Economic burden of resected squamous cell carcinoma of the head and neck in a US managed-care population

Abstract

Background:

Squamous cell carcinoma of the head and neck (SCCHN) places a high burden on society and poses complex challenges to healthcare providers.

Methods:

Retrospective claims-based analysis of commercially insured patients identified between 01-31-04 and 12-31-07 with diagnostic evidence of cancer of the lip, tongue, oral cavity, pharynx, or larynx who underwent surgical resection during identification period. Outcomes included treatment patterns, healthcare utilization, and costs. All study variables were analyzed descriptively.

Results:

Among the 1104 patients in the final study sample, 71.9% were male, with mean age 56.6 years. On average, patients were followed for 830 days (range of mean days: 805 for lip or tongue cancer to 847 for pharyngeal cancer). About half received radiation therapy during follow-up, whereas only 16.2% received chemotherapy. Patients with pharyngeal cancer were most likely to undergo chemotherapy. After their index surgery, 57.9% of patients had ≥1 inpatient stay, 44.9% had ≥1 ER visit, and all had ≥1 ambulatory visit. The percentage with ≥1 inpatient stay post-index was highest among patients with pharyngeal cancer (73.0%) and lowest in the laryngeal cancer cohort (49.5%). Mean number of hospitalized days, ER visits, and ambulatory visits was 0.45, 0.69, and 27.4, respectively, per-patient per-year. Overall, patients incurred ∼$94 million in cost following index surgery ($85,000 per-person, on average). Mean total healthcare cost was $34,450 per-patient per-year, the bulk of which comprised medical expenses ($32,401). The highest mean healthcare cost was incurred by the pharyngeal cancer cohort ($40,214).

Conclusions:

Patients with resected SCCHN incur substantial healthcare costs and have high utilization rates. Results of this analysis are primarily applicable to resected SCCHN in a managed-care setting, and therefore may not be generalizable to the entire US population. Furthermore, disease stage is an important factor impacting outcomes, but these analyses did not stratify patients according to disease stage.

Key words::

• Chemotherapy
• Claims analysis
• Cost of illness
• Head and neck cancer
• Outcomes assessment

Introduction

Head and neck cancer encompasses a group of cancers with similar etiology and pathogenesis that can arise at various sites including the lips, oral cavity, nasal cavity, salivary glands, paranasal sinuses, thyroid, pharynx, and larynx. It is the fifth most common neoplasm in the world1 and the sixth leading cause of cancer mortality worldwide2. In 2010, an estimated 49,260 new cases of head and neck cancer were diagnosed and 11,480 associated deaths reported in the United States3. More than 90% of these were squamous cell carcinomas of the head and neck (SCCHN) that originate from the epithelial mucosa lining1,3^--5.

Although many other types of cancer have considerably higher incidence rates, the significant economic burden that the various head and neck cancers place on society should not be underestimated. Furthermore, these cancers take a high toll on affected patients in terms of morbidity and complications, are associated with a poor prognosis, and pose complex treatment challenges to healthcare providers. Because of the association of ethanol and tobacco abuse as etiologic factors for these malignancies, patients with SCCHN often have comorbidities that contribute to their poor prognosis. In addition, the entire aerodigestive tract may be exposed to these carcinogens, rendering patients with SCCHN more susceptible to development of second primary tumors of the head and neck as well as other malignancies, especially of the esophagus or lung.

Patients diagnosed with localized (stage I or II) SCCHN generally fare better than those whose cancer is locally or regionally advanced (with or without lymph node metastases)6. For example, estimated average 5-year relative survival rates for cancer of the larynx, based on SEER data from 1999 to 2006, are 77.6% when the cancer is localized, 41.5% when it is regional, and 33.2% for distant SCCHN7. For cancer of the oral cavity and pharynx, the estimated 5-year survival rates are, respectively, 82.5%, 54.7%, and 32.2%8. About half of all SCCHN cases diagnosed each year are early-stage and are curable by conventional treatments such as surgery and radiotherapy9. Earlier diagnosis has been shown to increase the likelihood of single-modality therapy, lower the mortality risk, decrease medical expenditures, and improve patients’ quality of life1,10. For patients with metastatic or recurrent disease, however, treatment is usually palliative and only a small percentage derive long-term benefit from treatment due to the substantial toxicity and morbidity it confers5,11,12.

For the approximately 30–40% of patients with SCCHN who present with early-stage disease (stages I and II), the main treatment modalities are surgery and definitive radiotherapy, with similar local control and survival achieved with these two interventions13,14. For the ∼60% with locoregionally advanced SCCHN, treatment options include neoadjuvant chemotherapy (prior to surgery and/or radiotherapy), concurrent chemoradiotherapy, sequential chemotherapy followed by radiotherapy, or primary surgery followed by postoperative radiotherapy, often with concurrent chemotherapy15,16. Treatment options for individuals with recurrent disease are limited by the prior therapies that have been administered. Salvage surgical and radiotherapy intervention may be appropriate in the absence of metastatic foci, whereas systemic chemotherapy is the treatment of choice when metastatic disease is present. Palliative chemotherapy, as well as best supportive care, are other important treatment options for those with recurrent and metastatic disease16.

Regardless of the stage at which SCCHN is diagnosed, treatment is complex due to the anatomic constraints of the head and neck region and the importance of maintaining organ function in the face of treatment-related toxicity (e.g., mucositis, stomatitis, dermatitis, and myelosuppression17). According to the findings of a Longitudinal Oncology Registry of Head and Neck Carcinoma study, 55% of the 1524 patients enrolled in the study had a feeding tube placed, 13% needed a tracheotomy, 79% required opioid analgesics, 78% had antiemetics prescribed, and 34% reported grade 3–4 mucositis/stomatitis17. Sequelae of these treatment modalities – chemotherapy, radiation, and surgery – include enduring pain (including chemotherapy-related peripheral neuropathy), disfigurement, and disability that add to the burden of the disease1. The tumors themselves, as well as the treatments, can affect one’s ability to chew, swallow, and even breathe; senses of taste, smell, and hearing may be affected. The tumors, as well as the surgery, may be disfiguring, and sequelae such as inability to speak or hoarseness are readily apparent. Despite advances in the efficacy of treatment, the prognosis for SCCHN patients is poor, with a 5-year survival rate of about 50%18 and a high risk of local disease recurrence (10–15%) or disease progression (20–30%) within 1 year of treatment19. Studies have reported that within 2 years, 20–30% of patients develop metastatic disease and 50–60% experience local recurrence20^--22. In a prospective multivariate analysis of 446 consecutively treated SCCHN patients, median time to relapse was 8.3 months and 95% of relapses occurred within 3 years23.

Relatively few studies have evaluated the expenditures associated with SCCHN, and most cost studies published to date have pertained to individual cancers or particular treatments6,24. Furthermore, follow-up durations have been limited10 or patient samples have been small25.

To add to the existing literature, the present study was designed to examine patient demographic and clinical characteristics, post-surgery treatment patterns, and pre- and post-surgery all-cause and disease-specific healthcare utilization and costs in a sample of patients with resected SCCHN.

Materials and methods

Study design

This was a retrospective claims-based analysis using medical, pharmacy, and laboratory data and enrollment information from a large US database of commercially insured patients identified between 1 January 2004 and 31 December 2007 (identification period). At the time this study was conducted, approximately 14 million geographically diverse individuals with both medical and pharmacy benefit coverage were enrolled in the health plan. Data were de-identified in compliance with the Health Insurance Portability and Accountability Act (HIPAA); approval by an institutional review board was therefore not required.

Patient sample

Inclusion criteria

Health plan enrollees aged ≥18 years were included in the study if they had at least two SCCHN diagnoses 21 days apart and underwent a surgical resection during the identification period. SCCHN diagnoses were identified from International Classification of Diseases, Ninth Revision [ICD-9] and included lip (ICD-9 140.x), tongue (ICD-9 141.x), oral cavity (ICD-9 codes 143.x–145.x), pharynx (ICD-9 codes 146.x, 148.x, 149.x), and larynx (ICD-9 code 161.x). The date of the first observed surgical resection was defined as the index date. Patients were required to be continuously enrolled with medical and pharmacy benefits for 182 days prior to the index date (pre-index period), and for at least 12 months following the index date (post-index period).

Exclusion criteria

Patients with evidence of nasopharyngeal cancer (NPC) (ICD-9 147.x) in any position on the medical claim during the identification period were excluded from the study. NPC accounts for less than 1% of cancers in the United States and is more common in other geographic areas, especially Southeast Asia; only some NPCs are squamous cell and, unlike other head and neck cancers, all NPCs are associated with EBV-latent gene products. Patients with evidence of a malignancy other than head and neck cancer during the study period were removed from the analysis. Secondary malignancies, benign tumors, and tumors of unspecified site were not included as indicators of other cancer. Codes for other cancer included: 142.x, 150.x–160.x, 162.x–172.x, 174.x–195.x, and 200.xx–209.3x. Also excluded from the study were patients with evidence of the presence of two or more types of SCCHN, and patients with identifying information missing from the data.

Observation period

Patients were followed for at least 12 months from the index date or until death, disenrollment from the health plan, or 31 December 2008, whichever occurred first. Those who died during the follow-up period (in an inpatient setting only) were identified from facility claims, and the number of days from the date of index surgery until death was identified. The cancer diagnosis a patient received on the index date determined the disease-site cohort to which he or she was assigned.

Outcomes

Patients’ demographic and clinical characteristics were assessed, including their age as of the index year, the age group into which they were classified (18–44, 45–64, or 65+ years), gender, geographic region in which they were enrolled in their health plan (as categorized in accordance with US Census Bureau region designations), and the site of disease.

Post-index treatment patterns were evaluated, including radiation therapy (≥1 claim for radiation post-surgery), use of specific chemotherapies, administration of hormone therapy (such as for treatment after thyroid replacement), secondary surgeries performed, and treatment for related conditions such as anemia, neutropenia, thrombocytopenia, nausea/emesis, pain, xerostomia, and mucositis.

Pre- and post-index healthcare utilization and costs, categorized as either all-cause (overall) or SCCHN-attributable (as evidenced by claims with a primary diagnosis of cancer of the oral cavity, pharynx, lip, tongue, or larynx or claims with diagnoses of anemia or for reconstructive surgery or repeated surgical resection), were measured.

Healthcare utilization encompasses the total number of inpatient days and counts of emergency room (ER) visits and ambulatory visits (office and outpatient). Healthcare costs include expenditures for inpatient stays, ER visits, ambulatory visits, and other services, as well as medical costs, pharmacy costs, and total costs, defined as medical plus pharmacy costs computed from health plan-paid and patient-paid amounts and adjusted for inflation to reflect 2008 US dollars. In order to account for varying lengths of follow-up, costs were calculated over the entire follow-up period and annualized (total cost divided by number of years of follow-up). Costs for additional services related to administration of chemotherapy in an outpatient setting were also calculated.

Statistical analysis

This study was descriptive with no testable hypotheses; as such, all study variables, including pre-index and outcome measures, were analyzed descriptively and are presented by disease-site cohort. Numbers and percentages are provided for dichotomous and polychotomous variables, and mean values and standard deviations are provided for continuous variables.

Costs are presented per-patient per-year, as well as per cohort over the length of follow-up. Descriptive techniques were used that account for length of observation time, i.e. incidence rates for binary measures and annualized amounts for counts and costs.

Incidence rates (presented per-patient per-year) were calculated as the total number of patients who experienced an event divided by aggregate follow-up time. Patients with the event only contributed the time to their first event.

Results

Demographic and clinical characteristics

During the identification period, 15,785 patients were identified from claims as having SCCHN. After applying all inclusion and exclusion criteria, the final sample included 1104 patients: 323 (29.3%) with laryngeal cancer, 299 (27.1%) with tongue cancer, 229 (20.7%) with oral cavity cancer, 185 (16.8%) with pharyngeal cancer, and 68 (6.2%) with lip cancer. Figure 1 illustrates the step-by-step sample selection and attrition process.

Figure 1.  Sample selection and attrition.

The mean age of the total study population was 56.6 years, and the majority of patients (71.9%) were male (Table 1). Almost half of the study sample (49.1%) resided in the South, and almost a third (30.3%) lived in the Midwest. On average, patients were followed for 830 days (∼2.3 years), ranging from 805 days for patients with lip or tongue cancer to 847 days for those with pharyngeal cancer. More than half (56.8%) of patients had their follow-up censored because the end of the time period was reached, and 42.0% were censored because of disenrollment from the health plan. The remaining 13 patients (1.2%) died during the study period (as observed from hospital and skilled nursing facility discharge codes). The length of follow-up and reasons for end of patient follow-up were consistent across cohorts.

Table 1.  Patient demographic characteristics.

	Total (n = 1104)		Oral cancer (n = 229; 20.7%*)		Pharyngeal cancer (n = 185; 16.8%*)		Lip/tongue cancer (n = 367; 33.3%*)		Laryngeal cancer (n = 323; 29.3%*)
	n	%	n	%	n	%	n	%	n	%
Index year
2004	250	22.64	55	24.02	37	20.00	80	21.80	78	24.15
2005	272	24.64	50	21.83	58	31.35	77	20.98	87	26.93
2006	299	27.08	69	30.13	38	20.54	103	28.07	89	27.55
2007	283	25.63	55	24.02	52	28.11	107	29.16	69	21.36
Age group
18–4	149	13.50	32	13.97	21	11.35	67	18.26	29	8.98
45–4	730	66.12	152	66.38	145	78.38	233	63.49	200	61.92
65+	225	20.38	45	19.65	19	10.27	67	18.26	94	29.10
Gender (Male)	794	71.92	130	56.77	153	82.70	258	70.30	253	78.33
Geographic location
Northeast	92	8.33	27	11.79	12	6.49	27	7.36	26	8.05
South	542	49.09	95	41.48	99	53.51	171	46.59	177	54.80
Midwest	334	30.25	69	30.13	53	28.65	114	31.06	98	30.34
West	136	12.32	38	16.59	21	11.35	55	14.99	22	6.81
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Age	56.60	11.63	56.72	12.53	54.66	9.02	55.01	11.94	59.43	11.43

*Represents percentage contribution of each disease site in overall sample.

Pre-index healthcare utilization and costs

Almost all patients (97.6%) had at least one pre-index ambulatory visit, 13.9% had at least one ER visit and 8.4% had at least one inpatient stay during the 6-month baseline period (Table 2, Figure 2). The laryngeal cancer cohort had the highest percentages of patients with an ER visit (16.4%) and an inpatient stay (10.2%), and the oral cancer cohort had the lowest percentages (11.4% and 7.0%, respectively). On average, patients spent 0.13 days in hospital, had 0.35 ER visits, and had 14.7 ambulatory visits. On average, patients with lip/tongue cancer had the shortest hospital stay and patients with pharyngeal cancer had the longest stay.

Table 2.  Pre-index healthcare utilization and costs* (per-patient per-year).

	Total (n = 1104)		Oral cancer (n = 229)		Pharyngeal cancer (n = 185)		Lip/tongue cancer (n = 367)		Laryngeal cancer (n = 323)
	n	%	n	%	n	%	n	%	n	%
Inpatient stays	93	8.42	16	6.99	18	9.73	26	7.08	33	10.22
ER visits	153	13.86	26	11.35	26	14.05	48	13.08	53	16.41
Ambulatory visits	1,077	97.55	221	96.51	183	98.92	358	97.55	315	97.52
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Number of inpatient days	0.13	0.53	0.12	0.56	0.18	0.68	0.11	0.50	0.13	0.43
Number of ER visits	0.35	2.04	0.20	0.68	0.64	4.30	0.33	1.43	0.30	1.01
Number of ambulatory visits	14.66	15.06	12.93	11.75	19.73	18.49	14.76	17.08	12.87	11.51
Total healthcare cost ($)	12,572	30,253	10,090	22,197	20,071	34,960	12,097	33,841	10,575	27,298
Total medical cost ($)	11,517	29,634	9,274	21,823	18,882	34,342	10,963	32,920	9,520	26,946
Cost of Inpatient stays ($)	2,654	15,283	3,310	17,741	2,532	10,481	1,914	11,133	3,101	19,298
Cost of ER visits ($)	79	465	102	771	88	471	55	243	84	353
Cost of ambulatory visits ($)	8,404	20,458	5,441	8,215	15,875	30,285	8,633	22,789	5,966	15,069
Cost of other services ($)	380	2,021	421	1,857	387	1,385	362	2,530	369	1,782
Total pharmacy cost ($)	1,054	1,754	817	1,120	1,189	1,820	1,134	2,259	1,055	1,380

*The costs for subcategories shown in the result tables may not sum to the cost of the parent category due to rounding.

Figure 2.  Pre-index vs. post-index all-cause healthcare utilization (% of patients).

The mean total healthcare cost was $12,572 per-patient per-year during the 6-month pre-index period. The majority of the cost was for medical expenses ($11,517), and pharmacy costs comprised the remaining $1054 (Figure 3). Patients with pharyngeal cancer incurred the highest mean pre-index total healthcare cost ($20,071), and those with oral cancer had the lowest cost (mean $10,090). Patients with lip/tongue cancer incurred a mean cost of $12,097, and those with laryngeal cancer incurred a mean cost of $10,575.

Figure 3.  Pre-index vs. post-index all-cause costs.

Post-index treatment patterns

Approximately half of the sample received radiation therapy during the follow-up period, whereas only 16.2% received post-index chemotherapy (Table 3). About one-third of patients with oral or tongue/lip cancer (34.9%) and about two-thirds of patients with pharyngeal (67.6%) or laryngeal (66.3%) cancer received radiation. The incidence rate for radiation therapy was 0.3975 encounters per patient-year, and was highest among patients with pharyngeal and laryngeal cancer – 0.7570 and 0.7009, respectively (data not shown).

Table 3.  Post-index treatment patterns, overall and by type of SCCHN.

	Total (n = 1104)		Oral cancer (n = 229)		Pharyngeal cancer (n = 185)		Lip/tongue cancer (n = 367)		Laryngeal cancer (n = 323)
	n	%	n	%	n	%	n	%	n	%
Radiation treatment	555	50.27	80	34.93	125	67.57	136	37.06	214	66.25
Chemotherapy
Platinum chemotherapy	126	11.41	18	7.86	43	23.24	42	11.44	23	7.12
Antimitotic chemotherapy	75	6.79	11	4.80	29	15.68	23	6.27	12	3.72
Biologically directed therapies*	58	5.25	7	3.06	24	12.97	20	5.45	7	2.17
Antimetabolite chemotherapy	51	4.62	11	4.80	16	8.65	16	4.36	8	2.48
Non-platinum alkylating chemotherapy	11	1.00	2	0.87	2	1.08	4	1.09	3	0.93
Antibiotic chemotherapy	6	0.54	0	0.00	3	1.62	1	0.27	2	0.62
Podophyllotoxins	3	0.27	1	0.44	0	0.00	1	0.27	1	0.31
Miscellaneous agents	3	0.27	1	0.44	1	0.54	0	0.00	1	0.31
Immune therapies	2	0.18	0	0.00	0	0.00	1	0.27	1	0.31
Topoisomerase II inhibitors (anthracenedione)	1	0.09	0	0.00	1	0.54	0	0.00	0	0.00
Topoisomerase I inhibitors	0	0.00	0	0.00	0	0.00	0	0.00	0	0.00
Hormone use	62	5.62	17	7.42	11	5.95	23	6.27	11	3.41
Secondary surgeries	508	46.01	137	59.83	62	33.51	165	44.96	144	44.58
Treatment for related conditions
Pain	978	88.59	212	92.58	168	90.81	315	85.83	283	87.62
Mucositis	746	67.57	144	62.88	148	80.00	215	58.58	239	73.99
Nausea/emesis	564	51.09	97	42.36	121	65.41	173	47.14	173	53.56
Xerostomia	170	15.40	27	11.79	72	38.92	56	15.26	15	4.64
Total parenteral nutrition	111	10.05	19	8.30	22	11.89	33	8.99	37	11.46
Anemia	105	9.51	20	8.73	28	15.14	30	8.17	27	8.36
Neutropenia	59	5.34	8	3.49	24	12.97	17	4.63	10	3.10
Food supplements	26	2.36	6	2.62	8	4.32	7	1.91	5	1.55
Thrombocytopenia	5	0.45	1	0.44	1	0.54	1	0.27	2	0.62

*Biologically directed therapies include bevacizumab, cetuximab, and erlotinib.

In this sample, 16.2% (179 out of 1104) were treated with chemotherapy, and patients with pharyngeal cancer were the most likely to undergo chemotherapy, with 32.4% of the pharyngeal cancer cohort (n = 60) receiving some type of chemotherapy: Among pharyngeal cancer patients receiving chemotherapy, 71.7% were prescribed platinum-based treatment, 48.3% received antimitotic chemotherapy, and 40.0% were prescribed a biologically targeted agent (e.g., bevacizumab, cetuximab, erlotinib) (Table 4). Among all patients treated with chemotherapy, the most widely prescribed treatments were platinum-based, and were prescribed for 70.4% of the 179 patients who received chemotherapy.

Table 4.  Post-index treatment patterns among only those patients receiving chemotherapy.

Chemotherapy (any)	Total (n = 179)		Oral cancer (n = 27)		Pharyngeal cancer (n = 60)		Lip/tongue cancer (n = 57)		Laryngeal cancer (n = 35)
	n	%	n	%	n	%	n	%	n	%
Platinum chemotherapy	126	70.39	18	66.67	43	71.67	42	73.68	23	65.71
Antimitotic chemotherapy	75	41.90	11	40.74	29	48.33	23	40.35	12	34.29
Biologically directed therapies*	58	32.40	7	25.93	24	40.00	20	35.09	7	20.00
Antimetabolite chemotherapy	51	28.49	11	40.74	16	26.67	16	28.07	8	22.86
Non-platinum alkylating chemotherapy	11	6.15	2	7.41	2	3.33	4	7.02	3	8.57
Antibiotic chemotherapy	6	3.35	0	0.00	3	5.00	1	1.75	2	5.71
Podophyllotoxins	3	1.68	1	3.70	0	0.00	1	1.75	1	2.86
Miscellaneous agents	3	1.68	1	3.70	1	1.67	0	0.00	1	2.86
Immune therapies	2	1.12	0	0.00	0	0.00	1	1.75	1	2.86
Anthracenedione chemotherapy	1	0.56	0	0.00	1	1.67	0	0.00	0	0.00
Topoisomerase I Inhibitors	0	0.00	0	0.00	0	0.00	0	0.00	0	0.00

*Biologically directed therapies include bevacizumab, cetuximab, and erlotinib.

Almost half of the total study population (46.0%) underwent a second surgical procedure. Patients with oral cavity cancer had the highest rate of second surgeries (59.8%), and the pharyngeal cancer cohort underwent the fewest second surgeries (6.0%). Most of the patients (88.6%) received treatment for pain, 67.6% received treatment for mucositis, and 51.1% were treated for nausea. There was overlap between these treatment categories, i.e. some patients received more than one of these ancillary treatments.

Post-index healthcare utilization

During the post-index period (including the index date), 57.9% of patients had at least one (all-cause) hospitalization, 44.9% had at least one ER visit, and everyone had at least one ambulatory visit (Figure 4). The percentage of patients hospitalized at least once was highest among patients with pharyngeal cancer (73.0%) and lowest among those with laryngeal cancer (49.5%). The percentage of patients with at least one ER visit was about the same among patients with oral cavity cancer, pharyngeal cancer, and laryngeal cancer (46.1–49.2%), but was lower among those with lip/tongue cancer (39.5%). Overall, patients had a mean of 0.45 days per-patient per-year for all-cause inpatient hospitalizations during the post-index period (range of means across groups: 0.40 for patients with lip or tongue cancer to 0.57 for those with pharyngeal cancer). The mean number of all-cause ER visits was 0.69 per-patient per-year (range, 0.52 for patients with oral cancer to 1.2 for patients with pharyngeal cancer), and the mean number of all-cause ambulatory visits was 27.4 per-patient per-year (range, 25.1 for patients with oral cancer to 33.2 for those with pharyngeal cancer).

Figure 4.  Post-index all-cause and SCCHN-attributable healthcare utilization (% of patients).

Almost all of the study patients (99.6%) had an SCCHN-related ambulatory visit, about half (48.9%) had an SCCHN-related inpatient stay, and 6.7% had an SCCHN-related ER visit (Figure 4). The mean number of SCCHN-related inpatient days was 0.32 per-patient per-year for the overall study population, and the mean numbers of ER and ambulatory visits were 0.05 and 11.9, respectively. Among all cohorts, the incidences of all-cause inpatient stays and ambulatory visits were similar to the incidences of SCCHN-specific inpatient stays and ambulatory visits, but the percentages of patients with SCCHN-specific ER visits were much lower than the percentages who had all-cause ER visits, with fewer than 10% of each cohort having an SCCHN-attributable ER visit but about 40–50% having all-cause ER visits.

Incidence rates for all-cause inpatient stays, ER visits, and ambulatory visits were 0.5017, 0.2875, and 57.4918, respectively (detailed data not shown). The incidence rate for inpatient stays was at least twice as high among patients with pharyngeal cancer (1.048) as compared to the other three cohorts. The pharyngeal cancer cohort also had the highest incidence rate for ER visits, but rates were similar across cohorts. The incidence rate of ambulatory visits was much higher for laryngeal cancer (91.96) than for the other cohorts (range, 41.15–56.69).

Post-index healthcare costs

The mean total healthcare cost was $34,450 per-patient per-year (Table 5, Figure 3), almost 175% higher than pre-resection per-patient-per-year cost ($12,572). The bulk of the cost was for medical expenses ($32,401), and the remaining $2050 was for pharmacy costs. Inpatient stays accounted for almost $10,000 of the medical expenses, while ambulatory visits cost $21,195. After annualizing cost, the highest mean healthcare cost was observed in the cohort of patients with pharyngeal cancer, for which mean total cost was $40,214 per-patient per-year, double that observed per-patient-per-year during the pre-resection period ($20,071).

Table 5.  Post-index all-cause and SCCHN-related healthcare costs* (per-patient per-year, US$).

	Total (n = 1,104)		Oral cancer (n = 229)		Pharyngeal cancer (n = 185)		Lip/tongue cancer (n = 367)		Laryngeal cancer (n = 323)
	Mean per-patient per-year (SD)	Mean per patient follow- up^† (SD)	Mean per-patient per-year (SD)	Mean per-patient, entire follow-up^† (SD)	Mean per-patient per-year (SD)	Mean per-patient, entire follow-up^† (SD)	Mean per-patient per-year (SD)	Mean per-patient, entire follow-up^† (SD)	Mean per-patient per-year (SD)	Mean per-patient, entire follow-up^† (SD)
All-cause healthcare costs
Total healthcare cost	34,450 (44,720)	85,570 (105,130)	37,314 (51,470)	91,442 (122,437)	40,214 (34,307)	100,271 (78,517)	32,867 (46,800)	80,928 (111,735)	30,918 (42,099)	78,262 (96,456)
Total medical cost	32,401 (43,823)	79,909 (102,327)	35,351 (50,309)	86,008 (119,221)	38,147 (33,450)	94,843 (76,396)	30,767 (45,628)	75,169 (108,039)	28,874 (41,711)	72,417 (94,680)
Cost of inpatient stays	9,924 (23,858)	24,188 (55,287)	12,131 (23,607)	29,728 (61,414)	9,480 (14,492)	23,667 (35,349)	8,117 (16,911)	19,376 (37,610)	10,666 (33,200)	26,025 (73,663)
Cost of ER visits	150 (439)	384 (1,029)	114 (235)	312 (616)	166 (352)	418 (859)	137 (367)	339 (916)	181 (628)	465 (1,402)
Cost of ambulatory visits	21,195 (29,672)	52,439 (69,880)	21,627 (34,749)	52,089 (78,834)	27,425 (25,280)	68,096 (57,411)	21,403 (35,808)	52,634 (86,574)	17,084 (17,284)	43,498 (41,264)
Cost of other services	1,132 (3,042)	2,899 (8,092)	1,478 (4,560)	3,879 (13,227)	1,077 (2,156)	2,662 (4,969)	1,110 (3,027)	2,820 (7,483)	943 (1,944)	2,429 (4,662)
Total pharmacy cost	2,050 (3,303)	5,661 (9,544)	1,964 (3,085)	5,434 (8,489)	2,067 (2,857)	5,428 (7,483)	2,101 (4,109)	5,759 (12,278)	2,043 (2,601)	5,845 (8,058)
SCCHN-related healthcare costs
Total healthcare cost	19,699 (32,127)	46,759 (68,907)	19,229 (32,535)	45,082 (74,318)	24,050 (27,000)	57,509 (57,294)	18,748 (31,610)	43,637 (63,458)	18,621 (34,937)	45,338 (76,286)
Total medical cost	19,699 (32,127)	46,759 (68,907)	19,229 (32,535)	45,082 (74,318)	24,050 (27,000)	57,509 (57,294)	18,748 (31,610)	43,637 (63,458)	18,621 (34,937)	45,338 (76,286)
Cost of inpatient stays	7,655 (21,386)	18,061 (47,166)	9,450 (18,414)	22,113 (43,758)	7,570 (12,577)	18,536 (29,295)	6,434 (15,539)	14,833 (32,040)	7,820 (30,947)	18,584 (67,702)
Cost of ER visits	20 (160)	48 (370)	3 (25)	8 (55)	11 (44)	30 (112)	28 (200)	71 (506)	27 (200)	60 (410)
Cost of ambulatory visits	11,678 (19,185)	27,819 (41,008)	9,442 (20,569)	22,197 (46,379)	16,103 (20,450)	38,101 (42,593)	11,971 (21,766)	27,959 (44,262)	10,397 (12,931)	25,756 (29,771)
Cost of other services	346 (1,209)	831 (2,697)	334 (1,146)	764 (2,303)	365 (1,146)	842 (2,595)	316 (1,079)	774 (2,588)	377 (1,415)	938 (3,112)
Total chemotherapy cost	975 (4,714)	2,222 (10,090)	730 (3,593)	1,715 (8,793)	2,004 (5,804)	4,912 (13,761)	1,177 (6,110)	2,507 (11,912)	329 (2,043)	715 (4,215)
Total radiation cost	8,219 (14,168)	19,511 (30,748)	8,421 (18,766)	18,844 (38,790)	11,833 (13,846)	28,874 (30,304)	7,264 (13,751)	17,029 (29,649)	7,091 (10,192)	17,441 (24,231)

*The costs for subcategories shown in the result tables may not sum to the cost of the parent category due to rounding.

^†Patients were followed for an average of 2.3 years.

Total SCCHN-related health care cost was $19,699 per-patient per-year (Table 5). The mean cost of SCCHN-related inpatient stays was $7655 per-patient per-year, and cost for ambulatory visits was $11,678 per-patient per-year. Chemotherapy cost totaled $975 per-patient per-year, and radiation cost was $8219 per-patient per-year. SCCHN-related healthcare cost was highest among those with pharyngeal cancer (mean, $24,050 per-patient per-year).

Discussion

Our study results show high rates of post-index healthcare utilization and associated costs, especially among patients with pharyngeal cancer. The 1104 patients in this study were followed for approximately 2.3 years following their index surgery, on average, and incurred in excess of $94 million in expenditures following their index surgery ($85,000 per-person, on average). Post-index costs nearly tripled from the pre-index to the post-index period, with patients incurring mean total healthcare costs per-patient per-year of $34,450 during the post-index period versus $12,572 during the pre-index period. Total medical costs also almost tripled, from a mean of $11,517 to a mean of $32,401 per-patient per-year. Total pharmacy costs doubled from a mean of $1054 to a mean of $2050 per-patient per-year. These findings highlight the increasing economic burden of SCCHN with the passage of time from initial onset of the disease.

This study captured mean costs over an average follow up of 2.3 years following resection. In the longer term, costs are likely to be even greater, considering the high recurrence rates typically observed beyond 2 years among SCCHN patients after resection13,21. It has been reported that the highest incidence of local recurrence is in the first 3 years14, so the 2.3-year average follow-up time period does not capture costs related to recurrences that may have developed by the end of 3 years. The development of second primary tumors after 3 years is also common26.

Approximately half of the present study population received post-resection radiotherapy. Between 30% and 60% of the cohorts underwent a second surgery. Ancillary treatment for other conditions such as pain (88.6%), mucositis (67.6%), and nausea (51.1%) were considerable. A substantially lower proportion of patients (16.2%) received chemotherapy. The reasons for the low rate of chemotherapy are unclear. One possibility is that patients received chemotherapy preoperatively in an attempt to minimize the extent of the surgery and improve functional outcomes. A considerable proportion of patients presumably fit into this category. Some patients may have been diagnosed with later-stage disease, when surgical intervention with curative intent was not likely and palliative chemotherapy would have been the likely treatment. Others may have been enrolled in clinical trials that offered alternative treatments in lieu of chemotherapy. Many patients may have decided against chemotherapy after balancing quality-of-life issues against their survival prospects. However, the reasons are only speculative, since stage of diagnosis was not captured for this study population. The fact that patients who did not receive any chemotherapy were included in the analyses would explain the low cost of chemotherapy observed in the present study (mean of $975 per-patient per-year), i.e. the 84% who had zero costs were factored into the calculation of the means.

Comparison of these results to those reported in previously published studies assessing healthcare utilization and costs among SCCHN patients is largely inappropriate because of several major differences that preclude such comparisons, such as large differences in patient sample size and composition and in study design and methodology. For example, some studies have limited their analyses to costs or utilization associated with one type of SCCHN such as oropharyngeal or laryngeal cancer26,27, or were designed to compare two treatment modalities26,28,29. The present analysis differs from some previous studies in terms of study design (e.g., claims-based analysis versus a decision analytic model used to project outcomes and cost30). Studies conducted in other countries also did not seem appropriate for comparison31,32.

The study design and methodology most similar to that used in the present study is that reported by Lang et al.33, who conducted a retrospective cohort analysis using linked data from the SEER Program of the National Cancer Institute and Medicare claims from 1991–1993. They observed higher healthcare resource utilization rates in their cohort of SCCHN patients than in a group of matched controls without SCCHN, including hospitalization (82 vs. 55% for controls), home healthcare (48 vs. 26%), and outpatient visits (92 vs. 86%). The average per-patient cost was twice that of controls ($48,800 vs. $23,300), mostly due to higher hospitalization costs. Although these investigators compared a cohort of patients with SCCHN to a matched cohort of non-SCCHN Medicare beneficiaries (as opposed to comparing SCCHN cohorts by disease subtype as was done here), comparisons of outcomes can nevertheless be made with regard to several outcomes. A similar percentage of their SCCHN cohort (n = 4536) was male (66.1%, compared with 71.9% of the patient sample in the present study). As was observed with the present study population, their SCCHN cohort had a low rate of chemotherapy use, regardless of whether patients had been diagnosed with early-stage or advanced disease, but the authors observed high rates of surgical and radiation therapy. Approximately 82% of their study population was hospitalized during the 5-year follow-up period, with an average of 2.5 hospitalizations per patient, whereas only 8.4% of the present total SCCHN cohort had an inpatient stay. The mean (SD) number of inpatient days in their study was 24 ± 29 per patient over the entire (5-year) follow-up, which was much higher than the mean of 0.45 ± 0.61 inpatient days observed per-patient per-year for the total cohort (n = 1104) in the present study, suggesting that, overall, the patients studied by Lang et al. had more advanced disease and greater comorbidity. The mean total healthcare cost reported by Lang et al. was $48,847 ± 47,999 – very similar to the post-index mean all-cause total healthcare cost of $44,720 observed for the entire SCCHN cohort during the post-index period in the present study. However, their mean total cost was based on the 5-year follow-up period and included both pre- and post-surgical costs, whereas the $44,720 figure is based on variable follow-up times and includes only costs incurred after the patients’ first surgery.

An important difference between the present study and the analysis conducted by Lang et al. is that they evaluated patients identified during the period 1991–1993, when combined modality therapy had not come into play. If chemotherapy was administered to their patients, it would have been mostly infusional 5-FU or intermittent cisplatin during the 6-week course of radiotherapy. In the present study, treatments received during the pre-index period were not captured, but the outcomes reported here provide a more current assessment of real-world treatment patterns.

Study limitations

Study results should be interpreted with certain limitations in mind, some of which pertain to claims-based analyses in general. First, there are limits to the degree to which claims data can accurately capture an individual's medical history. Claims data are collected for the purpose of payment and not research, and are subject to possible coding errors, coding for the purpose of rule-out rather than actual disease, and undercoding. However, administrative data allow for the examination of healthcare utilization and expenditure patterns in a ‘real world’ setting, away from the highly controlled environment of clinical trials, and offer the advantage of large sample sizes with diverse medical histories.

Limitations specific to this study must also be considered. Results of this analysis are primarily applicable to resected SCCHN in a managed-care setting, and therefore may not be generalizable to the entire US population. The plans used for analysis, however, are discounted fee-for-service, IPA-network plans – not capitated or gatekeeper models – that encompass a wide US geographic distribution and therefore provide the capability for generalization to managed-care populations on a national level.

Another limitation is the fact that only hospital- or skilled nursing facility-reported patient deaths were captured (but not deaths that occurred at home); therefore, the number of patients who died during the study period may be greatly under-represented. Additionally, preoperative chemotherapy was not captured in this analysis. Finally, as mentioned previously, disease stage is an important factor impacting outcomes, but these analyses did not stratify patients according to disease stage.

Conclusions

While early-stage head and neck cancer is potentially curable with surgery and radiotherapy, long-term disease-free and overall survival remains a major concern for those diagnosed with more advanced stages of disease. The post-index treatment and healthcare utilization patterns and expenditures observed in this study confirm that despite the available treatment options of surgery, radiotherapy, and often chemotherapy, as well as molecularly targeted therapies, substantial healthcare costs are incurred by SCCHN patients. Effective treatment modalities for recurrent SCCHN following platinum-based therapy are limited, and as the existing literature and the present study suggest, there is a high unmet need for the development of efficacious novel treatments and for their integration into treatment paradigms. If such treatments were to increase cure rates, reduce toxicity, and bring about substantial improvements in survival, the increased costs might be outweighed by the clinical benefits. In terms of future research, studies that expand upon this one by stratifying results according to disease stage, and also according to age group and gender, would be a useful addition to the literature.

Transparency

Declaration of funding

This study was funded by GlaxoSmithKline.

Declaration of financial/other relationships

M.M.A. is employed by GlaxoSmithKline, the funder of this study. B.C. and N.S. are employed by Innovus, the company hired by GlaxoSmithKline to conduct this study, and A.T. is employed by i3 Research.

Acknowledgments

The authors thank Victoria Porter, medical writer, i3 Innovus, for her assistance with the preparation of this manuscript.

A poster based on this study, titled ‘Economic burden of resected squamous cell carcinoma of the head and neck (SCCHN) in a US managed-care population,’ was presented at the 2010 ASCO annual meeting, Chicago, Illinois, June 4–8, 2010.