Decreases in Radiation Oncology Medicare Reimbursement over time: Analysis by Billing Code

Abstract

Purpose:

Radiation oncology (RO) has seen declines in Medicare reimbursement (MCR). However, there are no recent studies analyzing the contributions of specific billing codes to overall RO reimbursement. We compared total MCR for specific Healthcare Common Procedure Coding System (HCPCS) codes in 2019 to MCR for those codes in 2010 and 2015, corrected for inflation, to see how the same basket of RO services in 2019 would have been reimbursed in 2010 and 2015 (adjusted MCR).

Methods:

The CMS Physician/Supplier Procedure Summary database was used to obtain MCR data for RO HCPCS codes in 2010, 2015 and 2019. For each code, total allowed charge was divided by the number of submitted claims to calculate average MCR per claim in 2010, 2015 and 2019. The 2019 billing frequency for each code was then multiplied by the inflation-adjusted average MCR for those codes in 2010 and 2015 to determine what the MCR would have been in 2010 and 2015 using 2019 dollars and utilization rates. Results were compared to actual 2019 MCR to calculate the projected difference.

Results:

Total inflation-adjusted RO MCR was $2,281 million (M), $1,991M, and $1,848M in 2010, 2015, and 2019 respectively. This represents a $433M (19%) and $143M (7%) cut from 2010 and 2015 respectively to 2019. After utilization adjustment, total reimbursement was $2,534M, $2,034M, and $1,848M for 2010, 2015, and 2019 respectively, representing a $686M (27%) and $186M (9%) cut from 2010 and 2015 respectively to 2019. IMRT treatment delivery and planning accounted for $917M (36%), $670M (33%), and $573 (31%) of adjusted MCR in 2010, 2015, and 2019 respectively.

Conclusions:

MCR has decreased substantially from 2010–2019. Decline in IMRT treatment reimbursement is the primary driver of MCR decline. Policymakers should consider these trends and their impact on quality and access to healthcare when considering further cuts.

Introduction:

Like other specialties, radiation oncology (RO) has seen declines in Medicare reimbursement (MCR). Under Medicare Access and CHIP Reauthorization Act of 2015 (MACRA), MCR is being reformatted as a bundled payment model instead of fee-for-service, and reimbursement amounts for individual specialties are being adjusted. While the Radiation Oncology Alternative Payment Model (RO-APM) was scheduled to be implemented on January 1, 2023 and would affect approximately 30% of RO practices, it has now been delayed to an undecided future date^1,2. Nearly all radiation oncologists have been participating in MACRA through the Merit-based Incentive Payment System (MIPS), a fee-for-service (FFS) reimbursement model, unless they are participating in an APM through an accountable care organization (ACO) or MIPS-APM through an employment model in an integrated health system. Those selected for participation in the RO-APM would still be required to participate in MIPS, despite MACRA’s intent for physicians to participate in either MIPS or an APM. Under the mandatory RO-APM model, participating RO providers are paid by 90-day episode for 15 different cancer types (instead of fee-for-service). The objective of the RO-APM is to decrease cost by discouraging longer course radiation treatments and the use of expensive technologies while the plan hopes to improve outcomes through several quality initiatives^3–5.

Under its current proposed model, RO-APM will lead to significant cuts in reimbursement. With a flat reimbursement by cancer type, centers treating patients with higher complexity, higher risk disease requiring more expensive and sophisticated treatments will lose income⁶. The RO-APM is also expected to disproportionately affect reimbursement for cancers more commonly seen in minority, under-served populations, with practices providing care for socioeconomically disadvantaged patients expected to see significant reductions in revenue, potentially reducing access to care for these patients.⁷ Discounts of 3.5% and 4.5% for the Professional and Technical components, respectively, are withheld from the provider, with further discounts of up to 4% for incorrect payments (1%), quality (2%), and patient experience (1%), some of which can be made up the following year based on performance, with the 5% incentive payment limited to 10–20% of reimbursement^3,8. Additionally, practices with multiple locations may be required to participate concurrently in both fee-for-service and RO-APM billing, creating tremendous operational and logistical burden for centers to maintain trained staff who can manage these different processes.

Adoption of the RO-APM model is estimated to result in $300M in cuts for RO reimbursement in coming years,^9,10 accelerating cuts in MCR that pre-date the RO-APM. A recent study has shown decreased allowed charge amounts for RO from 2012–2017¹¹. However, there are no recent studies analyzing how changes in MCR for specific billing codes contribute to changes over time in overall RO reimbursement. By quantifying changes in MCR from 2010–2019, our study aims to provide context as the RO APM model and current proposed cuts in reimbursement undergo revision and potential adjustment. We compared total MCR for specific Healthcare Common Procedure Coding System (HCPCS) codes in 2019 to MCR for those codes in 2010 and 2015, corrected for inflation and utilization to see how the same basket of RO services in 2019 would have been reimbursed in 2010 and 2015 (adjusted MCR). We hypothesized that the decline in MCR from 2010–2019 would be driven disproportionately by decreased reimbursement for IMRT codes.

Methods:

Dataset Acquisition

The publicly available Physician/Supplier Procedure Summary (PSPS) database was used to estimate reimbursement in 2010, 2015, and 2019¹². This database contains Medicare Part B fee-for-service claims, with one summary file for each year starting in 2010. The database was filtered by provider code 92 to identify charges submitted by radiation oncologists and results were filtered to include radiation-oncology specific procedures, including all codes outlined in the American Society for Radiation Oncology (ASTRO) process of care description except for 77750 (“Infusion or installation of radioelement solution”) which covers radiopharmaceutical treatment (Table S1). Submitted charges, estimated using the “allowed charge amount” column, were grouped by HCPCS code. Codes were further grouped into categories according to the ASTRO HCPCS Codes by Process of Care descriptions (Table S1) but all reimbursement analysis was calculated on a code-by-code basis^13,14. A list of 2019 codes, equivalent past codes (where applicable), and the ASTRO categories used for grouping codes are summarized in Table 1 and Table S2.

Table 1:

Current billing codes, their corresponding past codes if applicable, and category

Current Code(s)	Past Code(s) if applicable	Description	Category
G6013	77414	Radiation treatment delivery	3D treatment delivery
G6014	77416	Radiation treatment delivery	3D treatment delivery
G6011	77412	Radiation treatment delivery	3D treatment delivery
G6012	77413	Radiation treatment delivery	3D treatment delivery
77295		3-dimensional radiotherapy plan	3D Planning
77014		CT guidance	Daily Imaging
G6002	77421	IGRT	Daily Imaging
G6015	77418	IMRT treatment delivery	IMRT treatment delivery
G6016	0073T	IMRT treatment delivery	IMRT treatment delivery
77301		IMRT dose planning	IMRT planning
0394T, 0395T	0182T	Electronic Brachytherapy	Brachytherapy
77771, 77768	77786	HDR Brachytherapy	Brachytherapy
77778		LDR Brachytherapy	Brachytherapy
77772	77787	HDR Brachytherapy	Brachytherapy
Q3001		Radioelements for brachytherapy, any type, each	Brachytherapy
57155		Insert uteri tandems/ovoids	Brachytherapy
55875		Transperineal placement of needles or catheters into prostate	Brachytherapy
77317	77327	Brachytherapy isodose plan	Brachytherapy
77316	77326	Brachytherapy isodose plan	Brachytherapy
77318	77328	Brachytherapy isodose plan	Brachytherapy
77770, 77767	77785	HDR Brachytherapy	Brachytherapy
20555		Placement of needles or catheters into muscle and/or soft tissue	Brachytherapy
77522		Proton treatment delivery	Proton treatment delivery
77525		Proton treatment delivery	Proton treatment delivery
77520		Proton treatment delivery	Proton treatment delivery
77523		Proton treatment delivery	Proton treatment delivery
77334		Treatment devices	Treatment devices
77332		Treatment devices	Treatment devices
77333		Treatment devices	Treatment devices

Code Mapping

Since RO billing codes have changed in the past decade, old codes have been mapped to corresponding new codes. Codes that were split into numerous new codes were re-merged if necessary¹⁵. Since 2010 and 2015 were analyzed separately, codes that changed prior to 2015 were included without change in the 2015 analysis, and only merged or split for the 2010 analysis. A complete list of code mappings (from old to new) is provided in Table S3. After code mapping, codes with $0 reimbursement or unavailable reimbursement data in either 2010/2019 or 2015/2019 were excluded from their respective analyses. Details of HDR brachytherapy code mapping are discussed in the supplemental materials.

Reimbursement for teletherapy isodose plan codes were analyzed as follows. Codes 77305, 77310, and 77315 (simple, intermediate, and complex teletherapy dose plans respectively) were changed to 77306 and 77307 (simple and complex respectively) in 2015. In our analysis, former code 77305 (simple) is represented by current code 77306 (simple). Former codes 77305 and 77310 (intermediate and complex) are represented by current code 77307 (complex). This was done to avoid overestimating 2010/2015 reimbursement for simple treatment planning when compared to 2019 as overestimation of 2010/2015 reimbursement would artificially increase the adjusted difference in MCR.

After code mapping, codes 77293 (Respiratory Management Simulation) and 77469 (Intraoperative Treatment Management) had reimbursement values for 2015 and 2019 only. To avoid underestimating 2019 reimbursement, these codes were included accordingly. 77293 was created in 2012 as an add-on code for either IMRT dose planning or 3D treatment planning and can be billed once per treatment course. Since 77293 can be used for either 3D or IMRT treatment and is not daily imaging, it is classified as “other” in the 2015 analysis. However, since this procedure was included in billing for either IMRT or 3D before 2012, for the 2010 analysis the total reimbursement for 77293 was apportioned between IMRT dose planning (77295) and 3D treatment planning (77301) based on the number of times each code (3D planning or IMRT dose planning) was billed. This ensures that reimbursement for 3D planning and IMRT dose planning in 2019 includes planning work (77293) that would have been bundled with planning payments in 2010. While reimbursement for 3D and IMRT treatment planning were increased to include respiratory management simulation (77293), the billing counts for 3D and IMRT treatment planning were not increased, since 77293 is a new add-on code. 77469 was treated the same as code 77293 except that it was placed in the “other” reimbursement category for both the 2010 and 2015 analyses.

Adjusted Reimbursement Calculations

We compared total MCR for specific Healthcare Common Procedure Coding System (HCPCS) codes in 2019 to MCR for those codes in 2010 and 2015, corrected for inflation, to see how the same basket of RO services in 2019 would have been reimbursed in 2010 and 2015. Unless otherwise noted, adjusted MCR refers to inflation and utilization adjusted MCR, calculated using Equation 1c.

For each code, the total allowed charge was divided by the number of submitted claims to calculate the average MCR per claim in 2010, 2015 and 2019. The number of claims is determined by the number of episodes (times a code was billed) not by the number of patients. The 2019 billing frequency for each code was then multiplied by the inflation-adjusted average MCR for those codes in 2010 and 2015 to determine what the MCR would have been in 2010 and 2015 using 2019 dollars and utilization rates. These results were compared to actual 2019 MCR to calculate the adjusted difference. All results are expressed in 2019 dollars.

Adjusted MCR is defined as the amount of MCR in 2019 dollars that would have been paid in a given year for a given code if that code had been billed with 2019 frequency during the alternate year. Equation 1a is adjusted MCR for a specific year (either 2010 or 2015) for a single HCPCS code. $\frac{\text{Total dollars billed for a given code in a given year}}{\text{Code billing frequency in a given year}}$ is the average MCR per billing instance of a code in a given year. The Average MCR per billing was multiplied by the number of times that code was billed in 2019 to get adjusted MCR (equation 1a). Seen another way, equation 1a corrects for billing frequency over time.

Adjusted MCR for a given billing code for a given year (2010 or 2015) =

$$\$MC{R}_{\mathit{\text{year}},\mathit{\text{code}}}=(\frac{\mathit{\text{Total}}\$\mathit{\text{billed}}\mathit{\text{for}}a\mathit{\text{given}}\mathit{\text{code}}\mathit{\text{in}}a\mathit{\text{given}}\mathit{\text{year}}}{\mathit{\text{Code}}\mathit{\text{billing}}\mathit{\text{frequency}}\mathit{\text{in}}a\mathit{\text{given}}\mathit{\text{year}}}*\mathit{\text{Code}}\mathit{\text{Billing}}\mathit{\text{frequency}}\mathit{\text{in}}2019)$$ Equation 1a

The adjusted difference in MCR is the total actual 2019 reimbursement for a single code (total 2019) minus the adjusted MCR for a single code in another year (Equation 1b). Note that total 2019 reimbursement (unadjusted) is the same as adjusting 2019 reimbursement for 2019 billing frequency (support for use of unadjusted 2019 reimbursement in supplemental materials).

Total MCR for a given billing code in 2019 ($total_2019,code) minus adjusted MCR for that code in a previous year (2010 or 2015) =

$$\Delta \$MC{R}_{\mathit{\text{year}},\mathit{\text{code}}}=\${\mathit{\text{total}}}_{2019,\mathit{\text{code}}}-\$MC{R}_{\mathit{\text{year}},\mathit{\text{code}}}$$ Equation 1b

For example, code 77301 (IMRT treatment planning) was billed 152,016 times in 2019 for $141.5M. In 2010, this code was billed 91,675 times for inflation-adjusted total reimbursement of $124.5M in 2019 dollars for an average of $1,358 per billing, as compared to an average of $931 per billing in 2019¹⁶. If code 77301 were billed 152,016 times (2019 frequency) in year 2010, it would have been reimbursed for a total of [152,016 (2019 frequency)] * [$1,358 (inflation-adjusted MCR per billing in year 2010)], or $206M. Thus, $206M is the adjusted MCR in year=2010 for code=77301 (Equation 1a). To find the adjusted difference in reimbursement for code 77301 from 2010–2019, we take [$141.5M (the total 2019 reimbursement for 77301)] – [$206M (adjusted 2010 MCR for 77301)], which is -$65M, representing a $65M decrease in adjusted MCR (Equation 1b).

For both 2010 and 2015, adjusted difference in MCR was calculated for all n codes separately and summed by year (Equation 1c). Code mapping was performed as described above.

Adjusted difference in MCR for all billing codes in 2019 minus adjusted MCR for those codes in a previous year =

$$\sum _{\mathit{\text{code}}=1}^{n\mathit{\text{codes}}}\Delta \$MC{R}_{\mathit{\text{year}},\mathit{\text{code}}}$$ Equation 1c

After calculating adjusted differences in MCR by code, codes were combined based on clinical categories. The following categories were analyzed: IMRT treatment delivery, daily imaging, IMRT planning, brachytherapy planning and delivery, treatment devices, proton treatment delivery, 3D planning, and 3D treatment delivery. Codes not falling in these categories but included in the analysis were marked as other. A list of codes in each category is provided in Table 1 and Table S1. Note that categorization of codes does not affect the overall projected difference, which is calculated on a code-by-code basis. Adjusted difference for a category from 2010–2019 and 2015–2019 was calculated as the sum of adjusted difference for all codes in this category.

Sensitivity Analyses

Additional analyses were performed to assess the impact of possible confounders. In the 2010 analysis, we apportioned reimbursement for 77293 (respiratory management) between IMRT dose planning (77295) and 3D treatment planning (77301) based on the number of times 77295 and 77301 were billed. However, there is no information available as to whether 77293 was billed with IMRT or 3D treatment planning. 77293 apportionment does not affect 2010 billing frequency, as this would likely underestimate the reimbursement per billing count for IMRT and 3D planning. To account for uncertainty in 77293 billing, we ran two sensitivity analyses where we apportioned 100% of reimbursement for 77293 to either IMRT dose planning or 3D planning and re-ran the analysis. 100% of reimbursement for 77293 likely did not fall in either 3D or IMRT treatment planning completely, thus applying 100% of 77293 reimbursement to each treatment planning method separately provides a conservative estimate of the adjusted difference for that method. Since billing frequency is not affected, neither analysis changed total adjusted difference in MCR; however, these analyses did change the proportion of total adjusted difference attributable to IMRT planning and 3D planning.

The PSPS dataset does not have the allowed charge amount for billing instances with 10 or fewer submitted service counts. However, the National Claims History payment amount is available for these billing instances. Since Medicare pays 80% of the allowed charge amount, we imputed the allowed charge amount from the National Claims History payment amount when the allowed charge amount was unavailable and re-ran the analysis including these charges.

Some decrease in reimbursement may be due to improved technology or ease of treatment administration. To account for this, we gathered Relative Value Units (RVU) data for each billing code from the Physician Fee Schedule for 2010, 2015, and 2019 to represent treatment difficulty for procedures with available RVU data¹⁷, with RVU correction method details in the supplementary analysis. The RVU analysis was performed as a supplemental rather than primary analysis because RVU data are not available for all relevant CPT codes, and we were concerned that exclusion of CPT codes that lack RVU data could introduce systemic bias. We modified Equation 1a to account for changes in RVUs over time (Equation 1a) and repeated the analysis as described above (equations 1b and 1c). rvu₂₀₁₉ is the RVU amount for that code in 2019 and rvu_year is the RVU amount for that code in another year (either 2010 or 2015). For codes that were merged or split, mean RVU was used.

Adjusted MCR for a given billing code in a certain year (2010 or 2015) =

$$RVU\mathit{\text{adjusted}}\$MC{R}_{\mathit{\text{year}},\mathit{\text{code}}}=\$MC{R}_{\mathit{\text{year}},\mathit{\text{code}}}*\frac{rv{u}_{2019,\mathit{\text{code}}}}{rv{u}_{\mathit{\text{year}},\mathit{\text{code}}}}$$ Equation 2

Equation 2 was used to calculate inflation, utilization, and work value-adjusted MCR. To calculate differences in RVU-adjusted MCR, RVU adjusted MCR (equation 2) replaced MCR (equation 1a) in equations 1b and 1c. Once again, 2019 total reimbursement was used, as total 2019 reimbursement is equivalent to adjusting 2019 reimbursement for 2019 RVU values (support for use of unadjusted 2019 reimbursement in supplemental materials).

Results:

Total inflation-adjusted MCR for RO procedures was $2,281M, $1,991M, and $1,848M in 2010, 2015, and 2019 respectively (Figure 1). This represents a $433M (19%) and $143M (7%) cut from 2010 and 2015 respectively to 2019. After adjusting for utilization, total reimbursement was $2,534M, $2,034M, and $1,848M for 2010, 2015, and 2019 respectively, representing a $686M (27%) and $186M (9%) cut from 2010 and 2015 respectively to 2019.

Figure 1.

Changes in inflation-adjusted; inflation- and utilization-adjusted; and inflation-, utilization, and RVU-adjusted reimbursement in 2010, 2015, and 2019.

From 2010–2019, adjusted MCR for IMRT treatment delivery fell from $711M to $434M (39%), while adjusted MCR for IMRT treatment planning fell from $206M to $142M (31%). For daily imaging, adjusted MCR fell from $449M to $230M (49%). Adjusted MCR for 3D treatment delivery increased from $139M to $145M (4%) and adjusted MCR for 3D planning fell from $56M to $44M (21%). Adjusted MCR for treatment devices fell from $93M to $70M (25%). For brachytherapy and proton treatment delivery, adjusted MCR fell from $30M to $25M (18%) and from $80M to $77M (4%). For all other HCPCS codes in the analysis, adjusted MCR decreased from $769M to $681M (11%, Table 2).

Table 2.

Change in Medicare reimbursement by Healthcare Common Procedure Coding System (HCPCS) billing categories.

Category	2010 to 2019 Inflation & Utilization-adjusted ΔMCR (million $)	2010 to 2019 Inflation & Utilization-adjusted ΔMCR (%)	2015 to 2019 Inflation & Utilization -adjusted ΔMCR (million $)	2015 to 2019 Inflation & Utilization-adjusted ΔMCR (%)
3D Delivery	$6 M	4%	$8 M	6%
3D Planning	−$12 M	−21%	−$2 M	−5%
Brachytherapy	−$5 M	−18%	−$19 M	−44%
Daily Imaging	−$219 M	−49%	−$24 M	−9%
IMRT planning	−$65 M	−31%	−$21 M	−13%
IMRT treatment delivery	−$276 M	−39%	−$76 M	−15%
other	−$88 M	−11%	−$33 M	−5%
Proton treatment delivery	−$4 M	−4%	−$5 M	−6%
Treatment devices	−$23 M	−25%	−$13 M	−16%

From 2015–2019, adjusted MCR for IMRT treatment delivery fell from $511M to $434M (15%), while adjusted MCR for IMRT treatment planning fell from $160M to $138M (13%). For daily imaging, adjusted MCR fell from $254M to $230M (9%). Adjusted MCR for 3D treatment delivery increased from $137M to $145M (6%) and adjusted MCR for 3D Planning fell from $43M to $41M (5%). Adjusted MCR for Treatment devices fell from $83M to $70M (16%). For brachytherapy and proton treatment delivery, adjusted MCR fell from $44M to $25M (44%) and from $82M to $77M (6%). For all other HCPCS codes in the analysis, adjusted MCR decreased from $720M to $687M (5%, Table 2). The difference in adjusted 2019 MCR for IMRT planning, 3D planning, and other procedures between the 2010 and 2015 analyses is due to reclassification of code 77293 in the 2010–2019 analysis, which was reimbursed for $6M in 2019. Total 2019 reimbursement did not change with this adjustment.

IMRT treatment delivery and IMRT treatment planning accounted for $917M (36%), $670M (33%), and $573 (31%, not including 77293) of adjusted MCR in 2010, 2015, and 2019 respectively. Also, IMRT treatment delivery and planning account for $341M (50%) and 97M (52%) of the adjusted difference in MCR from 2010 and 2015 respectively to 2019.

The highest reimbursed code in 2019 was G6015 (Intensity modulated treatment delivery). This code was reimbursed for a total of $432M in 2019, a $295M (41%) and $79M (15%) decrease from 2010 and 2015 respectively (G6015 is combined with G6016 in Table 2 to describe IMRT treatment delivery). Average reimbursement per billing instance was $574, $413, and $351 in 2010, 2015, and 2019 respectively. Adjusted reimbursement for G6015 was $707M and $508M in 2010 and 2015, representing a $275M (39%) and $76M (15%) decrease from 2010 and 2015 to 2019 respectively.

The second most highly-reimbursed code was 77427 (radiation treatment management), which was reimbursed for $200M in 2019, a $56M (22%) and $9M (4%) decrease from 2010 and 2015 respectively. Average reimbursement per billing instance was $221, $191, and $189 in 2010, 2015, and 2019 respectively. Adjusted reimbursement for 77427 was $236M and $204M in 2010 and 2015, representing a $35M (15%) and $3M (2%) decrease from 2010 and 2015 to 2019 respectively. Other highly reimbursed codes for 2019 were 77014 (CT guidance), 77301 (IMRT dose planning, which has its own category in Table 2), and G6012 (radiation treatment delivery).

From 2010–2019, G6015 (IMRT treatment delivery), 77014 (CT guidance, daily imaging), 77301 (IMRT dose planning), and G6002 (IGRT, daily imaging) combined accounted for $559M (81%) of the decrease in adjusted MCR. For these codes, adjusted MCR fell from $1362M to $803M. From 2015–2019, G6015 (IMRT treatment delivery), 0394T and 0395T (electronic brachytherapy, formerly 0182T), 77301 (IMRT dose planning), and 77014 (CT guidance, daily imaging) combined accounted for $136M (73%) of the decrease in adjusted MCR. For these codes, adjusted MCR fell from $887M to $751M. Adjusted MCR by code for 2010–2019 and 2015–2019 is summarized in Tables S4 and S5.

Sensitivity analysis

When 100% of reimbursement for 77293 (respiratory management simulation) was classified as IMRT dose planning (with 77301) instead of being split between 77295 and 77301, IMRT treatment delivery and planning accounted for $338M (49%) of the total adjusted difference in MCR from 2010–2019 instead of $341M (50%). When 100% of reimbursement for 77293 (respiratory management simulation) was included with 77295 (3D radiotherapy planning) in 2019 instead of being split between 77295 and 77301, 3D treatment delivery and planning accounted for $8.5M (1%) of the total adjusted difference in MCR from 2010–2019 instead of $11.7M (1%). Total adjusted difference did not change.

The ratio of National Claims History payment amount to allowed reimbursement was .78, .77, and .78 for 2010, 2015, and 2019 respectively, and .78 overall. After using a ratio of .78 to calculate imputed charge amount and recalculating adjusted difference while imputing missing data, adjusted reimbursement increased from $2,534M to $2,561M (1% increase), $2,034M to $2,047M (0.7% increase), and $1,848M to $1,860M (0.7% increase) in 2010, 2015, and 2019 respectively. After imputing missing data, adjusted MCR fell $700M (27%) and $187M (9%) from 2010 and 2015 respectively to 2019. MCR and change in MCR by category after imputing missing data are provided in Tables S6 and S7.

For the RVU sensitivity analysis, RVU adjustment was possible for $2,253M (99%), $1,861M (93%), and $1,740M (94%) of inflation-adjusted reimbursement in 2010, 2015, and 2019 respectively. Codes 77520, 77522, 77523, 77525 (proton treatment delivery codes), 0394T and 0395T (electronic brachytherapy, formerly 0182T), G6017 (IGRT, formerly 0197T), and G0339 and G0340 (image-guided SRS) made up greater than 90% of the total reimbursement for codes with unavailable RVU data. Table S1 includes RVU availability for each code. The remainder of codes were adjusted for inflation, utilization, and RVU changes over time. The inflation-, frequency-, and RVU-adjusted MCR was $2,268M, $2,042M, and $1,848M in 2010, 2015, and 2019 respectively. This represents a $420M (19%) and $194M (10%) cut in MCR from 2010 and 2015 to 2019. MCR and change in MCR by category after RVU-adjustment are provided in Tables S8 and S9.

Discussion:

In this study, we report that MCR in 2019 declined substantially compared to actual inflation-corrected 2010 and 2015 MCR. MCR in 2019 declined even more when compared to 2010 and 2015 MCR corrected for both inflation and utilization, which holds constant the factors that could potentially increase MCR (e.g., population growth; increased use of IMRT) or decrease MCR (e.g., hypofractionation) over time. As this analysis accounts for the number of times a code is billed, the decrease in reimbursement observed in our study does not account for further potential reimbursement declines due to increasing adoption of hypofractionation. Declines in MCR remained significant when adjusting for relative difficulty using RVUs, and when imputing missing allowed charges using National Claims History payment amounts. Our results, which show a decrease in adjusted MCR when accounting for changes in RVU, are consistent with a recent study showing an increase in the total number of RVUs produced by radiation oncologists from 2012–2015, despite decreases in MCR during this time¹⁸. Additionally, it is notable that adjusted MCR declined from 2015–2019 despite the Patient Access and Medicare Protection Act (PAMPA) of 2015, which temporarily froze payment rates for G codes¹⁹.

A decline in IMRT reimbursement was the primary driver of MCR decline. While not adjusted for utilization rates, results of a recent study showing decreased RO MCR from 2012–2017 are consistent with our findings¹¹. If introduced without further modifications, the proposed changes to the physician fee schedule and the adoption of the current proposed RO-APM will introduce an estimated $300M in additional cuts to RO MCR.^9,10

This discussion focuses on the current proposed RO-APM. We acknowledge that the final implementation of the model may change from what is discussed. While the RO-APM will reduce cost, it is worthwhile to consider the potential impact of the RO-APM on the other stated cancer policy goals of the current administration: address inequities, personalized cancer treatment, care for patients and caregivers²⁰. To our knowledge, there are no studies quantifying the impact of decreased reimbursement and MACRA on patient outcomes and practice and provider well-being in radiation oncology. However, there are concerns that rather than address inequities, the RO-APM may further exacerbate them, limiting access to care for minority and under-served rural patient populations. Minority and under-served rural patient populations are already more likely to present with more advanced stage of disease than other patient groups in the US^21,22. In addition to reducing reimbursement, the RO-APM requires additional reporting requirements that will be challenging and costly for small practices in under-served areas to adopt²¹. Practices qualifying for exemptions under MIPS may not qualify for similar exemptions in the RO-APM model. Furthermore, the future requirement of treatment auditing by another radiation oncologist may penalize solo practices in under-served areas. A study published in JAMA found that practices with a high proportion of socially disadvantaged patients had lower MIPs scores²³. As quality metrics are carried over between MIPs and RO-APM²⁴, lower scores in MIPs may indicate lower scores in the RO-APM as well, reducing reimbursement²⁴. Consistent with these findings, a recent analysis using historical data projected disproportionate decreases in reimbursement under the RO-APM for cancers common in under-served, minority populations⁷. All of these factors will place greater financial strain on practices providing care to these under-served populations and may limit patient access-to-care. By discouraging more complex, individualized treatments such as IMRT or proton therapy which are more costly, but can reduce toxicity, the RO-APM may not advance the administration’s goal of personalized cancer treatment or care for patients and caregivers^25,26.

Although there is mixed evidence supporting APMs as ways to reduce cost, cost-savings and reducing unnecessary treatment are two potential benefits of the RO-APM²⁷. There are many opportunities to reduce cost and potentially unnecessary treatments. For example, a recent study showed that most radiation oncologists and surgical oncologists over-estimate the potential benefit of adjuvant radiation therapy for stage I ER+ breast cancer in older women, which may lead to overtreatment of some women^28,29. Navathe et al. posit that the recent plateau in healthcare spending as percent of GDP is a result of APMs³⁰. While US healthcare costs per capita exceed those of other high-income countries,³¹ radiation oncology made up small proportion of all US healthcare spending in 2012 and 2017 (1.6% and 1.4%, respectively)¹¹. Thus, significant further cuts to radiation oncology will likely have a small effect on overall healthcare spending. While fee-for-service is the norm in radiation oncology and complex surgical oncology in Europe, a recent European Union-supported multi-stakeholder study recommended a switch to APMs²⁷.

In examining radiation oncology reimbursement, it is important to understand trends in medical and surgical oncology reimbursement. Both of these specialties are facing cuts as well, with the American Society of Clinical Oncology (ASCO) projecting a 5% decrease in reimbursement, and a projected 9% decrease in reimbursement for surgery^31,32. Thus, efforts to cut physician reimbursement are not limited to radiation oncology; however, to our knowledge there are no comparable studies in surgical or medical oncology quantifying recent change in Medicare reimbursement to provide a framework for the proposed further cuts. A recent letter from ASTRO to CMS in response to the RO-APM noted that the proposed cuts for RO reimbursement are among the highest in any specialty³³.

A shift towards value-based care, as seen in RO-APM, has the potential to discourage unnecessary procedures and reduce costs, but carries the risk of decreasing access to innovative, high-quality therapy if they are more costly and reducing overall access-to-care, especially for vulnerable populations, due to significant decreases in reimbursement and higher associated costs. To address the access-of-care issues, medical oncologists have suggested a mixed model including episodic and fee for service reimbursement³⁴. In pursuing an alternative payment model for RO, we feel that policymakers should consider the significant cuts to reimbursement that have already occurred and the potential impact of resource allocation under the APM model when considering mandatory adoption of a new payment system with further cuts, and the negative impact on quality and access to care^35–38.

Study limitations include the fact that Medicare Part B does not provide reimbursement for all cancer patients. However, private healthcare reimbursement trends tend to follow Medicare, as Medicare is the largest insurance provider in the United States. Furthermore, per HIPAA regulations, Medicare data excludes reimbursement for submitted charges with ≤10 submitted service counts. To address this, we imputed adjusted allowed MCR using the National Claims History payment amount. While adjusting for inflation and utilization does not account for changes in relative procedure difficulty over time, we accounted for changes in RVU for codes where RVU data were available. The lack of RVU data availability introduces possible code selection bias in the RVU analysis, thus the RVU-adjusted change in MCR was performed as a supplemental analysis. Furthermore, HCPCS codes change over time. To address this, we mapped old codes to their current equivalents, merging codes where necessary. With respiratory management simulation (77293), ground truth for code mapping was not available. In this case, we performed a sensitivity analysis to provide a conservative estimate of the change in MCR attributable to either 3D or IMRT treatment planning. Additionally, separate dosimetry calculations (77300) were allowed to be billed alongside certain treatment isodose planning codes in 2010 but not 2019. While the changing frequency of 77300 over time is accounted for in our analysis, certain isodose treatment planning codes now account for the additional work that used to be billed under 77300, an increase in difficulty that may not be accounted for by changing reimbursement over time. Thus, the adjusted difference in reimbursement is likely a conservative one. An additional limitation of the discussion is that RO-APM is currently undergoing revisions and the details may be different from the current iteration discussed above. However, study methods and results remain unaffected by future changes as the study focuses on past reimbursement. Furthermore, by quantifying reimbursement cuts, this study may aid policy-makers in making evidence-based RO-APM adjustments.

The article is also limited in scope: it quantifies changes in reimbursement but does not quantify the effect of these changes. Recent studies have shown evidence of practice size consolidation in radiation oncology, yet it is unclear if this is directly linked to decreasing reimbursement placing additional strain on smaller practices^39,40. Examining how decreased reimbursement has affected care quality, access, and provider well-being could answer this question. This article only addresses outpatient reimbursement, not hospital-based reimbursement. While hospital reimbursement may be subject to additional factors, reimbursement trends are likely similar between the two care settings.

Conclusion:

2019 MCR was $1,848M, which was a $686M (27%) and $186M (9%) cut as compared to inflation- and utilization-corrected MCR in 2010 and 2015 for these same codes. These cuts remained significant when changes in RVU were accounted for and missing data were imputed. Policymakers should consider the significant declines in reimbursement that have already occurred when considering further cuts, and the negative impact additional decreases could have on quality and access to care. These issues are especially relevant with the upcoming adoption of the bundled payment Radiation Oncology Alternative Payment Model.

Figure 2.

Change in inflation and utilization adjusted MCR from 2010–2019 and 2015–2019 for each billing category.