Hypofractionated Radiotherapy for Breast Cancer: Financial Risk and Expenditures in the U.S. 2008–2017

Abstract

Background:

Rising cancer care expenditures and technological advancement of shorter radiotherapy regimens have drawn significant attention to the use of hypofractionated radiotherapy in clinical care. We examine the costs of hypofractionated (HF-WBI) compared to conventional whole breast irradiation (CF-WBI) in the U.S. and investigate the influences of patient characteristics and commercial insurance on HF-WBI use.

Methods:

In a retrospective study using private employer-sponsored insurance claims, a pooled cross-sectional evaluation of radiotherapy in patients with commercial insurance was performed from 2008–2017. The study population included female patients with early-stage breast cancer treated with lumpectomy and whole breast irradiation.

Results:

A total of 15,869 women received HF-WBI, and 59,328 received CF-WBI. HF-WBI use increased 2008–2017. Community-level factors like a higher proportion of college graduates and greater mixed racial composition were associated with increased HF-WBI use. Mean insurer-paid radiotherapy expenditures were significantly lower for HF-WBI vs. CF-WBI (adjusted difference $6,375, 95% CI $6,147–$6,603). Mean patient out-of-pocket expenditure for HF-WBI was $139 less than that of CF-WBI. Geographic variation existed across U.S. states in HF-WBI use (range: 9.6–36.2%) with no consistent relationship between HF-WBI use and corresponding average cost differences between HF-WBI and CF-WBI.

Conclusions:

If trends continue, HF-WBI will soon become the dominant form of radiation treatment in the U.S. Although HF-WBI represents significant savings to the health care system and individual patients, no evidence indicated that a financial disincentive had slowed adoption of HF-WBI. Therefore, multi-level approaches, including individuals, the community, and health policy, should be utilized to promote cost-effective cancer care.

Significance:

Innovations to policies on cost-effective radiotherapy treatment might consider non-financial incentives to promote HF-WBI use.

INTRODUCTION

Randomized trials [1–5] have shown that hypofractionated whole-breast irradiation (HF-WBI), a shorter treatment course with higher doses of radiation per fraction, is equivalent in local recurrence and disease-free survival to conventional fractionated whole-breast irradiation (CF-WBI), currently the standard early-stage breast cancer treatment after breast-conserving surgery [6]. Based on these trials, the American Society of Radiation Oncology (ASTRO) published guidelines in 2011 advising HF-WBI administration for patients older than 50 years, having a pT1–2N0 tumor, and not treated with chemotherapy [7]. Since 2011, rising cancer care expenditures and technological advancement of shorter radiotherapy regimens have drawn significant attention to HF-WBI use in clinical care. HF-WBI has been found to be cost-effective compared to CF-WBI for women with early-stage breast cancer [8]. Though its use has increased since 2011 [9–12], adoption of HF-WBI in the U.S. remains low compared to other countries that have more readily adopted HF-WBI into clinical practice [13–15].

Financial disincentives in the U.S. health care system have been indicated as potential barriers to uptake of HF-WBI, implying changes in reimbursement may accelerate adoption [16]. Catalyzed by health reform legislation, the momentum behind moving health care payment schemes from volume-based to value-based has motivated the development of new reimbursement mechanisms in oncology care [17]. The Centers for Medicare and Medicaid Services (CMS) has led the development of financial innovations to support value-based cancer care [17]. However, the extent to which private payers have emulated CMS’s more recent initiatives to align financial incentives facilitating HF-WBI use is currently unknown, while employer-sponsored private insurance is the dominant form of health coverage in the U.S.

Within private insurance, studies have examined radiotherapy-related insurer and patient out-of-pocket (OOP) expenses [12] or evaluated the costs of different radiation treatment strategies [18]. Private insurance is a diversified, fragmented market varying in benefit structure and financial risk exposure for healthcare stakeholders (e.g., insurers/payers and patients). Additionally, receipt of evidence-based cancer treatment varies by state, and considerable state variation exists in the economic burden of breast cancer to which alternative radiotherapy treatment costs may be relevant [19].

In this study, we assess the use and costs of HF-WBI and CF-WBI across commercial health plans for their beneficiaries with breast cancer between 2008 and 2017. We also investigate geographic variation across states in use and costs associated with the type of radiotherapy. To our knowledge, our study is the first to address within the private insurance market both variation in health plan type and the cost differences across U.S. states in HF-WBI/CF-WBI use.

METHODS AND MATERIALS

Data

The Health Care Cost Institute (HCCI, https://healthcostinstitute.org/about-hcci) is an independent, non-profit entity that consolidates employer-sponsored private insurance claims data from Aetna, Humana, Kaiser Permanente, and UnitedHealthcare. HCCI collects data covering about 55 million commercially insured individuals/year, making its dataset a valuable research tool to investigate the current use and costs of cancer care within private insurance. The data contain information on health services use, enrollment, health spending, and essential plan characteristics.

Study Population

We identified women with incident breast cancer from the HCCI database 2008–2017 if they had at least two insurance claims with breast cancer diagnosis codes from the International Classification of Diseases, Ninth [ICD-9] and Tenth [ICD-10] Revisions within a year and received whole breast irradiation after breast conserving surgery according to Current Procedural Terminology (CPT) codes(eTable 1 in the Supplement). We excluded patients who started radiotherapy after August 31, 2017, as their radiotherapy course may have extended beyond the study period, preventing an accurate determination of the type of radiotherapy received. We further excluded patients with missing/incomplete information on radiotherapy delivery, males, older adults (>64 years), and those under Medicare managed care plans(eFigure 1).

Outcomes

The study’s primary outcomes are separated into two categories: use and costs. The first outcome is the use of hypofractionated versus conventional WBI. We defined HF-WBI as the receipt of 15–24 radiation fractions and CF-WBI as the receipt of 25–40 fractions. If the number of radiation fractions was ambiguous (11–14 or >40 fractions), we used days on radiotherapy to define HF-WBI (21–31 days) versus CF-WBI (39–120 days) (eFigure 2). Radiotherapy type was determined for 96.6% of patients using radiotherapy fractions and 3.4% using radiotherapy days. The other outcomes are health plan expenditures and patient out-of-pocket expenses for respective radiotherapy strategies during 2008–2017. We calculated total healthcare expenditures for radiotherapy by aggregating total net health plan payments and patient out-of-pocket expenses, including deductibles, copayments, and coinsurance amounts. We examined total healthcare expenditures paid by commercial insurers and patient out-of-pocket expenses for radiotherapy initiated within one year of breast-conserving surgery and finished within 120 days after radiotherapy initiation. This time window was taken after consulting with clinicians and exploring the claims patterns: radiotherapy for the current breast cancer starts within a year of breast-conserving surgery for almost all patients, and once it is initiated, the radiotherapy usually finishes within 120 days. Total healthcare expenditures included the costs reflected in insurance claims for radiotherapy simulation, planning, physics, and delivery and management identified from CPT codes (eTable 1). Patient out-of-pocket expenses included deductible, copayment, and coinsurance amounts that are only related to radiotherapy costs listed above. All costs were adjusted for inflation to reflect expenditures in 2017 U.S. dollars.

Other Variables

Additional variables evaluated within the study population are shown in Table 1. Age categories represent the age of women at the time of radiotherapy. The type of breast cancer diagnosis was classified as either invasive or ductal carcinoma in situ (DCIS). Receipt of chemotherapy was determined from CPT codes(eTable 1). We evaluated comorbid diseases using an adaptation to the Charlson comorbidity index [20,21]. The relation to the employee, indicating the dominant health insurance policyholder, was categorized as the employee, spouse, or a child/unknown. The types of employer-sponsored health plans evaluated included: point-of-service plan (POS), preferred provider organization (PPO), health maintenance organization (HMO), exclusive provider organization (EPO), or others. We also assessed health plans designated as high deductible plans, fully-insured, or self-funded (contracting with external private insurers for administrative services only).

Table 1.

Characteristics of female breast cancer patients treated with radiotherapy, 2008–2017

	Type of Radiotherapy
Characteristic	CF-WBI	HF-WBI
Total, No.	59328	15869
Age in years, No. (%)
<35	936 (1.6)	97 (0.6)
35–44	8234 (13.9)	1240 (7.8)
45–54	23281 (39.2)	5539 (34.9)
55–64	26877 (45.3)	8993 (56.7)
Type of breast cancer diagnosis, No. (%)
Invasive	53170 (89.6)	13233 (83.4)
Ductal carcinoma in situ	6158 (10.4)	2636 (16.6)
Chemotherapy, No. (%)
Yes	26337 (44.4)	3862 (24.3)
No	32991 (55.6)	12007 (75.7)
Charlson comorbidity index, No. (%)
0	35314 (59.5)	6938 (43.7)
1	9526 (16.1)	2090 (13.2)
2	8116 (13.7)	4073 (25.7)
3+	6372 (10.7)	2768 (17.4)
Relation to employee, No. (%)
Employee	37528 (63.3)	9913 (62.5)
Spouse	21034 (35.5)	5676 (35.8)
Child/unknown	766 (1.3)	280 (1.8)
Type of health plan, No. (%)
POS	39921 (67.3)	11227 (70.7)
PPO	8307 (14.0)	2007 (12.6)
HMO	7058 (11.9)	1658 (10.4)
EPO	3658 (6.2)	872 (5.5)
Other	384 (0.6)	105 (0.7)
High deductible health plan, No. (%)
Yes	11363 (19.2)	3878 (24.4)
No	47965 (80.8)	11991 (75.6)
Insurance funding status, No. (%)
Administrative services only	14682 (26.1)	3925 (26.4)
Fully insured	41504 (73.9)	10934 (73.6)

Charlson comorbidity index: higher values indicate greater comorbidity burden. CF-WBI: Conventionally fractionated whole breast irradiation. HF-WBI: Hypofractionated whole breast irradiation. EPO: Exclusive Provider Organization. HMO: Health Maintenance Organization. POS: Point of Service. PPO: Preferred Provider Organization. Administrative services only (ASO) include completely self-funded employer plans.

Several community-level variables were determined according to zip code-based community demographic features of the patient’s residence, including race composition, education, median household income, proportion below the poverty line, unemployment rate, and female headed household. Data for these variables were determined from U.S. Census level data from 2018 [22]. If no racial group composed more than 50% population within a zip code, that zip code was labeled as a diverse community. We also calculated the “crowfly” distance between patients’ zip codes and treatment facilities.

Statistical Analysis

First, we performed multivariable logistic regression to evaluate temporal trends in the use of HF-WBI compared to CF-WBI from 2008–2017, adjusting for patient and community-level variables. We also used logistic regression to identify demographic, clinical, community factors, and insurance types as predictors for receiving HF-WBI versus CF-WBI. Adjusted odds ratios (ORs) with 95% confidence intervals (CI) were calculated to indicate association strength. Second, we examined health plan expenditures and out-of-pocket costs for HF-WBI and CF-WBI over time by year. We compared the mean cost difference between HF-WBI and CF-WBI using generalized linear models with gamma distribution and log-link function after adjusting for year, state, and patient features (e.g., community and insurance). Lastly, we calculated the average cost difference of CF-WBI and HF-WBI for each Census-designated core based statistical areas (CBSA) and for each radiation oncologist. To test the hypothesis that the CBSAs with a larger cost difference have a lower proportion of HF-WBI use because of the financial incentives to use conventional WBI, we fit mixed-effects logistic regressions to examine the association between within-CBSA or within-doctor cost difference and odds of receiving HF-WBI. If financial incentives impact the status quo use of conventional WBI, then larger cost differences between CF-WBI and HF-WBI would indicate a stronger (more influential) incentive. Thus, we would expect to observe an inverse association between within-doctor cost differences and the likelihood of HF-WBI adoption. Vermont (V.T.), Wyoming (W.Y.), and Hawaii (H.I.) were excluded from cost analyses due to unreliable data from these states within the HCCI database. All analyses were conducted using Stata 15 (StataCorp LLC). P-values <.05 were considered statistically significant.

RESULTS

Our final study cohort consisted of 75,197 patients, 15,869 receiving HF-WBI and 59,328 receiving CF-WBI during the study period. A slightly higher proportion of women with DCIS (30%) compared to women with invasive breast cancer (20%) received HF-WBI (Table 1). Most private insurance beneficiaries receiving either CF-WBI or HF-WBI treatment were employees. Point-of-service health plans were the most frequently held type of plan, and 20% of women were enrolled in high-deductible health plans.

The use of HF-WBI following breast-conserving surgery has steadily increased over time alongside a simultaneous decline in CF-WBI (eFigure 3). The adoption of HF-WBI markedly accelerated from 2013 to 2017 such that the gap between the two radiotherapy regimens has progressively diminished. Several patient demographic characteristics were associated with receipt of HF-WBI in multivariable analysis (Table 2). Those from older age groups were increasingly more likely to receive HF-WBI. Receipt of HF-WBI for women 55–64 years old was 1.88-fold higher than women 35–44 years old (adjusted OR 1.88, 95%CI 1.75–2.02; P<.001). Women who had been treated with chemotherapy were less likely to receive HF-WBI (adjusted OR 0.40, 95%CI 0.38–0.42; P<.001).

Table 2.

Multivariable logistic regression model on receiving hypofractionated radiotherapy in patients <65 years old^a

	Adjusted OR (95% CI)	P value
Age
<35	0.81 (0.63 − 1.02)
35−44	1 (Reference)	<.001
45−54	1.38 (1.28 − 1.49)
55−64	1.88 (1.75 − 2.02)
Type of breast cancer diagnosis
Invasive	1 (Reference)	<.001
Ductal carcinoma in situ	1.12 (1.06 − 1.19)
Chemotherapy
No	1 (Reference)	<.001
Yes	0.40 (0.38 − 0.42)
Distance from home to treatment facility
<10 miles	1 (Reference)	<.001
10−24.9 miles	1.17 (1.11 − 1.22)
25−49.9 miles	1.39 (1.29 − 1.50)
>=50 miles	1.56 (1.43 − 1.70)
% of college graduates in communityb
<23.4%	1 (Reference)	<.001
23.4% − 35.5%	1.21 (1.14 − 1.28)
35.6% − 51.0%	1.35 (1.27 − 1.43)
>51.0%	1.83 (1.73 − 1.95)
Race composition in communityb
White >50%	1 (Reference)	<.001
Black >50%	1.08 (0.98 − 1.19)
Asian >50%	1.02 (0.76 − 1.36)
Mixed community	1.20 (1.09 − 1.31)
Insurance type
HMO	1 (Reference)	0.18
EPO	1.00 (0.90 − 1.11)
POS	1.00 (0.93 − 1.07)
PPO	1.02 (0.93 − 1.11)
Other	1.36 (1.06 − 1.76)
High deductible plan
No	1 (Reference)	0.64
Yes	1.01 (0.96 − 1.06)
Insurance funding status
Administrative services only	1 (Reference)	0.99
Fully insured	1.00 (0.95 − 1.05)

^aMultivariable logistic regression model included all variables in the table at the time of radiotherapy and the year of radiotherapy.

^bCommunity demographic features based on zip code of patient’s residence. EPO: Exclusive Provider Organization.HMO: Health Maintenance Organization. POS: Point of Service. PPO: Preferred Provider Organization. OR: odds ratio. CI: confidence intervals. Administrative services only (ASO) include completely self-funded employer plans.

There was a dose-response trend positively associating the likelihood of HF-WBI treatment with the distance between the beneficiary residence and radiation clinic. In univariate analyses, the use of HF-WBI was related to multiple community-level characteristics, including education, race composition, median household income, proportion below the poverty line, unemployment rate, and female-headed household. However, only two community factors remained significant in the multivariable analysis(Table 2). First, there was a strong dose-response relationship between higher proportions of college graduates and a higher level of HF-WBI receipt. Second, a patient’s residence in a community with mixed racial composition was associated with higher odds of HF-WBI use when compared with residence in a community with a majority of one racial group.

Despite the importance of health insurance for beneficiary access to healthcare services, private insurance plan characteristics including 1) type of private insurance plan, some with narrower provider networks compared to others (e.g. HMO vs. PPO); 2) beneficiary enrollment in a high-deductible plan; and 3) health plan contracts with external insurers to comprehensively service their health plans (i.e., fully-insured) as opposed to enlisting administrative services only (i.e., self-funded), did not significantly alter a patient’s likelihood of receiving HF-WBI (Table 2).

Regardless of the variation in health plan characteristics, private insurers paid the majority of net expenses for radiotherapy from 2008–2017(Table 3). During this time, the mean radiotherapy expenditures insurers paid were significantly lower for their beneficiaries’ to receive HF-WBI relative to CF-WBI (adjusted difference $6,375, 95%CI $6,147–$6,603). Additionally, the mean patient out-of-pocket expenditure for HF-WBI was $139 less than that of CF-WBI. After adjusting for inflation, average costs for CF-WBI and HF-WBI increased over time. The absolute incremental growth per year in the costs for HF-WBI ($335+$9 per year) was lower than for CF-WBI ($540+$13 per year) over the study period (Table 3).

Table 3.

Radiotherapy-related expenditures for breast cancer patients aged <65 yrs with commercial insurance plans^a

Type of costs	Radiotherapy Type		Differences, U.S. $ b
	CF-WBI	HF-WBI
Insurer paid [Mean, US $b]
2008	19,167	15,654	3,513
2009	21,157	15,869	5,288
2010	21,498	17,136	4,362
2011	22,025	15,730	6,295
2012	22,517	15,948	6,569
2013	23,147	16,381	6,766
2014	23,286	17,311	5,975
2015	23,394	17,181	6,213
2016	24,603	18,352	6,251
2017	24,778	18,412	6,366
Increment per year	540	335
Patient OOP costs [Mean, US $b]
2008	388	232	156
2009	439	262	177
2010	492	361	131
2011	502	318	184
2012	539	342	197
2013	559	388	170
2014	575	397	179
2015	523	412	111
2016	525	353	172
2017	437	362	75
Increment per year	13	9
a2008–2017 [Adjusted Mean (95% CI), US $b]
Total costs	23,286 (23,158 to 23,415)	16,763 (16,583 to 16,945)	6,523 (6,294 to 6,751)
Insurer’s paid	22,751 (22,623 to 22,880)	16,376 (16,196 to 16,557)	6,375 (6,147 to 6,603)
Patient OOP costs	502 (491 to 513)	363 (206 to 229)	139 (119 to 160)

^aMultivariable generalized linear models with Gamma distribution and log link, with the adjustment for year of radiotherapy, state of beneficiary health plan, age at the time of radiotherapy, type of breast cancer diagnosis, receipt of chemotherapy, Charlson comorbidity index, community education level (% in the community with a college degree or higher), and type of insurance plan. All the differences were statistically significant with p<0.0001.

^bAll expenditures were rounded up to the nearest dollar amount in 2017 U.S. dollars after adjusting for inflation. CF-WBI: conventional fractionated whole breast irradiation therapy. HF-WBI: hypofractionated WBI. OOP: out of pocket. CI: confidence intervals.

Large geographic variation existed across U.S. states to use HF-WBI (Figure 1), ranging from 9.6% in Louisiana to 36.2% in New York. Variation presented in the costs of CF-WBI and HF-WBI treatment strategies across states, and the average cost of HF-WBI was lower than that of CF-WBI in all states with reliable data (Figure 1). There was no consistent relationship across states between the utilization of HF-WBI and the corresponding average cost differences for HF-WBI relative to CF-WBI. Furthermore, we calculated the within-CBSA difference in costs of CF-WBI and HF-WBI, and correlated these amounts with the proportion of HF-WBI use per CBSA (a geographic unit smaller than the state) to test the hypothesis that the CBSAs with larger within-CBSA cost difference have lower proportions of HF-WBI use (Figure 2A). Using mixed-effects logistic regression, we did not find a significant association between the within-CBSA difference in the expenditure of the two types of radiotherapies and use of HF-WBI (adjusted OR 0.97 per $6500, 95%CI 0.85–1.10; P=0.61). Similarly, we calculated within-doctor differences in costs of CF-WBI and HF-WBI and correlated those amounts with the proportion of HF-WBI use per doctor ordering radiation treatment (Figure 2B). We found no significant association between within-doctor cost differences and the use of HF-WBI (adjusted OR 1.03 per $6500, 95%CI 0.98–1.08; P=0.21).

Figure 1. Percent of HF-WBI Delivered and Average Costs in US Dollars^a of CF-WBI Relative to HF-WBI in Women with Breast Cancer by State, 2008–2017.

V.T., WY, and H.I. excluded because of unreliable data for these states. Colors on the map represent ranges for percent of HF-WBI delivered to women with breast cancer by state.

Figure 2. Scatter plots of HF-WBI use proportion against expenditure difference of CF-WBI and HF-WBI per Census-designated core based statistical area (A) or per radiation oncologist (B).

The size of the circle is proportional to the number of patients. Only core based statistical areas with at least 100 patients or doctors treating at least 80 patients were plotted.

DISCUSSION

The current study evaluates the use and costs of HF-WBI and CF-WBI across commercial health plans for women with breast cancer between 2008 and 2017. Our results indicate that the use of HF-WBI has increased over time, and, provided current rates continue, it may soon be the dominant form of radiation treatment for private insurance beneficiaries regardless of reimbursement rate. The U.S. adoption of HF-WBI has lagged behind that of other countries. The slow uptake of new evidence and concern for potential toxicities related to HF-WBI have been suggested as reasons for the delay in comparative uptake [23]. The most recent 2018 ASTRO guideline expanding eligibility for HF-WBI may increase HF-WBI adoption [24]. Our findings within the commercial insurance market reinforce the relationship previously demonstrated between patient characteristics like older age, type of breast cancer diagnosis (invasive vs. DCIS), and receipt of chemotherapy with a higher likelihood of HF-WBI use [9–12].

To our knowledge, ours is the first study to additionally evaluate both community-level and commercial health plan characteristics alongside these patient factors as predictive of HF-WBI utilization. We found a strong dose-response relationship between a higher proportion of college graduates in the patient’s community of residence with higher HF-WBI use. Mixed racial composition (i.e., lacking a racial majority group) of the patient’s resident community was also associated with a higher level of HF-WBI use. These community factors likely represent how neighborhoods and their residents’ socioeconomic characteristics contribute to shaping interactions with health care markets, especially related to the availability, type, and quality of health care resources [25]. Prior studies have demonstrated that facility-related factors (e.g., location, type, volume of cancer patients, etc.) across the country and even within a single state may also contribute to differences in HF-WBI use [10,26]. These previous studies highlight the importance of evaluating HF-WBI trends at multiple geographic levels in order to detect potential racial disparities in accessibility and adoption. Our results highlight that the specific mechanisms of how a community’s education level and racial/ethnic composition, particularly in relation to facility-level variation, might influence HF-WBI adoption merit further investigation.

Even though all breast cancer patients were privately insured in this study and the effectiveness of CF-WBI and HF-WBI treatment are similar, we observed that the costs for HF-WBI have been consistently lower than those for CF-WBI over time and across states. While our results within employer-sponsored insurance align with previous studies demonstrating lower radiotherapy-related healthcare costs paid by private insurers and by patients out-of-pocket for HF-WBI, our estimates of insurer costs were higher than previously reported [12]. This likely reflects the increased radiotherapy cost in more recent years (up to 2017) and the focus on the non-elderly population. In addition, from 2008–2017, absolute cost growth for insurers and patient expenditures for HF-WBI was slower than for CF-WBI while the relative increase of the two modalities are similar, lending further support that more widespread use of HF-WBI may be one additional area to curb the increase in health care costs in the U.S.

Fee-for-service (FFS) financial arrangements within the U.S. health care system incentivize providing more costly services and a larger quantity of services to maximize reimbursements. We hypothesized that this FFS financial incentive structure would drive use toward CF-WBI, the more costly service resulting in higher reimbursement, and smaller cost differences between the two treatments would increase HF-WBI use. Consequently, we postulated higher HF-WBI use among commercial enrollees of more capitated reimbursement arrangements, such as those within a health maintenance organization, where such FFS incentives would have less potential influence. However, no association was observed between any health plan characteristics and HF-WBI use. Neither type of health plan (e.g., PPO vs. HMO), whether a high deductible plan or not, or whether the plan was provided by a fully-insured vs. administrative services only employer was related to the likelihood HF-WBI treatment. Though we do not know the benefit structure of specific health plans from the commercial claims database, there was no evidence in our analysis that commercial enrollees of more restrictive health plans with cost-control measures such as placing limits on reimbursement amounts or narrowing provider networks were any more/less likely to receive HF-WBI.

There was large variation in both the use and relative costs of CF-WBI and HF-WBI across states and CBSA’s, a smaller geographic unit for commercial enrollees. Our results did not indicate a financial incentive biased towards the use of CF-WBI as larger cost differences between CF-WBI and HF-WBI were not associated with greater use of CF-WBI in either within-CBSA or within-doctor analyses. No consistent relationship could be observed between HF-WBI use and the cost differences between HF-WBI and CF-WBI across states. The geographical variations in HF-WBI use and costs observed are likely multifactorial and related to variations in regional health care markets [27,28]. To date, studies have focused on evaluating regional variation of HF-WBI use within the Medicare population [29]. Earlier research has demonstrated that commercial health care spending is related to aspects of market structure and that the reduction of provider market concentration would likely facilitate cost reductions within private markets [30]. Our study is one of few that evaluates regional factors in influencing trends in HF-WBI but the first to our knowledge that addresses multiple geographical levels.

Our results indicate that existing state policy and regulatory powers may be required to support cost-effective medical advancements like HF-WBI. States have the authority to regulate the commercial health insurance industry, which has historically targeted setting standards for insurers’ financial responsibility. Other state policy levers include encouraging price transparency, enhancing data collection/reporting requirements, and consolidating all-payer claims databases. Though the private insurance market may organically follow the example set by public health insurance programs like Medicare, states may have significant leverage to encourage policy to promote HF-WBI use and accrue cost savings [31]. The receipt of evidence-based treatment varies by state, and large state variation exists in the economic burden of breast cancer to which the costs of alternative radiotherapy may be relevant, suggesting possible missed opportunities to minimize the financial burden through the delivery of effective cancer prevention, screening, and treatment [19].

This study is not without limitations. First, our evaluation includes only commercial insurance claims for women receiving radiotherapy, which may not be generalizable to populations with other types of health insurance coverage. Second, lack of available data on the details about the benefit structure for each beneficiary’s health plan prevented deciphering any possible value-based reimbursement structures vs. fee-for-service. Though this lack of health plan information may present only a minor limitation, we acknowledge that value-based insurance design and alternative payment models are increasingly being incorporated into oncology care and radiation therapy to control accelerating costs [17,32]. These commercial payers’ strategies within alternative commercial health plans may provide insights into the adoption of HF-WBI and the costs incurred. Implemented policies like utilization management have enhanced the use of HF-WBI amongst commercially insured early-stage breast cancer patients [33]. Third, there may exist other clinical factors affecting the decision to use a HF-WBI such as prior radiotherapy, which we do not have access to through the HCCI database, so there is potential confounding. Lastly, the availability of HCCI data only extended through 2017 at this study, which prevented additional evaluation of the most recent 2018 clinical guidelines from ASTRO and the impact of COVID-19 pandemic. This 2018 guideline expanded the population of patients recommended to receive hypofractionated WBI [24] and could impact future use beyond our study period.

Emerging evidence in recent years further supports the safety, efficacy, and reduced costs of hypofractionated radiotherapy. We estimated that hypofractionated radiotherapy could save approximately $6,500 per one breast cancer patient covered by commercial insurance. Despite this large cost difference, our study did not provide any evidence that a financial disincentive is driving the slow adoption of the new radiotherapy modality. Instead, our study found that living in a higher-educated or racially diverse community was associated with the use of HF-WBI. Though we could not detect any influence of these specific reimbursement arrangements on HF-WBI use, the extent to which such financial innovations encouraging HF-WBI uptake have dispersed into employer-sponsored insurance remains unknown. Policy approaches like utilization management [33] are emerging within private insurance, which could accelerate greater HF-WBI adoption. The success of these new economic models to improve evidence-based cancer care will be encouraged by multiple level efforts targeting individuals, the community, and health policy to align financial incentives with appropriate HF-WBI use.

Data Availability:

Research data are not available at this time. The data underlying this article were provided by Health Care Cost Institute (HCCI, https://healthcostinstitute.org/about-hcci), and the study investigators do not have permission to share data with other entities. Interested researchers can apply for data from HCCI directly, and the study investigators could provide conceptual advice and help.