Mean complication-related costs and 6-month total health care costs for patients with treatment-related complications.
Lang K, Sussman M, Friedman M, Su J, Kan HJ, Mauro D, Tafesse E, Menzin J. Incidence and Costs of Treatment-Related Complications Among Patients With Advanced Squamous Cell Carcinoma of the Head and Neck. Arch Otolaryngol Head Neck Surg. 2009;135(6):582-588. doi:10.1001/archotol.125.11.1203
To evaluate the incidence and costs of complications due to radiotherapy alone vs platinum-based chemoradiotherapy among patients diagnosed as having advanced squamous cell carcinoma of the head and neck (ASCCHN) from a payer perspective.
Retrospective cohort study.
Data from the PharMetrics Patient-Centric Database from June 2000 through June 2006.
The study included patients with ASCCHN and an indication of a secondary malignant neoplasm (both identified based on International Classification of Diseases, Ninth Revision, Clinical Modification, diagnosis codes), 124 of whom were treated with radiotherapy alone and 77 of whom were treated with chemoradiotherapy (including 53 with cisplatin plus radiotherapy, 26 with carboplatin plus radiotherapy, and 2 with cisplatin and carboplatin plus radiotherapy). The patients were assigned to 1 of 2 cohorts based on treatment type—radiotherapy only and platinum-based chemoradiotherapy—and were followed up for 6 months.
Main Outcome Measures
Incidence and costs of treatment-related complications associated with chemotherapy in ASCCHN.
We found significantly (P < .001) higher rates of treatment-related complications among patients receiving chemoradiotherapy (86%) than among patients receiving only radiotherapy (51%). The mean per-patient costs associated with treatment-related complications were approximately $10 000 higher among patients who received chemoradiotherapy than among those treated with radiotherapy alone (P < .001). These costs represented 17% of the total costs during follow-up for patients who received chemoradiotherapy and 11% of costs for those who received radiotherapy. The most expensive complications were dehydration and/or electrolyte imbalance and oral complications.
Our study results suggest that the attributable incidence and costs of treatment-related complications associated with chemotherapy in ASCCHN are substantial. The emergence of safer treatments may have the advantage of alleviating this cost burden.
Squamous cell carcinomas of the head and neck (SCCHN) are malignant tumors that originate from the mucosal lining of the head and neck region.1 More specifically, SCCHN includes cancers of the oral cavity, salivary glands, sinuses, pharynx, and larynx and is the most common type of carcinoma in these sites.2,3 More than 50 000 new cases of SCCHN are diagnosed annually in the United States.1,2 Diagnosing SCCHN early in the progression of the disease has considerable clinical and economic benefits.4 In addition to lower mortality rates and medical expenditures, early diagnosis of SCCHN is associated with an increased likelihood of treatment with a single modality: surgery, radiotherapy, or chemotherapy.4 While 1 treatment modality is an option for early-stage cases, advanced-stage SCCHN (ASCCHN) cases usually require a more robust treatment regimen, including a combination of modalities. For instance, in advanced disease, chemotherapy is often used as an adjunct to radiotherapy either before or after surgery.2
Platinum-based chemotherapies, primarily cisplatin and carboplatin, are among the most common types of chemotherapy that are used in conjunction with radiotherapy. These chemotherapies, while effective in the treatment of ASCCHN, can be associated with several debilitating complications, including oral complications (eg, candidiasis of the mouth, acute pharyngitis, stomatitis, and esophagitis), nephrotoxicity, neurotoxicity, nausea and vomiting, and electrolyte disturbances, among others.5,6 Furthermore, radiotherapy carries a number of similar acute and long-term complications. Acute reactions can include mucositis, candidiasis, dysgeusia, osteoradionecrosis, radiation dermatitis, anemia, neutropenia, and soft-tissue necrosis.5,7
An exhaustive search of the literature identified a number of analyses assessing the efficacy of various treatment types for patients diagnosed as having SCCHN.5,8- 11 In fact, results from these analyses have shown high rates of treatment-related complications among patients with SCCHN. While treatment-related efficacy and outcomes data exist, to our knowledge there are no studies that have examined the associated costs or implications of treatment-related complications in this population. Bringing to the forefront the costs of complications may help third-party payers make informed decisions about innovative, oncological therapies and their potential to reduce complication rates and thus associated health care costs.
The primary objective of this study was to assess the health care costs associated with treatment-related complications among patients with ASCCHN from a managed care population undergoing only radiotherapy vs platinum-based chemoradiotherapy. A secondary objective was to confirm the results from previous studies regarding clinical practices and incidence rates of treatment-related complications.
This retrospective cohort study was conducted using the PharMetrics Patient-Centric Database, which contains complete health care information for approximately 55 million patients from 75 managed care organizations, as well as several million Medicare managed care enrollees from 4 US geographical regions. The database consists of 2 files: a claims detail file and an eligibility file. The claims detail file contains information on medical and pharmacy claims, including date of service, place of service, International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, procedure codes, physician specialty, national drug codes, drug quantity dispensed, days supplied, charged and paid amounts, and copayments. The eligibility file contains monthly medical and pharmacy eligibility flags and patient demographics. Data for this study included claims and eligibility covering the period January 1, 2000, through June 30, 2006, to allow for a 6-month baseline period and a 6-month follow-up period for each patient.
Using the PharMetrics database, we identified patients as having ASCCHN if they had a medical claim indicating SCCHN between June 30, 2000, and December 31, 2005 (allowing for 6-month baseline and follow-up periods), and a listed diagnosis of a secondary malignant neoplasm. The ICD-9-CM diagnosis codes for malignant neoplasms of the lip, tongue, major salivary gland, gum, floor of the mouth, other and unspecified parts of mouth, oropharynx, nasopharynx, hypopharynx, ill-defined sites within the lip/oral cavity/pharynx, nasal cavities/middle ear/accessory sinuses, larynx, or other ill-defined sites of the head, neck, and face were used to define SCCHN. Because the stage of diagnosis is not included in the PharMetrics database, we created a claims-based algorithm to identify patients with ASCCHN. The algorithm consisted of a diagnosis of at least 1 SCCHN and a diagnosis of at least 1 secondary metastasis in the lymph nodes (including lymph nodes of the head, face, and neck; intrathoracic lymph nodes; intra-abdominal lymph nodes; and lymph nodes of the axillae and upper limbs), respiratory system, or digestive system. These locations were chosen owing to their proximity to SCCHN anatomical sites.
To be included in the analysis, patients were required to be at least 35 years old on January 1, 2000, and to have been eligible for their health plan in the 6 months before and after their SCCHN diagnosis. Patients were excluded from the analysis if they had claims indicating any other primary cancer at any time during the study period. This criterion was applied to ensure that the study outcomes were a direct result of treatment for ASCCHN and not a result of treatment for other primary tumors. Patients with claims for lung, lymph, or basal cell carcinoma were allowed in the study sample because these sites are regional to SCCHN and thus would not be considered secondary sites. Also, basal cell carcinoma was allowed because patients with this diagnosis would not be treated with chemotherapy.
Once patient selection criteria were applied and the ASCCHN population was identified, we created 2 study cohorts based on treatment type. The radiotherapy-only cohort included all patients who had a medical claim for radiotherapy in the 3 months after SCCHN diagnosis, excluding patients who received any chemotherapy drug or drug administration. The platinum-based chemoradiotherapy cohort included all patients who had a medical claim for radiotherapy and at least 1 platinum-based chemotherapy drug (cisplatin or carboplatin) in the 3 months after SCCHN diagnosis. Radiotherapy and chemoradiotherapy were identified in the claims using Healthcare Common Procedure Coding System, Current Procedural Terminology, ICD-9-CM, and diagnosis-related group codes. Patients were followed up for 6 months after their diagnosis date to evaluate all study outcomes.
This study measured clinical practices, incidence rates, and costs associated with treatment-related complications among patients with ASCCHN who were undergoing radiotherapy alone vs platinum-based chemoradiotherapy. The ASCCHN-related treatments, including surgical interventions, radiotherapy, and chemoradiotherapy, were assessed separately among patients diagnosed in each study year in order to identify trends. Surgical interventions included endoscopic sinus surgery, gastrostomy, glossectomy, laryngectomy, mandibulectomy, maxillectomy, neck dissection, parotidectomy, and tracheostomy; these were identified using ICD-9-CM procedure codes.
Also, this study evaluated the incidence of treatment-related complications and the incidence of complications requiring hospitalization. The overall incidence of treatment-related complications was defined as the unique number of patients with a claim indicating 1 or more specified complications during follow-up divided by the total number of patients in each treatment cohort. The incidence of treatment-related complications resulting in hospitalization, including emergency department visits, was defined as the number of unique patients with a claim indicating hospitalization for 1 or more complications divided by the total number of patients in each treatment cohort. The complications we evaluated included oral complications, dehydration/electrolyte imbalance, nausea/emesis/diarrhea, infection and fever, anemia, malnutrition, neutropenia/thrombocytopenia, and hearing complications.6,7 Treatment-related complications were based on ICD-9-CM diagnosis codes (ICD-9-CM codes 112.0, 288.0, 462, 527.7, 528.0, 530.1x, 787.2, 276.1, 276.5, 276.8, 276.9, 780.79, 787.0x, 787.91, 038.9, 480-486, 490, 780.6, 785.5x, 790.7, 284.x, 783.0, 287.4, 288.0, 288.8, 288.9, 389.0x, 389.1x, 389.2, and 389.8).
Finally, this study assessed the differences in treatment-related complication costs per patient and total health care costs per patient with treatment-related complications among the chemoradiotherapy and the radiotherapy-only cohorts. All costs were calculated by component, including hospital inpatient, hospital outpatient, physician, and outpatient pharmacy. Costs represented the amount paid and were estimated based on health plan payments for medications, services rendered, and net of patient responsibility (ie, copayment, coinsurance, and/or deductible). Because patients were identified over multiple years, costs were adjusted to 2006 US dollars using the Consumer Price Index for Medical Care and were expressed as 6-month costs.
We evaluated treatment-related costs on both an unadjusted and an adjusted basis. For the latter, we adjusted costs for differences in potentially confounding factors, including age, sex, presence of comorbidities, whether surgery was performed, and anatomical site. Anatomical site was created as 5 categorical variables: pharynx, larynx, tongue, other specified (major salivary glands, floor of mouth, and nasal cavities/middle ear), and other unspecified (other and ill-defined sites/head/neck/face, other and unspecified parts of mouth, and other and ill-defined sites/lip/oral cavity/pharynx). Adjusted analyses were computed using a log link and a negative binomial distribution in the setting of generalized linear model, resulting in estimated effects on the log scale. To generate predicted values on the natural scale (ie, costs), we used the estimated linear predictor from the model with the mean covariate values. We then inverted the log link by exponentiating the resulting predictor. In other words, we multiplied each estimated regression coefficient from the model by the mean value for the associated covariate, summed the resulting products, and transformed these predicted mean values back to the natural scale through simple exponentiation.
Differences in study measures between the overall chemoradiotherapy cohort and the radiotherapy-only cohort were considered attributable to chemotherapy. The significance of differences in treatment-related complication rates was evaluated using χ2 tests, while the significance of differences in expenditures was evaluated using Wilcoxon nonparametric tests. All analyses of data were conducted using an SAS software package (Version 9.1; SAS Institute Inc, Cary, North Carolina).
We identified 124 patients with ASCCHN who were treated only with radiotherapy and 77 patients who were treated with chemoradiotherapy (53 with cisplatin plus radiotherapy, 26 with carboplatin plus radiotherapy, and 2 with cisplatin and carboplatin plus radiotherapy) (Table 1). The most common anatomical sites included the pharynx, the larynx, and the tongue. Patients undergoing chemoradiotherapy were slightly younger than patients undergoing radiotherapy only (mean age, 53.4 years for chemoradiotherapy and 56.2 years for radiotherapy) and were more likely to be male. Patients undergoing chemoradiotherapy were less healthy than patients in the radiotherapy-only cohort, as evidenced by higher Charlson comorbidity scores. Moreover, patients treated with chemoradiotherapy experienced higher rates of mandibulectomy and tracheostomy within 6 months of study entry than patients treated with radiotherapy alone (Table 2).
The incidence of treatment-related complications was significantly (P < .001) greater among patients treated with chemoradiotherapy than among those treated with radiotherapy alone, regardless of hospitalization status (Table 3). For instance, 86% of patients treated with chemoradiotherapy experienced a treatment-related complication, compared with 51% of patients treated only with radiotherapy. Similarly, the incidence of treatment-related complications requiring hospitalization was significantly (P < .001) greater among patients treated with chemoradiotherapy than among those treated only with radiotherapy (Table 3). Almost 59% of the patients treated with chemoradiotherapy experienced a complication requiring hospitalization compared with 35% of radiotherapy-only patients.
The most common treatment-related complications among patients undergoing chemoradiotherapy were oral complications, dehydration/electrolyte imbalance, and infection/fever. Incidence rates for most of these complications were significantly (P < .001) higher for the chemoradiotherapy cohort than for the radiotherapy-only cohort. The length of stay among hospitalized patients with complications was slightly longer for the chemoradiotherapy patients than for the radiotherapy-only patients (difference of 0.3 days, P = .10; Table 3).
Complication-related health care costs represented 17% of total health care costs for patients with ASCCHN who received chemoradiotherapy and 11% of total health care costs for patients who received only radiotherapy (Figure) during follow-up. Moreover, complication-related costs attributable to chemotherapy were estimated to be $9602 (P < .001, Table 4), with hospital inpatient costs representing the largest share. After potential confounding factors were adjusted for, complication-related costs were somewhat higher, at $14 714. For the purposes of the multivariate analyses, only age was significant at the P < .05 level.
Patients hospitalized with complications incurred substantially higher costs than nonhospitalized patients. Regarding costs for specific complications, dehydration and electrolyte imbalance and oral complications had the highest cost differences (excluding costs of neutropenia and thrombocytopenia, for which the sample sizes were quite small).
To our knowledge, this retrospective database analysis is the first to evaluate both the incidence and the costs associated with treatment-related complications in ASCCHN. We found that the use of chemoradiotherapy has been increasing in recent years, most likely because of its effectiveness. For instance, Adelstein et al11 found that platinum-based chemotherapy (cisplatin) and concurrent radiation therapy significantly improved survival among patients with squamous cell head and neck cancer despite higher levels of toxicity. Similar to Adelstein and colleagues' analysis, we found that patients receiving chemoradiotherapy had significantly (P < .001) higher rates of treatment-related complications than patients receiving only radiotherapy (86% and 51%, respectively, Table 3). The incidence of complications requiring hospitalization was also significantly (P < .001) greater among patients treated with chemoradiotherapy than among those treated only with radiotherapy (59% vs 35%, Table 3). We also found that mean per-patient 6-month costs associated with treatment-related complications were approximately $10 000 higher for patients treated with chemoradiotherapy than for those treated only with radiotherapy. The most expensive complications were oral complications, dehydration and electrolyte imbalance, and infection and fever.
While we are unaware of any previous studies of treatment-related complication costs in SCCHN, our findings are consistent with results from previous studies of the incidence of treatment-related complications in SCCHN. Using Surveillance, Epidemiology and End Results registry data linked to Medicare Claims, McLaurin et al5 assessed the incidence of treatment-related toxic effects in an elderly SCCHN Medicare population and found substantial excess rates of complications associated with chemoradiotherapy treatment (excess rates for myelosuppression, oral complications, dehydration, and nausea and vomiting ranging from 15% to 41%). Also, in a phase 2 study of concurrent cetuximab, cisplatin, and radiotherapy for the treatment of ASCCHN, Pfister et al9 found that the incidence of treatment-related adverse events (dehydration, fever, and anemia) was similar to our results. Al-Sarraf et al10 evaluated the rates of severe toxic effects due to concurrent radiotherapy and chemotherapy with cisplatin among patients with locally advanced and inoperable head and neck cancer and found similar incidence rates for anemia and nausea and vomiting. In a study of 55 male patients with locally advanced head and neck cancer, Nguyen et al8 found a 45% incidence rate of chronic dysphagia for patients undergoing chemoradiotherapy, which is slightly lower than our findings for oral complications overall.
Our findings are also in line with results from other studies of treatment-related complications rates and costs among patients diagnosed as having other types of cancers and treated with platinum-based therapies. For instance, Long et al12 studied women with carcinoma of the uterine cervix and found that patients undergoing a combination of cisplatin and topotecan had much higher incidence rates of neutropenia (70% vs 1%), thrombocytopenia (31% vs 3%), anemia (38% vs 23%), and infection (18% vs 8%) than patients undergoing treatment with cisplatin alone. In a phase 2 study of 89 patients with squamous cell cancer of the uterine cervix, Martinez et al13 found that 70% of patients undergoing carboplatin therapy experienced a complication, with vomiting (43%), infection (11%), and diarrhea (6%) among the most common. Hassett et al14 assessed treatment-related complications among a population of patients with breast cancer, finding incidence rates for neutropenia, thrombocytopenia, nausea, emesis, and diarrhea that were similar to our findings. They also found incremental expenditures for treatment-related complications of $1271 and expenditures of $12 907 for treatment-related complications leading to hospitalization. Finally, Lyman et al15 assessed excess costs of anemia among patients undergoing chemotherapy for a variety of cancers. They found mean 6-month total expenditures to be $36 781 for patients without anemia and $62 499 for patients with anemia, leading to an excess anemia-related cost of approximately $26 000.
Our study had a number of limitations related to difficulties with identifying ASCCHN cases. First, we were unable to confirm that study patients had advanced-stage cancer because the stage of cancer was not available in the PharMetrics database. We therefore created a patient selection algorithm that required a diagnosis of SCCHN as well as a diagnosis of a metastasis in the lymph nodes, respiratory system, or digestive system. While we expect that this algorithm ensured identification of patients with ASCCHN, we were unable to unequivocally determine that this is the case. Because of this limitation, we cannot state a definitive and causal relationship between chemoradiotherapy and higher costs and rates of complications. Instead, we can only ascertain an association between the combination therapy and outcomes. Second, as is the case with all retrospective database analyses, our study is subject to inaccuracies inherent in ICD-9-CM coding.16 Underestimating or overestimating the number of ASCCHN cases may have led to inaccuracies in measuring and reporting study outcomes.
Other limitations dealt with the characteristics of the study population. Because the date of cancer diagnosis was not available, we were unable to distinguish between newly diagnosed patients and patients further along in their cancer. Also, the age distribution of patients with head and neck cancer in our sample is representative of the PharMetrics database only and does not necessarily reflect the general population of patients with SCCHN. In fact, according to the Surveillance, Epidemiology, and End Results Program SEER-Stat Database, the fraction of Medicare-eligible patients (ie, persons aged ≥65 years) diagnosed as having SCCHN is 45%.17 Underrepresentation of this age group in our analysis may have had a downward effect on the estimates of treatment-related complication rates and mean health care costs. As a result, further investigation of older patients is warranted. A final limitation concerns our inability to confirm with certainty that complications were treatment related. However, the complications we evaluated are common among patients who are treated with radiotherapy and chemotherapy.5,6 Furthermore, the average length of time between treatment and complication in our analysis was about 1 month, giving us confidence in our approach.
Despite its limitations, this study showed that chemoradiotherapy is associated with substantially higher complication-related costs as compared with radiotherapy only among patients diagnosed as having ASCCHN. New treatments that have the ability to maintain similar efficacy profiles and to elicit fewer complications7,11,18 should be further explored for their potential to reduce the cost burden presented by current therapies.
Correspondence: Joseph Menzin, PhD, Boston Health Economics, Inc, 20 Fox Rd, Waltham, MA 02451 (firstname.lastname@example.org).
Submitted for Publication: March 20, 2008; final revision received August 26, 2008; accepted September 21, 2008.
Author Contributions: All authors 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: Lang, Sussman, Friedman, Su, Kan, Mauro, Tafesse, and Menzin. Acquisition of data: Kan and Menzin. Analysis and interpretation of data: Lang, Sussman, and Menzin. Drafting of the manuscript: Lang and Sussman. Critical revision of the manuscript for important intellectual content: Lang, Sussman, Friedman, Su, Kan, Mauro, Tafesse, and Menzin. Statistical analysis: Lang, Sussman, Kan, and Menzin. Obtained funding: Sussman,Su, Tafesse, and Menzin. Administrative, technical, and material support: Mauro. Study supervision: Lang, Sussman, Friedman, Su, and Menzin.
Financial Disclosure: Drs Su and Kan are employees of the sponsor (Bristol-Myers Squibb, Wallingford, Connecticut); Dr Su has an equity stake of more than $10 000 in the sponsor, and Dr Kan has an equity stake of less than $10 000 in the sponsor; Drs Mauro and Tafesse were employed by the sponsor at the time of the study.
Role of the Sponsor: The sponsor had no role in the design and conduct of the study; in the collection, analysis, and interpretation of the data; or in the preparation, review, or approval of the manuscript aside from the contributions of Drs Su, Kan, Mauro, and Tafesse. The publication of the results was not contingent on the sponsor's approval of the manuscript.
Financial Support: Funding for this study was provided by Bristol-Myers Squibb.
Additional Contributions: Jonathan Korn, BA, and Lisa M. Lines, BS, ELS, provided valuable assistance with manuscript preparation.