Direct-Acting Antiviral Agents for Patients With Hepatitis C Virus Genotype 1 Infection Are Cost ...

41 Pages·2016·0.68 MB·English

Accepted Manuscript
Direct-Acting Antiviral Agents for Patients With Hepatitis C Virus Genotype 1
Infection are Cost Saving
Jagpreet Chhatwal, PhD, Tianhua He, Chin Hur, MD, MPH, Maria A. Lopez-Olivo,
MD, PhD
PII: S1542-3565(16)30673-5
DOI: 10.1016/j.cgh.2016.09.015
Reference: YJCGH 54917
To appear in: Clinical Gastroenterology and Hepatology
Accepted Date: 6 September 2016
Please cite this article as: Chhatwal J, He T, Hur C, Lopez-Olivo MA, Direct-Acting Antiviral Agents for
Patients With Hepatitis C Virus Genotype 1 Infection are Cost Saving, Clinical Gastroenterology and
Hepatology (2016), doi: 10.1016/j.cgh.2016.09.015.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Direct-Acting Antiviral Agents for Patients With Hepatitis C Virus Genotype 1
Infection are Cost Saving

Jagpreet Chhatwal, PhD1,2,3 Tianhua He,4 Chin Hur, MD, MPH1,2,3 Maria A. Lopez-Olivo, MD,
PhD5
T

1Institute for Technology Assessment, Massachusetts General Hospital, Boston, MPA, USA
2Harvard Medical School, Boston, MA, USA
I
3Liver Center and Gastrointestinal Division, Massachusetts General Hospital, Boston, MA, USA
R
4Tsinghua University School of Medicine, Beijing, China.
5Department of General Internal Medicine, The University of Texas, MDC Anderson Cancer
Center, Houston, TX, USA
S

U
Abbreviations: HCV, hepatitis C virus; DAA, direct-acting antiviral; CEA, cost-effectiveness
analysis; ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life year; PRISMA,
Preferred Reporting Items for Systematic Reviews and MetNa-Analyses; DARE, Database of
Abstracts of Reviews of Effects; NHS EED, National Health Service Economic Evaluation
Database; CHEERS, The Consolidated Health EconomAic Evaluation Reporting Standards;
WTP, willingness to pay.
M
Word count: 3,976

Number of figures and tables: 6
D

Correspondence: Jagpreet Chhatwal, PhD
E
MGH Institute for Technology Assessment
101 Merrimac Street, Floor 10th
T
Boston, MA 02114
Email: [email protected]
P
Phone: [phone]
Fax: [phone]
E

Grant support: This project was funded by Gilead Sciences, Inc. Funding source had no
C
involvement in study design; in the collection, analysis and interpretation of data; in the writing
of the report; and in the decision to submit the article for publication.
C

Disclosures: Chhatwal has received a research grant from Gilead Sciences, and consulting fee
A
from Gilead Sciences, Merck & Co. Inc. Other authors have no conflict of interest.

Authors contributions: study concept and design: Chhatwal, He, Hur, Angeles-Lopez,
experiments and procedures: Chhatwal and He; writing of article: Chhatwal, Angeles-Lopez
1
ACCEPTED MANUSCRIPT
Abstract
Background & Aims: Direct-acting antivirals (DAAs) are effective in treatment of hepatitis C
virus (HCV) genotype 1 infection, but their cost and value have been debated. We performed a
systematic review of published cost-effectiveness analyses of DAAs, synthesized their results
with updated drug prices, and calculated the maximum price at which DAA therapy for HCV
T
genotype 1 infection is cost effective (increased quality-adjusted life-years [QALYs]) and
increased cost that the society is willing to pay) and cost saving (increased QALYs and
P
decreased costs).
I
Methods: We conducted a systematic review of the Pubmed, Medline, EMBRASE, Cochrane
library, EconLit, Database of Abstracts of Reviews of Effects, National Health Service Economic
Evaluation Database, and Health Technology Assessment, and Tufts University databases for
C
cost-effectiveness analyses published from 2011 through 2015. Our analysis included cost
effectiveness of DAAs vs previous standard-of-care regimens (peginterferon and ribavirin,
S
boceprevir and telaprevir), or no treatment, performed for patients with HCV genotype 1
infection. We excluded studies that were not written in English or those that did not report
U
QALYs. Reported incremental cost-effectiveness ratios (ICERs) and treatment costs for each
comparison were extracted; the threshold price was estimated for each analysis in which
N
regimens were found to be cost effective (ICER≤$100,000/QALY) or cost saving (ICER<$0)—
those that decreased costs and increased QALYs.
A

Results: We identified 24 cost effectiveness studies that reported 170 ICERs for combinations
M
of 11 drugs, from 11 countries. Of those, 81 ICERs were determined for 1st generation DAAs
(boceprevir and telaprevir) and 89 ICERs were determined for 2nd generation DAAs (drugs
approved after the 1st generation DAAs) as a p rimary intervention. The median threshold prices
at which 1st generation and 2nd generation DDAAs became cost-effective were estimated as
$120,100 (inter-quartile range, $90,700–$176,800) and $227,200 (inter-quartile range,
$142,800–$355,800), respectively. At the discounted price of $60,000, 71% of the analyses
E
found 2nd generation DAAs to be cost saving and 22% to be cost effective.
T
Conclusions: In a systematic review, we found treatment of HCV genotype 1 infection with 2nd
generation DAAs to be cost effePctive when they cost less than and $227,200; these drugs
produced cost savings at current discounts.
E
KEY WORDS: cost effectiveness, budget impact, public health, health care expenses
C
C
A
2
ACCEPTED MANUSCRIPT
INTRODUCTION
Worldwide, more than 170 million people are chronically infected with hepatitis C virus (HCV).1
HCV infection is the leading cause of hepatocellular carcinoma and is the most common
T
indication for liver transplantation in the US and Europe.2 Of the six HCV genotypes, genotype 1
P
is the most prevalent in the Western world and accounts for at least 70% of all chronic
I
infections.3 R

C
HCV treatment has rapidly evolved over the past few years—beginning with the launch of two
S
direct-acting antivirals (DAAs)—boceprevir and telaprevir, in 2011 for genotype 1; followed by
U
the availability of several new oral DAAs from 2013 onward, including sofosbuvir, simeprevir,
N
ledipasvir, daclatasvir, and Viekira Pak. New oral therapies are superior, with efficacy above
A
95% in the majority of patients, and require shorter duration of treatment and have fewer
M
adverse effects than the older therapies.

D
However, the high price coupled with the high demand for oral DAAs has created concerns
E
about their impact on health care buTdgets, delaying timely treatment to several HCV patients.
The high price of these drugs hPas led to a national debate about the value and affordability of
HCV treatment in the U.S. aEnd elsewhere. Citing these concerns, several payers, including state
Medicaid programs, havCe restricted these treatments to patients with advanced stages of
hepatic fibrosis.4 C
A
Cost-effectiveness analysis (CEA) can inform stakeholders regarding the value of HCV
treatment and allow them to compare its value with other medical interventions. Using a
decision-analytic modeling approach, such analysis can project the long-term health benefits of
HCV treatment, predict long-term costs of HCV sequelae, and weigh them against the cost of
up-front treatment.
3
ACCEPTED MANUSCRIPT

Several cost-effectiveness analyses using models of HCV were recently published that
assessed the value of treatment by estimating the incremental cost-effectiveness ratios
T
(ICERs), which provide the cost of gaining one additional quality-adjusted life-years (QALY).
P
However, different modeling assumptions, including those regarding the costs of DAAs, may
I
have influenced their results. Furthermore, drug prices have come down conRsiderably after the
publication of these cost-effectiveness studies, so their results are outdated and may not
C
accurately depict the current real-world value. Our objective was two-fold: (i) to systematically
S
review the published cost-effectiveness analyses of DAAs and synthesize published results
U
after updating their cost assumptions, and (ii) to estimate the threshold drug prices below which
N
HCV treatment is cost-effective and/or cost-saving.
A

M
METHODS

We synthetized the cost-effectiveness resultDs by controlling for the drug price, HCV genotype,
treatment history and geographic region. We followed the Preferred Reporting Items for
E
Systematic Reviews and Meta-AnalTyses (PRISMA) statement to report our results.
P
Information Sources and SeEarch
We conducted a literatuCre search in general (i.e., Pubmed, Medline, EMBASE, The Cochrane
library) and contentC specific electronic databases (i.e, EconLit, Database of Abstracts of
Reviews of EffeActs (DARE), National Health Service Economic Evaluation Database (NHS-
EED) and Health Technology Assessment (HTA) database, Tufts CEA registry). The search
covered peer-reviewed original articles published from January 1st 2011 until September 8th
2015. The list of references from potentially relevant articles were also searched. All results
were imported into a reference manager software and merged, and then duplicates were
removed. Supplementary Table S1 shows the search strategy used for the Medline database.
4
ACCEPTED MANUSCRIPT

Inclusion and Exclusion Criteria
We included any article that reported an economic model to estimate the cost-effectiveness of
T
treating HCV patients with DAAs in comparison with old standard of care, which was dependent
P
on the primary intervention as well as the target population. We included all approved DAAs
I
categorized as first-generation and second-generation DAAs. The 1st-generRation DAAs
included boceprevir and telaprevir, and 2nd-generation DAAs included simeprevir, paritaprevir,
C
asunaprevir, ledipasvir, ombitasvir, sofosbuvir, daclatasvir, and dasabuvir. The comparators for
S
1st-generation DAAs were: peginterferon-ribavirin or no treatment; and the comparators for 2nd-
U
generation DAAs were: 1st-generation DAAs, peginterferon-ribavirin or no treatment.
N

A
We excluded articles that were: not published in English, not original studies (i.e., reviews,
M
opinion articles), did not provide modeling techniques used, did not report QALYs, reported only

budget impact or cost-of-illness analysis, didD not report any results for genotype 1 patients,
reported a drug regimen that has not been approved by the FDA or recommended by a
E
professional organization, did not reTport cost-effectiveness results compared to the old standard
of care, did not report sensitivityP analysis on the cost of HCV treatment, or only included patients
co-infected with other viral inEfections (e.g., human immunodeficiency virus, hepatitis B virus).
C
Study Selection C
Two reviewers sAcreened the titles and abstracts of the unique citations independently. The first
step was to assess whether inclusion criteria were met. Any disagreements regarding whether
or not a particular analysis fulfilled the initial inclusion criteria were resolved by discussion
leading to consensus. We then retrieved the full-text of those relevant citations and excluded
unsuitable analyses according to our exclusion criteria, resulting in the final group of studies
analyzed.
5
ACCEPTED MANUSCRIPT

Data collection process and data items
We used a standardized extraction form to collect information. One reviewer abstracted data
T
from the included studies and another reviewer crosschecked the abstracted information. We
P
collected basic information regarding the study including: study year; country for which the
I
analysis was performed; drug regimen as primary intervention, which was fuRrther categorized as
1st-generation DAAs (boceprevir- and telaprevir-based therapies) and 2nd-generation DAAs;
C
and the comparator regimen. Additional information such as characteristics of the modeled
S
population including presence of cirrhosis, HCV genotype, and treatment experience were
U
extracted. Model features relevant to the cost-effectiveness of HCV treatment such as
N
perspective, discount rate, treatment-as-prevention benefits, extra-hepatic benefits, and re-
A
infection after SVR were also noted.
M

Quality Appraisal D
One investigator (TH) assessed the quality of reporting using The Consolidated Health
E
Economic Evaluation Reporting StaTndards (CHEERS) statement,5 and another investigator
(MLO) cross-checked the entriePs. The CHEERS statement is a 24-item checklist evaluating six
sections of an economic evaEluation. Each of the items was explicitly judged using: Yes =
“information reported” oCr No = “information not reported.” The quality of the included studies was
reported as the numCber of missing information across studies per item and converted to
percentages. A

Synthesis of results: Re-analysis of ICERs and Threshold Drug Price
To account for differences in treatment costs used by published studies, we re-calculated each
ICER at the wholesale acquisition cost of DAAs and over a range of $20,000–$100,000. We
converted all costs to 2014 US Dollar value using the Consumer Price Index for each country
6
ACCEPTED MANUSCRIPT
and the average currency conversion rate in 2014.6, 7 Because treatment costs and ICERs have
a linear relationship, we used the linear interpolation approach to find the ICER at any treatment
cost. For that purpose, we first extracted all reported costs, QALYs, and ICERs from all studies.
T
For each analysis, we also extracted an ICER at a different price either using the published 1-
P
way sensitivity analysis or price-threshold analysis. If the sensitivity analysis on the cost of HCV
I
treatment was not reported, we removed that ICER from our analysis. The bRelow equation
provides the ICER (denoted by ) when the price of DAAs is X:
(cid:1)(cid:2)(cid:3)(cid:4) C
(cid:5)
(cid:1)(cid:2)(cid:3)(cid:4) = (cid:7)(cid:8)(cid:9)(cid:10)(cid:11)(cid:12)(cid:13)(cid:8)(cid:9)(cid:10)(cid:11)(cid:14)(cid:20)× ((cid:24)−(cid:26)(cid:27)(cid:28)(cid:29)(cid:30) )+(cid:1)S(cid:2)(cid:3)(cid:4)
(cid:5) (cid:31) (cid:31)
(cid:15)(cid:16)(cid:17)(cid:18)(cid:19)(cid:12)(cid:13)(cid:15)(cid:16)(cid:17)(cid:18)(cid:19)(cid:14)
U
where Price and Price along with their corresponding ICERs (ICER and ICER ) were
A B A B
extracted from the published studies. N
A
For each study, we further estimated the threshold Mprice below which treatment with DAAs
would become cost-effective using the willingness-to-pay (WTP) threshold of $100,000-per-

D
QALY and cost-saving (i.e. WTP threshold of $0-per-QALY). The following equation provides
E
the relationship between the threshold drug price and WTP:
T
(cid:26)(cid:27)(cid:28)(cid:29)(cid:30)+−(cid:26)(cid:27)(cid:28)(cid:29)(cid:30)(cid:31)
(cid:26)(cid:27)(cid:28)(cid:29)(cid:30)"#(cid:16)(cid:19)$#%&' = (cid:26)(cid:27)(cid:28)(cid:29)(cid:30)(cid:31)+(()(cid:26)−(cid:1)(cid:2)(cid:3)(cid:4)(cid:31))×* ,
P (cid:1)(cid:2)(cid:3)(cid:4) −(cid:1)(cid:2)(cid:3)(cid:4)
+ (cid:31)
E
RESULTS C
Our search yielded 304 records. After exclusion of duplicates and two-step screening, 24 unique
C
studies were incAluded in our analysis (Figure 1).

Overview of Studies and Patient Characteristics
Twenty-four studies were included in the systematic review. Supplementary Table S2 provides
an overview of the study characteristics. These studies evaluated 170 ICERs of 11 different
drug combinations from 11 different countries (Supplementary Table S3). Among them, 81
7
ACCEPTED MANUSCRIPT
ICERs evaluated first-generation DAAs as the primary intervention; 67 were compared with
peginterferon-ribavirin and 14 with no treatment. Furthermore, 89 ICERs evaluated second-
generation DAAs as the primary intervention; 67 were compared with the old standard-of-care
T
(peginterferon-ribavirin or first-generation DAAs) and 22 with no treatment. Table 1 summarizes
P
key modeling features of the included studies.
I
R
Region
C
Among the selected studies, 110 ICERs (65%) were reported by 10 (42%) studies conducted in
S
the U.S., 51 (26%) ICERs by 12 (44%) studies in Europe, 7 (4%) ICERs by 1 (4%) study in Asia,
U
and 2 (1%) ICERs by 1 (4%) study in Australia (Supplementary Table S4).
N

A
Treatment strategies
M
The cost-effectiveness models evaluated the following DAAs as primary interventions: first-

generation DAAs—boceprevir or telaprevir, iDn combination with peginterferon and ribavirin;
second-generation DAAs consisting of sofosbuvir and/or simeprevir with peginterferon and
E
ribavirin, and oral DAAs consisting oTf different combinations of sofosbuvir, ledipasvir, daclatasvir
with or without ribavirin, and omPbitasvir/paritaprevir/ritonavir with dasabuvir and/or ribavirin
(Figure 2). E
C
Modeling assumpCtions
We noted structAural assumptions made by published studies that are important in the context of
HCV modeling and could impact the cost-effectiveness results. With the exception of one
study,8 which used a societal perspective, all other studies used the payer’s perspective. We
found that none of the studies captured the benefits of HCV treatment in reducing HCV
transmission (i.e. treatment as prevention); no study considered extra-hepatic benefits
associated with HCV treatment; only one study considered indirect economic benefits resulting
8
ACCEPTED MANUSCRIPT
from HCV cure;8 and only 4 out of 24 studies8-13 modeled the possibility of re-infection after
achieving SVR.

T
Quality of reporting
P
Supplementary Figure S1 summarizes the percentage of studies adequately reporting (and
I
not reporting) each CHEERS methodological item evaluated. In 29% of the sRtudies an adequate
structured abstract according to CHEERS standards was not provided, 20% did not adequately
C
describe target population and subgroups, and 54% did not adequately describe the population
S
and methods used to elicit preferences for outcomes. In addition, 17% did not provide the
U
methodologic approach used to estimate resource use, and 29% did not explore all
N
characteristics of uncertainty associated with patient-level data or model parameters.
A

M
Incremental Cost-effectiveness Ratios

We estimated the ICER of each reported anal ysis by varying the price of DAAs from $20,000 to
D
$100,000 and summarized the results by plotting the percentage of analyses that would be cost-
E
saving, cost-effective, and not cost-effective for a given drug price. Figure 3A shows the results
T
of 81 ICERs reported for first-generation DAAs. For instance, at a $60,000 price for DAAs, 2%
P
analyses found first-generation DAAs to be cost-saving, 91% were found to be cost-effective
E
and 7% were found to be not cost-effective. Similarly, Figure 3B summarizes the results of 89
C
ICERs reported for 2nd-generation DAAs. At a $60,000 price for these DAAs: 71% of the
C
analyses found them to be cost-saving, 22% found them cost-effective, and 7% of the analyses
A
found them not cost-effective.

We further conducted a subgroup analysis of second-generation DAAs. The cost-effectiveness
of DAAs was similar in treatment-naïve versus -experienced patients (Figures 4A and 4B) and
cirrhotic versus non-cirrhotic patients (Figure 4C and 4D). However, results differed
9