Impact of Intensity-Modulated Radiotherapy on Health Care Costs of Patients With Anal Squamous Cell Carcinoma

Abstract

Purpose:

Drivers of variation in the cost of care after chemoradiotherapy for the management of anal squamous cell carcinoma (SCC) have not been fully elucidated. We sought to characterize the direct and indirect impact of radiotherapy modality on health care costs among patients with anal SCC.

Patients and Methods:

A retrospective cohort study was performed using the 2014 linkage of the SEER-Medicare database. We identified 1,025 patients with anal SCC diagnosed between 2001 and 2011 and treated with chemoradiotherapy. Propensity score matching was used to balance baseline differences between patients treated with intensity-modulated radiotherapy (IMRT) and those treated with three-dimensional conformal radiotherapy (3D-CRT). Differences in total, cancer-attributable, and procedure-specific costs between groups were measured.

Results:

Radiation-related, patient out-of-pocket, and total costs in the 1-year period after radiotherapy start were all higher for the IMRT group than the 3D-CRT group (median total cost, $35,890 v $27,262, respectively; P < .001). Patients who received IMRT had lower cumulative costs associated with urgent hospitalizations and emergency department visits at both 9 months and 1 year after treatment start compared with a matched cohort of patients who received 3D-CRT (median, $711 v $4,957 at 1 year, respectively; P = .021).

Conclusion:

Although total costs of care were higher for IMRT compared with 3D-CRT, primarily as a result of higher radiotherapy-specific costs, IMRT was associated with decreased unplanned health care utilization costs starting at 9 months after treatment start. Radiotherapy-centered episodes of care may need to encompass a longer time horizon to capture the full cost savings associated with more advanced radiation modalities.

INTRODUCTION

The cost of cancer care is a significant contributor to the increasing costs of health care in the United States. The total cost of cancer care was $125 billion in 2010 and is projected to grow to $173 billion in 2020.¹ Technologic advances, including intensity-modulated radiotherapy (IMRT), image guidance, respiratory-motion management, and better treatment planning and delivery systems, have resulted in potentially safer and more effective delivery of radiation but with added cost to the overall course of treatment. Reflecting these advances, the total Medicare payments for external-beam radiotherapy increased 322% between 2000 and 2009, from $256 million to $1.08 billion, with most of that increase attributable to IMRT reimbursement.²

IMRT is a complex radiation technique that has had a revolutionary impact on the field of radiation therapy because of its ability to deliver a highly conformal radiation dose to tumor targets. It can thereby minimize dose to adjacent normal tissue and mitigate the significant toxicities associated with treatment. Definitive chemoradiotherapy has been the standard of care for the treatment of anal cancer for four decades with good clinical outcomes.^3,4 Because of the proximity of sensitive normal tissue, IMRT has been increasingly used for the treatment of anal cancer, with improved acute and late toxicity rates as demonstrated by a phase II trial and multiple retrospective studies.^5-9 Although the initial cost of IMRT is expected to be higher, downstream cost savings may be realized as a result of these decreased toxicities. However, the total cost of care related to chemoradiotherapy for the management of anal cancer and the drivers of variation in cost remain unknown.

Therefore, we used the SEER-Medicare database to characterize the impact of this increasingly used, resource-intensive treatment modality on real-world health care costs among patients with anal squamous cell carcinoma (SCC) treated with chemoradiotherapy. Improved understanding of the total and procedure-specific costs of care in the setting of modern radiotherapy techniques could help to inform the development of episode-based bundled payment models for managing patients with anal cancer.

PATIENTS AND METHODS

Cohort Selection

We performed a retrospective cohort study using the 2014 linkage of the SEER-Medicare database to examine the comprehensive financial impact of IMRT from the patient and payer perspective among elderly patients with anal SCC. We identified patients with nonmetastatic anal SCC diagnosed between 2001 and 2011 who were treated with both radiation and chemotherapy, defined as mitomycin, cisplatin, fluorouracil, or capecitabine, in the period from 1 month before through 6 months after diagnosis. All patients were required to have continuous Medicare Part A and Part B enrollment from 1 year before diagnosis to 1 year after radiotherapy start to fully capture costs during and after radiotherapy administration. Additional details of the cohort selection process were previously described in an earlier publication.¹⁰ The Institutional Review Board of Stanford University (Stanford, CA) deemed this study exempt from review.

Radiotherapy Treatment Modality

The primary exposure of interest was receipt of IMRT. Patients were identified as having received IMRT if Current Procedural Terminology (CPT) codes for IMRT planning or delivery were present. Patients with CPT codes for external-beam radiation therapy were deemed to have received three-dimensional conformal radiation therapy (3D-CRT).

Outcome Measures

Costs were estimated using Medicare claims data and were calculated as the sum of Medicare Part A and Part B reimbursements, third-party payer reimbursements, and patient liability amounts (sum of patient deductible and coinsurance amounts). Patient out-of-pocket costs were equated to patient liability amounts for the purposes of this study, although the portion of a patient’s liability covered by supplemental insurance (eg, Medigap) could not be determined using the Medicare claims data. Inpatient costs were calculated using claims associated with hospitalizations in the Medicare Provider Analysis and Review hospital stay file, excluding admissions to skilled nursing facilities. Outpatient costs were calculated using data from the Outpatient and National Claims History (NCH) files.

Costs attributable to anal cancer were estimated on the patient level by subtracting baseline health care costs before diagnosis from the total costs in the 1-year period after diagnosis.¹¹ A patient’s baseline annual health care costs were calculated using all Medicare claims in the 9-month period from 12 months to 3 months before diagnosis and scaling it to a 12-month period. The 3-month period before diagnosis was excluded from the baseline calculation to minimize the inclusion of costs related to cancer workup.

We defined the costs related to unplanned health care use as the sum of reimbursements for emergency department visits not resulting in a hospital admission and for inpatient hospitalizations with nonzero emergency department charges or with an admission type of urgent or emergent.¹² Given the relatively high cost of hospital-based care in relation to outpatient care,¹³ visits to non–hospital-based urgent care centers were not included in this calculation. Procedure-specific costs related to diagnostic imaging, endoscopy, and intravenous fluid hydration were calculated as the sum of reimbursements related to their respective CPT codes in the Outpatient and NCH files. In accordance with previously published methods,^14,15 radiation-related costs were calculated as the sum of reimbursements related to radiotherapy simulation, treatment planning, treatment delivery, and physician management, defined as CPT codes between 77261 and 77999, within 15 days before and 90 days after date of first radiotherapy treatment. Nominal prices were converted to real 2011 dollars by using the Inpatient Prospective Payment System Hospital Market Basket for Medicare Part A claims and the Medicare Economic Index for Medicare Part B claims.¹⁶

Additional Study Covariates

Clinical and demographic characteristics, including tumor stage, socioeconomic status, and Charlson comorbidity index, were obtained from the SEER database, as previously described.¹⁰ Chemotherapy administration and type were identified though Medicare claims.¹⁷ Radiotherapy was determined to have been delivered at a freestanding facility if radiotherapy claims were present only in the NCH file.¹⁴ If claims were present in the Outpatient and NCH files or in only the Outpatient file, treatment was considered to have been delivered at a hospital-associated facility.

Statistical Analyses

Differences in baseline characteristics between patients who did and did not receive IMRT were assessed using Pearson’s χ² test. Propensity score matching was used to balance observed differences in baseline covariates between the two groups. Patients who were diagnosed between 2001 and 2003 were excluded given the low use of IMRT in that time period. Propensity scores were estimated using a logistic regression model with receipt of IMRT as the dependent variable and baseline covariates thought to affect both treatment selection and outcome as the independent variables. Independent variables in our logistic regression model included year of diagnosis, age, sex, marital status, race, Charlson comorbidity index, disability status, HIV status, chemotherapy type, disease stage, socioeconomic status, geographic region, metropolitan status, treatment facility type, and National Cancer Institute cancer center designation. Patients were matched 1:1 without replacement using a greedy algorithm¹⁸ and a caliper distance of 0.2 standard deviations of the propensity score.¹⁹ Balance of baseline covariates after matching was assessed using standardized differences, with imbalance defined as an absolute value > 0.2.²⁰

Differences in total and cancer-attributable costs after diagnosis among the entire cohort over time were assessed using the Kruskal-Wallis rank sum test. Differences in total and procedure-specific costs after radiotherapy start between the matched cohorts were assessed using the Wilcoxon rank sum test. Kaplan-Meier survival analyses and log-rank testing were used to estimate differences in time to death after treatment. All tests were two-sided with an α = .05. Statistical analyses were performed using SAS software (version 9.4; SAS Institute, Cary, NC).

RESULTS

Patient Characteristics

We identified 1,025 patients with nonmetastatic anal SCC diagnosed between 2001 and 2011 and treated with chemotherapy and radiotherapy. Baseline characteristics are listed in Table 1. Most patients were women (65%) and had good disability status (93%). The most common chemotherapy regimen was mitomycin based (76%), and among the entire cohort of patients, 405 (39%) received IMRT, with IMRT use significantly increasing over time from 5% in 2001 to 2003 to 78% in 2010 to 2011 (P < .001).

Table 1.

Baseline Characteristics and Association With Receipt of IMRT in Patients With Anal Squamous Cell Cancer Treated With Chemotherapy and Radiotherapy in the SEER-Medicare Database

Historical Trends in Costs After Anal Cancer Diagnosis

Median total health care costs in the year after anal SCC diagnosis increased from $33,682 (interquartile range [IQR], $25,251 to $45,367) in 2001 to 2003 to $47,759 (IQR, $37,523 to $64,720) in 2010 to 2011 (P < .001), corresponding to an increase in median patient out-of-pocket costs from $8,001 to $9,187 (P = .001). After accounting for prediagnosis baseline health care spending, median cancer-attributable costs increased significantly from $30,100 (IQR, $21,981 to $45,367) to $41,171 (IQR, $31,723 to $56,492) over the same period (P < .001), representing an estimated compound annual growth rate of 3.8%. Costs related to radiation treatment planning, delivery, and management increased almost two-fold, from $11,304 in 2001 to 2003 to $20,275 in 2010 to 2011, and consequently accounted for an increasing proportion of cancer-attributable costs (38% in 2001 to 2003 v 49% in 2010 to 2011; Appendix Fig A1, online only).

Impact of Radiotherapy Modality on Costs

We matched 188 patients who received IMRT to similar patients who received 3D-CRT. There was no significant difference in any measured baseline characteristic as assessed using standardized differences. The cohorts were balanced with respect to year of diagnosis, age, sex, comorbidity index, chemotherapy type, disease stage, and other demographic and treatment characteristics as described earlier.

Radiation-related, patient out-of-pocket, and total costs in the 1-year period after radiotherapy start were all significantly higher for the IMRT group than for the 3D-CRT group (Table 2). However, patients who received IMRT had significantly lower cumulative costs associated with urgent hospitalizations and emergency department visits at both 9 months and 1 year after treatment start compared with patients who received 3D-CRT (median, $711 v $4,957 at 1 year, respectively; P = .021). A trend toward higher costs for the 3D-CRT cohort was found at the 3- and 6-month time points but did not reach statistical significance (P = .056 and P = .051, respectively; Fig 1). In addition, there were no significant differences in costs associated with intravenous fluid administration or endoscopy procedures, but there was a trend toward higher diagnostic imaging costs for IMRT patients compared with 3D-CRT patients (median, $1,376 v $1,642, respectively; P = .059). This finding was driven in part by slightly increased use of positron emission tomography (PET) among IMRT patients, with a mean of 0.9 scans in the 1-year period after radiotherapy start compared with 0.5 scans for the 3D-CRT group; however, the median number of PET scans in both groups was zero scans.

Table 2.

Cumulative Costs in the 1-Year Period After Radiotherapy Start Among Propensity Score–Matched Cohorts of Patients Who Received IMRT and 3D-CRT

Fig 1.

Costs associated with emergency department visits and urgent hospitalizations. (*) Difference is statistically significant with P < .05 on Wilcoxon rank sum test. 3D-CRT, three-dimensional conformal radiation therapy; IMRT, intensity-modulated radiotherapy.

Survival Analysis

Although the difference in survival was not significant between the matched patients, there was a trend toward improved overall survival in patients who received IMRT compared with patients who received 3D-CRT, with 5-year overall survival rates of 73.4% and 65.8%, respectively (log-rank P = .177), on Kaplan-Meier analysis.

DISCUSSION

The cost of care related to anal SCC among Medicare beneficiaries in the United States has increased significantly over the past decade. We estimated that the median cancer-attributable costs in the 1-year postdiagnosis period increased from $30,100 per patient in 2001 to 2003 to $41,171 in 2010 to 2011. In addition, as the use of IMRT has increased, costs related to radiation therapy have accounted for an increasing proportion of cancer-attributable costs, almost 50% in 2010 to 2011. Although IMRT represents a more expensive treatment modality than 3D-CRT, the true cost savings of decreased treatment-related toxicity had not yet been elucidated. Our study characterized the direct and indirect impact of radiotherapy modality on total and procedure-specific costs of care. We chose anal SCC for this analysis because the standard of care has not significantly changed over the past four decades with respect to the therapeutic modalities and systemic agents used, with the primary advancements during this time period being driven by radiation technology.

In a propensity score–matched analysis, patients who received IMRT had median total costs of $35,890 in the 1-year period after radiotherapy start, compared with $27,262 for patients who received 3D-CRT. Median radiation-related and patient out-of-pocket costs were approximately $9,600 and $2,000 higher, respectively, for patients who received IMRT. On the basis of 2011 national Medicare payment rates,²¹ we estimate that this difference in radiation-related costs was nearly fully accounted for by the higher payments for IMRT treatment delivery than for 3D-CRT treatment delivery, rather than differences in other factors, such as image guidance, treatment planning, or simulation. However, patients in the IMRT cohort had significantly reduced costs related to urgent unplanned health care use, as characterized by emergency department visits and urgent or emergent hospitalizations.

Although the cause of the observed difference in unplanned health care use in our study cannot be directly verified in claims data, we presume that the primary driver of this difference is the reduced toxicity with IMRT, especially given the lack of difference in survival outcomes immediately after the 1-year period between the two cohorts. We previously reported IMRT’s impact on reducing hospitalizations after radiation start in this cohort,¹⁰ which reflects the reduced treatment-associated acute toxicity with IMRT reported in the Radiation Therapy Oncology Group (RTOG) 0529 trial.⁸ In the current study, we found that IMRT did not translate into significantly lower cumulative costs associated with unplanned health care use until 9 months after radiotherapy start, suggesting that there may also be differences in late complication rates between IMRT and 3D-CRT. In the long-term update of RTOG 9811, Gunderson et al²² reported overall late grade 3 or 4 toxicity rates of 13.1% and 10.7% for those who received 3D-CRT with mitomycin and with cisplatin, respectively. Late toxicity rates were not reported in RTOG 0529.⁸ However, in a single-institution retrospective study, Mitchell et al²³ reported a late grade 3 GI toxicity rate of 2.4% and no instances of late dermatologic, genitourinary, or hematologic toxicity after IMRT treatment with concurrent chemotherapy.

The delayed effect of radiotherapy treatment modality on urgent health care use has potential implications on policy development. Successful shift from a fee-for-service payment system to one that incentivizes high-value care requires an understanding of not only the drivers of cost variation, but also the timing of these costs. Many episode-based payment models are structured around a time period only a few weeks to months after treatment to encompass acute and subacute treatment-related complications.^24-26 The Centers for Medicare and Medicaid Services Oncology Care Model is the latest iteration of a specialty-based global payment system that encompasses a 6-month period of care after the initiation of chemotherapy.²⁷ However, our study suggests that this approach may be insufficient for radiotherapy-centric episodes because cost savings as a result of decreased toxicity from more advanced radiation techniques may require a longer time period to be observed.

Hodges et al²⁸ performed a cost-effectiveness analysis of radiotherapy modalities for anal cancer and determined that IMRT is a cost-ineffective strategy compared with 3D-CRT. Regarding input parameters into their Markov decision model, the authors modeled differential rates of acute toxicities of IMRT and 3D-CRT as per clinical data extracted from RTOG 0529⁸ and RTOG 9811,³ respectively, and assumed that all acute toxicities resolved after 3 months. Interestingly, the authors chose to model rates of long-term toxicities for both IMRT and 3D-CRT based on data from RTOG 9811, setting equivalent rates for each modality. Therefore, the cost implications of differential toxicity profiles of IMRT versus 3D-CRT in this analysis were largely driven by differences in acute toxicities. Thus, this methodology could potentially underestimate the cost savings associated with reduced treatment-related toxicities from IMRT at time points beyond 3 months after treatment start.

Another interesting finding in our study was that there was a trend toward higher diagnostic imaging costs in the IMRT cohort, although the absolute difference in median costs at 1 year was relatively small. Notably, these costs were calculated in the outpatient setting and thus not related to differences in inpatient hospitalization rates. This finding suggests an increased likelihood for patients who received IMRT to also receive slightly more frequent or higher cost follow-up imaging, such as PET scans. Certain unmeasured confounders, such as patient or physician preference, may have contributed to this result and warrant further investigation.

The analyses in our study were limited to a time horizon of 1 year after treatment given the nature of the Medicare claims data. Although the overall costs at this time point were higher for patients who received IMRT in our Medicare cohort, it is possible that this difference might diminish given a longer time horizon. Several additional limitations of our study exist. We used propensity score matching to try to balance the observed covariates between our IMRT and 3D-CRT cohorts, but it did not address the impact of unmeasured confounders. Furthermore, limitations inherent in the use of Medicare claims include lack of data regarding a significant proportion of patients younger than age 65 years or care reimbursed by private insurers, inability to measure true patient out-of-pocket costs, and potential errors in billing and procedural coding. In addition, we were unable to differentiate outpatient clinic visits by provider specialty or by urgency, as we did in the hospital setting. As such, our study does not estimate the incremental outpatient visit costs associated with each treatment modality. We were also unable to analyze the effect of treatment on prescription drug use, which has been recognized as a significant component of overall health care costs.²⁹

In conclusion, total 1-year costs of care were higher for IMRT compared with 3D-CRT, primarily as a result of higher radiotherapy-specific costs. However, IMRT was associated with significantly decreased costs related to urgent unplanned health care use at 9 and 12 months after start of treatment, which may be a result of lower rates of late treatment-related toxicity. Successful development of alternative payment models will require additional investigation into the drivers of cost variation. In addition, radiotherapy-centered episodes of care may need to encompass a longer time horizon to capture the full cost savings associated with more advanced radiation treatment modalities.

Appendix

Fig A1.

Median costs in the 1-year period after diagnosis of anal squamous cell carcinoma by year of diagnosis. Error bars represent interquartile range.

AUTHOR CONTRIBUTIONS

Conception and design: Alexander L. Chin, Erqi L. Pollom, Albert C. Koong, Daniel T. Chang

Financial support: Erqi L. Pollom, Daniel T. Chang

Administrative support: Alexander L. Chin, Erqi L. Pollom, Daniel T. Chang

Provision of study materials or patients: Erqi L. Pollom

Collection and assembly of data: Alexander L. Chin, Erqi L. Pollom

Data analysis and interpretation: Alexander L. Chin, Erqi L. Pollom, Yushen Qian

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

Impact of Intensity-Modulated Radiotherapy on Health Care Costs of Patients With Anal Squamous Cell Carcinoma

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/jop/site/ifc/journal-policies.html.

Alexander L. Chin

No relationship to disclose

Erqi L. Pollom

No relationship to disclose

Yushen Qian

No relationship to disclose

Albert C. Koong

No relationship to disclose

Daniel T. Chang

Stock or Other Ownership: ViewRay

Honoraria: Varian Medical Systems

Research Funding: Varian Medical Systems