www.elsevier.com/locate/jhep

Journal of Hepatology 50 (2009) 89–99

The burden of illness associated with hepatocellular carcinomain the United Statesq

Kathleen Lang1,#, Natalya Danchenko1, Kathleen Gondek2, Sonalee Shah2,David Thompson1,*

1i3 Innovus, 10 Cabot Road, Suite 304, Medford, MA, USA2Bayer Healthcare Pharmaceuticals, Montville, NJ, USA

See Editorial, pages 10–12

Background/Aims: Despite the rising prevalence of hepatocellular carcinoma (HCC), data on its economic consequences

are limited. This study was undertaken to estimate the aggregate annual financial burden associated with HCC in the Uni-

ted States, including healthcare costs and the value of lost productivity.Methods: Annual prevalence of HCC and incidence and survival were estimated using SEER data. The linked SEER-

Medicare database was used to estimate distributions of healthcare utilization, quantities of treatment, and unit costs

among 392 HCC patients. An age- and sex-matched cohort of non-cancer controls was used to estimate background

non-cancer-related resource use and costs.

Results: We determined the annual cost of HCC in the United States to be $454.9 million, with per-patient costs of

$32,907. Healthcare costs and lost productivity accounted for 89.2% and 10.8% of total cost, respectively. Costs associ-

ated with localized HCC accounted for the highest portion (44.5%) of the total cost of illness, at $202.5 million. Regional,

distant, and unstaged HCC accounted for 31.0%, 13.9%, and 10.6%, respectively.Conclusions: Our results exhibit a considerable economic impact of HCC and substantial national spending on this

disease.

� 2008 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Keywords: Hepatocellular carcinoma (HCC); Cost of illness; Burden of illness; Prevalence; Outcomes; Treatment pattern

0168-8278/$34.00 � 2008 European Association for the Study of the Liver.

doi:10.1016/j.jhep.2008.07.029

Received 25 February 2008; received in revised form 15 July 2008;

accepted 19 July 2008; available online 1 October 2008

Associate Editor: J.M. Llovetq This study has been supported by an educational grant from Bayer

Healthcare Pharmaceuticals, Montville, NJ to obtain access to thedatabase (SEER-Medicare) and published articles, and to cover otherexpenses associated with the project. KG and SS are employed byBayer Healthcare Pharmaceuticals, Montville, NJ. KL, ND and DTare employed by i3 Innovus, Medford, MA.

* Corresponding author. Tel.: +781 338 9700x222; fax: +781 338 9522.E-mail address: dthompson@i3innovus.com (D. Thompson).

# Presently at Boston Health Economics, Waltham, MA, USA.Abbreviations: HCC, hepatocellular carcinoma; SEER, surveillance

epidemiology and end results; ICD, international classification of di-seases; NCI, National Cancer Institute; TACE, transarterial chemo-embolization; HCPS, Health Care Financing Administration’sCommon Procedure Coding System.

1. Introduction

Hepatocellular carcinoma (HCC) is the predominanthistologic subtype of liver cancer in adults, comprisingapproximately 65% of all cases of primary liver cancer[1]. Worldwide, it is the sixth most common neoplasm,with more than half a million new cases diagnosed eachyear [2]. HCC is the third most common cause of can-cer mortality [3] and the main cause of death amongcirrhotic patients [2]. The overall incidence rate ofHCC has been estimated at 2.99 per 100,000 [1], withaccruing evidence over the past 5–8 years, indicatingthat the incidence is rising in several countries [4]. In

Published by Elsevier B.V. All rights reserved.

90 K. Lang et al. / Journal of Hepatology 50 (2009) 89–99

2000, approximately 564,000 HCC cases were reportedglobally, with 11,500 cases in the United States and50,000 in Europe [5].

Although HCC has predominantly affected personsin the developing world (where >80% of liver cancercases occur) [6,7], Western countries have seen risingnumbers of cases over the past two decades – an increaselargely attributed to the high prevalence of hepatitis Band C virus (HBV and HCV) infections [1,8]. Chronicinfection with hepatitis C virus is the predominant riskfactor for HCC in Western countries and Japan, andhepatitis B virus is the predominant risk factor forHCC in Southeast Asia and Africa [9]. Incidence andmortality rates associated with HCC have doubled inthe United States over the past 25 years, and given thecurrent prevalence of HCV among persons aged 30–50years, the US incidence and mortality rates of HCCare expected to double over the next 10–20 years [10].

Historically, treatment of HCC has been problematic,primarily due to tumor identification at an advancedstage—although the more recent trend has been encour-aging, as HCC is being diagnosed at an earlier stage (i.e.,when liver function is preserved and cancer-relatedsymptoms are absent) in many patients [4]. Anotherobstacle to effective treatment is that most patients withHCC also have cirrhosis, which limits the use of surgicalresection [11]. Although liver transplantation hasachieved increasing success in the treatment of HCC,many clinicians consider the high risk of tumor recur-rence and hepatitis infection to be a contraindicationto transplantation—and moreover its cost-effectivenesshas been questioned when it is performed in patientswho do not meet the appropriate selection criteria [11].Nevertheless, due to advances over the past two decades– such as the use of transarterial catheter chemoemboli-zation (TACE) and radiofrequency ablation, transplan-tation is now recognized as one of the few curativetreatment modalities for HCC patients [12]. Encourag-ingly, significant survival benefit has been shown witharterial chemoembolization in a subset of patients withunresectable HCC [13]. The treatment response to sys-temic chemotherapy among HCC patients has been esti-mated to be lower than 10%, and no clear impact onoverall survival has been observed [11]. Only recently,angiogenesis inhibitors such as sorafenib, which wasrecently approved by the US Food and Drug Adminis-tration for the treatment of unresectable HCC, havebeen shown to increase overall survival in advanced-stage HCC patients [14,15].

Given the rising prevalence of HCC and the limitedtreatment options for patients with advanced disease,it is important to know the financial burden associatedwith HCC in the United States. However, data on itseconomic consequences are limited. Most of the litera-ture on HCC-related costs pertains to cost-effectivenessof screening for the disease [16–20] or cost-analyses of

specific treatments [21–24]. Other studies have assessedthe economic burden of several types of cancer [25–27], but these studies either did not include data onHCC or addressed inpatient care only or did not includedata on lost productivity. Our study was undertaken toestimate the aggregate annual financial burden associ-ated with HCC in the United States from a societal per-spective, including healthcare costs and the value of lostproductivity (i.e., lost workdays due to cancer).

2. Materials and methods

2.1. General model structure

Key relationships presented in the model include annual numbersof patients treated for HCC by age group and cancer stage; utilizationof various cancer-specific treatments among these patients; unit costs ofthese treatments; workdays missed; and wage rates.

The general modeling approach is illustrated in Figs. 1 and 2.The model begins with all HCC patients alive in a given year (i.e.,the prevalence of HCC), stratified by factors expected to influenceresource utilization and burden of illness, including age (<65,65+), gender, and cancer stage as categorized in the SEER database(i.e., localized, regional, distant [metastasized], and unstaged). Local-ized cancer is defined as disease that is limited to the organ in whichit originated; regional cancer is defined as having spread beyond theprimary site to nearby lymph nodes, organs, and tissues; and distantcancer is defined as having spread beyond the primary site to distantorgans and lymph nodes (i.e., metastatic) [28]. Patients in each age-,gender- and stage-specific stratum are assumed to use cancer-relatedhealthcare resources, which translate to healthcare costs attributableto cancer. They also are assumed to have missed workdays (e.g.,absenteeism and unemployment), which translate to lost productivityattributable to cancer.

These figures constitute simplified versions of the structuralequations estimated in the model. Multiplying stratum- (i.e., age-, gen-der-, and cancer stage-) specific distributions of treatment by annualquantities of HCC treatments and unit costs yields stratum-specificestimates of annual healthcare costs associated with HCC. Similarly,multiplying stratum-specific estimates of annual workdays missed byaverage wage rates yields stratum-specific estimates of lost productivityassociated with HCC. Summing these estimates across strata yields anestimate of the total annual burden attributable to HCC.

2.2. Data sources

Data sources used to estimate the model include the SurveillanceEpidemiology and End Results (SEER)-Stat data file, the linkedSEER-Medicare database, the US Bureau of Labor Statistics, pub-lished literature, and other secondary sources.

The SEER Program of the National Cancer Institute (NCI) [29] isan epidemiologic surveillance system designed to track cancer inci-dence and survival in the United States. It comprises population-basedtumor registries that collect demographic and diagnostic informationon cancer patients in geographically defined metropolitan areas, cur-rently representative of roughly 25% of the US population [30]. TheSEER-Stat database contains summary data on cancer incidence andsurvival from population-based cancer registries belonging to theSEER Program, The linked SEER-Medicare database is a collabora-tive effort of the NCI, the SEER cancer registries, and the Centersfor Medicare and Medicaid Services. The Medicare administrativeclaims files include detailed information on the use of inpatient, outpa-tient, home health, hospice, and skilled nursing facility (SNF) services,including dates and type of service, diagnosis codes, Medicare paymentamounts, primary insurer payments, and patient deductibles andcopayments. Age- and sex-specific salary data from the Bureau ofLabor Statistics were used to estimate wage rates.

Cancer patient strata

X

Annual numbers of work-days missed by cancer patients

X

Daily wage rates

=Annual value

of lost productivity

Fig. 2. Estimation of annual value of lost productivity among HCC patients. [This figure appears in colour on the web.]

Cancer patients

XDistribution of cancer treatments

X

Annual quanitites of

cancer treatments

XUnit costs of cancer treatments

=

Annual cost of cancer

treatments

Fig. 1. Estimation of annual healthcare costs among HCC patients. [This figure appears in colour on the web.]

Table 1

Prevalence of HCC in the US, by cancer stage, age, sex, and overall,

2005

Sex/Age Cancer stage

K. Lang et al. / Journal of Hepatology 50 (2009) 89–99 91

We used 1999 as our year of analysis, since this was the latest yearfor which new cancer cases could be identified in the SEER-Medicaredata at the time we conducted our study, and we inflated all costs toUS$2006.

groupLocalized Regional Distant Unstaged All patients

Men

<25 29 0 24 0 5425–34 50 40 16 8 11435–44 186 97 65 40 38845–54 808 978 428 331 254655–64 1043 808 428 251 253065–74 962 695 356 299 231275–84 606 598 226 226 165785+ 105 73 81 65 323

Total men 3789 3290 1625 1220 9923

Women

<25 0 0 8 0 825–34 0 0 0 0 035–44 99 24 32 0 15645–54 294 97 105 40 53755–64 380 234 105 81 80065–74 614 251 186 178 122975–84 348 218 162 137 86585+ 89 113 40 65 307

Total women 1824 938 639 501 3901

Total menand women

5612 4227 2263 1722 13,824

3. Model estimation

3.1. HCC prevalence

Using SEER-Stat software, the annual prevalence ofHCC in each stratum was calculated as the product ofincidence and median survival. HCC cases were selectedusing the cancer site variable (coded as ‘‘liver”) andInternational Classification of Diseases, Ninth Revision

(ICD-9-CM) code for hepatocellular carcinoma (ICD-0-3 histology code 8170). Median survival was estimatedas the time point at which 50% of patients in a particularstratum were alive. For all strata in which median sur-vival was less than 1 year, prevalence was assumed tobe equal to incidence in order to ensure that all patientswho were alive at least 1 day during the study year werecaptured in the analysis. Annual age-, sex-, and stage-specific HCC prevalence estimates are shown in Table 1.

3.2. Healthcare utilization and unit costs

The linked SEER-Medicare database was the pri-mary source used to estimate distributions of healthcareutilization, quantities of treatment, and unit costsamong HCC patients.

3.3. Sample selection for SEER-Medicare analyses

3.3.1. HCC cohort

All patients diagnosed with HCC as a first cancer in aSEER registry between 1991 and 1999 and who were stillalive on January 1, 1999 were selected for inclusion in theHCC cohort. Patients were excluded if they were youngerthan 65 during the study year, had incomplete data, orwere enrolled in an HMO or not eligible for Part A or PartB Medicare benefits in any study month. In total, 392HCC patients at various stages of disease were identified.

3.3.2. Control cohort

An age- and sex-matched cohort of non-cancer con-trols alive on January 1, 1999 was used to estimate back-ground non-cancer-related resource use and costs. Thiscohort was selected from Medicare enrollment filesusing a 5% sample of Medicare beneficiaries residingin SEER areas who were not reported as having anycancer during the study year or in any prior year; theywere subject to the same exclusion criteria as the HCCcohort. One control patient of identical age and sexwas matched to each HCC patient; if more than 1 matchwas available, the control was selected randomly. Con-trol patients were not required to have used services inorder to be selected for inclusion and, as with HCCpatients, could continue in the study even if they devel-oped other cancers after study entry.

92 K. Lang et al. / Journal of Hepatology 50 (2009) 89–99

3.4. Estimation of healthcare utilization and unit costs

Patients in both cohorts were followed for the entirestudy year to track resource use and costs, with thesemeasures set to zero in months following death, if appli-cable. Differences in resource use and costs betweenHCC cases and matched non-cancer controls were con-sidered attributable to HCC.

Due to a lack of detailed healthcare utilization andcost data for patients younger than 65, we assumedresource use and costs for this age group to be equiv-alent to those observed for the youngest patients in theSEER-Medicare database (i.e., those aged 65–69years). This assumption was based on literature indi-cating that there is greater healthcare resource utiliza-tion among younger cancer patients than among theirolder counterparts [25,31], and was subjected to sensi-tivity analysis.

Table 2 displays estimates for the components ofhealthcare utilization covered in the model. SpecificHCC procedures and treatments were identified by thepresence of relevant ICD-9-CM diagnosis codes, ICD-9-CM procedure codes, and the Health Care FinancingAdministration’s Common Procedure Coding System(HCPCS) codes in the claims files (see Appendix). Eachcomponent of additional cancer-related resource utiliza-tion was estimated as the difference between HCCpatients and matched non-cancer controls (after sub-tracting out the resource use attributable to the majorprocedures noted above). Oral prescription drug use,which is not covered by Medicare and therefore notavailable in the SEER-Medicare database, was esti-mated using data from the published literature [32–35].We did not observe any liver transplants in the SEER-Medicare data, likely because these transplants aregenerally performed on younger patients [36–38]. Wetherefore used an estimate from a report by EastonAssociates, LLC, which indicated a transplant rate of4% among patients with localized HCC [39]. We appliedthis rate only to younger patients (665 years), assumingno transplants were performed among patients aged 65years and older.

It was assumed that 100% of patients at any cancerstage would receive oral non-opioid pain medication,whereas oral opioids were assumed to be utilized onlyby patients with metastatic HCC [32]. Since antiemetictherapy is administered on the day of chemotherapy orradiotherapy [33], the percentage of patients receivingthese medications was estimated as the maximum ofthe percentage of patients receiving chemotherapy andthe percentage of patients receiving radiation therapyfor the given cancer stage and age group. The percentageof patients receiving antidepressants was estimated onthe basis of a published finding that 16% of advancedcancer patients were prescribed antidepressants [34].We therefore assumed that 16% of regional, distant,

and unstaged HCC patients were treated with antide-pressants, and we assumed a lower percentage (10%)for patients with localized cancer. Nutritional supple-ment use was estimated on the basis of a study thatreported utilization rates of 66.2%, 55.9%, and 53.3%,respectively, among patients with localized, regional,and distant cancer [35]. Patients with unstaged cancerwere assumed to be similar to the median with respectto use of nutritional supplements (55.9%).

Unit costs of major cancer-related procedures andadditional cancer-related resource utilization were esti-mated for each stratum using SEER-Medicare data(Table 3). Estimates of unit costs of liver transplantationwere based on results from the literature [40]. Unit costsof oral prescription medications (i.e., costs per therapyday) were estimated on the basis of daily dosages indi-cated for the drugs of interest and dosages specified inthe Red Book [41], with costs inflated to US$2006.

3.5. Workdays missed and average wage rates

We derived our estimate of the annual numbers ofworkdays lost from the published finding by Yabroffand colleagues that patients with short-survival cancerslost an average of 94 days over a 1-year period [42]. Thisestimate was adjusted for age- and stage-specific follow-up during the year (i.e., fraction of the year alive). Fol-low-up length was estimated for each HCC age andstage stratum using SEER-Medicare data, with follow-up among those younger than 65 assumed to be thesame as for the group aged 65–69 years.

Due to limited data availability for specific age, sexand stage strata, the following assumptions were made:(1) patients aged 65 years and older were assumed tohave a work schedule reduced by approximately 25%,and therefore were assumed to miss fewer workdaysthan patients younger than 65; and (2) patients aged75 years and older were assumed to be retired and thusto incur no lost workdays.

3.6. Average wage rates

It was assumed that, in the absence of disease, per-sons with HCC would earn the same hourly wage asthe average wage earner of similar age and sex. Wagerates were estimated according to age- and sex-specificsalary data from the US Bureau of Labor Statistics [43].

4. Results

We estimated the aggregate annual cost of HCC inthe United States (US$2006) to be $454.9 million, withper-patient costs equal to $32,907 (Table 4). Healthcarecosts and lost productivity account for 89.2% ($405.8million) and 10.8% ($49.1 million) of the total cost,

Table 2

Annual percentages of HCC patients receiving selected components of healthcare utilization, and annual quantities of resource utilization among HCC

patients, by age group and cancer stage

Component of healthcareutilization

Percentage of patients and annual quantity by age group and stage

<65 years 65+ years

Localized Regional Distant Unstaged Localized Regional Distant Unstaged

% (mean number of occurrences)Major inpatient procedures

Major cancer-relatedsurgical procedure

12.5% (1.0) 6.5% (1.0) 1.2% (1.0) 0.4% (1.0) 10.5% (1.1) 0.9% (1.0) 3.6% (1.0) 1.1% (1.0)

Transplant 4.0% (1.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0)Percutaneous injection

therapy3.3% (1.0) 4.1% (1.3) 2.7% (1.0) 1.6% (1.0) 2.2% (1.0) 2.8% (1.7) 1.8% (1.0) 1.1% (1.0)

Radiofrequency ablation 0.9% (1.0) 1.1% (1.0) 4.2% (1.0) 1.3% (1.0) 0.8% (1.0) 0.9% (1.0) 3.6% (1.0) 1.1% (1.0)Chemotherapy 7.6% (1.0) 15.1% (1.2) 5.5% (1.0) 6.5% (1.4) 7.5% (1.3) 14.8% (1.6) 5.4% (1.3) 6.4% (1.8)Arterial embolization 7.6% (1.2) 15.0% (1.5) 9.1% (1.3) 10.8% (1.1) 7.5% (1.1) 14.8% (1.4) 8.9% (1.2) 10.6% (1.0)

Supporting physician services

Surgical resection post-opsupport

3.6% (1.0) 1.0% (1.0) 0.0% (0.0) 0.0% (0.0) 6.7% (1.2) 1.9% (1.0) 0.0% (0.0) 0.0% (0.0)

Transcatheter arterialembolization support

8.5% (1.3) 11.7% (1.6) 4.1% (1.0) 3.6% (1.4) 7.5% (1.6) 10.2% (1.9) 3.6% (1.0) 3.2% (1.7)

Percutaneous injectiontherapy

0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.9% (2.0) 0.0% (0.0) 2.1% (1.5)

Chemotherapy 13.6% (30.4) 16.9% (17.3) 8.7% (30.8) 12.9% (56.0) 11.2% (19.2) 13.9% (10.9) 7.1% (19.5) 10.6% (35.4)Cisplatin 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 6.2% (9.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 1.1% (9.0)Paclitaxel 4.4% (21.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.8% (21.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0)

Radiation therapy 2.2% (4.3) 2.8% (2.2) 10.6% (7.7) 0.0% (0.0) 2.2% (4.7) 2.8% (2.3) 10.7% (8.3) 0.0% (0.0)Megace 0.9% (2.7) 2.2% (56.2) 0.0% (0.0) 2.5% (42.5) 0.8% (1.0) 1.9% (20.5) 0.0% (0.0) 2.1% (15.5)Zometa 3.6% (11.2) 5.5% (5.4) 4.2% (2.2) 11.4% (13.7) 3.0% (5.0) 4.6% (2.4) 3.6% (1.0) 9.6% (6.1)Morphine 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.9% (1.0) 0.0% (0.0) 0.0% (0.0)

Supporting outpatient hospital services

Minor cancer-relatedsurgical procedure

1.5% (1.0) 1.8% (1.0) 3.5% (1.0) 0.0% (0.0) 0.8% (1.0) 0.9% (1.0) 1.8% (1.0) 0.0% (0.0)

Transcatheter arterialembolization support

0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.8% (1.0) 0.0% (0.0) 0.0% (0.0) 3.2% (0.0)

Percutaneous injectiontherapy

1.5% (1.0) 1.8% (1.0) 0.0% (0.0) 6.3% (1.0) 0.8% (1.0) 0.9% (1.0) 0.0% (0.0) 3.2% (1.0)

Radiofrequency ablationsupport

0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.9% (1.0) 1.8% (52.0) 3.2% (0.0)

Chemotherapy 8.6% (3.2) 7.6% (1.7) 8.9% (1.1) 5.3% (3.0) 5.2% (3.9) 4.6% (2.0) 5.4% (1.3) 3.2% (3.7)Cisplatin 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.9% (1.0) 1.8% (1.0) 0.0% (0.0)

Biotherapy/ImmunotherapyInterferon-alpha 2.2% (1.3) 0.0% (0.0) 0.0% (0.0) 3.2% (2.7) 0.8% (1.0) 0.0% (0.0) 0.0% (0.0) 1.1% (2.0)

Arterial embolizationsupport

0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 0.8% (1.0) 1.9% (1.0) 0.0% (0.0) 2.1% (1.0)

Radiation therapy 0.6% (1.0) 1.6% (1.0) 6.1% (1.6) 0.0% (0.0) 0.8% (1.0) 1.9% (1.5) 7.1% (15.0) 0.0% (0.0)Megace 0.0% (0.0) 0.0% (0.0) 3.5% (1.0) 4.2% (2.6) 0.0% (0.0) 0.0% (0.0) 1.8% (1.0) 2.1% (3.5)Fentanyl 3.0% (1.0) 11.0% (1.0) 7.1% (1.0) 4.2% (1.0) 0.8% (1.0) 2.8% (1.0) 1.8% (1.0) 1.1% (1.0)

% (mean number of days/visits)Additional cancer-related resource utilization

Inpatient hospitalization 63.1% (6.1) 75.4% (7.7) 75.5% (8.2) 63.0% (7.7) 59.7% (4.7) 71.3% (5.9) 71.4% (6.3) 59.6% (5.9)Skilled nursing facility 11.0% (1.1) 8.2% (0.3) 10.6% (2.6) 12.6% (2.2) 14.9% (1.7) 11.1% (0.5) 14.3% (3.9) 17.0% (3.4)Hospice care 22.7% (8.4) 27.6% (4.7) 29.5% (7.4) 32.3% (9.9) 34.3% (13.7) 41.7% (7.7) 44.6% (12.0) 48.9% (16.2)Home health care 14.7% (4.1) 18.3% (0.2) 25.7% (4.7) 25.0% (4.0) 19.4% (5.5) 24.1% (0.3) 33.9% (6.4) 33.0% (5.4)Physician services 100.0% (17.8) 98.1% (24.1) 100.0% (22.6) 100.0% (21.2) 96.3% (12.2) 94.4% (16.4) 96.4% (15.4) 97.9% (14.5)Outpatient hospital services 77.5% (3.1) 81.3% (3.7) 79.4% (2.3) 89.8% (3.1) 67.9% (1.6) 71.3% (1.8) 69.6% (1.1) 78.7% (1.6)

% (mean number of therapy days)Oral prescription medication

Pain medication (non-opioid)

100.0% (156.8) 100.0% (122.5) 100.0% (76.0) 100.0% (123.9) 100.0% (156.8) 100.0% (122.5) 100.0% (76.0) 100.0% (123.9)

Opioids 0.0% (0.0) 0.0% (0.0) 90.0% (106.0) 0.0% (0.0) 0.0% (0.0) 0.0% (0.0) 90.0% (106.0) 0.0% (0.0)Antinausea medication 20.4% (8.2) 28.7% (7.0) 13.3% (12.3) 18.5% (9.9) 16.4% (7.2) 23.2% (6.6) 10.7% (16.3) 14.9% (6.6)Antidepressants 10.0% (156.8) 16.0% (122.5) 16.0% (76.0) 16.0% (123.9) 10.0% (156.8) 16.0% (122.5) 16.0% (76.0) 16.0% (123.9)Nutritional

supplementation66.2% (156.8) 55.9% (122.5) 53.3% (76.0) 55.9% (123.9) 66.2% (156.8) 55.9% (122.5) 53.3% (76.0) 55.9% (123.9)

K. Lang et al. / Journal of Hepatology 50 (2009) 89–99 93

Table 3

Unit costs of treatments among HCC patients, by age group and cancer stage

Unit cost by age group and stage

<65 years 65+ years

Localized regional distant unstaged Localized regional distant unstaged

Unit cost per occurrenceMajor inpatient procedures

Major cancer-related surgical procedure $44,572 $23,922 $33,231 $43,291 $40,657 $21,820 $30,312 $39,488Transplant $247,679 $0 $0 $0 $0 $0 $0 $0Percutaneous injection therapy $10,433 $6823 $5375 $6403 $14,046 $4186 $7236 $8621Radiofrequency ablation $13,600 $4986 $10,105 $4306 $27,230 $9982 $20,231 $8621Chemotherapy $6909 $11,100 $8583 $7894 $8353 $13,421 $10,377 $9544Arterial embolization $20,399 $19,012 $27,161 $23,872 $16,859 $15,713 $22,448 $19,731

Supporting physician services

Surgical resection post-op support $2736 $1249 $0 $0 $1749 $799 $0 $0Transcatheter arterial embolization support $1195 $1070 $1412 $1302 $1142 $1022 $1349 $1244Percutaneous injection therapy $0 $0 $0 $0 $0 $106 $0 $59Chemotherapy $53 $47 $31 $73 $55 $48 $32 $76

Cisplatin $0 $0 $0 $289 $0 $0 $0 $289Paclitaxel $752 $0 $0 $0 $752 $0 $0 $0

Radiation therapy $71 $75 $116 $0 $95 $100 $155 $0Megace $0 $2 $0 $5 $0 $1 $0 $4Zometa $36 $38 $30 $26 $34 $36 $28 $24Morphine $0 $0 $0 $0 $0 $12 $0 $0

Supporting outpatient hospital services

Minor cancer-related surgical procedure $8165 $1530 $24,732 $0 $3789 $709 $11,475 $0Transcatheter arterial embolization support $0 $0 $0 $0 $17,469 $0 $0 $0Percutaneous injection therapy $166 $105 $0 $1084 $714 $452 $0 $4649Radiofrequency ablation support $0 $0 $0 $0 $0 $3184 $346 $0Chemotherapy $4868 $5457 $1864 $1081 $3332 $3736 $1275 $739

Cisplatin $0 $0 $0 $0 $0 $1336 $4227 $0Biotherapy/immunotherapy

Interferon-alpha $2968 $0 $0 $2527 $3227 $0 $0 $2748Arterial embolization support $0 $0 $0 $0 $3227 $1302 $0 $1129Radiation therapy $683 $1278 $749 $0 $890 $1667 $977 $0Megace $0 $0 $877 $550 $0 $0 $633 $398Fentanyl $1828 $4089 $14,886 $4162 $1409 $3152 $11,475 $3208

Unit cost per day/visitAdditional resource utilization

Inpatient hospitalization $1734 $2036 $1546 $1699 $1563 $1836 $1395 $1532Skilled nursing facility $581 $805 $324 $439 $655 $907 $365 $495Hospice care $174 $178 $184 $186 $171 $175 $181 $183Home health care $57 $1212 $59 $70 $72 $1524 $75 $87Physician services $173 $159 $169 $166 $165 $152 $161 $159Outpatient hospital services $591 $583 $452 $646 $711 $702 $545 $778

Oral prescription medication

Pain medication (non-opioid) $0.57 $0.57 $0.87 $0.57 $0.57 $0.57 $0.87 $0.57Opioids $0.00 $0.00 $1.34 $0.00 $0.00 $0.00 $1.34 $0.00Anti-nausea medication $68.22 $68.22 $68.22 $68.22 $68.22 $68.22 $68.22 $68.22Antidepressants $2.17 $2.17 $2.17 $2.17 $2.17 $2.17 $2.17 $2.17Nutritional supplementation $0.04 $0.04 $0.04 $0.04 $0.04 $0.04 $0.04 $0.04

94 K. Lang et al. / Journal of Hepatology 50 (2009) 89–99

respectively. Reflecting its higher prevalence, the totalcost associated with localized HCC accounts for thehighest portion (44.5%) of the total cost of illness, at$202.5 million. Regional, distant, and unstaged HCCaccount for 31.0% ($141.0 million), 13.9% ($63.3 mil-lion), and 10.6% ($48.2 million), respectively.

The most expensive inpatient procedures are majorcancer-related surgical procedures, liver transplantation,and arterial embolization, which account for 8.6%

($34.8 million), 7.1% ($28.6 million), and 8.8% ($35.7million) of the total healthcare cost, respectively. A totalof $17.7 million is spent on chemotherapy/antineoplas-tic agents. Inpatient hospitalizations cost $91.9 millionfor patients <65 years (extrapolated from the youngestpatients in the SEER Medicare database, as previouslydescribed in the ‘‘Estimation of Healthcare Utilizationand Unit Costs” section) and $61 million for those aged65 and older. Patients younger than 65 incur $24.9 mil-

Table 4

Annual cost of HCC, by cancer stage

Cost component Annual total cost by stage

Localized Regional Distant Unstaged All patients

Healthcare cost

Total $178,140,973 $126,282,736 $58,497,459 $42,872,328 $405,793,496Cost per patient $31,740 $29,874 $25,848 $24,903 $29,354

Value of lost productivity

Total $24,309,583 $14,670,999 $4,793,956 $5,335,186 $49,109,725Cost per patient $4332 $3470 $2118 $3099 $3553

Overall cost of illness

Total $202,450,556 $140,953,735 $63,291,415 $48,207,514 $454,903,221Cost per patient $36,071 $33,344 $27,967 $28,002 $32,907

K. Lang et al. / Journal of Hepatology 50 (2009) 89–99 95

lion in physician services, and the older age group $15.7million. Oral prescription medications account foralmost $4 million, with patients <65 years spending�$1.9 million. In general, across age strata, youngerpatients incur higher healthcare costs than olderpatients.

Major cancer-related procedures contribute equallyto the healthcare costs of younger and older patients(8.4% and 8.8%, respectively). Chemotherapy accountsfor a higher portion of healthcare cost among olderpatients (aged P65 years) versus younger (6.4% vs3.0%), while arterial embolization contributes a similarportion to healthcare costs of the younger and olderage strata (8.6% vs 9.1%, respectively).

Lost productivity is highest among patients withlocalized HCC, comprising 49.5% of the total value oflost productivity. Patients with regional, distant, andunstaged HCC contribute 29.9%, 9.8%, and 10.9%,respectively, to the total value of lost productivity.Across age and sex strata, peak productivity loss isobserved between the ages of 45 and 74 years, whenearnings are generally highest. Healthcare costs are alsohighest for patients with localized HCC, who incur$178.1 million, compared with $126.3 million for

$49

$380.3

$454.9

0

100

200

300

400

500

600

Base-Case Same health-careutilization for <65,

65+

Healutilizathigher

Ann

ual b

urde

n of

HC

C ($

mill

ions

)

Health-care cost

Fig. 3. Sensitivity of annual burden of HCC to changes in key pa

patients with regional HCC, $58.5 million for patientswith distant HCC, and $42.9 million for patients withunstaged HCC.

4.1. Sensitivity analyses

We conducted several sensitivity analyses to test therobustness of our findings (Fig. 3). First, we tested theimpact of our assumption that resource utilizationamong patients younger than 65 is similar to that forpatients aged 65–69 years. The assumption that resourceutilization and unit costs of procedures and services areequal across age strata resulted in an 18.3% decrease inhealthcare costs and a 16.4% decrease in the overall costof illness. Our assumption that resource utilization is15% higher among patients younger than 65 relative toour base case resulted in a 9.0% increase in healthcarecosts and an 8.0% increase in the overall cost of illness.Increasing wage rates by 20% resulted in a 20% increasein the value of lost productivity, but only a 2.2% impacton the total burden of illness. Finally, we tested theimpact of assuming that older patients with localizedHCC would receive liver transplants at a rate similarto that of younger patients, at a similar unit cost. This

$481.7$464.31.0

th-careion 15% for <65

Wage rates 20%higher

Same livertransplant rates for

<65, 65+

s Lost productivity

rameter values. [This figure appears in colour on the web.]

96 K. Lang et al. / Journal of Hepatology 50 (2009) 89–99

would increase healthcare costs and overall cost of ill-ness by 6.6% and 5.9%, respectively. However, youngerpatients would still incur higher healthcare costs ($243.4million vs $189.3 million), shouldering 56.2% of the totalcosts.

5. Discussion

This study is the first to evaluate the annual economicburden of HCC in the United States, including health-care costs and lost productivity. Our 1-year preva-lence-based methodology, commonly used to studycosts associated with other diseases as pioneered byDorothy Rice in 1966 [44], included newly diagnosedHCC patients as well as patients diagnosed in previousyears who were still alive and using resources at anypoint during the study year. Other prevalence-basedstudies have been conducted, often with less detaileddata, to evaluate the annual burden of other cancers(generally ignoring lost productivity) [25–27]. Our meth-odology was advantageous in that we factored in allcomponents of healthcare resource utilization as wellas data on lost productivity.

We found the overall per-patient cost of HCC to be$32,907. With an average annual HCC prevalence of13,824 cases in the United States, the total annual bur-den of HCC was thus estimated to be $454.9 million.This is at the low end of estimates of annual nationalspending on cancers, which range from $728 millionfor esophageal cancer to $8.3 billion for breast cancer(inflated to $2006) [30]. Adjusting for average follow-up, our findings indicate per-patient monthly costs of$7845. These results are in line with reported meanmonthly cancer-related cost estimates ranging from$2358 for prostate cancer to $9,310 for pancreatic can-cer (inflated to $2006) [27]. Finally, our results areconsistent with other reported estimates of the costof terminal cancer. Krahn reported an annual per-patient cost of terminal care of $65,405 (inflated to$2006) [45], while Taplin estimated a 6-month costof terminal cancer of �$23,412 (updated to $2006and rounded) [46].

In actual clinical practice, HCC is found almostexclusively in cirrhotic HCV patients [11,47]. Therefore,our study accounts for excess resource use and costs dueto cirrhosis and its complications among HCC patients.By calculating additional cancer-related resource utiliza-tion as the difference between HCC patients andmatched non-cancer controls (after subtracting out theresource use attributable to the major procedures notedabove), we captured the excess use and costs attributableto cirrhosis as part of this remaining difference betweencancer patients and controls. This methodology of iden-tifying specific treatments as well as remaining differ-ences between cancer patients and controls has

allowed us to capture all possible resource use and costsattributable to HCC.

It should be noted that our study is subject to severallimitations. First, we relied on administrative Medicareclaims data for patients aged P65 years to assess treat-ment patterns and unit costs for HCC patients of allages. To the extent that utilization and cost profiles dif-fered between patients younger than 65 and those aged65–69, our results may not be accurate. However, otherstudies have documented increased resource use amongyounger patients [26,28], using methods similar to oursto adjust resource use among the elderly to be reflectiveof younger patients [28]. Recognizing that this assump-tion, although based on the literature, may not reflectactual clinical care because older patients may havemore comorbidities, we conducted sensitivity analysesaround this assumption. Since we selected patients aged65–69 years to reflect the entire younger patient popula-tion, our base-case results can be viewed as conservative.

Second, the treatments used in this study may not bean accurate reflection of current treatment patterns forHCC, as some of the medications and procedures mayhave been prescribed for conditions other than HCCthat patients were concomitantly experiencing. Forexample, it could not be determined whether patientswere receiving radiation therapy for other types of can-cer or as an experimental therapy for HCC. Similarly,those patients who were taking Megace (megestrol)and Zometa (zoledronic acid) may have had anothercancer (e.g., breast) in addition to HCC. Furthermore,although the use of nutritional supplements is contrain-dicated in patients with late-stage HCC due to theireffect on fluid and salt overload, we included estimatesof nutritional supplement use based on literature indi-cating such use among cancer patients. Excluding thiscomponent would change our results by much less than1%. It should also be noted that at the time this studywas conducted, there was no standard therapy foradvanced-stage HCC, and some treatment modalitiesthat have since been determined to be ineffective forlate-stage HCC were being administered to patients withadvanced disease in the population being studied – suchas radiofrequency ablation, arterial embolization, orsurgical resection. The differences in past and currenttreatment patterns may also account for the high surgi-cal costs observed in the study patients with regionaland distant disease.

Third, due to the difficulty of assessing caregiver bur-den and because large studies of costs to caregivers forHCC patients have not been conducted, we did notaccount for this variable. Future work aimed at assess-ing caregiver burden would lend useful insights into thispotentially important component of disease burden.

The SEER-Medicare data are subject to additionallimitations, including potential inaccuracy of the diag-nostic and procedural coding, demographic coding

Table A1

Codes for identifying specific HCC treatments in SEER-Medicare data

Cancer treatment Codes

ICD-9-CMdiagnosis code(s)

ICD-9-CMprocedure code(s)

HCPCS/CPT code(s)

Surgical procedure to liver 502x, 503x, 504x 47120, 47122, 47125, 47126, 47127, 47128, 47129, 47130, 47133,47134

Transplant 505x 47135Transcatheter arterial embolization 37204Percutaneous injection therapy 509x 47380, 47500Radiofrequency ablation 922x 47380, 47382Chemotherapy V581, V662, V672 9925 964xx, 965xx, J9045, J9190, J9300, Q0083, Q0084, Q0085,

J9035, J9201, J8520, J8521, J8530–J9999Cisplatin J9060, J9062Paclitaxel J9265

Radiation therapy V580 9221, 9222, 9223 77431, 77399, 77401, 77402, 77403, 77404, 77406, 77407, 77408,77409, 77411, 77412, 77413, 77414, 77416, 77417, 77419, 77420,77425, 77430, 77431, 77432, 77470, 77499, 61793, 77418

Arterial embolization 3979, 380x, 386x, 387x,388x,3950, 3891, 3893

Biotherapy/immunotherapyInterferon-alpha J9212, J9213, J9214, J9215

Megace S0179, Z1835, Z1836, 90782Zometa J3487, 90780Fentanyl J1810, J3010Morphine S0093, J2270, J2271, J2272, J2273, J2274, J2275Nutritional supplements 99561

K. Lang et al. / Journal of Hepatology 50 (2009) 89–99 97

errors, potentially incomplete data on Medicare claims[48], as well as the fact that not all relevant healthcareservices are covered by Medicare (e.g., prescription med-ications). We have attempted to supplement the datawith findings from published literature on use of pre-scription medications in order to capture this compo-nent. And finally, healthcare claims were not availablefor all Medicare beneficiaries (e.g., those enrolled inHMOs).

Despite its limitations, the linked SEER-Medicaredatabase has proven extremely useful for case selectionand profiling cancer treatment patterns and survivalfor a multitude of cancers [49]. Combining the SEERand Medicare data provides information on both initialcancer diagnosis and later cancer treatment, as well asdownstream medical care for cancer patients.

Our results are based on a sample of HCC patientsalive and treated in 1999, the latest year for which datawere available. Since that time, treatment has changedconsiderably, especially with regard to surgical treatment,liver transplantation, and radiological therapy. Ourresults are therefore reflective of treatment practices thatwere current at the time the study was conducted, andmay significantly underreport the current therapies forHCC. Nevertheless, they can be used as a baseline againstwhich to evaluate the impact of emerging pharmacologictreatments for advanced HCC. Future studies shouldincorporate the impact of such emerging treatments.

The general nature of our model makes it easily ame-nable to adaptation. Using local-area treatment patternand unit cost data, the model can be applied to generate

country-specific burden of illness estimates. Such adap-tation to multiple countries would enable assessmentof the global burden of HCC and comparison acrosscountries.

In conclusion, our results exhibit a considerable eco-nomic impact of HCC in the US and substantialnational spending each year on this disease. Interven-tions to reduce the prevalence of HCC have the poten-tial to yield sizable economic benefits.

Acknowledgement

The authors thank Victoria Porter for her valuableassistance with manuscript preparation.

Appendix A

See Table A1.

References

[1] McGlynn KA, Tarone RE, El-Serag HB. A comparison of trendsin the incidence of hepatocellular carcinoma and intrahepaticcholangiocarcinoma in the United States. Cancer EpidemiolBiomarkers Prev 2006;15:1198–1203.

[2] Erhardt A, Theobald I, Petry W, Rohrborn A, Heintges T,Wettstein M, et al. Hepatocellular carcinoma. Dtsch MedWochenschr 2002;127:2665–2668.

[3] Verhoef C, Holman FA, de Man RA, de Wilt JHW, IjzermansJNM. Resection of extrahepatic hepatocellular carcinoma metas-tasis can result in long-term survival. Acta Chir Belg2005;105:533–536.

98 K. Lang et al. / Journal of Hepatology 50 (2009) 89–99

[4] Bruix J, Sherman M. Practice Guidelines Committee, AmericanAssociation for the Study of Liver Diseases Management ofhepatocellular carcinoma. Hepatology 2005;42:1208–1236.

[5] Bayer Corporate Web site. Investor News 2004: Bayer and OnyxAnnounce Phase II Results of Treatment with BAY 43-9006 inpatients with advanced primary liver cancer. Available from:<http://www.investor.bayer.com/15602_investornews2004/inves-tornews2004.php?action=detail&newsId=239> [accessed July2008].

[6] Wild CP, Hall AJ. Primary prevention of hepatocellular carci-noma in developing countries. Mutat Res 2000;462:381–393.

[7] Stuver SO. Towards global control of liver cancer? Semin CancerBiol 1998;8:299–306.

[8] Laurent C, Blanc F, Nobili S, Sa Cunha A, le Bail B, Bioulac-SageP, et al. Prognostic factors and long-term survival after hepaticresection for hepatocellular carcinoma originating from noncirrh-otic liver. J Am Coll Surg 2005;201:656–662.

[9] Llovet JM, Di Bisceglie AM, Bruix J, Kramer BS, Lencioni R,Zhu AX, et al. Panel of experts in HCC design clinical trials designand endpoints of clinical trials in hepatocellular carcinoma. J NatlCancer Inst 2008;100:698–711.

[10] El-Serag HB. Hepatocellular carcinoma and hepatitis C in theUnited States. Hepatology 2002;36:S74–S83.

[11] Rizzetto M. Hepatocellular carcinoma: screening and therapy.Minerva Gastroenterol Dietol 2000;46:143–147.

[12] Taketomi A, Soejima Y, Yoshizumi T, Uchiyama H, YamashitaY, Maehara Y. Liver transplantation for hepatocellular carci-noma. J Hepatobiliary Pancreat Surg 2008;15:124–130.

[13] Llovet JM, Bruix J. Systematic review of randomized trials forunresectable hepatocellular carcinoma: chemoembolizationimproves survival. Hepatology 2003;37:429–442.

[14] Llovet J, Ricci S, Mazzaferro V, Hilgard P, Gane E, Blanc JF,et al. Sorafenib in advanced hepatocellular carcinoma. N Engl JMed 2008;359:378–390.

[15] FDA. Center for Drug Evaluation and Research. Press release:FDA Approves Sorafenib (Nexavar) for the Treatment ofUnresectable Hepatocellular Carcinoma. November 20, 2007.Available from: <http://www.fda.gov/cder/Offices/OODP/whats-new/sorafenib.htm> [accessed July 2008].

[16] Patel D, Terrault NA, Yao FY, Bass NM, Ladabaum U. Cost-effectiveness of hepatocellular carcinoma surveillance in patientswith hepatitis C virus-related cirrhosis. Clin Gastroenterol Hep-atol 2005;3:75–84.

[17] Ruelas-Villavicencio AL, Vargas-Vorackova F. In whom, howand how often is surveillance for hepatocellular carcinoma cost-effective? Ann Hepatol 2004;3:152–159.

[18] Lin OS, Keeffe EB, Sanders GD, Owens DK. Cost-effectiveness ofscreening for hepatocellular carcinoma in patients with cirrhosisdue to chronic hepatitis C. Aliment Pharmacol Ther2004;19:1159–1172.

[19] Arguedas MR, Chen VK, Eloubeidi MA, Fallon MB. Screeningfor hepatocellular carcinoma in patients with hepatitis C cirrhosis:a cost-utility analysis. Am J Gastroenterol 2003;98:679–690.

[20] Yuen M-F, Lai C-L. Screening for hepatocellular carcinoma:survival benefit and cost-effectiveness. Ann Oncol2003;14:1463–1467.

[21] Wong JB. Cost-effectiveness of treatments for chronic hepatitis C.Am J Med 1999;107:74S–78S.

[22] Sarasin FP, Majno PE, Llovet JM, Bruix J, Mentha G, HadenqueA. Living donor liver transplantation for early hepatocellularcarcinoma: a life-expectancy and cost-effectiveness perspective.Hepatology 2001;33:1073–1079.

[23] Seror O, N’Kontchou G, Tin Tin Htar M, Durand-Zaleski I,Trinchet JC, Sellier N, et al. Ethanol versus radiofrequencyablation for the treatment of small hepatocellular carcinoma inpatients with cirrhosis: a retrospective study of efficacy and cost.Gastroenterol Clin Biol 2006;30:1265–1273.

[24] Ikeda K, Kobayashi M, Saitoh S, Someya T, Hosaka T, Sezaki H,et al. Cost-effectiveness of radiofrequency ablation and surgicaltherapy for small hepatocellular carcinoma of 3 cm or less indiameter. Hepatol Res 2005;33:241–249.

[25] Wilson LS, Lightwood JM. Pancreatic cancer: total costs andutilization of health services. J Surg Oncol 1999;71:171–181.

[26] Chang S, Long SR, Kutikova L, Bowman L, Finley D, CrownWH, et al. Estimating the cost of cancer: results on the basis ofclaims data analyses for cancer patients diagnosed with seventypes of cancer during 1999 to 2000. J Clin Oncol2004;22:3524–3530.

[27] Kim WR, Gores GJ, Benson JT, Therneau RM, Melton LJ.Mortality and hospital utilization for hepatocellular carcinoma inthe United States. Gastroenterology 2005;129:486–493.

[28] Sloane D, Chen H, Howell C. Racial disparity in primaryhepatocellular carcinoma: tumor stage at presentation, surgicaltreatment and survival. J Natl Med Assoc 2006;98:1934–1939.

[29] Surveillance, Epidemiology, and End Results (SEER) Program(www.seer.cancer.gov) SEER*Stat Database: Incidence – SEER13 Regs Public-Use, Nov 2004 Sub (1973–2002 varying), NationalCancer Institute, DCCPS, Surveillance Research Program, CancerStatistics Branch, released April 2005, based on the November2004 submission.

[30] Wang X, Hershman DL, Abrams JA, Feingold D, Grann VR,Jacobson JS, et al. Predictors of survival after hepatic resectionamong patients with colorectal liver metastasis. Br J Cancer2007;97:1606–1612.

[31] Brown ML, Riley GF, Schussler N, Etzioni R. Estimating healthcare costs related to cancer treatment from SEER-Medicare data.Med Care 2002;40:IV104–IV117.

[32] Berger A, Smith M, Lidsky L, Seifeldin R, Oster G. Opioid useand health-care charges at the end of life in patients withmetastatic cancer. Manag Care Interface 2004;17:28–34.

[33] RxList. Zofran indications and usage, dosage and administration.Available from: <http://www.rxlist.com/cgi/generic4/zofran_ids.htm>[accessed July 2008].

[34] Fisch M, Callahan CM, Kesterson JG, Nichols C, Tierney WM.The use of an electronic patient record system to identifyadvanced cancer patients and antidepressant drug use. J PalliatMed 1999;2:403–409.

[35] Jazieh AR, Kopp M, Foraida M, Ghouse M, Khalil M, SavidgeM, et al. The use of dietary supplements by veterans with cancer. JAltern Complement Med 2004;10:560–564.

[36] Best JH, Veenstra DL, Geppert J. Trends in expenditures forMedicare liver transplant recipients. Liver Transpl2001;7:858–862.

[37] Fernandez JA, Robles R, Marin C, Sanchez-Bueno F,Ramirez P, Pons JA, et al. Can we expand the indicationsfor liver transplantation among hepatocellular carcinomapatients with increased tumor size? Transplant Proc 2003;35:1818–1820.

[38] Yao FY, Ferrell L, Bass NM, Watson JJ, Bacchetti P, Venook A,et al. Liver transplantation for hepatocellular carcinoma: expan-sion of the tumor size limits does not adversely impact survival.Hepatology 2001;33:1394–1403.

[39] Easton Associates Interviews and Analysis. Unpublished datacollected by Easton Associates, LLC, 2004.

[40] Evans RW, Manninen DL, Dong FB. An economic analysis ofliver transplantation. Costs, insurance coverage, and reimburse-ment. Gastroenterol Clin North Am 1993;22:451–473.

[41] Drug Topics. Red book. Montvale (NJ): Medical Economics;2004.

[42] Yabroff KR, Lawrence WF, Clauser S, Davis WW, Brown ML.Burden of illness in cancer survivors: findings from a popula-tion-based national sample. J Natl Cancer Inst 2004;96:1322–1330.

K. Lang et al. / Journal of Hepatology 50 (2009) 89–99 99

[43] US Bureau of Labor Statistics. Median usual weekly earnings offull-time wage and salary workers. Available from: <http://www.bls.gov/news.release/wkyeng.t02.htm> [accessed July2008].

[44] Rice DP. Estimating the cost of illness. Health Economics Series,No. 6, DHEW Pub No. (PHS) 947-6. Rockville (MD): U.S.Department of Health Education and Welfare; 1966.

[45] Krahn MD, Mahoney JE, Eckman MH, Trachtenberg J, PaukerSG, Detsky AS. Screening for prostate cancer. A decision analyticview. JAMA 1994;272:773–780.

[46] Taplin SH, Barlow W, Urban N, Mandelson MT, Timlin DJ,Ichikawa L, et al. Stage, age, comorbidity, and direct costs of

colon, prostate, and breast cancer care. J Natl Cancer Inst1995;87:417–426.

[47] But DY, Lai CL, Yuen MF. Natural history of hepatitis-relatedhepatocellular carcinoma. World J Gastroenterol 2008;14:1652–1656.

[48] Du X, Freeman JL, Goodwin JS. Information on radiationtreatment in patients with breast cancer: the advantages of thelinked Medicare and SEER data. J Clin Epidemiol1999;52:463–470.

[49] National Cancer Institute. SEER-Medicare Publications, 2005.Available from: <http://healthservices.cancer.gov/seermedicare/overview/publications.html> [accessed October 2007].