Publication
Article
Evidence-Based Oncology
Author(s):
Worldwide, more than half a million new cases of hepatocellular carcinoma (HCC) are diagnosed annually. The incidence of HCC in the United States is rising with an estimated 31,000 new cases in 2018. Disease prognosis remains poor, with a 5-year survival rate across all disease stages estimated between 10%-20%, and 3% for those diagnosed with distant disease. Although morbidity is significant, especially among patients with advanced-stage disease, limited information exists on the humanistic and economic burden of HCC.
OBJECTIVE: Worldwide, more than half a million new cases of hepatocellular carcinoma (HCC) are diagnosed annually. The incidence of HCC in the United States is rising with an estimated 31,000 new cases in 2018. Disease prognosis remains poor, with a 5-year survival rate across all disease stages estimated between 10%-20%, and 3% for those diagnosed with distant disease. Although morbidity is significant, especially among patients with advanced-stage disease, limited information exists on the humanistic and economic burden of HCC.
STUDY DESIGN: Systematic literature review.
METHODS: A systematic literature search was conducted using MEDLINE and Embase computerized databases January 1, 2007, to November 1, 2017) to identify studies that evaluated adult HCC populations and quantified humanistic (utility or patient-reported) or economic (costs or resource-utilization) outcomes.
RESULTS: Fifty-seven studies met the inclusion criteria. Overall quality of life (QOL) reported by patients with HCC is poor; those with advanced disease have lower health status/QOL scores compared with those diagnosed at earlier stages of disease. HCC imposes a substantial healthcare resource utilization and cost impact on both patients and payers in the United States. Direct costs of HCC reported in the reviewed literature varied considerably.
CONCLUSIONS: The economic and humanistic burden of HCC in the United States is substantial. Patients need effective new therapies that prolong survival and positively affect QOL. Healthcare payers need to consider clinical outcomes while balancing economic and QOL implications. With the advent of new therapies, particularly immuno-oncology (I-O) therapies, additional research is needed to gain understanding of the economic and humanistic aspects of HCC and its treatment.
METHODS:
A systematic literature search was conducted using the MEDLINE and Embase computerized databases from January 1, 2007, to November 1, 2017. This time frame allowed identification of studies performed in the past decade, while excluding older studies using outdated or nonguideline recommended treatments. The search strategy is available in eAppendix 1.
Two investigators independently reviewed all citations and screened all potentially relevant, full-text articles for inclusion using a priori defined criteria, with disagreement resolved through discussion. To be eligible for inclusion, studies had to: employ an observational or experimental study design; evaluate a treated adult HCC population; provide data quantifying economic (costs or resource utilization) or QOL outcome data; and be an English-language, full-text publication. Studies investigating treatments for HCC were required to use an evidence-based or standard-of-care approach. Additional inclusion/exclusion criteria are available in eAppendix 2. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram is depicted in eAppendix 3.6
RESULTS:
The initial search yielded 2147 nonduplicate citations (eAppendix 3), of which 57 studies representing 59 analyses met the inclusion criteria. Of these, 5 were included in health utility analyses (Table 17-11), 24 in patient- reported outcome (PRO) and QOL analyses (Tables 2 [part A and part B]7,9,12-33 and 3 [part A, part B, part C, and part D]7,9,12-33), 14 in costs and resource utilization analyses (Tables 4 [part A and part B]34-47 and 5 [part A and part B]34-47), and 16 in cost-effectiveness analyses (eAppendices 4-848-53).
Humanistic BurdenHealth Utility Analyses/Measures. Utility measures are expressed as a numeric value from 0 to 1, with 0 representing death and 1 representing perfect health.54 Five studies reported health utility values among patients with HCC, with lower scores among patients with advanced HCC (aHCC), even while on systemic treatment.7-11 Although clinical trial data evaluating utilities in HCC exist,7-9 there is a lack of comparable real-world utility data for these patients (Table 1).
Levy et al interviewed patients with hepatitis B virus (HBV) infection and uninfected respondents using health-state descriptions related to HBV infection (HBV combined with compensated and decompensated cirrhosis, HCC, and liver trans- plantation) to elicit health utility scores.11 HCC was among the health states with the lowest mean health utility score in these patients.11 Similarly, a study in Canadian patients with HCC and HBV found that lower QOL scores in this population were associated with cirrhosis and HCC, rather than with the infection itself.10 This study reported values of 0.77 to 0.85 for patients with HCC and HBV, depending on the utility assessment instrument used.10
Three multicenter, randomized controlled trials (RCTs) reported utility values in patients with aHCC categorized as Child-Pugh class A with Eastern Cooperative Oncology Group performance status (ECOG PS) of 0 or 1.7-9 Although no statistical analysis was provided, utility values declined as treatment continued.
QOL. Twenty-four studies reported results related to PRO measures evaluating QoL and symptoms in patients with HCC7,9,12-33; US-only populations were included in 10 studies (Tables 2 and 3). A detailed summary of the data can be found in eAppendix 9.
QOL in aHCC Populations. Patients with aHCC have lower global health scores compared with patients who are diagnosed at earlier stages of disease; the median European Organization for Research and Treatment of Cancer Quality of Life Questionnaire Core-30 (EORTC QLQ-C30) score in Barcelona Clinic Liver Cancer (BCLC) stage A disease is 54, declining to 47 and 39 for stages B/C and D, respectively.23 Kaiser et al reported that 90% of patients with aHCC identified pain as an important concern, which often caused significant functional limitations. In most cases, pain was ongoing and began within 6 months of diagnosis. Other concerns included common symptoms of HCC: diarrhea, fatigue, and loss of appetite; skin toxicity also appeared to be a key issue. It is important to note that most patients in this study received treatment with sorafenib.22 A small study (n = 18) found that outpatients with aHCC nearing end-of-life reported pain and lack of energy as their most frequent and distressing symptoms; other symptoms included drowsiness and problems with sexual interest or activity.12
QOL in HCC Therapies. QOL can reflect benefits or harms associated with therapeutic interventions from a patient perspective. Using the Medical Outcomes Study Short Form 36 (SF-36), previously untreated patients with HCC reported improved mental health during the first 4 months of treatment with transarterial chemoembolization (TACE) and showed improvement in bodily pain, but worsening vitality scores, after the first TACE procedure.31 Hinrichs et al prospectively evaluated QOL in 79 patients before and 2 weeks after TACE. Patients reported a 12.1% decrease in global health, as well as major decreases in role, physical, and social functioning; life-impairing symptoms including pain, loss of appetite, and fatigue increased, compared with baseline.13 Doxorubicin drug-eluting bead TACE (DEB-TACE) demonstrated long-term preservation of QOL among previously untreated patients with HCC with no significant change in any SF-36 domain at 3, 6, and 12 months post—DEB-TACE compared with baseline.17 A prospective observational study suggested that, despite having more advanced disease (higher tumor burden and BCLC stage), patients with HCC who received yttrium-90 (Y90) radioembolization showed significant increases in several QOL domains, driven in part by social and functional well-being, compared with TACE-treated patients.28 Chie et al reported that QOL outcomes were similar in HCC patients receiving surgery or embolization, while ablation was less effective at maintaining QOL.17
In an RCT comparing sorafenib to selective internal radiotherapy (SIRT) with Y90 in patients with locally advanced or intermediate-stage HCC previously treated with TACE, the global health status subscore was significantly better in the SIRT group versus the sorafenib group.15 In a multicenter, multinational phase 2 trial, HCC patients receiving combination therapy with sorafenib and everolimus reported a greater decrease in physical well-being and mood, compared with those receiving sorafenib alone, despite worse baseline mood scores in the latter group.16 Finally, a small, real-world study of sorafenib in patients with aHCC reported significantly decreased overall and domain QOL scores and severe drug-related adverse events (eg, fatigue, hand-foot-skin reaction, throm- bocytopenia), leading to a cumulative therapy discontinuation rate of 33%.25
Economic Burden Costs and Healthcare Resource Utilization. Fourteen studies reported results related to costs or resource utilization associated with HCC in the United States (Table 4, eAppendix 10).34-47
Direct Costs. Direct cost outcomes varied markedly across the studies (Table 5). For studies reporting overall direct costs for patients with HCC regardless of stage or treatment, per patient per year (PPPY) costs ranged from $29,35447 to $58,529,45 with median overall costs of up to $176,456 per patient.39
Reported costs varied by stage of disease at diagnosis and by age. In the Surveillance, Epidemiology, and End Results (SEER) registries and linked Medicare (SEER-Medicare) database, 15-year direct costs for HCC were estimated at $54,829, with the highest costs incurred by those with localized disease ($78,553), followed by regional ($49,492) and distant disease ($34,352).44 In contrast, patients with distant HCC had the highest mean total per patient per month (PPPM) costs ($9585) followed by regional ($8072) and localized disease ($7265); inpatient stays and physician visits were the primary cost drivers.44 Another SEER-Medicare study found that localized disease accounted for the highest proportion (44.5%) of the total cost of illness, followed by regional (31%) and distant disease (14%).47 The authors reported that, across age strata, younger patients generally incur higher healthcare costs than older patients.47
Two studies in the SEER-Medicare database examined costs stratified by treatment. Shaya et al reported that patients who received no treatment incurred cumulative expenditures of $23,600 to $38,300 in medical costs; this figure nearly doubled for systemic chemotherapy and radiation, while patients treated with liver-directed therapies (ablation and/or TACE) incurred costs of $69,000 to $97,500.41 There was a general trend towards decreasing costs as disease staging increased; the highest costs were associated with stage I disease, and the lowest with stage IV.41 These increased costs are likely due to longer survival or increased healthcare utilization associated with specific treatment options, such as surgical resection and associated inpatient stays, compared to the lack of intervention and treatments beyond first-line therapy for patients with more advanced disease. Breunig et al found that for patients receiving TACE, cumulative per-patient Medicare expenditures were $74,788 for 1 course of TACE, increasing with each additional course of therapy.42 For patients undergoing 4 courses of TACE, cumulative Medicare-paid expenditures were $148,878; those receiving ≥4 TACE courses lived at least 1 additional year versus patients receiving 1 course of TACE.42 Electronic health data from a transplant center showed that patients with liver transplants incurred higher overall monthly costs ($7492) than nontransplant patients ($4802).39
A more recent study using the SEER-Medicare database found that patients with HCC had increased levels of resource utilization compared with noncancer controls, with both inpatient and outpatient charges between 1.15 and 1.55 times higher than charges incurred by non-HCC patients.34 A study using National (Nationwide) Inpatient Sample (NIS) data from 2005 to 2009 found an increase in the number of inpatient cases of HCC; although inpatient mortality decreased and length of stay remained stable, HCC-associated inpatient charges continued to increase, with total national HCC charges rising from $1.0 billion in 2005 to $2.0 billion in 2009. These findings are consistent with reports confirming an increasing prevalence of HCC in the United States.43 Another NIS analysis using data from 2002 to 2011 found a decreasing inpatient mortality rate, from 13.5% to 9.9%, during this period; this trend was more prominent for patients admitted with a primary diagnosis of HCC (17% vs 9.8%, respectively) and length of stay (6.5 days vs 5.6 days, respectively).36 Nevertheless, HCC-related resource utilization continued to increase, with 24,024 hospital admissions in 2002 (10,762 related to HCC as the primary diagnosis) and 50,609 in 2011 (16,350 related to HCC as the primary diagnosis).36
This may be due not only to increased use of hospital-based healthcare resources, but also to expanding treatment options across the HCC treatment continuum.
Indirect Costs. Two studies included data on indirect costs, specifically reporting costs of lost productivity (eg, lost workdays due to cancer). When calculated using published estimates and salary data, annual estimates of lost productivity were $3553 per patient47; when calculated using employee claims and employer payroll data, indirect costs were estimated to be $3594 per patient per 6-month period.40 In the SEER-Medicare database, the overall (direct and indirect) annual cost of HCC in the United States was estimated to be $454.9 million, with lost productivity accounting for 10.8% ($49.1 million) of the total.47 Lost productivity was reported to be highest among patients with localized HCC.47
Additional Resource Utilization. Despite evidence-based guidelines, patients receiving no treatment are common in studies examining the overall HCC population. A US MarketScan study reported that about 39% of cases did not receive any proven HCC treatment,46 while a SEER-Medicare database study found that approximately 60% of those diagnosed received no treatment.41 Shaya et al estimated that more than 33% of untreated patients were diagnosed with early disease (stage I or II), and untreated patients were prevalent at all ages and levels of comorbidity and liver dysfunction; this figure may be falsely elevated since sorafenib use could not be captured in the data set. Of note, the study found a clear survival advantage among patients who had undergone any treatment versus no treatment.41
Hospice care is an underutilized measure in patients with HCC, which affords significant savings in healthcare expenditure and resource utilization. In a US MarketScan study, approx- imately 15% of patients treated from 2002 to 2008 received hospice care.46 In a SEER-Medicare study involving 7992 patients who died after HCC diagnosis from 2004 to 2011, 63% received hospice care.37 Although the median time from the first claim for hospice care to death was only 18 days, these patients reported lower rates of emergency department visits not resulting in hospitalization (6.1% vs 16.2%), hospitalization (7.9% vs 47.8%), intensive care unit stays (2.8% vs 25.3%), and inpatient mortality (3.5% vs 58.4%).37 A study performed using the National Cancer Database, which retrospectively captures approximately 70% of all patients treated for cancer in the United States, included 3267 patients with unresectable HCC receiving palliative therapy from 1998 to 2011.35 Of these, 8.8% received surgical palliation, 25.3% received palliative radiotherapy, 26.8% received palliative chemotherapy, and 32.6% received pain management therapy; 6.4% received a combination of therapies.35
Economic Models. We identified and reviewed 6 articles that focused on economic models in patients with HCC from a US perspective (eAppendices 4-5)48-53 and identified 10 additional articles that evaluated economic models in other countries (eAppendices 6-7). Details of the included economic model studies are reported in eAppendix 8.
DISCUSSION
The recent introduction of the tyrosine kinase inhibitors lenvatinib and regorafenib, and nivolumab, a novel immuno-oncology (I-O) therapy, has ended a decade-long hiatus in new aHCC treatment options. As the HCC treatment landscape expands, healthcare providers and payers will need firm data about the clinical, economic and humanistic value of new therapies. This systematic review underlines the substantial humanistic and economic burdens associated with aHCC—a condition with a rising prevalence and limited treatment options.
Humanistic Burden
Evidence from this review illustrates the negative impact of HCC on QOL, most notably due to symptoms and effects of the disease itself, disease progression, and adverse effects of treatment. The impact on QOL appears to be greatest in aHCC, where therapy goals include relief of symptoms and improvements in well-being.15 The literature suggests that continual evaluation of QOL throughout treatment can preserve or improve QOL in patients with aHCC, as knowledge of timing, frequency, and duration of symptoms can improve communication and expectations between patients and their healthcare providers.13,22,41
Although utility and QOL end points are increasingly being included in cancer studies, this review found reporting on these metrics for HCC to be limited. Such information is a critical component of value-based oncology care and, in addition to clinical and health economic data, stakeholders are increasingly interested in humanistic aspects of therapies to guide clinical, regulatory, and reimbursement decisions.55 Consequently, QOL and other PROs will play an important role in health technology assessments, ultimately impacting value-based care in patients with HCC.55,56 This review revealed important limitations of existing cancer-specific PRO measures for evaluating I-O therapies. Many PRO measures were not designed to evaluate I-O therapies with their unique attributes (eg, delayed response, longer treatment duration, prolonged survival), and may not conform to traditional benchmarks and other measures that would accurately account for the benefits and tolerability of I-O therapies. Accurate measurement of QOL using validated disease- and treatment-specific instruments will be critical in evaluating therapies, creating a potential need to develop unique QOL instruments as novel HCC therapies (eg, I-O therapy) emerge.
Economic Burden
This review has confirmed that aHCC imposes a substantial healthcare resource utilization and cost impact on both patients and payers in the United States. Direct costs of HCC varied considerably between studies, likely due to differences in expenditures across practices and healthcare systems, treatment (including the increased use of sorafenib as standard of care), study design, and cost identification, measurement, and collection. Costs also varied by stage of disease at diagnosis.
This review has also revealed important knowledge gaps. Most of the HCC cost analysis studies are outdated, especially in light of recent developments in the HCC treatment landscape. Moreover, while most of the reviewed studies reported direct overall costs of HCC, there are few data on indirect costs, such as time required to obtain care, caregiver costs, out-of-pocket expenses, and reduced work productivity. Fewer than 20% of the reviewed studies evaluated indirect costs, with limited data showing an increase in lost productivity costs associated with HCC diagnosis. This highlights a clear need for additional studies to fully understand the economic burden associated with disease stage, advanced disease, and therapeutic options offered to patients with HCC. Future studies should aim to integrate both cost and QOL end points to more fully recognize the complexities and value of I-O treatment.
Limitations
This review has several potential limitations. The evaluation of humanistic and economic outcomes was not the primary objective of many of the reviewed studies and thus these outcomes may have been selectively and inconsistently reported. Due to inconsistent or incomplete reporting of PRO and economic end points, many studies failed to meet the inclusion criteria. While this review included over 50 studies, there are potential issues with external validity. Heterogeneity of study populations, study design, methods for data collection (across all outcomes), and differing methods in end point measurement preclude a more rigorous description of the humanistic and economic burden of HCC. Finally, as with any review, there is the possibility of publication bias, although we decreased this risk through our systematic methodology and broad inclusion criteria that did not exclude any observational or experimental study design.
CONCLUSIONS
The economic and humanistic burden of HCC in the United States is substantial; patients need effective new therapies that not only prolong survival, but also positively affect QOL. Likewise, healthcare payers need to consider patient clinical outcomes, including QOL, when determining value. However, the literature on the economic and humanistic burden of HCC is limited and outdated, and it does not accurately reflect current guidelines or the rapidly changing therapeutic landscape. The advent of novel therapies for aHCC, and promising emerging HCC therapies, may bring meaningful improvements in survival and QOL to patients. Additional research, including comparative effectiveness studIes, will be needed to develop a better understanding of the economic and humanistic aspects of HCC and these emerging therapies.AUTHOR INFORMATION:
IBM Watson Health (CGK, JDM), Cambridge, MA; Department of Medicine, University of Connecticut School of Medicine (CGK, JTC), Farmington, CT; Evidence-Based Practice Center, UConn/Hartford Hospital (CGK), Hartford, CT; Bristol-Myers Squibb (PS, BK, BES, ACM), Princeton, NJ; Department of Internal Medicine, University of Michigan (NDP), Ann Arbor, MI.
Christine G. Kohn, PharmD, is a senior scientist at the AHRQ-funded University of Connecticut/Hartford Hospital Evidence-Based Practice Center.
Prianka Singh, PharmD, MPH, is currently the worldwide Health Economics and Outcomes Research (HEOR) lead for head and neck cancer at Bristol-Myers Squibb.
Beata Korytowsky, MA, is a medical director at Bristol-Myers Squibb with extensive experience in real-world data generation and health economics.
Jonathan T. Caranfa, PharmD, is currently pursuing his medical degree at the University of Connecticut School of Medicine.
Jeffrey D. Miller, MS, is practice leader of economic modeling in the Global Value and Market Access group at IBM Watson Health.
Bruce Sill, PharmD,MS, is a field director of HEOR with Bristol-Myers Squibb, providing scientific, pharmacoeconomic and health outcomes information to support decision-makers in national and regional payer markets.
Alexander Marshall, PharmD, MPH, works at Bristol-Myers Squibb in the US HEOR group, where he currently directs evidence generation strategies and studies for hepatocellular carcinoma.
Neehar Parikh, MD, MS, is an assistant professor in the Division of Gastroenterology and Hepatology at the University of Michigan, where he is the medical director of the Liver Tumor program and the Living Donor Liver Transplantation program.
DISCLOSURES:
CDK and JDM are employed by IBM Watson Health, which received contracted funding from Bristol-Myers Squibb to conduct the study and prepare the manuscript. PS, BK, BES, ACM are employed by Bristol Myers-Squibb, which funded this study. NDP reports consultancies with Bristol-Myers Squibb. JTC reports no conflicts of interest.
ACKNOWLEDGEMENTS:
Editorial assistance was provided by PAREXEL International, funded by Bristol-Myers Squibb.
REFERENCES: