Determinants of Medicare Costs for Elderly Patients With Oral Cavity and Pharyngeal Cancers

Importance  In the United States, nearly 8400 patients die each year from oral cavity and pharynx cancers, most of whom are 65 years and older; however, the costs attributable to these cancers are not well described.

Objective  To identify the primary determinants of cost in patients with oral and pharyngeal cancer.

Design, Setting, and Participants  In this retrospective cohort analysis of data from Medicare and Surveillance, Epidemiology, and End Results hospitals (January 1, 1995, through December 31, 2005), we studied patients 66 years and older with newly diagnosed oral cavity (n = 6724) and pharyngeal (n = 3987) cancers.

Main Outcomes and Measures  Five-year cumulative costs, defined as Medicare Parts A and B payments, were estimated using inverse probability weighting. Linear regression analysis with inverse probability weighting was used in multivariate analyses of costs to estimate the effects of covariates on cumulative costs.

Results  In multivariate analyses, costs were significantly increased by demographics, comorbidities, and treatment selection. Compared with white patients, African Americans accumulated $11 450 (95% CI, $1320-$22 299) and $25 093 (95% CI, $14 917-$34 985) more in costs for oral cavity and pharyngeal cancers, respectively. The presence of 1 or 2 comorbidities increased the mean 5-year cumulative costs by $13 342 (95% CI, $6248-$19 186) and $14 139 (95% CI, $6009-$22 217) for patients with oral cavity and pharyngeal cancers, respectively. For 3 or more comorbidities, the mean 5-year cumulative costs increased by $22 196 (95% CI, $15 319-$28 614) and $27 799 (95% CI, $19 139-$36 702) for patients with oral cavity and pharyngeal cancers, respectively. Patients who received chemotherapy accumulated a mean of $26 919 (95% CI, $18 309-$35 056) and $37 407 (95% CI, $29 971-$44 644) more in costs by 5 years for oral cavity and pharyngeal cancers, respectively.

Conclusions and Relevance  Oral and pharyngeal cancer is burdensome to elderly patients from a Medicare cost perspective. Several factors were associated with 5-year costs, including some modifiable factors that may be potential targets for interventions to reduce overall costs.

Tumors of the head and neck account for approximately 650 000 newly diagnosed cancers worldwide annually.¹ Although the peak incidence lies between 50 and 70 years of age, it is estimated that as many 25% of these cancers are diagnosed in patients 70 years or older.^2-5 Tumors of the oral cavity and pharynx are among the most affected sites in elderly patients.^1,3-5 Survival and treatment modalities are largely determined by age and stage at the time of diagnosis. Other factors, such as race/ethnicity, comorbidities, and marital status, have also been found to be associated with outcomes.^6-8 In early-stage cancers, survival ranges from 70% to 90%, and single-modality treatments are feasible; however, less than 50% of those with advanced disease survive more than 5 years.^1,9,10

Many cases of oral cavity cancer and most cases of pharyngeal cancer are diagnosed at advanced stages, when management is complex and requires multidisciplinary, multimodality, and aggressive care. The most common treatment options, including surgery, radiotherapy, and chemotherapy, carry a high burden of morbidity. Recurrence, ultimately leading to death, is relatively common.^9,11 Costs associated with management of head and neck cancers are also high.^9,10,12-18 The estimated financial impact of head and neck cancer annually in the United States is nearly $1.8 billion (adjusted to year 2012 dollars).^19,20 Reports of annual costs per incident case have ranged from approximately $25 000 (year 2002 dollars) to $154 000 (year 2009 dollars).^9,12-14,20

There are many reasons for the wide range of reported costs, including heterogeneity of data sources, the distribution of stage of disease in samples, and methodologic approaches for cost computations. In the United States, because of the separation between the institutions that finance and provide health care, relatively few data sources include the detailed clinical information and billing records necessary to estimate the costs of cancer. The Medicare program, which is publicly funded health insurance for elderly patients, provides an opportunity to estimate costs of care for elderly patients with oral cavity and pharyngeal cancers. Our aim was to estimate the attributable costs of oral cavity and pharyngeal cancer in elderly Medicare patients and identify the primary determinants of costs of care during the 5-year period after initial diagnosis.

Data for this study were obtained from the Surveillance, Epidemiology, and End Results (SEER)–Medicare linked database. The SEER-Medicare database combines tumor registry data from the National Cancer Institute’s SEER program for patients who are covered by Medicare with their Medicare billing records.²¹ The SEER-Medicare data contain tumor registry data on 54 955 patients who were diagnosed as having an initial primary tumor of the oral cavity (lip, oral tongue [excluding base of tongue and lingual tonsil], salivary gland, floor of mouth, gum, and other mouth) or pharynx (base tongue [base of tongue and lingual tonsil only], nasopharynx, tonsil, oropharynx, hypopharynx, and other) from January 1, 1973, through December 31, 2005. Claims data are only available beginning in 1986 and are only available for certain types of services since 1994. Therefore, we limited our sample to patients who were diagnosed as having an initial primary tumor of the oral cavity or pharynx from January 1, 1995, through December 31, 2005. We further restricted cases to patients 66 years and older and excluded patients for whom the diagnosis of cancer was obtained on autopsy or from a death certificate only. In addition, patients enrolled in a health maintenance organization at or after the cancer diagnosis were excluded because these patients do not have any Medicare billing records. The final sample contained 10 711 patients. This study was approved by the internal review board of The Pennsylvania State University College of Medicine. Informed consent was not required given the retrospective nature of the study design.

Our primary aims were to estimate the total cumulative 5-year costs for patients with oral cavity and pharyngeal cancers and to determine patient and disease characteristics that were associated with costs. The perspective adopted for the cost analyses was Medicare as the payer. Therefore, costs represent actual payments made by Medicare for all-cause treatments, including inpatient stays, emergency department visits, outpatient procedures, office visits, home health visits, durable medical equipment, and hospital care but excluding outpatient prescription drugs. Hospital costs were adjusted using the Hospital and Related Services component of the Consumer Price Index. Outpatient costs, physician payments, and hospice and home care costs were adjusted using the Medical Care Services component of the Consumer Price Index. Medical equipment costs were adjusted using the Medical Care Commodities component of the Consumer Price Index. All costs were adjusted for inflation to year 2005 dollars.

Our analyses estimated the effect of demographic characteristics, disease characteristics, and comorbidities on treatment choice. Demographic variables included patient age at diagnosis, sex, race/ethnicity (black, white, or other), year of diagnosis, marital status, and geographic location. Disease variables included cancer site and morphologic extent of malignant disease defined using the SEER historic stage (local, regional, distant metastasis, or unstaged).²² We controlled for comorbidities using the sum of the number of Agency for Health Care Research and Quality comorbidities for each patient up to 1 year before diagnosis.^23,24 Treatment indicators were mutually exclusive for surgery alone, radiation alone, a combination of surgery and radiation, chemotherapy in combination with any other treatment, and absence of any treatment. Radiation and surgery were defined according to the SEER tumor registry data. Chemotherapy was determined from Medicare billing records. Specifically, all Medicare claims files were searched for any record with a Healthcare Common Procedure Coding System code of 964.xx, 965.xx, or Q0083-Q0085; an International Classification of Diseases, Ninth Revision (ICD-9),diagnosis code of V58.1, V66.2, or V67.2; or an ICD-9 procedure code of 99.25.²¹ Patients with at least one of these claims in the window ranging from 1 month before diagnosis up to 4 months after diagnosis were classified as having received chemotherapy. Patients in which a chemotherapy claim was made at least 2 months before diagnosis were excluded from the analysis.

The statistical analysis was designed to determine the associations between cumulative costs and covariates, including demographic characteristics, comorbidities, and tumor features. Descriptive statistics were used to summarize the sample populations. Mean cumulative costs were estimated using inverse probability weighting to account for patient censoring.²⁵ Patients were censored if they were alive and had been diagnosed as having cancer less than 5 years before December 31, 2006, the SEER cut-off date for follow-up in our data set. We used the partitioned estimator based on monthly costs, resulting in a total of 60 partitions. Uncensored patients contributed costs for all 60 months; censored patients contributed costs for all months before the SEER cut-off date. For a given partition, uncensored patients were given a weight inversely proportional to the estimated probability of being uncensored at that time point. Cumulative costs were also assessed in a multivariable regression model using inverse probability weighting and the same monthly partitions.²⁶ Bootstrapping with 1000 replicates was used to estimate 95% CIs.

All analyses were performed using SAS statistical software, version 9.2 (SAS Institute Inc), and R software, version 2.13.0 (open source software available at http://www.r-project.org/).

A summary of the characteristics of Medicare patients with oral cavity and pharyngeal cancers are given in Table 1. Patients with oral cavity cancers were older and more likely to be female and white compared with patients with pharyngeal cancers. Among patients with oral cavity and pharyngeal cancers, 17.9% and 18.7%, respectively, had at least one comorbidity. The distribution of comorbidities is presented in Table 2. The most common comorbidities for oral and pharyngeal cancers were hypertension, chronic pulmonary disease, fluid and electrolyte disorders, and diabetes mellitus. The disease characteristics of the cancers are presented in Table 3. Patients with oral cavity cancer were more likely to present with localized disease than patients with pharyngeal cancer (44.0% vs 12.6%), and patients with pharyngeal cancer were much more likely to present with regional disease or distant metastases (80.6% vs 44.2%). Treatment patterns were also different; patients with oral cavity cancer were more likely to be treated with single-modality surgery (49.4%), whereas patients with pharyngeal cancer were more likely to be treated with chemotherapy (34.1%) or radiation therapy alone (27.0%).

Multivariate analysis of costs using inverse probability weighting is presented in Table 4. After controlling for other factors, costs for patients with oral cavity cancers who were 76 to 80 years old were $9677 more than patients 66 to 70 years old. Other age groups were not statistically significant for oral cavity cancer, and the effect of age was not statistically significant for pharyngeal cancer. Black race was associated with large and significant effects on 5-year costs. African Americans with oral cavity cancer accumulated $11 450 (95% CI, $1320-$22 299) more in costs than white patients. African American patients with pharyngeal cancers accumulated $25 093 (95% CI, $14 917-$34 985) more during 5 years than white patients. There were systematic effects of geography on 5-year costs; costs decreased consistently from the largest metropolitan areas to the smallest rural areas for patients with oral and pharyngeal cancers. Stage was not a significant contributor to 5-year cumulative costs, with the exception of regional disease in oral cavity cancers, which was associated with an increased mean 5-year cumulative cost of $10 677. Comorbidities were a significant determinant of 5-year costs. The presence of 1 or 2 comorbidities increased the mean 5-year cumulative costs by $13 342 (95% CI, $6248-$19 186) for patients with oral cavity cancer and $14 139 (95% CI, $6009-$22 217) for patients with pharyngeal cancer (Table 4). For 3 or more comorbidities, the mean 5-year cumulative costs increased by $22 196 (95% CI, $15 319-$28 614) for patients with oral cavity cancer and $27 799 (95% CI, $19 139-$36 702) for patients with pharyngeal cancer (Table 4). Treatment selection was also a significant determinant of cumulative costs. Patients with oral cavity cancer who received chemotherapy accumulated a mean of $26 919 (95% CI, $18 309-$35 056) more in costs by 5 years, and patients with pharyngeal cancer accumulated a mean of $37 407 (95% CI, $29 971-$44 644) more in costs by 5 years. Treatment with surgery and radiation was associated with increased mean 5-year cumulative costs of $21 463 (95% CI, $15 104-$27 480) for oral cavity cancer and $29 624 (95% CI, $21 023-$38 642) for pharyngeal cancer.

Oral cavity and pharyngeal cancers are aggressive diseases that are often diagnosed in advanced stages and require complex management. Primary treatment for oral cavity cancer generally begins with surgery. Radiation and chemotherapy are given postoperatively in locally advanced (stage III-IVb) disease or administered as a substitute for surgery if the tumor is locoregionally confined but inoperable. Chemotherapy, with or without radiotherapy administered with palliative intent, is used if metastatic (stage IVc) disease is diagnosed. Primary treatment for pharyngeal cancers, particularly when diagnosed in advanced stages, generally begins with radiation and chemotherapy, with the goal of sparing organ function such as swallowing, speaking, and airway protection.¹ Complications associated with treatments and recurrences after curative intent interventions in advanced stages tend to elevate the burden associated with the disease.^11,27,28 Given the high health care resource use associated with the treatment and downstream care of patients with oral cavity and pharyngeal cancers, efforts have been directed toward understanding and characterizing the economic burden associated with this disease.^{9,11-14,16-18,20}

In comparing cumulative costs for patients with oral cavity cancer to patients with pharyngeal cancers, these results suggest that patients with pharyngeal tumors incur higher costs. By year 5, the difference in the mean cumulative cost between the 2 cancer sites was $19 000 ($72 000 for oral cavity cancer vs $91 000 for pharyngeal cancer, P < .001). This finding appears to be largely due to differences in treatment approach. More patients with oral cavity cancers receive surgery alone as the primary modality of treatment, whereas most patients with pharyngeal cancers were more likely to receive chemotherapy and radiation, which are more costly. Notably, these costs also encompass any downstream complications of the treatment modalities, such as neutropenia. A substantial number of patients with oropharyngeal cancer received chemotherapy, significantly increasing incremental costs. Although it is commonly recommended that advanced age per se should not be considered a contraindication for receiving concurrent chemotherapy with radiotherapy,²⁸ a large individual-patient meta-analysis²⁹ of 87 randomized clinical trials found a decreased benefit with increasing age and no benefit for patients older than 70 years.

After adjusting for confounders in multivariate analysis, the treatment modalities were the largest drivers of cost, with surgery combined with radiation and chemotherapy (alone or in combination with other modalities) associated with significantly increased 5-year costs compared with surgery alone. Because radiation therapy alone is often used as a palliative measure in unresectable oral cavity cancer, the decreased 5-year cumulative costs associated with this treatment are secondary to the decreased survival rates at 5 years. This phenomenon also likely explains why, with the exception of those with oral cavity cancer with regional disease, advanced stages were not found to be a significant predictor of costs at 5 years compared with localized or in situ disease. The predominance of treatment modality as the major driver of costs has also been found in other studies^9,13 evaluating this association among different payer systems for head and neck cancer.

The effect of comorbidities on prognosis, treatment selection, and perioperative complications in pharyngeal cancers has been well documented.^11,30-34 Our results also reveal the increased incremental 5-year cumulative costs associated with comorbidities. This association has been found in other studies that used Medicaid and Medicare databases.^9,14 In particular, using 2 years of Medicare-SEER data (1991-1993), Lang et al¹⁴ reported that, within each stage, patients with the highest comorbidity scores had costs that were more than 2 times those of patients with the lowest comorbidity scores. We similarly have found that costs increased approximately 2-fold for oral cavity and pharyngeal cancers between patients with 1 or 2 comorbidities compared with those with 3 or more comorbidities. In contrast to these previous studies, we used inverse probability weighting to account for patient censoring and provide adjusted dollar estimates for the cumulative 5-year costs attributable to incremental increases in patient comorbidities in those with oral cavity and pharyngeal cancers compared with matched noncancer cohorts.

In evaluating the economic burden of oral cavity and pharyngeal cancer, comparing estimates among different studies is difficult because of the heterogeneity in definitions of disease, methods, and data sources. Early focus on quantifying the costs associated with the diagnosis and treatment of these cancers drew from US and international experiences from the Netherlands and Greece.^12,17,18 Although these studies described the higher costs associated with increasing stage of disease and the influence of treatment modalities on cumulative costs, their generalizability was limited by institutional experiences and varying health care infrastructures. In addition, the reported costs in these studies were derived from charges, which in a US context are not satisfactory proxies for cost of care.

Previous literature used a variety of national databases to characterize population-based costs associated with oral cavity and pharyngeal cancer cohorts.^{9,11,13,14,19} Similar to our study, most costs across different US payers are incurred within the first year of diagnosis, reflecting the treatment period.^9,13,14 Using the California Medicaid database (1995-2003), Epstein et al⁹ estimated that the median 1-year cost of care after initial diagnosis of oral and pharyngeal cancer was $25 319 (year 2002 dollars). In a study by Jacobson et al¹³ that used the Thomas Reuters MarketScan database (2004-2008), the authors compared costs across US commercial, Medicare, and Medicaid insurance. They reported 1-year annual increased costs in the Medicare population of $35 890 (year 2008 dollars) compared with propensity score–matched patients in the noncancer control group. Notably, the costs encountered in the commercially insured population were approximately double this value. Finally, Lang et al¹⁴ conducted a similar study to our report by using a Medicare-SEER linked database (1991-1993). They report mean Medicare payments among those with squamous cell cancer of the head and neck of $25 542 (year 1998 dollars) higher than those of matched comparison patients. In addition, they estimate that the costs to Medicare for incident cases in a given year followed up to 5 years is more than $250 million.¹⁴ The fundamental difference between our study and previous studies is in how we have dealt with patient censoring. For example, the MarketScan data do not differentiate between death and disenrollment, and the Medicare database is characterized by variable follow-up. By using inverse probability weighting, our statistical models account for patient censoring in our estimates of cost.³⁵ In comparison, we estimate the 5-year cumulative costs to Medicare to be approximately $27 000 for oral cavity cancers and approximately $40 000 for pharyngeal cancers.

This study is limited by use of administrative data. In addition, because this research uses data from the Medicare population, these costs cannot be generalized to younger patients. The costs reported are only for fee-for-service patients; patients enrolled in a health maintenance organization (accounting for 16% of Medicare overall) are not included in the analysis. Furthermore, we could not measure the incidence of complications and their contribution to cumulative costs. Another limitation is that there are characteristics of patients that may be important but are not available in the data set. For example, we might expect smoking to be more prevalent among patients with oral cavity and pharyngeal cancers than the noncancer comparison group. However, smoking is not available in the data set. The database also does not include human papillomavirus status, and we are not able to link human papillomavirus with pharyngeal cancer using this data set, although conventional wisdom would consider this an increasingly important risk factor. Finally, and most importantly, outpatient drug costs are not included in these estimates because Medicare did not provide reimbursements for drugs during the study interval, including chemotherapy drugs, if they were used for treatment. Costs for administration of chemotherapy are available, but the drug costs themselves are not.

Despite these limitations, our study reveals the significant economic burden associated with oral cavity and pharyngeal cancers in a contemporary cohort of elderly patients from a Medicare cost perspective. Using inverse probability weighting, we were able to account for patient censoring to provide a more accurate estimate of costs. With an attributable cost of $27 000 for patients with oral cavity cancer and $40 000 for pharyngeal cancers, there are substantial costs attributable to this disease. Although any comprehensive effort to reduce the economic burden of oral cavity and pharyngeal cancers must emphasize preventive and early diagnostic measures given the role of comorbidities and treatment modality as the primary determinants of cumulative costs at 5 years, further research will be needed to understand the cost-effectiveness of different modalities of treatment in patients with varying levels of comorbidities and stage of disease.

Submitted for Publication: October 30, 2014; final revision received March 12, 2015; accepted April 15, 2015.

Corresponding Author: Christopher S. Hollenbeak, PhD, Departments of Surgery and Public Health Sciences, College of Medicine, The Pennsylvania State University, 500 University Dr, Mailbox H151, PO Box 850, Hershey, PA 17033 (chollenbeak@psu.edu).

Published Online: June 4, 2015. doi:10.1001/jamaoto.2015.0940.

Author Contributions: Dr Hollenbeak and Mr Schaefer had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Hollenbeak, Mackley, Koch, Schaefer, Goldenberg.

Acquisition, analysis, or interpretation of data: Hollenbeak, Kulaylat, Mackley, Schaefer.

Drafting of the manuscript: Hollenbeak, Kulaylat, Schaefer, Goldenberg.

Critical revision of the manuscript for important intellectual content: Hollenbeak, Mackley, Koch, Schaefer.

Statistical analysis: Kulaylat, Schaefer.

Obtained funding: Hollenbeak, Mackley.

Administrative, technical, or material support: Hollenbeak.

Study supervision: Hollenbeak, Mackley, Koch, Goldenberg.

Conflict of Interest Disclosures: None reported.

Funding/Support: This study was supported by grant 5 R03 DEOOO019511 from the National Institute of Dental and Craniofacial Research, National Institutes of Health (Drs Hollenbeak, Mackley, Koch, and Goldenberg and Mr Schaefer).

Role of the Funder/Sponsor: The funding source had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and the decision to submit the manuscript for publication.

Disclaimer: This study used the linked SEER-Medicare database. The interpretation and reporting of these data are the sole responsibility of the authors.

Additional Contributions: The Applied Research Program of the National Institutes of Health, the Office of Research, Development and Information of the Centers for Medicare & Medicaid Statistics, Information Management Services Inc, and the SEER Program tumor registries participated in the creation of the SEER-Medicare database.