplementary Figure 1, available online), including the increase in early-onset CRC. Modeling assumptions that were varied included initial and ongoing screening participation and participation in follow-up and surveillance colonoscopy.

Primary screening was modeled as decennial colonoscopy screening or annual FIT screening, in line with current recommendations for average-risk individuals (10,19) (see Supplementary Table 1, available online, for test characteristics). In the “current state” without cost-sharing, 20% of the population starts screening at age 45 (20). Screening parameters in MISCAN-Colon were calibrated to match observed screening data (20,21), resulting in a gradual increase in the proportion of the population in the model who were screened at least once with either FIT or colonoscopy from 52.5% at age 50 to 88% at age 75 (Supplementary Tables 2 and 3, available online). The calibrated proportion of the population up-to-date with screening recommendations increases from 20% at age 45, to 65% at age 60, and to 66% at age 75. Participation in colonoscopy after a positive FIT and guideline-recommended surveillance after polyp removal (19) was based on claims and electronic health record data in the United States (60%) (22).

As the USPSTF recommendations to start screening at age 45 and explicit mention of diagnostic colonoscopy as part of the screening process were only introduced in 2021 and elimination of cost-sharing was clarified by the US Departments of Labor, Health, and Human Services and Treasury in 2022 (13), we assumed that screening rates before 2021 in the “current state” scenario were equal to the screening rates in the Braidwood scenario.

In the Braidwood scenario, we assumed that the initial Braidwood ruling was upheld, leading to decreased screening participation from 2025 onward. Because the lower court decision was paused during appeal, the true effect of Braidwood on screening participation is uncertain. For individuals who already reached the screening start age in 2025, we assumed that the reintroduction of cost-sharing decreased the probability of continuing participation by 5 percentage points (PPT). This was based on both the reimplementation of cost-sharing and the assumption that individuals who follow up abnormal noninvasive tests with colonoscopy or undergo screening colonoscopy and receive an unexpected bill due to polyp removal may be discouraged from future screening. For individuals who did not reach the starting age yet, we assumed that reversing the mandate to cover the grade “B” recommendation to start screening at age 45 decreased screening participation at age 45 by 50%, because some health plans might continue to cover screening in this age group. Together, this resulted in a 7-PPT drop in individuals with at least once screening by age 75 (Supplementary Table 3, available online). Concurrently, the proportion of individuals who were up-to-date with screening recommendations decreased by 8 PPTs. Participation in follow-up and surveillance colonoscopies was also assumed to decrease by 10 PPTs because of patient cost-sharing.

We considered multiple sensitivity analyses. Specifically, we evaluated the change in CRC outcomes per 2-PPT decrease in initial screening participation and per 1-PPT decrease in continuing screening rates and adherence rates to surveillance and follow-up screenings (ranging from 2 PPTs to 20 PPTs and from 1 PPT to 10 PPTs, respectively). Subsequently, the proportion of the population who were screened at least once at age 75 ranged from 64.3% to 86.2%, and the proportion up-to-date with screening recommendations ranged from 39.4% to 64.3% (Supplementary Table 4, available online). Next, we considered a scenario in which the Braidwood ruling reinstates cost-sharing for diagnostic colonoscopies after a positive FIT, but not for screening colonoscopies with polyp removal, as cost-sharing for diagnostic colonoscopies may be more directly related to the Braidwood ruling. Finally, we considered a similar scenario where cost-sharing for diagnostic colonoscopies applies only to individuals younger than 65, as Medicare has eliminated cost-sharing under a separate set of rules. In both scenarios, initial screening was affected as in the base case scenario.

Modeled outcomes included CRC incidence and CRC mortality, quality-adjusted life-years (QALYs) gained, and associated costs. Total costs related to CRC screening and treatment were computed from a payer perspective (Supplementary Table 5, available online). In the Braidwood scenario, we assumed a coinsurance of 20% for all colonoscopies with polyp removal and colonoscopies performed after a positive FIT, resulting in a cost reduction from the payer perspective by 20% (23).

In the current state, screening resulted in 142 annual CRC cases and 45 annual CRC deaths per 100 000 individuals in 2055, with a total cost of $2.6 million and 492 QALYs gained compared with no screening (Figure 1). If the initial Braidwood ruling is upheld, we project an additional 7 (+5.0%) annual CRC cases and 4 (+8.7%) annual CRC deaths per 100 000 individuals in 2055 compared with the current state. Moreover, the number of QALYs gained from screening would decrease by 44 (-8.9%). The increase in CRC incidence and mortality was largely attributable to lower screening participation (+3.1% and +6.0%, respectively). Moreover, later screening initiation increased the number of CRC cases by 1.9% and CRC deaths by 2.3%.

Figure 1.

Colorectal cancer (CRC) incidence (A), CRC mortality (B), total costs related to CRC (C), and number of quality-adjusted life-years (QALYs) gained compared with no screening (D) per 100 000 individuals with 1) current screening participation, and 2) screening participation if the Braidwood ruling is upheld. This would reverse USPSTF-recommended services that were made after the implementation of the ACA: 1) lowering the screening start age to 45, leading to later screening initiation, and 2) waiving cost-sharing for diagnostic colonoscopies, leading to lower ongoing screening rates across all screening ages. The dashed lines represent scenarios that isolate the impact of Braidwood on initial screening participation and ongoing screening participation.

Cost from a payer perspective under the initial Braidwood ruling would decrease in 2025 due to lower screening costs and increased out-of-pocket expenses for patients. However, in the long term, total costs are expected to be higher than in the current scenario because of increased CRC incidence and delayed diagnoses, both of which substantially increase corresponding treatment costs.

Incidence, mortality, and costs in 2055 linearly increased, whereas QALYs decreased linearly with the decreasing screening participation (Figure 2). Even if the initial Braidwood ruling resulted in only a small decrease in screening participation of 5 PPTs, CRC incidence and mortality increases would be around 5%. Moreover, costs only slightly increased for participation drops greater than 10 PPTs, and potential cost savings did not outweigh the substantial decrease in QALYs gained. If cost-sharing was reinstated only for diagnostic colonoscopies, CRC incidence and mortality increased by 2.1% and 2.9%, respectively. If cost-sharing applies only to individuals younger than 65, the increases in CRC incidence and mortality were slightly lower (+2.0% and 2.6%, respectively).

Figure 2.

Change in colorectal cancer incidence, mortality, quality-adjusted life-years (QALYs) gained, and costs in 2055 by percentage point (PPT) decrease in screening participation under the Braidwood ruling compared with the current screening state.

In this study, we show that if the initial Braidwood ruling is upheld, the benefits of CRC screening would be substantially reduced through later screening initiation and lower ongoing screening participation. Effects would likely be greatest for individuals with limited financial resources, potentially widening existing disparities in CRC outcomes in the United States. CRC screening is only one of many evidence-based preventive services with USPSTF recommendations since 2010. The overall adverse public health effects of upholding the initial Braidwood ruling are likely much larger when considering, for example, the effects of reintroducing cost-sharing for tobacco cessation and screening for anxiety, hypertension, and hepatitis B and C.

Our findings are especially relevant, as there is clear evidence that early-onset CRC incidence in the United States is increasing before age 50 (11,24), and previous analyses have illustrated that reducing the screening start age to 45 provides an efficient balance of colonoscopy burden and life-years gained from screening (18,25,26). Moreover, it was estimated that waiving Medicare cost-sharing for colonoscopy screening is already cost-effective from a payer perspective if screening rates increase even slightly, from 60% to 60.6%, suggesting a very favorable balance of health and cost impact (27). The increased treatment costs and decreases in QALYs gained from screening underscore the potential harms of upholding the initial Braidwood ruling for CRC burden.

Although a limitation of this study is that the immediate and long-term effect of the initial Braidwood ruling on screening participation is unknown, modeling provides insights into its potential impact. Although individuals will likely initiate screening at a later age, it is uncertain how cost-sharing will affect ongoing screening participation, and it depends, in part, on how many health plans abandon current USPSTF-recommended coverage without cost-sharing. To comprehensively evaluate the potential impact of the initial Braidwood decision, we conducted a range of sensitivity analyses. Although this range is consistent with changes observed in post-ACA screening rates (4,5,28-30), our study specifically models coverage changes for insured individuals. The observed post-ACA changes in screening rates appear to have a smaller affect across this population. Another limitation is that we did not consider the multitarget stool-DNA test, because there are limited national data on the type of stool tests people received. This test was added to the USPSTF recommendations only after 2010, which would affect coverage mandates under Braidwood. Given its higher sensitivity compared with FIT, and the likelihood that individuals switched to multitarget stool-DNA following the recommendation, we may have underestimated the effect of Braidwood. Finally, private health insurer timelines for evaluating returns on investment for coverage decisions can be substantially shorter than our study timeline. Nonetheless, the Medicare program is the primary payer for the population aged 65 years and older in the United States and will ultimately bear many of the costs associated with higher CRC incidence, later-stage diagnoses, and more intensive treatment, underscoring the utility and policy relevance of our findings.

Recently, the US Court of Appeals for the Fifth Circuit partially reversed the initial court decision, reducing its potential impact to employees of the plaintiff. However, the Fifth Circuit still found the underlying structure of the USPSTF violates the US Constitution for the ACA provision that prohibits cost-sharing for recommended services. The US Department of Justice has since filed a motion indicating that it will be appealing this decision to the US Supreme Court. Concerns remain because the Fifth Circuit decision left open the possibility of lawsuits by different plaintiffs, and it is unclear what will happen at the US Supreme Court. In this study, we showed that upholding the initial Braidwood ruling and reversing USPSTF recommendations made after 2010 could reverse the progress made in CRC prevention and control and will likely increase treatment costs and CRC deaths and widen existing health disparities.

Author contributions

Rosita van den Puttelaar, MS (Conceptualization; Formal analysis; Methodology; Writing—original draft; Writing—review & editing), Kewei Sylvia Shi, MPH (Conceptualization), Robert Smith, PhD (Writing—review & editing), Jingxuan Zhao, MPH (Conceptualization), Margaret Katana Ogongo, PhD (Conceptualization), Matthias Harlass, MS (Writing—review & editing), Anne I. Hahn, MPH (Conceptualization; Project administration; Writing—review & editing), Ann G. Zauber, PhD (Conceptualization; Funding acquisition), Robin Yabroff, PhD, MBA (Conceptualization; Methodology; Visualization; Writing—review & editing), Iris Lansdorp-Vogelaar, PhD (Supervision; Writing—review & editing).

Supplementary material

Supplementary material is available at JNCI: Journal of the National Cancer Institute online.

Funding

This research was supported by grant U01CA253913 from the National Cancer Institute (NCI) as part of the Cancer Intervention and Surveillance Modeling Network (CISNET). Additional funding was obtained through National Institutes of Health (NIH)/NCI Cancer Center Support grant P30CA008748 (Zauber and Hahn) and a contract from the American Cancer Society. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health and the American Cancer Society.

Conflict of interest

The authors have no conflict of interest to disclose.

Data availability

The data generated in this study are available within the article and its supplementary data files.

Acknowledgments

We thank V. Paul Doria-Rose, Mary Rouvelas, and Anna Howard for their valuable comments and suggestions on a draft of the manuscript.