If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Address correspondence to: Yasmin A. Saeed, BScPhm, Toronto Health Economics and Technology Assessment Collaborative, Toronto General Hospital, 10th Floor Eaton North, Room 248, 200 Elizabeth St, Toronto, ON M5G 2C4.
Institute of Health Policy, Management and Evaluation, University of Toronto, Toronto, ON, CanadaBiostatistics Research Unit, University Health Network, Toronto, ON, Canada
Toronto Health Economics and Technology Assessment Collaborative, University Health Network, Toronto, ON, CanadaInstitute of Health Policy, Management and Evaluation, University of Toronto, Toronto, ON, Canada
Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, ON, CanadaToronto Health Economics and Technology Assessment Collaborative, University Health Network, Toronto, ON, CanadaInstitute of Health Policy, Management and Evaluation, University of Toronto, Toronto, ON, Canada
Chronic hepatitis C is associated with a significant impairment in global health status, as measured by health utility instruments. Curative therapy can alleviate this burden, although further research is needed in certain subgroups of CHC patients, as well as on the long-term impacts of treatment on utilities.
•
We found that experimental study designs yielded higher health utilities than observational studies—an effect that has not been previously documented, and has implications for the use and interpretation of health utility measurements.
•
This review will allow future cost-effectiveness analyses of hepatitis C screening and treatment programs to incorporate the best available utility data, enhancing the quality of results and enabling better health policy decisions to guide treatment access and strategies for hepatitis C elimination.
Abstract
Background
Chronic hepatitis C (CHC) is among the most burdensome infectious diseases in the world. Health utilities are a valuable tool for quantifying this burden and conducting cost-utility analysis.
Objective
Our study summarizes the available data on utilities in CHC patients. This will facilitate analyses of CHC treatment and elimination strategies.
Methods
We searched MEDLINE, Embase, and the Cochrane Library for studies measuring utilities in CHC patients. Utilities were pooled by health state and utility instrument using meta-analysis. A further analysis used meta-regression to adjust for the effects of clinical status and methodological variation.
Results
Fifty-one clinical studies comprising 15 053 patients were included. Based on the meta-regression, patients’ utilities were lower for more severe health states (predicted mean EuroQol-5D-3L utility for mild/moderate CHC: 0.751; compensated cirrhosis: 0.671; hepatocellular carcinoma: 0.662; decompensated cirrhosis: 0.602). Patients receiving interferon-based treatment had lower utilities than those on interferon-free treatment (0.647 vs 0.733). Patients who achieved sustained virologic response (0.786) had higher utilities than those with mild to moderate CHC. Utilities were substantially higher for patients in experimental studies compared to observational studies (coefficient: +0.074, P < .05). The time tradeoff instrument was associated with the highest utilities, and the Health Utilities Index 3 was associated with the lowest utilities.
Conclusion
Chronic hepatitis C is associated with a significant impairment in global health status, as measured by health utility instruments. Impairment is greater in advanced disease. Experimental study designs yield higher utilities—an effect not previously documented. Curative therapy can alleviate the burden of CHC, although further research is needed in certain areas, such as the long-term impacts of treatment on utilities.
Despite a declining rate of new infections, CHC morbidity and mortality are increasing owing to the progression of the disease in those already infected.
This poses a challenge for healthcare systems worldwide as more patients will require hospitalization and/or liver transplantation. For example, in the United States, annual healthcare costs associated with CHC are expected to increase from $6.5 billion to $9.1 billion between 2011 and 2024.
Chronic hepatitis C also poses a profound burden at the patient level. Symptoms ranging from fatigue and mild cognitive impairment to severe ascites and delirium can impair quality of life and functioning.
Additionally, misconceptions about how hepatitis C is transmitted and stigma surrounding its association with injection drug use can adversely affect social interactions and relationships.
Antiviral therapies can alleviate the burden of CHC at the patient and population levels, but at high costs—warranting cost-effectiveness analysis.
Health utility is a preference-based measure of global health status that can be used to quantify this health burden. Utilities are anchored at 0 (dead) and 1 (perfect health), and they not only describe a health state but also indicate how it is valued.
They are an important tool for quantifying disease burden, measuring the potential benefits and harms of clinical and policy interventions, and making comparisons to guide decision making—such as in cost-utility analysis.
Multiple cost-utility analyses have concluded that the new direct-acting antiviral (DAA) therapies are cost-effective,
Evaluating the cost-effectiveness of different prioritization schemes for DAA therapy remains an important exercise to guide funding decisions. Additionally, with the announcement of the World Health Organization's 2030 hepatitis elimination targets,
governments are shifting their objectives toward hepatitis C elimination. Strategies such as screening and harm reduction scale-up will need to be evaluated to determine the optimal path towards these targets, and modeling costs and health outcomes—including health utilities—will be critical. This review will facilitate such analyses by providing a comprehensive and up-to-date summary of the hepatitis C health utility literature.
Methods
This review was conducted and reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines.
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration.
We included any study reporting original utility data for patients diagnosed with current or past hepatitis C virus infection regardless of age, disease severity, comorbidities and coinfections, or treatment status. All study designs (experimental or observational) and publication types (article or abstract) were included. Studies had to report a utility estimate, a measure of uncertainty (eg, standard deviation) or sample size, and sufficient information to classify patients into a health state (ie, liver disease severity and/or treatment status and/or major comorbidity). Studies reporting utilities for a mixed cohort of patients with and without hepatitis C were excluded.
Search Strategy
We searched MEDLINE, Embase, and the Cochrane Library. The search was limited to English-language papers published from the year the hepatitis C virus was discovered (1989) to March 9, 2017.
The search strategy, designed in consultation with a medical research librarian, combined medical subject headings and text words related to hepatitis C, quality of life, and health utilities—including common utility instruments (EuroQol-5D,
Our search strategy for MEDLINE is shown in Table 1 (see Appendix 1 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005 for Embase and Cochrane search strategies). To ensure the comprehensiveness of our search results, we scanned the reference lists of included studies and relevant reviews identified through the search.
Table 1MEDLINE search strategy.
#
Search terms
1
exp hepatitis C/ or (“non-a non-b hepatitis” or (hepatitis adj3 c) or HCV).ti,ab.
2
(EQ-5D$ or EQ5D$ or Euro-Qol$ EuroQol$ or EQ-VAS or EQ-SDQ or “health utilit$” or HUI$ or time-trade-off or time-tradeoff or TTO or “Standard Gamble” or ((“Visual Analog$ Scale” or VAS) adj4 (“quality of life” or QOL or “general health”)) or short-form-6D or shortform-6D or SF-6D or SF6D or 15-D or 15D or “Quality of Well-Being” or “Quality of Wellbeing” or QWB).ti,ab.
3
Quality of life/ or Patient Preference/ or Quality-Adjusted Life Years/ or ((qualit$ adj2 life) or HRQL or HRQOL or QOL or (adjusted adj4 life-year$) or QALY$ or DALY$ or (“Healthy Year$” adj3 Equivalent$) or cost-effective$ or CEA or cost-utility or CUA or (patient adj preference$) or (preference adj2 measure$) or preference-based or “preference elicitation”).ti,ab.
4
utilit$.ti,ab.
5
(1 and 2) or (1 and 3 and 4)
6
limit 5 to (english language and yr=”1989 –Current”)
Note. Databases searched: Ovid MEDLINE: Epub Ahead of Print, In-Process & Other Non-Indexed Citations, Ovid MEDLINE® Daily, and Ovid MEDLINE 1946-Present.
Date of search: March 9, 2017.
ab indicates abstract; CEA, cost-effectiveness analysis; CUA, cost-utility analysis; DALY, disability-adjusted life years; EQ-5D, EuroQol-5D; HRQOL, health-related quality of life; HCV, hepatitis C virus; HUI, Health Utilities Index; QALY, quality-adjusted life years; QOL, quality of life; QWB, Quality of Well-Being; SF-6D, Short Form-6D; ti, title; TTO, time-tradeoff; VAS, visual analog scale.
De-duplication was performed using EndNote reference management software. Two independent reviewers (not blinded to journal names, authors, or institutions) screened titles and abstracts for inclusion. Reviewers resolved disagreements by consensus. Full texts of records not excluded at this stage were retrieved and further screened against the inclusion criteria; data were extracted from included studies. A subset of full texts was screened, extracted, and assessed for quality by a second reviewer to ensure consistency.
The following data were extracted: study characteristics (country, year, study design), patient demographics (age, sex, race), clinical characteristics (disease severity, treatment status, major comorbidities), and health utility estimates (including measures of uncertainty and utility instrument used).
Quality Assessment
Study quality was assessed using the criteria outlined in the NICE guidance document on systematic reviews of utilities
Utility scores were categorized into the following health states based on patients’ liver disease severity: mild to moderate (no cirrhosis; METAVIR F0-F3 fibrosis), compensated cirrhosis (METAVIR F4 with no symptoms of decompensation), decompensated cirrhosis (history of symptoms such as jaundice, ascites, encephalopathy, or variceal bleeding), hepatocellular carcinoma (HCC), or post-liver transplant. Health state classification was also based on treatment status (not on treatment, on interferon-based treatment, on interferon-free treatment, sustained virologic response [SVR] post-treatment, or no SVR post-treatment [treatment failure]) and presence of significant comorbidities (eg, HIV). All health states were defined as having chronic hepatitis C (hepatitis C virus RNA-positive) except for the SVR health state (hepatitis C virus antibody-positive, RNA-negative).
For studies that did not report a measure of uncertainty, a standard error was estimated based on their sample size and the mean squared standard deviation across the remaining studies (stratified by utility instrument).
If more than 1 study described the same health state using the same instrument, the results were pooled using meta-analysis. If there were few (<4) studies in a pooled subgroup, a fixed-effects model was used owing to insufficient information to estimate the between-study variance. Otherwise, a DerSimonian-Laird random effects model was used to incorporate the expected between-study heterogeneity owing to factors such as differences in study design or setting. Studies were weighted by inverse squared standard error.
A meta-regression was performed to pool utility estimates across all instruments and health states into 1 model and examine the effects of the following variables on utility estimates (for studies that reported on all variables): study design (experimental [defined as studies with a controlled or uncontrolled intervention] or observational), liver disease severity, treatment status, and utility instrument. A multilevel, multivariate meta-regression model was used to account for within-study and within-study-subgroup clustering of utility estimates (multiple utility measures from the same study and multiple utility measures from the same patient subgroup). This model was used to predict mean utility estimates for each health state, standardized to each study design and utility instrument. A secondary meta-regression model examined the effects of variables that were reported in only a subset of the studies.
Statistical heterogeneity was assessed using the I2 statistic. Publication bias was assessed through funnel plots and Egger’s regression test for funnel plot asymmetry. Two sensitivity analyses were performed to examine the effects of excluding abstracts and of excluding studies that did not report a measure of uncertainty. All analyses were performed in R
Health-related quality of life in HIV-infected and at-risk women: the impact of illicit drug use and hepatitis C on a community preference weighted measure.
Psychometric evaluation of the hepatitis C virus patient-reported outcomes (HCV-PRO) instrument: validity, responsiveness, and identification of the minimally important difference in a phase 2 clinical trial.
Health-related quality of life (HRQoL), health state, function and wellbeing of chronic HCV patients treated with interferon-free, oral DAA regimens: patient reported outcome (PRO) results from the AVIATOR study.
Fatigue during treatment for hepatitis C virus: results of self-reported fatigue severity in two Phase IIb studies of simeprevir treatment in patients with hepatitis C virus genotype 1 infection.
Cost effectiveness of daclatasvir/asunaprevir versus peginterferon/ribavirin and protease inhibitors for the treatment of hepatitis C genotype 1B in Chile.
Health-related quality of life in genotype 1 treatment-naive chronic hepatitis C patients receiving telaprevir combination treatment in the ADVANCE study.
Changes in quality of life, healthcare use, and substance use in HIV/hepatitis C coinfected patients after hepatitis C therapy: a prospective cohort study.
(Fig 1) (see Appendix 3 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005 for excluded full texts and reasons for exclusion). These publications reported on 51 clinical studies and included 33 481 individual utility measurements from 15 053 patients (Table 2). Most patients were white (79%) and male (57%), with a mean age of 49. The most commonly used utility instruments were the EuroQol-5D-3L (EQ-5D-3L) (n = 21) and SF-6D (n = 19). Study-level characteristics are presented in Appendix 4 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005.
Figure 1PRISMA flow diagram. The number of records identified, included, and excluded at each stage of the review; and the reasons for exclusions.
EQ-5D-3L indicates EuroQol-5D 3 levels; EQ-5D-5L, EuroQol-5D 5 levels; HUI, Health Utilities Index; SF-6D, Short Form-6D; SG, standard gamble; TTO, time tradeoff; VAS, visual analog scale.
∗ Some publications reported on more than 1 study.
Many studies reported insufficient information to be assessed on a number of quality assessment criteria (see Appendix 5 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005). Multiple studies failed to report patient characteristics, liver disease severity, and measures of uncertainty around utility estimates.
Results of Subgroup Meta-analysis by Health State and Utility Instrument
The largest amount of data was available for patients with mild to moderate CHC who were not undergoing treatment (range of pooled sample sizes: n = 45-4339 patients) (Table 3). (Forest plots and I2 values are presented in Appendix 6 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005.) Compared to mild to moderate CHC (pooled mean ± standard error SF-6D utility: 0.690 ± 0.007), utilities were lower for interferon-based treatment (0.628 ± 0.007), treatment failure (0.650 ± 0.014), compensated cirrhosis (0.615 ± 0.009), decompensated cirrhosis (0.630 ± 0.015), HCC (0.610 ± 0.027), post-transplant (0.650 ± 0.016), and HIV coinfection (0.667 ± 0.003). Utilities were higher for SVR (0.735 ± 0.013) and similar to mild to moderate CHC for the interferon-free treatment health state (0.679 ± 0.027). These trends held across most utility instruments.
Table 3Results of meta-analysis by liver disease severity, treatment status, and utility instrument (mean ± standard error utilities).
Subgroup
Utility instrument
Indirect
Direct
EQ-5D-3L
EQ-5D-5L
HUI2
HUI3
SF-6D
SG
TTO
VAS
Mild to moderate CHC
0.829 ± 0.021 14 studies 4339 patients
0.823 ± 0.019 4 studies 717 patients
0.754 ± 0.008 3 studies 563 patients
0.715 ± 0.010 3 studies 635 patients
0.690 ± 0.007 10 studies 4015 patients
0.810 ± 0.040 1 study 45 patients
0.811 ± 0.011 3 studies 573 patients
0.737 ± 0.022 8 studies 3095 patients
On IFN-based treatment
0.756 ± 0.018 6 studies 1848 patients
NA 0 studies 0 patients
0.710 ± 0.023 1 study 64 patients
0.520 ± 0.040 1 study 64 patients
0.628 ± 0.007 4 studies 571 patients
0.840 ± 0.060 1 study 14 patients
0.816 ± 0.025 2 studies 78 patients
0.702 ± 0.012 4 studies 834 patients
On IFN-free treatment
NA 0 studies 0 patients
0.858 ± 0.009 2 studies 170 patients
NA 0 studies 0 patients
NA 0 studies 0 patients
0.679 ± 0.027 4 studies 2454 patients
NA 0 studies 0 patients
NA 0 studies 0 patients
NA 0 studies 0 patients
SVR
0.858 ± 0.023 13 studies 1348 patients
0.877 ± 0.010 2 studies 152 patients
0.800 ± 0.010 2 studies 282 patients
0.709 ± 0.015 3 studies 318 patients
0.735 ± 0.013 5 studies 2580 patients
0.860 ± 0.039 1 study 36 patients
0.885 ± 0.011 2 studies 282 patients
0.774 ± 0.031 8 studies 695 patients
No SVR post- treatment
0.818 ± 0.026 5 studies 373 patients
NA 0 studies 0 patients
0.740 ± 0.020 1 study 103 patients
0.580 ± 0.034 1 study 103 patients
0.650 ± 0.014 1 study 103 patients
NA 0 studies 0 patients
0.840 ± 0.024 1 study 103 patients
0.852 ± 0.057 1 study 13 patients
Compensated cirrhosis
0.717 ± 0.021 8 studies 414 patients
0.600 ± 0.042 1 study 68 patients
0.707 ± 0.017 3 studies 162 patients
0.585 ± 0.023 3 studies 178 patients
0.615 ± 0.009 3 studies 191 patients
0.814 ± 0.035 2 studies 41 patients
0.790 ± 0.019 3 studies 171 patients
0.571 ± 0.017 3 studies 142 patients
Decompensated cirrhosis
0.595 ± 0.062 6 studies 169 patients
0.460 ± 0.053 1 study 11 patients
0.708 ± 0.026 2 studies 68 patients
0.588 ± 0.037 2 studies 66 patients
0.630 ± 0.015 1 study 57 patients
0.649 ± 0.077 2 studies 17 patients
0.757 ± 0.035 2 studies 65 patients
0.616 ± 0.050 2 studies 17 patients
Hepatocellular carcinoma
0.788 ± 0.039 3 studies 42 patients
NA 0 studies 0 patients
0.720 ± 0.056 1 study 20 patients
0.545 ± 0.064 2 studies 35 patients
0.610 ± 0.027 1 study 20 patients
0.720 ± 0.047 1 study 15 patients
0.780 ± 0.054 1 study 20 patients
0.550 ± 0.072 1 study 15 patients
Post-liver transplant
0.701 ± 0.044 5 studies 68 patients
0.570 ± 0.033 1 study 28 patients
0.750 ± 0.024 1 study 50 patients
0.673 ± 0.025 2 studies 80 patients
0.650 ± 0.016 1 study 50 patients
0.728 ± 0.046 2 studies 40 patients
0.801 ± 0.036 2 studies 60 patients
0.571 ± 0.017 3 studies 55 patients
HIV co-infection
NA 0 studies 0 patients
0.670 ± 0.034 1 study 27 patients
NA 0 studies 0 patients
NA 0 studies 0 patients
0.667 ± 0.003 2 studies 1037 patients
0.764 ± 0.035 2 studies 86 patients
0.811 ± 0.034 2 studies 86 patients
0.673 ± 0.036 4 studies 263 patients
Note. Health utilities were pooled separately for each health state (subgroup) and utility instrument. For n<4 studies, a fixed effects meta-analysis model was used; otherwise, a DerSimonian-Laird random effects model was used.
CHC indicates chronic hepatitis C; EQ-5D-3L, EuroQol-5D 3 levels; EQ-5D-5L, EuroQol-5D 5 levels; HIV, human immunodeficiency virus; HUI, Health Utilities Index; IFN, interferon; NA, not available; SF-6D, Short Form-6D; SG, standard gamble; SVR, sustained virologic response; TTO, time tradeoff; VAS, visual analog scale.
Studies reporting utilities for the HCC, decompensated cirrhosis, and post-transplantation health states had small samples (range: n = 11-169). Larger samples were available for treatment-associated health states across several utility instruments (range: n = 13-2580)—although data on interferon-free treatment utilities were available for only 2 instruments (Table 3).
The TTO instrument produced the highest pooled mean utilities across most health states (range: 0.757-0.885), followed by the SG (0.649-0.860). The Health Utilities Index 3 produced the lowest utilities (0.520-0.709), followed by the SF-6D instrument (0.610-0.735).
Results of Meta-regression: Impacts of Sociodemographic and Methodological Variables
A meta-regression was performed on 200 subgroups of patients from 44 studies, representing 31 445 utility measurements (Fig 2). All studies reported study design (experimental or observational). Some studies did not report patients' age, sex, race, or HIV status (Table 2). Hence, age, sex, and race were not included as covariates in the main model but were incorporated into a secondary meta-regression model that included only the studies that reported these characteristics. Patients’ HIV status was not reported in enough studies to be included as a coefficient in either model (and studies with >20% HIV coinfected patients were excluded from both models to avoid confounding).
Figure 2Results of meta-regression. Chronic hepatitis C health utilities were synthesized using meta-regression to examine the effects of disease severity, treatment status, methodological factors, and patient characteristics on utilities. Model 1: k = 200; Model 2: k = 109. Method = restricted maximum likelihood. Predicted utilities for severity and treatment-associated health states are standardized to the EQ-5D-3L utility instrument and an observational study design.
The primary meta-regression model had an intercept (± standard error) of 0.751 ± 0.017, representing the pooled utility estimate for mild to moderate CHC patients who are not on treatment. (This intercept and all of the following predicted utilities are standardized to the EuroQol-5D-3L instrument and an observational study design; see Table 4 for a complete list of predicted utilities for all combinations of utility instruments and study designs). Compared with mild to moderate CHC, utility estimates were lower for compensated cirrhosis (0.671 ± 0.021), decompensated cirrhosis (0.602 ± 0.026), HCC (0.662 ± 0.034), and post-transplantation (0.657 ± 0.026).
Table 4Predicted utility estimates by utility instrument, health state, and study design.
Utility instrument
Health state
Observational study design mean (95% CI)
Experimental study design mean (95% CI)
EQ-5D-3L
Mild to moderate
0.751 (0.718-0.785)
0.825 (0.780-0.870)
EQ-5D-3L
Compensated cirrhosis
0.671 (0.630-0.713)
0.745 (0.692-0.798)
EQ-5D-3L
Decompensated cirrhosis
0.602 (0.551-0.653)
0.676 (0.612-0.739)
EQ-5D-3L
Hepatocellular carcinoma
0.662 (0.595-0.730)
0.736 (0.660-0.813)
EQ-5D-3L
Post-transplant
0.657 (0.605-0.708)
0.731 (0.668-0.793)
EQ-5D-3L
On IFN-based treatment
0.647 (0.612-0.682)
0.721 (0.675-0.766)
EQ-5D-3L
On IFN-free treatment
0.733 (0.697-0.768)
0.807 (0.760-0.853)
EQ-5D-3L
SVR post-treatment
0.786 (0.752-0.820)
0.860 (0.814-0.905)
EQ-5D-3L
No SVR post-treatment
0.676 (0.631-0.721)
0.750 (0.698-0.802)
EQ-5D-5L
Mild to moderate
0.806 (0.767-0.845)
0.880 (0.831-0.930)
EQ-5D-5L
Compensated cirrhosis
0.726 (0.680-0.772)
0.800 (0.743-0.857)
EQ-5D-5L
Decompensated cirrhosis
0.657 (0.602-0.711)
0.731 (0.664-0.798)
EQ-5D-5L
Hepatocellular carcinoma
0.717 (0.647-0.788)
0.791 (0.712-0.870)
EQ-5D-5L
Post-transplant
0.712 (0.657-0.767)
0.786 (0.720-0.851)
EQ-5D-5L
On IFN-based treatment
0.702 (0.662-0.741)
0.776 (0.726-0.826)
EQ-5D-5L
On IFN-free treatment
0.788 (0.747-0.828)
0.862 (0.811-0.912)
EQ-5D-5L
SVR post-treatment
0.841 (0.801-0.880)
0.915 (0.865-0.964)
EQ-5D-5L
No SVR post-treatment
0.731 (0.682-0.780)
0.805 (0.749-0.861)
HUI2
Mild to moderate
0.818 (0.782-0.855)
0.892 (0.845-0.940)
HUI2
Compensated cirrhosis
0.738 (0.695-0.782)
0.812 (0.758-0.867)
HUI2
Decompensated cirrhosis
0.669 (0.616-0.722)
0.743 (0.678-0.809)
HUI2
Hepatocellular carcinoma
0.730 (0.660-0.799)
0.804 (0.726-0.882)
HUI2
Post-transplant
0.724 (0.671-0.778)
0.798 (0.734-0.863)
HUI2
On IFN-based treatment
0.714 (0.676-0.752)
0.788 (0.740-0.836)
HUI2
On IFN-free treatment
0.800 (0.762-0.838)
0.874 (0.825-0.923)
HUI2
SVR post-treatment
0.853 (0.816-0.890)
0.927 (0.879-0.975)
HUI2
No SVR post-treatment
0.743 (0.695-0.791)
0.817 (0.762-0.872)
HUI3
Mild to moderate
0.689 (0.652-0.727)
0.763 (0.715-0.812)
HUI3
Compensated cirrhosis
0.609 (0.565-0.653)
0.683 (0.628-0.739)
HUI3
Decompensated cirrhosis
0.540 (0.486-0.594)
0.614 (0.548-0.680)
HUI3
Hepatocellular carcinoma
0.601 (0.531-0.670)
0.675 (0.596-0.753)
HUI3
Post-transplant
0.595 (0.541-0.649)
0.669 (0.604-0.734)
HUI3
On IFN-based treatment
0.585 (0.546-0.624)
0.659 (0.610-0.708)
HUI3
On IFN-free treatment
0.671 (0.632-0.710)
0.745 (0.695-0.794)
HUI3
SVR post-treatment
0.724 (0.686-0.762)
0.798 (0.749-0.847)
HUI3
No SVR post-treatment
0.614 (0.566-0.663)
0.688 (0.632-0.744)
SF-6D
Mild to moderate
0.734 (0.699-0.770)
0.808 (0.762-0.854)
SF-6D
Compensated cirrhosis
0.654 (0.612-0.696)
0.728 (0.675-0.782)
SF-6D
Decompensated cirrhosis
0.585 (0.533-0.637)
0.659 (0.594-0.723)
SF-6D
Hepatocellular carcinoma
0.646 (0.577-0.714)
0.719 (0.642-0.797)
SF-6D
Post-transplant
0.640 (0.587-0.693)
0.714 (0.650-0.777)
SF-6D
On IFN-based treatment
0.630 (0.593-0.667)
0.704 (0.657-0.751)
SF-6D
On IFN-free treatment
0.716 (0.679-0.753)
0.790 (0.742-0.837)
SF-6D
SVR post-treatment
0.769 (0.733-0.805)
0.843 (0.796-0.889)
SF-6D
No SVR post-treatment
0.659 (0.612-0.706)
0.733 (0.679-0.787)
SG
Mild to moderate
0.816 (0.765-0.867)
0.890 (0.831-0.950)
SG
Compensated cirrhosis
0.736 (0.680-0.792)
0.810 (0.745-0.876)
SG
Decompensated cirrhosis
0.667 (0.603-0.731)
0.741 (0.666-0.816)
SG
Hepatocellular carcinoma
0.728 (0.652-0.803)
0.802 (0.718-0.885)
SG
Post-transplant
0.722 (0.657-0.787)
0.796 (0.722-0.870)
SG
On IFN-based treatment
0.712 (0.660-0.764)
0.786 (0.726-0.846)
SG
On IFN-free treatment
0.798 (0.746-0.850)
0.872 (0.811-0.932)
SG
SVR post-treatment
0.851 (0.800-0.902)
0.925 (0.865-0.985)
SG
No SVR post-treatment
0.741 (0.682-0.801)
0.815 (0.750-0.881)
TTO
Mild to moderate
0.897 (0.860-0.934)
0.971 (0.923-1.019)
TTO
Compensated cirrhosis
0.817 (0.773-0.860)
0.891 (0.836-0.946)
TTO
Decompensated cirrhosis
0.747 (0.694-0.801)
0.821 (0.756-0.887)
TTO
Hepatocellular carcinoma
0.808 (0.739-0.878)
0.882 (0.804-0.960)
TTO
Post-transplant
0.803 (0.748-0.857)
0.876 (0.812-0.941)
TTO
On IFN-based treatment
0.793 (0.754-0.831)
0.866 (0.818-0.915)
TTO
On IFN-free treatment
0.878 (0.839-0.917)
0.952 (0.903-1.001)
TTO
SVR post-treatment
0.931 (0.894-0.969)
1.005 (0.957-1.054)
TTO
No SVR post-treatment
0.822 (0.773-0.870)
0.896 (0.840-0.951)
VAS
Mild to moderate
0.697 (0.663-0.731)
0.771 (0.726-0.816)
VAS
Compensated cirrhosis
0.617 (0.575-0.658)
0.691 (0.637-0.744)
VAS
Decompensated cirrhosis
0.547 (0.496-0.599)
0.621 (0.558-0.685)
VAS
Hepatocellular carcinoma
0.608 (0.540-0.676)
0.682 (0.606-0.758)
VAS
Post-transplant
0.602 (0.551-0.654)
0.676 (0.614-0.739)
VAS
On IFN-based treatment
0.592 (0.557-0.627)
0.666 (0.621-0.712)
VAS
On IFN-free treatment
0.678 (0.643-0.714)
0.752 (0.706-0.799)
VAS
SVR post-treatment
0.731 (0.697-0.766)
0.805 (0.760-0.851)
VAS
No SVR post-treatment
0.622 (0.576-0.667)
0.696 (0.643-0.748)
Note. Predicted utility estimates by liver disease severity and treatment status, standardized to each utility instrument and study design (based on primary meta-regression model).
EQ-5D-3L indicates EuroQol-5D 3 levels; EQ-5D-5L, EuroQol-5D 5 levels; HUI, Health Utilities Index; IFN, interferon; SF-6D, Short Form-6D; SG, standard gamble; SVR, sustained virologic response; TTO, time tradeoff; VAS, visual analog scale.
The utility estimate for interferon-free treatment (0.733 ± 0.018) was similar to mild to moderate CHC (0.751 ± 0.017) but substantially higher than interferon-based treatment (0.647 ± 0.018). Sustained virologic response (0.786 ± 0.017) was associated with higher utilities than mild to moderate CHC, whereas treatment failure (0.676 ± 0.023) was associated with lower utilities.
Experimental studies had substantially higher utilities than observational studies (coefficient: +0.074, P < .05). The relative effects of the utility instruments were similar to the results from the unadjusted meta-analysis. All coefficients in the model were statistically significant (P < .05) except for use of the SF-6D instrument.
The secondary meta-regression model incorporated 18 936 utility measurements from 109 subgroups of patients. This model suggested that age has a minimal impact on health utilities in CHC patients (coefficient: -0.004/year, not statistically significant), while female sex has a large but nonsignificant negative effect (coefficient: -0.151). Racial minorities may have lower utilities than white patients (-0.260, P < .05) (see Appendix 7 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005 for all model coefficients).
Heterogeneity, Funnel Plots, and Sensitivity Analysis
The random effects meta-analyses by health state and utility instrument had I2 index values ranging from 33% to 99% (see Appendix 6 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005 for all I2 values). The primary meta-regression model had an I2 index of 98%, indicating that a large proportion of the observed variance was due to variance in true effect sizes rather than sampling error.
Two sensitivity analyses were performed to assess the impacts of excluding studies that did not report a measure of uncertainty and those that were reported as conference abstracts. Neither had a substantial impact on the results (see Appendix 7 in Supplemental Materials found at https://doi.org/10.1016/j.jval.2019.07.005).
Discussion
Impact of CHC on Health Utilities
Consistent with previous literature, we found utilities for patients with mild to moderate CHC to be relatively high (0.751 for pooled EQ-5D-3L estimate), but substantially lower than the population norm of 0.87 for adults aged 45 to 54 in the United States.
The observed decrement might be partially explained by the fact that CHC, although often asymptomatic, can be associated with nonspecific symptoms such as fatigue and mild cognitive impairment as well as stigma.
It is unclear whether increasing fibrosis affects health utility in noncirrhotic patients because only 1 of the studies included in our review reported utilities separately by METAVIR fibrosis score.
From a clinical perspective, it is unlikely that quality of life differs substantially from METAVIR F0 to F3 because symptoms are often minimal until decompensation occurs.
Further research with adjustment for confounding factors is needed to explore this relationship.
There is a large utility decrement associated with compensated cirrhosis, and an additional large decrement associated with decompensated cirrhosis. These results are expected because symptoms increase with disease severity. Unexpectedly, HCC had a similar utility to compensated cirrhosis. This may be attributable to small sample sizes. Additionally, patients with curable HCC may experience an improvement in quality of life once they become aware of their favorable prognosis. The decompensated cirrhosis and HCC health states each encompass a broad spectrum of morbidity, ranging from mild ascites to severe hepatic encephalopathy.
The entirety of these spectrums may not have been captured: very advanced patients would likely be underrepresented because their significant impairments could interfere with their ability to complete utility assessments.
Decompensated cirrhosis and HCC can be indications for liver transplantation. Transplantation was associated with a higher utility than decompensated cirrhosis, as expected, but represented minimal improvement compared with the unexpectedly high utilities for HCC. The beneficial effects of transplantation on health utility may have been underestimated because most studies did not distinguish between the first year post-transplant—which may be associated with a lower quality of life owing to medication side effects and potential complications
Cost effectiveness of daclatasvir/asunaprevir versus peginterferon/ribavirin and protease inhibitors for the treatment of hepatitis C genotype 1B in Chile.
Despite their small sample sizes (n = 5 and n = 1), these studies suggest that utilities in the first year may be substantially lower than in subsequent years (eg, 0.46 vs 0.80 EQ-5D-3L utility reported by Pol et al
). A longitudinal study reporting CHC utilities before and after transplantation could help to clarify its impacts on health utility, but no such study was identified.
Utilities for patients on interferon-free treatment were higher than for interferon-based treatment while representing a minimal decrement compared with pretreatment utilities. These new DAA-based interferon-free regimens are thought to be associated with minimal side effects, representing a significant improvement over interferon-based therapy. Our analysis supports this expectation.
Mean utilities for patients who achieved SVR were higher than for mild to moderate CHC patients by a clinically significant margin, whereas utilities for those who failed treatment were substantially lower. These differences could be attributable to CHC symptoms or knowledge of treatment outcome. Notably, studies tended to measure patients’ utilities 24 weeks after the end of treatment; the long-term effects of SVR on health utility, and the long-term impacts of failed therapy, have not been well studied and should be explored in future research.
Impacts of Methodological Factors
Our findings showed that enrollment in an experimental study such as a randomized controlled trial (RCT) is an important predictor of higher utilities. Although differences in health outcomes between RCT participants and the general population have been noted in previous literature,
we were unable to find studies that examined the relationship between study design and health utility. This may be an important finding that could inform the use and interpretation of health utility measurements.
The reasons behind this observation are likely tied to a more pronounced “healthy volunteer effect”
in RCTs. Research suggests that narrow inclusion/exclusion criteria; recruitment targeting healthier individuals; self-selection by individuals with a higher income, education, or level of health; and self-exclusion by those in relatively poor health who meet inclusion criteria are some of the key reasons for pronounced differences in the health of patients enrolled in RCTs compared with the general population.
In our analysis, the comparator was observational study participants rather than a general population; the difference might be even greater if a general population was compared. This could limit the generalizability of utilities collected in RCTs, and consequently, their results should be used with caution in cost-effectiveness analyses. This may be especially true for diseases such as CHC which disproportionately affect marginalized patients, who face barriers to participating in clinical research.
Instruments that directly measured a patient's preference for their current health (SG and TTO) produced higher utilities than most indirect instruments (which incorporate societal preferences). This suggests that CHC patients' valuation of their own health states may be higher than the general population's preferences for the same health states. These findings have been reported for other diseases and may be due to factors such as patients adapting to their health state.
Impacts of Sociodemographic Factors and Comorbidities
Women with CHC may have substantially lower utilities than men with CHC. This finding is consistent with quality-of-life research in CHC, although the reasons behind it are not well understood.
It has been suggested that women may experience stigma differently than men and may experience concerns that CHC or substance use issues will negatively affect their families.
Studies with a higher proportion of non-white patients had significantly lower utilities. In previous literature on other diseases, some of these racial differences in health-related quality of life have been explained by demographic and socioeconomic factors
; however, these data were not available to be analyzed in our review. These patients may also experience aspects of CHC such as stigma differently.
HIV coinfection caused a varying degree of reduction in health utility, potentially reflecting heterogeneity in this subpopulation. One study found extremely low utilities in HIV-positive patients on opioid substitution therapy who had missed multiple medical appointments, some of whom were coinfected with HCV
(not included in this review). By contrast, a study on HIV/HCV coinfected patients who may be in more regular contact with the healthcare system reported higher utilities.
Changes in quality of life, healthcare use, and substance use in HIV/hepatitis C coinfected patients after hepatitis C therapy: a prospective cohort study.
Table 4 contains predicted utility estimates from the meta-regression model, adjusted to each utility instrument and study design. These may be useful as model inputs for cost-effectiveness analyses because they avoid the challenges that can arise when the utilities in a model come from different study designs or utility instruments. For example, our predicted results for treatment-associated health states adjusted to an observational study design may be closer to real-world treatment-associated utilities than utilities from RCTs. Utilities for combined health states not generated in our analysis can be estimated using the multiplicative method outlined in the NICE Decision Support Unit's guidance document on using health utilities in decision models.
Our meta-regression model decreases the chances of finding logically inconsistent utilities from different studies and methods (eg, severe health states having higher health utilities than milder ones) by drawing on a large pool of studies and controlling for methodological factors. The model also provides estimates of uncertainty that can be incorporated into sensitivity analyses.
Conclusion
Chronic hepatitis C can have a profound effect on patients' health-related quality of life. Our findings provide a comprehensive summary of the available data on health utilities in CHC patients at different stages of disease. We also show that experimental study designs are associated with higher utilities—an effect that has not been previously documented, and has implications for the use and interpretation of health utility measurements. This work will facilitate future research to improve health outcomes for CHC patients and achieve elimination targets.
Acknowledgments
Source of financial support: This study was supported by the Canadian Institutes of Health Research (Operating Grant #137490 and #PJT-148970). Yasmin Saeed was supported by a Trainee Fellowship from the Canadian Network on Hepatitis C (CanHepC). CanHepC is funded by a joint initiative of the Canadian Institutes of Health Research (#NHC-142832) and the Public Health Agency of Canada. William W.L. Wong was supported by an Ontario Ministry of Research, Innovation, and Science Early Researcher Award. Murray Krahn was supported by a Tier 1 Canada Research Chair in Health Technology Assessment. These organizations were not involved in the design, conduct, or writing of this review.
The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration.
Health-related quality of life in HIV-infected and at-risk women: the impact of illicit drug use and hepatitis C on a community preference weighted measure.
Psychometric evaluation of the hepatitis C virus patient-reported outcomes (HCV-PRO) instrument: validity, responsiveness, and identification of the minimally important difference in a phase 2 clinical trial.
Health-related quality of life (HRQoL), health state, function and wellbeing of chronic HCV patients treated with interferon-free, oral DAA regimens: patient reported outcome (PRO) results from the AVIATOR study.
Fatigue during treatment for hepatitis C virus: results of self-reported fatigue severity in two Phase IIb studies of simeprevir treatment in patients with hepatitis C virus genotype 1 infection.
Cost effectiveness of daclatasvir/asunaprevir versus peginterferon/ribavirin and protease inhibitors for the treatment of hepatitis C genotype 1B in Chile.
Health-related quality of life in genotype 1 treatment-naive chronic hepatitis C patients receiving telaprevir combination treatment in the ADVANCE study.
Changes in quality of life, healthcare use, and substance use in HIV/hepatitis C coinfected patients after hepatitis C therapy: a prospective cohort study.
32300-9&id=gr1.jpg){kind=link}
32300-9&id=gr2.jpg){kind=link}