- Email: [email protected]
Tackling the burden of the hepatitis C virus in the UK: characterizing and assessing the clinical and economic consequences
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Available online at www.sciencedirect.com
Public Health journal homepage: www.elsevier.com/puhe
Original Research
Tackling the burden of the hepatitis C virus in the UK: characterizing and assessing the clinical and economic consequences T. Ward a,*, J. Gordon a,b,c, H. Bennett a, S. Webster a, D. Sugrue a, B. Jones a, M. Brenner d, P. McEwan a,e a
Health Economics and Outcomes Research Ltd, Cardiff, UK Department of Public Health, University of Adelaide, Australia c School of Medicine, University of Nottingham, UK d UK HEOR, Bristol-Myers Squibb Pharmaceuticals Ltd, Uxbridge, UK e School of Human & Health Sciences, Swansea University, Swansea, UK b
article info
abstract
Article history:
Objectives: The hepatitis C virus (HCV) remains a significant public health issue. This study
Received 4 March 2016
aimed to quantify the clinical and economic burden of chronic hepatitis C in the UK,
Received in revised form
stratified by disease severity, age and awareness of infection, with concurrent assessment
3 August 2016
of the impact of implementing a treatment prioritization approach.
Accepted 5 August 2016
Study design and methods: A previously published back projection, natural history and costeffectiveness HCV model was adapted to a UK setting to estimate the disease burden of chronic hepatitis C and end-stage liver disease (ESLD) between 1980 and 2035. A published
Keywords:
meta-regression analysis informed disease progression, and UK-specific data informed
Back projection
other model inputs.
Public health
Results: At 2015, prevalence of chronic hepatitis C is estimated to be 241,487 with 22.20%,
Fibrosis
33.72%, 17.22%, 16.67% and 10.19% of patients in METAVIR stages F0, F1, F2, F3 and F4,
HCV
respectively, but is estimated to fall to 193,999 by 2035. ESLD incidence is predicted to peak
Chronic hepatitis C
in 2031. Assuming all patients are diagnosed and treatment is prioritized in F3 and F4 using highly efficacious direct-acting antiviral (DAA) regimens, a 69.85% reduction in ESLD incidence is predicted between 2015 and 2035, and the cumulative discounted medical expenditure associated with the lifetime management of incident ESLD events is estimated to be £1,202,827,444. Conclusions: The prevalence of chronic hepatitis C is expected to fall in coming decades; however, the ongoing financial burden is expected to be high due to an increase in ESLD incidence. This study highlights the significant costs of managing ESLD that are likely to be incurred without the employment of effective treatment approaches. © 2016 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Health Economics & Outcomes Research Ltd, 9 Oak Tree Court, Mulberry Drive, Cardiff Gate Business Park, CF23 8RS, UK. E-mail address: [email protected] (T. Ward). http://dx.doi.org/10.1016/j.puhe.2016.08.002 0033-3506/© 2016 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Introduction Patients with chronic hepatitis C are at risk of developing long-term, life-threatening sequelae, including decompensated cirrhosis and hepatocellular carcinoma (HCC), precursors to liver transplant, and death.1e3 While chronic hepatitis C can be successfully diagnosed and treated, significant numbers of patients are unaware of their infection due to the asymptomatic nature of the condition until the development of end-stage liver disease (ESLD); therefore, the expected financial burden of these undiagnosed cases is not well categorized, but is expected to be substantial.4 In the UK, the two major routes of HCV transmission have historically been the sharing of needles and paraphernalia among people who inject drugs (PWID) and transfusion of infected blood or blood products.5,6 Screening of blood products and heat inactivation of the virus have virtually eliminated the latter as a source of newly acquired HCV infection, and the HCV epidemic in the UK is now largely driven by injecting drug use.5,6 Latest estimates from UK public health bodies suggest that up to 90% of current HCV infections have occurred as a direct result of injection drug use,6 it is estimated that between 20% and 68% of PWID in the UK are infected with HCV, depending on geographical location and injection practices.4,7,8 Recent advances in the therapeutic landscape have led to the introduction of direct-acting antiviral (DAA) regimens with sustained virologic response (SVR) rates consistently >90%.9e12 These treatments are associated with improved safety profiles and reduced therapy durations compared to historical interferon-based regimens, which provides the potential for improved adherence.9e12 In principle, these therapies introduce the possibility of eradicating HCV; however, this depends not only on the effectiveness and acceptability of treatments, but on implementation approaches that are both clinically feasible and economically viable. Related research has suggested that concerns around the affordability of treating a large number of patients with chronic hepatitis C may be alleviated by prioritizing treatment in certain patient groups, such as those most likely to transmit infection to others.13 However, assessment of the impact of targeted implementation at a national level has not been undertaken. With this in mind, the principle aims of this research were to derive estimates of the prevalence of chronic hepatitis C in the UK, stratified by disease severity, age and awareness of infection, predict the onward presentation of ESLD and its economic burden, and assess the impact of implementing a targeted approach to treatment, prioritizing patients with advanced disease, on incidence of ESLD at the population level.
Methods Analysis plan Using a previously published back projection, natural history and cost-effectiveness HCV model14e18 adapted to a UK setting, a three-stage approach was taken:
43
Stage 1: The back-projection component of the model was used to predict the UK prevalence of chronic hepatitis C between 1980 and 2014 with constant infection rates applied to extrapolate to years 2015e2035. These prevalence estimates were then stratified by METAVIR fibrosis stage, age at infection and awareness of infection. Stage 2: Utilizing the results from stage 1 and the model's natural history component, the incidence of ESLD events (decompensated cirrhosis, HCC and liver transplant) and associated life expectancy and total lifetime costs were estimated over the period 2015e2035. Discounted and undiscounted results were reported at both the per-patient and population levels. Additionally, the total costs accumulated for the population were reported for 5-, 10- and 20-year horizons from 2015. Stage 3: The incidence of ESLD events and related mortality over the period 2015e2035 applying four different treatment implementation approaches was estimated.
Back-projection methodology Back projection is a mathematical methodology employed to estimate historical rates of disease incidence or prevalence subject to a known contemporary rate of disease progression or mortality. The underlying mathematical methodology applied remains independent of the specific disease area, with adaptation relying only on the accurate modelling of disease progression specific to each disease and geographical area. As such, the appropriateness of the methodology depends on the availability of robust input data and the accurate estimation of disease progression. Back projection techniques have been employed extensively to estimate historic HIV infection levels based on known AIDS incidence data.19 The technique has analogously been applied to estimating the incidence and prevalence of chronic hepatitis C in England,6,20 France,21 Australia22 and Taiwan15 by utilizing reported HCV-related HCC incidence data, together with an estimated incubation period, to calculate the total number of unobserved HCV infections that must have occurred in order to give rise to the observed number of HCC complications. Back projection can be implemented in a number of ways, ranging from statistical to deterministic.23e25 This study adopted an approach previously taken by Brookmeyer and Damiano24 to estimate HIV incidence, which has previously been applied to estimating the prevalence of chronic hepatitis C in Taiwan.15 The approach adopted was tailored to the UK setting in order to estimate HCV incidence and prevalence by utilizing UK-specific HCC data and a previously published disease progression model, validated to UK data.18 This approach estimates the UK incidence and prevalence of chronic hepatitis C via the maximization of Poisson likelihood, utilizing a smoothed expectation maximization (EM) algorithm. The approach assumes that the incidence of chronic hepatitis C occurs as an independent random process, where the number of individuals developing chronic hepatitis C in period t is denoted Nt and the number of subjects with HCV-related HCC is denoted Yt, where t ¼ 1,2,…,T. Specifically, t represents 1 calendar year and T the last year of recorded HCC included in the analysis (2013).
44
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
In year t, the expected number of HCV-related HCC cases is modelled by Equation (1) and the mean number of chronic hepatitis C cases then calculated by Equation (2). Assuming each infection is an independent Poisson variate, the loglikelihood function is represented by Equation (3). E½YjN1 ; N2 ; …; Nt ¼
t X
Ni fti
(1)
i¼1
mt ¼
t X
li fti
where; mi ¼ E½Yt and li ¼ E½Ni :
(2)
i¼1
log LðljyÞ ¼
Subsequently, the back projection aspect of the model utilizes this information to estimate the number of historical infections required to incur observed levels of HCC incidence. After the model has estimated the number of historical infections in each given year, the natural history aspect of the model is used to predict the natural disease progression of each set of annual incident infections. Results are accumulated, and an explicit estimate of the number of patients in each disease stage over the model time horizon is made available.
Data T X
yt logðmt Þ mt
(3)
t¼1
The EM algorithm was coded in Visual Basic for Applications within Microsoft Excel. The time between infection and development of HCC in those with chronic hepatitis C was estimated using the natural history component of the model. The model iterates in annual cycles, with subjects progressing through METAVIR fibrosis stages F0 (no fibrosis), F1 (portal fibrosis with no septa), F2 (portal fibrosis with few septa), F3 (portal fibrosis with numerous septa) and F4 (compensated cirrhosis), and on to ESLD events and death. The model structure is illustrated in Fig. 1. Progression through fibrosis stages is controlled via stagespecific transition probabilities, whilst progression to ESLD events is controlled using published static transition rates (Supplementary Table 1). All-cause mortality from fibrosis stages F0eF4 was incorporated through the use of UK-specific annual mortality estimates,26 and mortality from the decompensated cirrhosis, HCC and liver transplant states was modelled using according to published transition rates.27 As such, estimates of the proportion of patients reaching a specific disease stage at each future time point are available; consequently, estimates of time to disease progression are available for each specific stage of the disease.
Published UK rates of hospital admissions for HCC in patients infected with HCV were sourced. Data for the period 1998e2013 were available for England and Wales; however, data for Northern Ireland and Scotland were limited to the periods 2009e2013 and 1998e2012, respectively. Therefore, these were extrapolated to provide estimates of HCC incidence for 2013 (Scotland) and 1998e2009 (Northern Ireland).28 Missing values were linearly interpolated and a smoothing function was applied to the raw incidence data, as the back projection process represents an ‘ill-posed inverse problem’. Supplementary Fig. 1 presents a comparison of the unadjusted HCC data, the smoothed trajectory and HCC incidence predicted by the back projection model between 1980 and 2013. As back projection provides little insight into recent infection rates due to the long length of time between infection and progression to ESLD, it was required to calculate a minimum incident rate of chronic infection. In the absence of data for the whole of the UK, this was based on available data for Scotland, scaled to the UK: the incidence rate of chronic hepatitis C in Scottish PWID (419 per 16,00029) as of 2012 was scaled to the estimated number of PWID in the UK (122,894).30 As PWID account for 90% of incident infections in the UK,6 the derived figure was inflated to account for all incident infections, giving rise to an annual incidence of 3576. The age at infection was set to 24 years, based on data from the UK studies included in the meta-regression analysis that
Fig. 1 e Flow diagram of the HCV simulation model. HCV, hepatitis C virus.
45
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
provided the fibrosis stage-specific transition probabilities used within the disease progression model.31e37 In order to partition chronic hepatitis C prevalence by awareness of infection, it was assumed that 58% of those infected in the UK had been diagnosed at 2014; this is also based on Scottish data, in the absence of UK-wide data.6 The use of Scottish data was deemed the most appropriate approximation of UK-wide data due to their shared use of the UKs health system (the National Health Service [NHS]) and their employment of broadly similar and consistent health-related guidelines and policies (the value used was varied between 48% and 68% in sensitivity analysis). To inform the distribution of awareness of infection across fibrosis stages, prevalence output from the model was combined with the distribution of patients across fibrosis stages observed in the meta-regression analysis, assuming such data provide a proxy for those identified nationwide.31e37 Such an adjustment was undertaken based on the assumption that the UK studies used to inform the Thein et al.31 meta-regression analysis selected patients for inclusion at random and that the resultant distribution of patients across fibrosis stages is more representative of diagnoses than arbitrarily assuming an equal distribution. The proportion of patients diagnosed in each fibrosis stage was subsequently inflated so that it replicated the absolute distribution of patients across fibrosis stages, as estimated by the pooled UK-specific studies incorporated in the meta-regression analysis; the overall diagnosis rate of 58% was retained. An annual identification rate of 3.02% was applied (independently of HCV genotype or fibrosis stage) to undiagnosed infections from 2015 onwards; this rate was calculated based on the derived chronic hepatitis C prevalence estimate, and the annual average number of observed laboratory confirmed cases of HCV in the UK between 2010 and 2012,6 with the assumption that 75% of these cases resulted in chronic infection.1,38 Incorporation of treatment was restricted to HCV genotypes 1 and 3, due to their dominance of the prevalent population in the UK (47% and 44%, respectively).6 Treatment rates over the period 2002e2014 were assumed to remain constant, based on observed rates from 2006e2011,6 and treatment regimens modelled were based on predominating standard of care regimens at appropriate time points (referred from this point forwards as ‘historical standard of care’); i.e., pegylated interferon-alfa and ribavirin for both HCV genotypes 1 and 3 until 2012, when telaprevir (used in combination with peg-
interferon and ribavirin) was introduced for the treatment of HCV genotype 1. HCV genotype- and fibrosis stage-specific SVR rates applied can be found in Table 1. From 2015 on, 5000 patients were assumed to be treated per year, and a number of alternative treatment strategies were compared including the use of a hypothetical ‘novel DAA regimen’, i.e., one with high efficacy (assumed SVR of 95%, independent of genotype and fibrosis stage). Retreatment of those failing to achieve SVR was not included in the analysis. Annual costs of decompensated cirrhosis (£11,729), HCC (£10,452) and liver transplant (initial event cost of £47,311 applied in incident year and maintenance cost of £1781 applied in subsequent years) were derived from UK-specific sources27,39 and inflated to 2012/13 values using the NHS Hospital and Community Health Service index.40
Treatment implementation scenarios Scenario 1: Continuation of historical standard of care (baseline) for all diagnosed patients (F0eF4), with distribution of treatment applied proportionally across METAVIR fibrosis stages as determined by prevalence estimates. Scenario 2: Introduction of a novel DAA regimen (upscaled linearly to 100% from 2015 to 2017) for diagnosed patients with METAVIR scores of F3 and F4, and historical standard of care for all other diagnosed patients (F0eF2), with distribution of treatment applied proportionally across METAVIR fibrosis stages as determined by prevalence estimates. Scenario 3: Introduction of a novel DAA regimen (upscaled linearly to 100% from 2015 to 2017) for all diagnosed patients (F0eF4), with treatment preferably allocated to those with more advanced disease (METAVIR fibrosis stages F3 and F4). Scenario 4: Introduction of a novel DAA regimen (upscaled linearly to 100% from 2015 to 2017), assuming that all prevalent patients have been diagnosed and with treatment preferably allocated to those with more advanced disease (METAVIR fibrosis stages F3 and F4).
Results The back projection analysis produced a chronic hepatitis C prevalence estimate of 241,487 in 2015 (Fig. 2). The trend
Table 1 e Efficacy data for modelled chronic hepatitis C treatment regimens. METAVIR fibrosis stage
HCV genotype 1 TVR þ PR
F0 F1 F2 F3 F4
HCV genotype 3 PR
SVR (%)
Source
SVR (%)
81a 81a 75 62 62
ADVANCE 47
57b 57b 57b 41b 41b
PR Source NV15942
48
SVR (%)
Source
70b 70b 70b 49b 49b
Shiffman et al. 49
HCV, hepatitis C virus, P, pegylated interferon-alfa; R, ribavirin; SVR, sustained virologic response; TVR, telaprevir. a Patients with ‘no or minimal fibrosis’ were pooled in the ADVANCE study.47 b Patients in the NV15942 48 and Shiffman et al. 49 studies were pooled by ‘with cirrhosis or bridging fibrosis’ and ‘without cirrhosis or bridging fibrosis’.
46
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Fig. 2 e Predicted chronic hepatitis C prevalence, stratified by METAVIR fibrosis stage.
indicated that the most rapid period of growth occurred during the 1980s. At 2015, it was estimated that 53,603 (22.20%), 81,422 (33.72%), 41,588 (17.22%), 40,266 (16.67%), and 24,608 (10.19%) of those chronically infected were in METAVIR fibrosis stages F0, F1, F2, F3 and F4, respectively. However, while the overall prevalence of chronic hepatitis C is predicted to decline to 193,999 by 2035, the proportion of patients with more advanced disease (F3 and F4) is expected to increase, resulting in fibrosis stage-specific prevalence estimates of 42,038 (21.67%), 52,478 (27.05%), 27,586 (14.22%), 35,683 (18.39%) and 36,214 (18.67%) in F0, F1, F2, F3 and F4, respectively. Patients in more advanced disease stages were estimated to be older, with a larger proportion aware of their infection (Fig. 3). At 2015, it was estimated that the average age of patients in METAVIR fibrosis stage F0 was 38 years, increasing to 53 years in those in F4. Over the subsequent 20 years, it was predicted that the average age of patients in F0 would decrease to 36 years, whilst the average age of patients with fibrosis stage F4 would increase to 69 years due to disease progression and the incidence of HCV infection being predominantly in younger patients. Similar patterns were observed when considering the proportion of patients aware of their infection; for patients in F0, the proportion of patients aware of their infection was expected to decrease from 55.51% to 25.42% between 2015 and 2035, whilst the opposite was observed in patients in F4, with an increase from 58.05% to 69.97% over the same period. Assuming the continuation of historical standard of care treatments past 2015 (i.e., assuming novel DAA regimens are not made available), it was estimated that ESLD event incidence would peak in 2031 with 1930 annual events, decreasing to 1883 in 2035 (Table 2). Due to the increasing average age of patients, the predicted undiscounted life expectancy of an
average incident ESLD patient decreased from 20.57 to 18.52 between 2015 and 2035; similarly, the expected undiscounted lifetime cost associated with these patients decreased from £76,785 to £69,329 over the same period. It was estimated that the total undiscounted medical expenditure associated with the lifetime management of patients progressing to ESLD between 2015 and 2035 would accumulate to £2,647,335,555, with an equivalent discounted value of £1,202,827,444. From 2015, the total costs associated with ESLD complications for the UK population were estimated at approximately £46.3, £207.7 and £915.5 million, over 5-, 10- and 20-year horizons, respectively. Fig. 4 shows the effect on ESLD incidence and mortality of the four treatment implementation scenarios assessed. The incidence of ESLD-related mortality assuming the continuation of historical standard of care was estimated to rise from 859 in 2015 to 1792 in 2035. The introduction of a novel DAA regimen with an SVR of 95% for diagnosed patients in F3 or F4 resulted in an average reduction in ESLD event incidence of 8.95% over the period 2015e2035 and a similar reduction of 6.17% in ESLD-related mortality over the same period. Prioritizing patients with more advanced disease resulted in a significantly greater reduction in the incidence of ESLD events and related mortality, with average reductions of 43.91% and 35.70%, respectively, compared to the historical standard of care scenario. However, ESLD event incidence plateaued at approximately 988 patients per year due to the inability to treat those without knowledge of their infection and the assumption that retreatment of those failing to achieve SVR is not undertaken. Assuming all patients were diagnosed, and combining this with the prioritization of treatment in those with advanced disease, resulted in reductions of ESLD event incidence of 69.85% and ESLD mortality of 56.56%, when compared to the historical standard of care scenario over the period 2015e2035. Varying the proportion of patients initially
47
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Fig. 3 e Estimated average age and distribution of awareness of infection in patients with chronic hepatitis C, stratified by METAVIR fibrosis stage.
Table 2 e Predicted incidence of ESLD events and associated life expectancy and lifetime costs. Year
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 Total
ESLD incidencea
1289 1363 1433 1500 1564 1623 1677 1726 1770 1809 1842 1870 1892 1909 1921 1928 1930 1925 1914 1901 1883 36,670
Undiscounted Per-patient
Discounted Total cohort
Per-patient
Total cohort
Life expectancy
Lifetime cost (£)
Life years
Lifetime cost (£)
Life expectancy
Lifetime cost (£)
Life years
Lifetime cost (£)
20.57 20.34 20.34 20.11 20.11 19.87 19.87 19.62 19.62 19.36 19.36 19.09 19.09 19.09 18.81 18.81 18.81 18.52 18.52 18.52 18.52 e
76,785 75,969 75,969 75,120 75,120 74,241 74,241 73,324 73,324 72,376 72,376 71,395 71,395 71,395 70,381 70,381 70,381 69,329 69,329 69,329 69,329 e
26,519 27,722 29,157 30,171 31,450 32,244 33,320 33,861 34,717 35,014 35,664 35,701 36,122 36,437 36,139 36,270 36,294 35,653 35,456 35,206 34,870 e
99,004,877 103,527,191 108,885,383 112,705,305 117,482,112 120,485,165 124,506,683 126,568,296 129,767,817 130,923,111 133,352,798 133,536,221 135,109,570 136,288,670 135,221,811 135,710,148 135,800,401 133,449,489 132,714,533 131,775,896 130,520,079 2,647,335,555
13.61 13.07 12.63 12.13 11.72 11.24 10.86 10.42 10.07 9.65 9.32 8.93 8.63 8.34 7.99 7.72 7.46 7.14 6.89 6.66 6.44 e
47,815 45,928 44,375 42,606 41,165 39,509 38,173 36,618 35,380 33,924 32,777 31,412 30,350 29,324 28,089 27,139 26,221 25,102 24,253 23,433 22,641 e
17,552 17,816 18,105 18,193 18,323 18,246 18,217 17,983 17,815 17,456 17,179 16,710 16,335 15,920 15,345 14,880 14,386 13,736 13,198 12,662 12,117 e
61,651,761 62,588,598 63,601,890 63,923,165 64,379,160 64,118,407 64,017,905 63,208,019 62,614,353 61,365,548 60,390,702 58,753,300 57,435,307 55,977,338 53,967,095 52,330,426 50,594,423 48,318,523 46,427,454 44,540,185 42,623,883 1,202,827,444
ESLD, end-stage liver disease. a Includes decompensated cirrhosis, hepatocellular carcinoma and liver transplant.
48
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Fig. 4 e Estimated annual incidence of ESLD events (decompensated cirrhosis, HCC and liver transplant) and related mortality according to treatment implementation strategy. ESLD, end-stage liver disease; HCC, hepatocellular carcinoma; DAA, direct-acting antiviral; SOC, standard of care.
identified (58%) by 10% in either direction resulted in little change (<1%) in estimates of ESLD incidence and associated mortality across all treatment scenarios, with the exception of scenario 3. In this scenario, a decrease in the proportion of patients initially identified resulted in average reductions in the incidence of ESLD events and related mortality of 36.43% and 29.42%, respectively, whereas an increase led to reductions of 51.26% and 41.81%, respectively, compared to the historical standard of care scenario.
Discussion This study utilized a contemporary HCV natural history disease progression model to produce estimates of chronic hepatitis C prevalence that are consistent with previously predicted estimates.6,20 The predicted prevalence of HCC events at 2012 is supported by observational record-linkage data on liver-related sequelae.6 This study has provided
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
further granularity by partitioning according to METAVIR fibrosis stage, age and knowledge of infection status. This information is valuable as fibrosis stage and age are predictive of both complication development and treatment efficacy; those at a higher age and with a greater degree of fibrosis (i.e., METAVIR fibrosis stages F3 and F4) are at highest risk of progression to ESLD and have historically been less likely to achieve SVR with conventional treatments.31 Improved knowledge of the prevalent chronic hepatitis C population characteristics will better equip policy makers to consider effective treatment prioritization. With recent advances in the therapeutic landscape and introduction of new therapies with SVRs consistently >90%, even in the most difficult to treat patients, cure of any diagnosed individual is becoming a realistic achievement. The introduction of such treatments, combined with clinicians' knowledge of prognostic factors in individual patients, may allow for further optimization of approaches to treatment, in terms of resource utilization and treatment success. European clinical guidelines advocate the prioritization of treatment for patients with significant fibrosis or cirrhosis (METAVIR score F3 or F4).3 This study is supportive of this approach as it has demonstrated that introducing treatment with a novel DAA regimen capable of achieving a 95% SVR and prioritizing treatment in more advanced patients has the potential to significantly reduce onward ESLD event rates. However, the high cost of newly available DAA regimens has been regarded as a barrier to their use and attaining the benefits associated with their improved efficacy.9e12 This study highlights the significant costs of managing ESLD complications without the use of DAAs or the prioritization of treatment, at both patient and population levels. Moreover, this study suggests that the benefits of implementation strategies that target patients with advanced disease can only be fully realized with the use of newer therapies with higher rates of SVR and improved tolerability. However, the potential effect of such prioritization is limited as issues remain in relation to treatment uptake and numbers lost to follow-up along the HCV treatment cascade. Only 22% of PWID in England who stated seeing a specialist nurse or doctor regarding their infection in 2013 reported receiving any HCV-related medication,28 and similar levels of imperfect follow-up (27%) have been shown to greatly reduce the real-world effectiveness of HCV therapy.41 Even modestly effective interventions to improve follow-up, such as peer navigators or integrated case management, would likely provide significant value in terms of improving the costeffectiveness of interventions in the real world. The current annual rate of treatment uptake in the UK is approximately 3%,28 a rate shown to have little impact on the incidence of infection, or on the incidence of ESLD and HCC complications in the near term.18,27,42e44 However, previously published studies have demonstrated that even modest increases in the rate of treatment uptake may have a significant effect upon the prevalence of hepatitis C in the UK.42,45 As such, this study did not attempt to quantify the relationship between treatment uptake and the burden of ESLD but endeavoured to approximate the relationship between various targeted
49
treatment strategies and their effect upon the burden of ESLD, subject to credible assumptions made in relation to modelling the UK HCV population. Further, in order to reduce the future burden of ESLD, it is critical that those chronically infected are identified and successfully treated. Diagnostic challenges remain in the UK, with significant numbers unaware of their infection status. Additional complexity in identifying those infected relates to the high prevalence (~90%) of PWID amongst incident infections, a subgroup of patients that do not commonly access health care via traditional settings and are often associated with poverty, homelessness and an unwillingness to engage services.46,47 Combined with a relatively high degree of contraindications in this patient group, including psychiatric disorders, and alcohol and substance addiction, there are considerable challenges to the successful identification, engagement and monitoring of these patients.47 There are inherent limitations to the modelling approach presented. The back projection methodology requires accurate estimates of HCC end points; the source data used for this analysis were routine hospital episode statistics, which are subject to potential under-reporting and incorrect use of diagnostic codes. The degree of bias associated with these limitations is difficult to quantify; however, their effect may result in the under prediction of HCV-related ESLD. Due to the nature of the back projection methodology, predictions of prevalence are largely dependent upon the HCV disease transition rates utilized in the natural history model, which are subject to a high degree of uncertainty due to the small numbers of studies from which they are derived.48 Though such analyses are subject to limitations, their inference of statistical relationships is a valuable tool in economic evaluation and to inform decision-making. In conclusion, despite the expected fall in the prevalence of chronic hepatitis C in coming decades, the overall burden is estimated to substantially increase due to the rising incidence of ESLD. A targeted treatment approach that prioritizes the use of novel, highly effective DAA regimens in those patients at more advanced stages is likely to lead to a considerable reduction in ESLD rates. This impact may be augmented via improved rates of diagnosis and treatment initiation. Given the large number of emerging HCV therapies, a preoccupation with the relative cost-effectiveness of these treatments should arguably be replaced with an emphasis on identifying the most appropriate treatment for individual patients. Such a strategy is suggested to improve treatment uptake, adherence and, ultimately, success in tackling the population burden of HCV, ensuing in the reduced health and economic burden of HCV-related ESLD complications.
Author statements Ethical approval Not required, as the manuscript describes an economic analysis using publicly available data.
50
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Funding Writing and research was funded by Bristol-Myers Squibb Pharmaceuticals Ltd; however, the publication of study results is not contingent on the sponsor's approval or censorship of the manuscript.
Competing interests Thomas Ward, Phil McEwan, Jason Gordon, Hayley Bennett Wilton, Beverley Jones, Daniel Sugrue and Samantha Webster are employees of Health Economics and Outcomes Research Ltd, who received funding from Bristol-Myers Squibb Pharmaceuticals Ltd. Michael Brenner is an employee of the study funder.
references
1. Chen SL, Morgan TR. The natural history of hepatitis C virus (HCV) infection. Int J Med Sci 2006;3(2):47e52. 2. Kanwal F, Hoang T, Kramer JR, Asch SM, Goetz MB, Zeringue A, et al. Increasing prevalence of HCC and cirrhosis in patients with chronic hepatitis C virus infection. Gastroenterology 2011;140(4):1182e8. 3. European Association For The Study Of The Liver. Clinical practice guidelines: EASL recommendations on treatment of hepatitis C 2015. J Hepatol 2015;62(4). 4. Public Health England. Data tables of the unlinked anonymous monitoring survey of HIV and hepatitis in people who inject drugs. 2013. Available from: http://webarchive.nationalarchives.gov. uk/20140629102627/http:/hpa.org.uk/web/ HPAweb&HPAwebStandard/HPAweb_C/1202115519183. (last accessed February 2015). 5. Goldberg D, Anderson E, Hepatitis C. Who is at risk and how do we identify them? J Viral Hepat 2004;11(s1):12e8. 6. Public Health England. Health Protection Scotland, Public Health Wales, Public Health Agency. Hepatitis C in the UK: 2015 report. 2015. Available from: https://www.gov.uk/government/ publications/hepatitis-c-in-the-uk. (last accessed 11 November 2015). 7. Wiessing L, Ferri M, Grady B, Kantzanou M, Sperle I, Cullen KJ, et al. Hepatitis C virus infection epidemiology among people who inject drugs in Europe: a systematic review of data for scaling up treatment and prevention. PloS One 2014;9(7):e103345. 8. Martin NK, Foster GR, Vilar J, Ryder S, E Cramp M, Gordon F, et al. HCV treatment rates and sustained viral response among people who inject drugs in seven UK sites: real world results and modelling of treatment impact. J Viral Hepat 2014;22(4):399e408. 9. Sulkowski MS, Gardiner DF, Rodriguez-Torres M, Reddy KR, Hassanein T, Jacobson I, et al. Daclatasvir plus sofosbuvir for previously treated or untreated chronic HCV infection. N Engl J Med 2014;370(3):211e21. 10. Lawitz E, Mangia A, Wyles D, Rodriguez-Torres M, Hassanein T, Gordon SC, et al. Sofosbuvir for previously untreated chronic hepatitis C infection. N Engl J Med 2013;368(20):1878e87. 11. Lawitz E, Sulkowski MS, Ghalib R, Rodriguez-Torres M, Younossi ZM, Corregidor A, et al. Simeprevir plus sofosbuvir, with or without ribavirin, to treat chronic infection with hepatitis C virus genotype 1 in non-responders to pegylated interferon and ribavirin and treatment-naive patients: the COSMOS randomised study. Lancet 2014;384(9956):1756e65.
12. Lawitz E, Poordad FF, Pang PS, Hyland RH, Ding X, Mo H, et al. Sofosbuvir and ledipasvir fixed-dose combination with and without ribavirin in treatment-naive and previously treated patients with genotype 1 hepatitis C virus infection (LONESTAR): an open-label, randomised, phase 2 trial. Lancet 2014;383(9916):515e23. 13. Innes H, Goldberg D, Dillon J, Hutchinson SJ. Strategies for the treatment of Hepatitis C in an era of interferon-free therapies: what public health outcomes do we value most? Gut 2014;64(11):1800e9. 14. McEwan P, Ward T, Yuan Y, Kim R, L'Italien G. The impact of timing and prioritization on the cost-effectiveness of birth cohort testing and treatment for hepatitis C virus in the United States. Hepatology 2013;58(1):54e64. 15. McEwan P, Ward T, Chen C-J, Lee M-H, Yang H-I, Kim R, et al. Estimating the incidence and prevalence of chronic hepatitis C infection in Taiwan using back projection. Value Health Regional Issues 2013;3:5e11. 16. McEwan P, Kim R, Yuan Y. Assessing the cost utility of response-guided therapy in patients with chronic hepatitis C genotype 1 in the UK using the MONARCH model. Appl Health Econ Health Policy 2013;11(1):53e63. 17. McEwan P, Ward T, Webster S, Yuan Y, Kalsekar A, Broglio K, et al. Estimating the long-term clinical and economic outcomes of Daclatasvir plus Asunaprevir in difficult-to-treat Japanese patients chronically infected with hepatitis C genotype 1b. Value Health Regional Issues 2014;3:136e45. 18. McEwan P, Ward T, Bennett H, Kalsekar A, Webster S, Brenner M, et al. Estimating the clinical and economic benefit associated with incremental improvements in sustained virologic response in chronic hepatitis C. PLoS One 2015;10(1):e0117334. 19. Bacchetti P, Segal MR, Jewell NP. Backcalculation of HIV infection rates. Stat Sci; 1993::82e101. 20. Sweeting MJ, De Angelis D, Brant LJ, Harris HE, Mann AG, Ramsay ME. The burden of hepatitis C in England. J Viral Hepat 2007;14(8):570e6. 21. Deuffic S, Buffat L, Poynard T, Valleron AJ. Modeling the hepatitis C virus epidemic in France. Hepatology 1999;29(5):1596e601. 22. Razali K, Thein HH, Bell J, Cooper-Stanbury M, Dolan K, Dore G, et al. Modelling the hepatitis C virus epidemic in Australia. Drug Alcohol Depend 2007;91(2):228e35. 23. Becker NG, Watson LF, Carlin JB. A method of non-parametric back-projection and its application to AIDS data. Stat Med 1991;10(10):1527e42. 24. Brookmeyer R, Damiano A. Statistical methods for short-term projections of AIDS incidence. Stat Med 1989;8(1):23e34. 25. Isham V. Estimation of the incidence of HIV infection. Philosophical Trans R Soc Lond Biol Sci 1989;325(1226):113e21. 26. UK office for National Statistics. England and Wales life tables: 1980e82 to 2012e2014. Available from: http://www. ons.gov.uk/ons/taxonomy/index.html?nscl¼LifeþTables-tabdata-tables. (last accessed 16 October 2015). 27. Martin NK, Vickerman P, Miners A, Foster GR, Hutchinson SJ, Goldberg DJ, et al. Cost-effectiveness of hepatitis C virus antiviral treatment for injection drug user populations. Hepatology 2012;55(1):49e57. 28. Public Health England. Hepatits C in the UK: 2014 report. 2014. Available from: https://www.gov.uk/government/publications/ hepatitis-c-in-the-uk. (last accessed October 2014). 29. Palmateer NE, Hutchinson SJ, McLeod A, Codere G, Goldberg DJ. Comparison of deaths related to hepatitis C and AIDS in Scotland. J Viral Hepat 2007;14(12):870e4. 30. European Monitoring Centre for Drugs and Drug additicon. European drug report 2014. 2015. Available from: http://www. emcdda.europa.eu/edr2014. (last accessed 2 March 2015).
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
31. Thein HH, Yi Q, Dore GJ, Krahn MD. Estimation of stagespecific fibrosis progression rates in chronic hepatitis C virus infection: a meta-analysis and meta-regression. Hepatology 2008;48(2):418e31. 32. Forrest EH, Thorburn D, Spence E, Oien KA, Inglis G, Smith CA, et al. Polymorphisms of the renineangiotensin system and the severity of fibrosis in chronic hepatitis C virus infection. J Viral Hepat 2005;12(5):519e24. 33. Freeman A, Hamid S, Morris L, Vowler S, Rushbrook S, Wight DGD, et al. Improved detection of hepatocyte proliferation using antibody to the pre-replication complex: an association with hepatic fibrosis and viral replication in chronic hepatitis C virus infection. J Viral Hepat 2003;10(5):345e50. 34. Mohsen AH, Murad S, Easterbrook PJ. Prevalence of hepatitis C in an ethnically diverse HIV-1-infected cohort in south London. HIV Med 2005;6(3):206e15. 35. Ryder SD. Progression of hepatic fibrosis in patients with hepatitis C: a prospective repeat liver biopsy study. Gut 2004;53(3):451e5. 36. Wong V, Caronia S, Wight D, Palmer CR, Petrik J, Britton P, et al. Importance of age in chronic hepatitis C virus infection. J Viral Hepat 1997;4(4):255e64. 37. Patel K, Norris S, Lebeck L, Feng A, Clare M, Pianko S, et al. HLA class I allelic diversity and progression of fibrosis in patients with chronic hepatitis C. Hepatology 2006;43(2):241e9. 38. Micallef JM, Kaldor JM, Dore GJ. Spontaneous viral clearance following acute hepatitis C infection: a systematic review of longitudinal studies. J Viral Hepat 2006;13(1):34e41. 39. Shepherd J, Jones J, Hartwell D, Davidson P, Price A, Waugh N. Interferon alpha (pegylated and non-pegylated) and ribavirin for the treatment of mild chronic hepatitis C: a systematic review and economic evaluation. Health Technol Assess Winch Engl 2007;11(11). 1e205, iii. 40. Personal Social Services Research Unit. Unit costs of health & social care 2013. In: Section 16.2: the hospital & community health services (HCHS) index; 2013. 41. Linas BP, Barter DM, Leff JA, Assoumou SA, Salomon JA, Weinstein MC, et al. The hepatitis C cascade of care: identifying priorities to improve clinical outcomes. PloS One 2014;9(5):e97317. 42. Martin NK, Vickerman P, Foster GR, Hutchinson SJ, Goldberg DJ, Hickman M. Can antiviral therapy for hepatitis C
43.
44.
45.
46.
47.
48.
49.
51
reduce the prevalence of HCV among injecting drug user populations? A modeling analysis of its prevention utility. J Hepatol 2011;54(6):1137e44. Martin NK, Vickerman P, Grebely J, Hellard M, Hutchinson SJ, Lima VD, et al. Hepatitis C virus treatment for prevention among people who inject drugs: modeling treatment scale-up in the age of direct-acting antivirals. Hepatology 2013;58(5):1598e609. Martin NK, Hickman M, Hutchinson SJ, Goldberg DJ, Vickerman P. Combination interventions to prevent HCV transmission among people who inject drugs: modeling the impact of antiviral treatment, needle and syringe programs, and opiate substitution therapy. Clin Infect Dis 2013;57(Suppl. 2):S39e45. Bennett H, McEwan P, Sugrue D, Kalsekar A, Yuan Y. Assessing the long-term impact of treating hepatitis C virus (HCV)-Infected people who inject drugs in the UK and the relationship between treatment uptake and efficacy on future infections. PLoS One 2015;10(5):e0125846. Harris M, Rhodes T. Hepatitis C treatment access and uptake for people who inject drugs: a review mapping the role of social factors. Harm Reduct J 2013;10:7. World Health Organization. Barriers and facilitators to hepatitis C treatment for people who inject drugs. A Qual study. 2012. Available from: http://www.euro.who.int/__data/ assets/pdf_file/0011/179750/Barriers-and-facilitators-tohepatitis-C-treatment-for-PWID-A-qualitative-study-June2012-rev-5.pdf. (last accessed 2 December 2015). Townsend R, McEwan P, Kim R, Yuan Y. Structural frameworks and key model parameters in cost-effectiveness analyses for current and future treatments of chronic hepatitis C. Value Health 2011;14(8):1068e77. R, Shiffman ML, Suter F, Bacon BR, Nelson D, Harley H, Sola et al. Peginterferon alfa-2a and ribavirin for 16 or 24 weeks in HCV genotype 2 or 3. N Engl J Med 2007;357(2):124e34. http:// dx.doi.org/10.1056/NEJMoa066403.
Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.puhe.2016.08.002.
Available online at www.sciencedirect.com
Public Health journal homepage: www.elsevier.com/puhe
Original Research
Tackling the burden of the hepatitis C virus in the UK: characterizing and assessing the clinical and economic consequences T. Ward a,*, J. Gordon a,b,c, H. Bennett a, S. Webster a, D. Sugrue a, B. Jones a, M. Brenner d, P. McEwan a,e a
Health Economics and Outcomes Research Ltd, Cardiff, UK Department of Public Health, University of Adelaide, Australia c School of Medicine, University of Nottingham, UK d UK HEOR, Bristol-Myers Squibb Pharmaceuticals Ltd, Uxbridge, UK e School of Human & Health Sciences, Swansea University, Swansea, UK b
article info
abstract
Article history:
Objectives: The hepatitis C virus (HCV) remains a significant public health issue. This study
Received 4 March 2016
aimed to quantify the clinical and economic burden of chronic hepatitis C in the UK,
Received in revised form
stratified by disease severity, age and awareness of infection, with concurrent assessment
3 August 2016
of the impact of implementing a treatment prioritization approach.
Accepted 5 August 2016
Study design and methods: A previously published back projection, natural history and costeffectiveness HCV model was adapted to a UK setting to estimate the disease burden of chronic hepatitis C and end-stage liver disease (ESLD) between 1980 and 2035. A published
Keywords:
meta-regression analysis informed disease progression, and UK-specific data informed
Back projection
other model inputs.
Public health
Results: At 2015, prevalence of chronic hepatitis C is estimated to be 241,487 with 22.20%,
Fibrosis
33.72%, 17.22%, 16.67% and 10.19% of patients in METAVIR stages F0, F1, F2, F3 and F4,
HCV
respectively, but is estimated to fall to 193,999 by 2035. ESLD incidence is predicted to peak
Chronic hepatitis C
in 2031. Assuming all patients are diagnosed and treatment is prioritized in F3 and F4 using highly efficacious direct-acting antiviral (DAA) regimens, a 69.85% reduction in ESLD incidence is predicted between 2015 and 2035, and the cumulative discounted medical expenditure associated with the lifetime management of incident ESLD events is estimated to be £1,202,827,444. Conclusions: The prevalence of chronic hepatitis C is expected to fall in coming decades; however, the ongoing financial burden is expected to be high due to an increase in ESLD incidence. This study highlights the significant costs of managing ESLD that are likely to be incurred without the employment of effective treatment approaches. © 2016 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
* Corresponding author. Health Economics & Outcomes Research Ltd, 9 Oak Tree Court, Mulberry Drive, Cardiff Gate Business Park, CF23 8RS, UK. E-mail address: [email protected] (T. Ward). http://dx.doi.org/10.1016/j.puhe.2016.08.002 0033-3506/© 2016 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Introduction Patients with chronic hepatitis C are at risk of developing long-term, life-threatening sequelae, including decompensated cirrhosis and hepatocellular carcinoma (HCC), precursors to liver transplant, and death.1e3 While chronic hepatitis C can be successfully diagnosed and treated, significant numbers of patients are unaware of their infection due to the asymptomatic nature of the condition until the development of end-stage liver disease (ESLD); therefore, the expected financial burden of these undiagnosed cases is not well categorized, but is expected to be substantial.4 In the UK, the two major routes of HCV transmission have historically been the sharing of needles and paraphernalia among people who inject drugs (PWID) and transfusion of infected blood or blood products.5,6 Screening of blood products and heat inactivation of the virus have virtually eliminated the latter as a source of newly acquired HCV infection, and the HCV epidemic in the UK is now largely driven by injecting drug use.5,6 Latest estimates from UK public health bodies suggest that up to 90% of current HCV infections have occurred as a direct result of injection drug use,6 it is estimated that between 20% and 68% of PWID in the UK are infected with HCV, depending on geographical location and injection practices.4,7,8 Recent advances in the therapeutic landscape have led to the introduction of direct-acting antiviral (DAA) regimens with sustained virologic response (SVR) rates consistently >90%.9e12 These treatments are associated with improved safety profiles and reduced therapy durations compared to historical interferon-based regimens, which provides the potential for improved adherence.9e12 In principle, these therapies introduce the possibility of eradicating HCV; however, this depends not only on the effectiveness and acceptability of treatments, but on implementation approaches that are both clinically feasible and economically viable. Related research has suggested that concerns around the affordability of treating a large number of patients with chronic hepatitis C may be alleviated by prioritizing treatment in certain patient groups, such as those most likely to transmit infection to others.13 However, assessment of the impact of targeted implementation at a national level has not been undertaken. With this in mind, the principle aims of this research were to derive estimates of the prevalence of chronic hepatitis C in the UK, stratified by disease severity, age and awareness of infection, predict the onward presentation of ESLD and its economic burden, and assess the impact of implementing a targeted approach to treatment, prioritizing patients with advanced disease, on incidence of ESLD at the population level.
Methods Analysis plan Using a previously published back projection, natural history and cost-effectiveness HCV model14e18 adapted to a UK setting, a three-stage approach was taken:
43
Stage 1: The back-projection component of the model was used to predict the UK prevalence of chronic hepatitis C between 1980 and 2014 with constant infection rates applied to extrapolate to years 2015e2035. These prevalence estimates were then stratified by METAVIR fibrosis stage, age at infection and awareness of infection. Stage 2: Utilizing the results from stage 1 and the model's natural history component, the incidence of ESLD events (decompensated cirrhosis, HCC and liver transplant) and associated life expectancy and total lifetime costs were estimated over the period 2015e2035. Discounted and undiscounted results were reported at both the per-patient and population levels. Additionally, the total costs accumulated for the population were reported for 5-, 10- and 20-year horizons from 2015. Stage 3: The incidence of ESLD events and related mortality over the period 2015e2035 applying four different treatment implementation approaches was estimated.
Back-projection methodology Back projection is a mathematical methodology employed to estimate historical rates of disease incidence or prevalence subject to a known contemporary rate of disease progression or mortality. The underlying mathematical methodology applied remains independent of the specific disease area, with adaptation relying only on the accurate modelling of disease progression specific to each disease and geographical area. As such, the appropriateness of the methodology depends on the availability of robust input data and the accurate estimation of disease progression. Back projection techniques have been employed extensively to estimate historic HIV infection levels based on known AIDS incidence data.19 The technique has analogously been applied to estimating the incidence and prevalence of chronic hepatitis C in England,6,20 France,21 Australia22 and Taiwan15 by utilizing reported HCV-related HCC incidence data, together with an estimated incubation period, to calculate the total number of unobserved HCV infections that must have occurred in order to give rise to the observed number of HCC complications. Back projection can be implemented in a number of ways, ranging from statistical to deterministic.23e25 This study adopted an approach previously taken by Brookmeyer and Damiano24 to estimate HIV incidence, which has previously been applied to estimating the prevalence of chronic hepatitis C in Taiwan.15 The approach adopted was tailored to the UK setting in order to estimate HCV incidence and prevalence by utilizing UK-specific HCC data and a previously published disease progression model, validated to UK data.18 This approach estimates the UK incidence and prevalence of chronic hepatitis C via the maximization of Poisson likelihood, utilizing a smoothed expectation maximization (EM) algorithm. The approach assumes that the incidence of chronic hepatitis C occurs as an independent random process, where the number of individuals developing chronic hepatitis C in period t is denoted Nt and the number of subjects with HCV-related HCC is denoted Yt, where t ¼ 1,2,…,T. Specifically, t represents 1 calendar year and T the last year of recorded HCC included in the analysis (2013).
44
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
In year t, the expected number of HCV-related HCC cases is modelled by Equation (1) and the mean number of chronic hepatitis C cases then calculated by Equation (2). Assuming each infection is an independent Poisson variate, the loglikelihood function is represented by Equation (3). E½YjN1 ; N2 ; …; Nt ¼
t X
Ni fti
(1)
i¼1
mt ¼
t X
li fti
where; mi ¼ E½Yt and li ¼ E½Ni :
(2)
i¼1
log LðljyÞ ¼
Subsequently, the back projection aspect of the model utilizes this information to estimate the number of historical infections required to incur observed levels of HCC incidence. After the model has estimated the number of historical infections in each given year, the natural history aspect of the model is used to predict the natural disease progression of each set of annual incident infections. Results are accumulated, and an explicit estimate of the number of patients in each disease stage over the model time horizon is made available.
Data T X
yt logðmt Þ mt
(3)
t¼1
The EM algorithm was coded in Visual Basic for Applications within Microsoft Excel. The time between infection and development of HCC in those with chronic hepatitis C was estimated using the natural history component of the model. The model iterates in annual cycles, with subjects progressing through METAVIR fibrosis stages F0 (no fibrosis), F1 (portal fibrosis with no septa), F2 (portal fibrosis with few septa), F3 (portal fibrosis with numerous septa) and F4 (compensated cirrhosis), and on to ESLD events and death. The model structure is illustrated in Fig. 1. Progression through fibrosis stages is controlled via stagespecific transition probabilities, whilst progression to ESLD events is controlled using published static transition rates (Supplementary Table 1). All-cause mortality from fibrosis stages F0eF4 was incorporated through the use of UK-specific annual mortality estimates,26 and mortality from the decompensated cirrhosis, HCC and liver transplant states was modelled using according to published transition rates.27 As such, estimates of the proportion of patients reaching a specific disease stage at each future time point are available; consequently, estimates of time to disease progression are available for each specific stage of the disease.
Published UK rates of hospital admissions for HCC in patients infected with HCV were sourced. Data for the period 1998e2013 were available for England and Wales; however, data for Northern Ireland and Scotland were limited to the periods 2009e2013 and 1998e2012, respectively. Therefore, these were extrapolated to provide estimates of HCC incidence for 2013 (Scotland) and 1998e2009 (Northern Ireland).28 Missing values were linearly interpolated and a smoothing function was applied to the raw incidence data, as the back projection process represents an ‘ill-posed inverse problem’. Supplementary Fig. 1 presents a comparison of the unadjusted HCC data, the smoothed trajectory and HCC incidence predicted by the back projection model between 1980 and 2013. As back projection provides little insight into recent infection rates due to the long length of time between infection and progression to ESLD, it was required to calculate a minimum incident rate of chronic infection. In the absence of data for the whole of the UK, this was based on available data for Scotland, scaled to the UK: the incidence rate of chronic hepatitis C in Scottish PWID (419 per 16,00029) as of 2012 was scaled to the estimated number of PWID in the UK (122,894).30 As PWID account for 90% of incident infections in the UK,6 the derived figure was inflated to account for all incident infections, giving rise to an annual incidence of 3576. The age at infection was set to 24 years, based on data from the UK studies included in the meta-regression analysis that
Fig. 1 e Flow diagram of the HCV simulation model. HCV, hepatitis C virus.
45
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
provided the fibrosis stage-specific transition probabilities used within the disease progression model.31e37 In order to partition chronic hepatitis C prevalence by awareness of infection, it was assumed that 58% of those infected in the UK had been diagnosed at 2014; this is also based on Scottish data, in the absence of UK-wide data.6 The use of Scottish data was deemed the most appropriate approximation of UK-wide data due to their shared use of the UKs health system (the National Health Service [NHS]) and their employment of broadly similar and consistent health-related guidelines and policies (the value used was varied between 48% and 68% in sensitivity analysis). To inform the distribution of awareness of infection across fibrosis stages, prevalence output from the model was combined with the distribution of patients across fibrosis stages observed in the meta-regression analysis, assuming such data provide a proxy for those identified nationwide.31e37 Such an adjustment was undertaken based on the assumption that the UK studies used to inform the Thein et al.31 meta-regression analysis selected patients for inclusion at random and that the resultant distribution of patients across fibrosis stages is more representative of diagnoses than arbitrarily assuming an equal distribution. The proportion of patients diagnosed in each fibrosis stage was subsequently inflated so that it replicated the absolute distribution of patients across fibrosis stages, as estimated by the pooled UK-specific studies incorporated in the meta-regression analysis; the overall diagnosis rate of 58% was retained. An annual identification rate of 3.02% was applied (independently of HCV genotype or fibrosis stage) to undiagnosed infections from 2015 onwards; this rate was calculated based on the derived chronic hepatitis C prevalence estimate, and the annual average number of observed laboratory confirmed cases of HCV in the UK between 2010 and 2012,6 with the assumption that 75% of these cases resulted in chronic infection.1,38 Incorporation of treatment was restricted to HCV genotypes 1 and 3, due to their dominance of the prevalent population in the UK (47% and 44%, respectively).6 Treatment rates over the period 2002e2014 were assumed to remain constant, based on observed rates from 2006e2011,6 and treatment regimens modelled were based on predominating standard of care regimens at appropriate time points (referred from this point forwards as ‘historical standard of care’); i.e., pegylated interferon-alfa and ribavirin for both HCV genotypes 1 and 3 until 2012, when telaprevir (used in combination with peg-
interferon and ribavirin) was introduced for the treatment of HCV genotype 1. HCV genotype- and fibrosis stage-specific SVR rates applied can be found in Table 1. From 2015 on, 5000 patients were assumed to be treated per year, and a number of alternative treatment strategies were compared including the use of a hypothetical ‘novel DAA regimen’, i.e., one with high efficacy (assumed SVR of 95%, independent of genotype and fibrosis stage). Retreatment of those failing to achieve SVR was not included in the analysis. Annual costs of decompensated cirrhosis (£11,729), HCC (£10,452) and liver transplant (initial event cost of £47,311 applied in incident year and maintenance cost of £1781 applied in subsequent years) were derived from UK-specific sources27,39 and inflated to 2012/13 values using the NHS Hospital and Community Health Service index.40
Treatment implementation scenarios Scenario 1: Continuation of historical standard of care (baseline) for all diagnosed patients (F0eF4), with distribution of treatment applied proportionally across METAVIR fibrosis stages as determined by prevalence estimates. Scenario 2: Introduction of a novel DAA regimen (upscaled linearly to 100% from 2015 to 2017) for diagnosed patients with METAVIR scores of F3 and F4, and historical standard of care for all other diagnosed patients (F0eF2), with distribution of treatment applied proportionally across METAVIR fibrosis stages as determined by prevalence estimates. Scenario 3: Introduction of a novel DAA regimen (upscaled linearly to 100% from 2015 to 2017) for all diagnosed patients (F0eF4), with treatment preferably allocated to those with more advanced disease (METAVIR fibrosis stages F3 and F4). Scenario 4: Introduction of a novel DAA regimen (upscaled linearly to 100% from 2015 to 2017), assuming that all prevalent patients have been diagnosed and with treatment preferably allocated to those with more advanced disease (METAVIR fibrosis stages F3 and F4).
Results The back projection analysis produced a chronic hepatitis C prevalence estimate of 241,487 in 2015 (Fig. 2). The trend
Table 1 e Efficacy data for modelled chronic hepatitis C treatment regimens. METAVIR fibrosis stage
HCV genotype 1 TVR þ PR
F0 F1 F2 F3 F4
HCV genotype 3 PR
SVR (%)
Source
SVR (%)
81a 81a 75 62 62
ADVANCE 47
57b 57b 57b 41b 41b
PR Source NV15942
48
SVR (%)
Source
70b 70b 70b 49b 49b
Shiffman et al. 49
HCV, hepatitis C virus, P, pegylated interferon-alfa; R, ribavirin; SVR, sustained virologic response; TVR, telaprevir. a Patients with ‘no or minimal fibrosis’ were pooled in the ADVANCE study.47 b Patients in the NV15942 48 and Shiffman et al. 49 studies were pooled by ‘with cirrhosis or bridging fibrosis’ and ‘without cirrhosis or bridging fibrosis’.
46
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Fig. 2 e Predicted chronic hepatitis C prevalence, stratified by METAVIR fibrosis stage.
indicated that the most rapid period of growth occurred during the 1980s. At 2015, it was estimated that 53,603 (22.20%), 81,422 (33.72%), 41,588 (17.22%), 40,266 (16.67%), and 24,608 (10.19%) of those chronically infected were in METAVIR fibrosis stages F0, F1, F2, F3 and F4, respectively. However, while the overall prevalence of chronic hepatitis C is predicted to decline to 193,999 by 2035, the proportion of patients with more advanced disease (F3 and F4) is expected to increase, resulting in fibrosis stage-specific prevalence estimates of 42,038 (21.67%), 52,478 (27.05%), 27,586 (14.22%), 35,683 (18.39%) and 36,214 (18.67%) in F0, F1, F2, F3 and F4, respectively. Patients in more advanced disease stages were estimated to be older, with a larger proportion aware of their infection (Fig. 3). At 2015, it was estimated that the average age of patients in METAVIR fibrosis stage F0 was 38 years, increasing to 53 years in those in F4. Over the subsequent 20 years, it was predicted that the average age of patients in F0 would decrease to 36 years, whilst the average age of patients with fibrosis stage F4 would increase to 69 years due to disease progression and the incidence of HCV infection being predominantly in younger patients. Similar patterns were observed when considering the proportion of patients aware of their infection; for patients in F0, the proportion of patients aware of their infection was expected to decrease from 55.51% to 25.42% between 2015 and 2035, whilst the opposite was observed in patients in F4, with an increase from 58.05% to 69.97% over the same period. Assuming the continuation of historical standard of care treatments past 2015 (i.e., assuming novel DAA regimens are not made available), it was estimated that ESLD event incidence would peak in 2031 with 1930 annual events, decreasing to 1883 in 2035 (Table 2). Due to the increasing average age of patients, the predicted undiscounted life expectancy of an
average incident ESLD patient decreased from 20.57 to 18.52 between 2015 and 2035; similarly, the expected undiscounted lifetime cost associated with these patients decreased from £76,785 to £69,329 over the same period. It was estimated that the total undiscounted medical expenditure associated with the lifetime management of patients progressing to ESLD between 2015 and 2035 would accumulate to £2,647,335,555, with an equivalent discounted value of £1,202,827,444. From 2015, the total costs associated with ESLD complications for the UK population were estimated at approximately £46.3, £207.7 and £915.5 million, over 5-, 10- and 20-year horizons, respectively. Fig. 4 shows the effect on ESLD incidence and mortality of the four treatment implementation scenarios assessed. The incidence of ESLD-related mortality assuming the continuation of historical standard of care was estimated to rise from 859 in 2015 to 1792 in 2035. The introduction of a novel DAA regimen with an SVR of 95% for diagnosed patients in F3 or F4 resulted in an average reduction in ESLD event incidence of 8.95% over the period 2015e2035 and a similar reduction of 6.17% in ESLD-related mortality over the same period. Prioritizing patients with more advanced disease resulted in a significantly greater reduction in the incidence of ESLD events and related mortality, with average reductions of 43.91% and 35.70%, respectively, compared to the historical standard of care scenario. However, ESLD event incidence plateaued at approximately 988 patients per year due to the inability to treat those without knowledge of their infection and the assumption that retreatment of those failing to achieve SVR is not undertaken. Assuming all patients were diagnosed, and combining this with the prioritization of treatment in those with advanced disease, resulted in reductions of ESLD event incidence of 69.85% and ESLD mortality of 56.56%, when compared to the historical standard of care scenario over the period 2015e2035. Varying the proportion of patients initially
47
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Fig. 3 e Estimated average age and distribution of awareness of infection in patients with chronic hepatitis C, stratified by METAVIR fibrosis stage.
Table 2 e Predicted incidence of ESLD events and associated life expectancy and lifetime costs. Year
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 Total
ESLD incidencea
1289 1363 1433 1500 1564 1623 1677 1726 1770 1809 1842 1870 1892 1909 1921 1928 1930 1925 1914 1901 1883 36,670
Undiscounted Per-patient
Discounted Total cohort
Per-patient
Total cohort
Life expectancy
Lifetime cost (£)
Life years
Lifetime cost (£)
Life expectancy
Lifetime cost (£)
Life years
Lifetime cost (£)
20.57 20.34 20.34 20.11 20.11 19.87 19.87 19.62 19.62 19.36 19.36 19.09 19.09 19.09 18.81 18.81 18.81 18.52 18.52 18.52 18.52 e
76,785 75,969 75,969 75,120 75,120 74,241 74,241 73,324 73,324 72,376 72,376 71,395 71,395 71,395 70,381 70,381 70,381 69,329 69,329 69,329 69,329 e
26,519 27,722 29,157 30,171 31,450 32,244 33,320 33,861 34,717 35,014 35,664 35,701 36,122 36,437 36,139 36,270 36,294 35,653 35,456 35,206 34,870 e
99,004,877 103,527,191 108,885,383 112,705,305 117,482,112 120,485,165 124,506,683 126,568,296 129,767,817 130,923,111 133,352,798 133,536,221 135,109,570 136,288,670 135,221,811 135,710,148 135,800,401 133,449,489 132,714,533 131,775,896 130,520,079 2,647,335,555
13.61 13.07 12.63 12.13 11.72 11.24 10.86 10.42 10.07 9.65 9.32 8.93 8.63 8.34 7.99 7.72 7.46 7.14 6.89 6.66 6.44 e
47,815 45,928 44,375 42,606 41,165 39,509 38,173 36,618 35,380 33,924 32,777 31,412 30,350 29,324 28,089 27,139 26,221 25,102 24,253 23,433 22,641 e
17,552 17,816 18,105 18,193 18,323 18,246 18,217 17,983 17,815 17,456 17,179 16,710 16,335 15,920 15,345 14,880 14,386 13,736 13,198 12,662 12,117 e
61,651,761 62,588,598 63,601,890 63,923,165 64,379,160 64,118,407 64,017,905 63,208,019 62,614,353 61,365,548 60,390,702 58,753,300 57,435,307 55,977,338 53,967,095 52,330,426 50,594,423 48,318,523 46,427,454 44,540,185 42,623,883 1,202,827,444
ESLD, end-stage liver disease. a Includes decompensated cirrhosis, hepatocellular carcinoma and liver transplant.
48
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Fig. 4 e Estimated annual incidence of ESLD events (decompensated cirrhosis, HCC and liver transplant) and related mortality according to treatment implementation strategy. ESLD, end-stage liver disease; HCC, hepatocellular carcinoma; DAA, direct-acting antiviral; SOC, standard of care.
identified (58%) by 10% in either direction resulted in little change (<1%) in estimates of ESLD incidence and associated mortality across all treatment scenarios, with the exception of scenario 3. In this scenario, a decrease in the proportion of patients initially identified resulted in average reductions in the incidence of ESLD events and related mortality of 36.43% and 29.42%, respectively, whereas an increase led to reductions of 51.26% and 41.81%, respectively, compared to the historical standard of care scenario.
Discussion This study utilized a contemporary HCV natural history disease progression model to produce estimates of chronic hepatitis C prevalence that are consistent with previously predicted estimates.6,20 The predicted prevalence of HCC events at 2012 is supported by observational record-linkage data on liver-related sequelae.6 This study has provided
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
further granularity by partitioning according to METAVIR fibrosis stage, age and knowledge of infection status. This information is valuable as fibrosis stage and age are predictive of both complication development and treatment efficacy; those at a higher age and with a greater degree of fibrosis (i.e., METAVIR fibrosis stages F3 and F4) are at highest risk of progression to ESLD and have historically been less likely to achieve SVR with conventional treatments.31 Improved knowledge of the prevalent chronic hepatitis C population characteristics will better equip policy makers to consider effective treatment prioritization. With recent advances in the therapeutic landscape and introduction of new therapies with SVRs consistently >90%, even in the most difficult to treat patients, cure of any diagnosed individual is becoming a realistic achievement. The introduction of such treatments, combined with clinicians' knowledge of prognostic factors in individual patients, may allow for further optimization of approaches to treatment, in terms of resource utilization and treatment success. European clinical guidelines advocate the prioritization of treatment for patients with significant fibrosis or cirrhosis (METAVIR score F3 or F4).3 This study is supportive of this approach as it has demonstrated that introducing treatment with a novel DAA regimen capable of achieving a 95% SVR and prioritizing treatment in more advanced patients has the potential to significantly reduce onward ESLD event rates. However, the high cost of newly available DAA regimens has been regarded as a barrier to their use and attaining the benefits associated with their improved efficacy.9e12 This study highlights the significant costs of managing ESLD complications without the use of DAAs or the prioritization of treatment, at both patient and population levels. Moreover, this study suggests that the benefits of implementation strategies that target patients with advanced disease can only be fully realized with the use of newer therapies with higher rates of SVR and improved tolerability. However, the potential effect of such prioritization is limited as issues remain in relation to treatment uptake and numbers lost to follow-up along the HCV treatment cascade. Only 22% of PWID in England who stated seeing a specialist nurse or doctor regarding their infection in 2013 reported receiving any HCV-related medication,28 and similar levels of imperfect follow-up (27%) have been shown to greatly reduce the real-world effectiveness of HCV therapy.41 Even modestly effective interventions to improve follow-up, such as peer navigators or integrated case management, would likely provide significant value in terms of improving the costeffectiveness of interventions in the real world. The current annual rate of treatment uptake in the UK is approximately 3%,28 a rate shown to have little impact on the incidence of infection, or on the incidence of ESLD and HCC complications in the near term.18,27,42e44 However, previously published studies have demonstrated that even modest increases in the rate of treatment uptake may have a significant effect upon the prevalence of hepatitis C in the UK.42,45 As such, this study did not attempt to quantify the relationship between treatment uptake and the burden of ESLD but endeavoured to approximate the relationship between various targeted
49
treatment strategies and their effect upon the burden of ESLD, subject to credible assumptions made in relation to modelling the UK HCV population. Further, in order to reduce the future burden of ESLD, it is critical that those chronically infected are identified and successfully treated. Diagnostic challenges remain in the UK, with significant numbers unaware of their infection status. Additional complexity in identifying those infected relates to the high prevalence (~90%) of PWID amongst incident infections, a subgroup of patients that do not commonly access health care via traditional settings and are often associated with poverty, homelessness and an unwillingness to engage services.46,47 Combined with a relatively high degree of contraindications in this patient group, including psychiatric disorders, and alcohol and substance addiction, there are considerable challenges to the successful identification, engagement and monitoring of these patients.47 There are inherent limitations to the modelling approach presented. The back projection methodology requires accurate estimates of HCC end points; the source data used for this analysis were routine hospital episode statistics, which are subject to potential under-reporting and incorrect use of diagnostic codes. The degree of bias associated with these limitations is difficult to quantify; however, their effect may result in the under prediction of HCV-related ESLD. Due to the nature of the back projection methodology, predictions of prevalence are largely dependent upon the HCV disease transition rates utilized in the natural history model, which are subject to a high degree of uncertainty due to the small numbers of studies from which they are derived.48 Though such analyses are subject to limitations, their inference of statistical relationships is a valuable tool in economic evaluation and to inform decision-making. In conclusion, despite the expected fall in the prevalence of chronic hepatitis C in coming decades, the overall burden is estimated to substantially increase due to the rising incidence of ESLD. A targeted treatment approach that prioritizes the use of novel, highly effective DAA regimens in those patients at more advanced stages is likely to lead to a considerable reduction in ESLD rates. This impact may be augmented via improved rates of diagnosis and treatment initiation. Given the large number of emerging HCV therapies, a preoccupation with the relative cost-effectiveness of these treatments should arguably be replaced with an emphasis on identifying the most appropriate treatment for individual patients. Such a strategy is suggested to improve treatment uptake, adherence and, ultimately, success in tackling the population burden of HCV, ensuing in the reduced health and economic burden of HCV-related ESLD complications.
Author statements Ethical approval Not required, as the manuscript describes an economic analysis using publicly available data.
50
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
Funding Writing and research was funded by Bristol-Myers Squibb Pharmaceuticals Ltd; however, the publication of study results is not contingent on the sponsor's approval or censorship of the manuscript.
Competing interests Thomas Ward, Phil McEwan, Jason Gordon, Hayley Bennett Wilton, Beverley Jones, Daniel Sugrue and Samantha Webster are employees of Health Economics and Outcomes Research Ltd, who received funding from Bristol-Myers Squibb Pharmaceuticals Ltd. Michael Brenner is an employee of the study funder.
references
1. Chen SL, Morgan TR. The natural history of hepatitis C virus (HCV) infection. Int J Med Sci 2006;3(2):47e52. 2. Kanwal F, Hoang T, Kramer JR, Asch SM, Goetz MB, Zeringue A, et al. Increasing prevalence of HCC and cirrhosis in patients with chronic hepatitis C virus infection. Gastroenterology 2011;140(4):1182e8. 3. European Association For The Study Of The Liver. Clinical practice guidelines: EASL recommendations on treatment of hepatitis C 2015. J Hepatol 2015;62(4). 4. Public Health England. Data tables of the unlinked anonymous monitoring survey of HIV and hepatitis in people who inject drugs. 2013. Available from: http://webarchive.nationalarchives.gov. uk/20140629102627/http:/hpa.org.uk/web/ HPAweb&HPAwebStandard/HPAweb_C/1202115519183. (last accessed February 2015). 5. Goldberg D, Anderson E, Hepatitis C. Who is at risk and how do we identify them? J Viral Hepat 2004;11(s1):12e8. 6. Public Health England. Health Protection Scotland, Public Health Wales, Public Health Agency. Hepatitis C in the UK: 2015 report. 2015. Available from: https://www.gov.uk/government/ publications/hepatitis-c-in-the-uk. (last accessed 11 November 2015). 7. Wiessing L, Ferri M, Grady B, Kantzanou M, Sperle I, Cullen KJ, et al. Hepatitis C virus infection epidemiology among people who inject drugs in Europe: a systematic review of data for scaling up treatment and prevention. PloS One 2014;9(7):e103345. 8. Martin NK, Foster GR, Vilar J, Ryder S, E Cramp M, Gordon F, et al. HCV treatment rates and sustained viral response among people who inject drugs in seven UK sites: real world results and modelling of treatment impact. J Viral Hepat 2014;22(4):399e408. 9. Sulkowski MS, Gardiner DF, Rodriguez-Torres M, Reddy KR, Hassanein T, Jacobson I, et al. Daclatasvir plus sofosbuvir for previously treated or untreated chronic HCV infection. N Engl J Med 2014;370(3):211e21. 10. Lawitz E, Mangia A, Wyles D, Rodriguez-Torres M, Hassanein T, Gordon SC, et al. Sofosbuvir for previously untreated chronic hepatitis C infection. N Engl J Med 2013;368(20):1878e87. 11. Lawitz E, Sulkowski MS, Ghalib R, Rodriguez-Torres M, Younossi ZM, Corregidor A, et al. Simeprevir plus sofosbuvir, with or without ribavirin, to treat chronic infection with hepatitis C virus genotype 1 in non-responders to pegylated interferon and ribavirin and treatment-naive patients: the COSMOS randomised study. Lancet 2014;384(9956):1756e65.
12. Lawitz E, Poordad FF, Pang PS, Hyland RH, Ding X, Mo H, et al. Sofosbuvir and ledipasvir fixed-dose combination with and without ribavirin in treatment-naive and previously treated patients with genotype 1 hepatitis C virus infection (LONESTAR): an open-label, randomised, phase 2 trial. Lancet 2014;383(9916):515e23. 13. Innes H, Goldberg D, Dillon J, Hutchinson SJ. Strategies for the treatment of Hepatitis C in an era of interferon-free therapies: what public health outcomes do we value most? Gut 2014;64(11):1800e9. 14. McEwan P, Ward T, Yuan Y, Kim R, L'Italien G. The impact of timing and prioritization on the cost-effectiveness of birth cohort testing and treatment for hepatitis C virus in the United States. Hepatology 2013;58(1):54e64. 15. McEwan P, Ward T, Chen C-J, Lee M-H, Yang H-I, Kim R, et al. Estimating the incidence and prevalence of chronic hepatitis C infection in Taiwan using back projection. Value Health Regional Issues 2013;3:5e11. 16. McEwan P, Kim R, Yuan Y. Assessing the cost utility of response-guided therapy in patients with chronic hepatitis C genotype 1 in the UK using the MONARCH model. Appl Health Econ Health Policy 2013;11(1):53e63. 17. McEwan P, Ward T, Webster S, Yuan Y, Kalsekar A, Broglio K, et al. Estimating the long-term clinical and economic outcomes of Daclatasvir plus Asunaprevir in difficult-to-treat Japanese patients chronically infected with hepatitis C genotype 1b. Value Health Regional Issues 2014;3:136e45. 18. McEwan P, Ward T, Bennett H, Kalsekar A, Webster S, Brenner M, et al. Estimating the clinical and economic benefit associated with incremental improvements in sustained virologic response in chronic hepatitis C. PLoS One 2015;10(1):e0117334. 19. Bacchetti P, Segal MR, Jewell NP. Backcalculation of HIV infection rates. Stat Sci; 1993::82e101. 20. Sweeting MJ, De Angelis D, Brant LJ, Harris HE, Mann AG, Ramsay ME. The burden of hepatitis C in England. J Viral Hepat 2007;14(8):570e6. 21. Deuffic S, Buffat L, Poynard T, Valleron AJ. Modeling the hepatitis C virus epidemic in France. Hepatology 1999;29(5):1596e601. 22. Razali K, Thein HH, Bell J, Cooper-Stanbury M, Dolan K, Dore G, et al. Modelling the hepatitis C virus epidemic in Australia. Drug Alcohol Depend 2007;91(2):228e35. 23. Becker NG, Watson LF, Carlin JB. A method of non-parametric back-projection and its application to AIDS data. Stat Med 1991;10(10):1527e42. 24. Brookmeyer R, Damiano A. Statistical methods for short-term projections of AIDS incidence. Stat Med 1989;8(1):23e34. 25. Isham V. Estimation of the incidence of HIV infection. Philosophical Trans R Soc Lond Biol Sci 1989;325(1226):113e21. 26. UK office for National Statistics. England and Wales life tables: 1980e82 to 2012e2014. Available from: http://www. ons.gov.uk/ons/taxonomy/index.html?nscl¼LifeþTables-tabdata-tables. (last accessed 16 October 2015). 27. Martin NK, Vickerman P, Miners A, Foster GR, Hutchinson SJ, Goldberg DJ, et al. Cost-effectiveness of hepatitis C virus antiviral treatment for injection drug user populations. Hepatology 2012;55(1):49e57. 28. Public Health England. Hepatits C in the UK: 2014 report. 2014. Available from: https://www.gov.uk/government/publications/ hepatitis-c-in-the-uk. (last accessed October 2014). 29. Palmateer NE, Hutchinson SJ, McLeod A, Codere G, Goldberg DJ. Comparison of deaths related to hepatitis C and AIDS in Scotland. J Viral Hepat 2007;14(12):870e4. 30. European Monitoring Centre for Drugs and Drug additicon. European drug report 2014. 2015. Available from: http://www. emcdda.europa.eu/edr2014. (last accessed 2 March 2015).
p u b l i c h e a l t h 1 4 1 ( 2 0 1 6 ) 4 2 e5 1
31. Thein HH, Yi Q, Dore GJ, Krahn MD. Estimation of stagespecific fibrosis progression rates in chronic hepatitis C virus infection: a meta-analysis and meta-regression. Hepatology 2008;48(2):418e31. 32. Forrest EH, Thorburn D, Spence E, Oien KA, Inglis G, Smith CA, et al. Polymorphisms of the renineangiotensin system and the severity of fibrosis in chronic hepatitis C virus infection. J Viral Hepat 2005;12(5):519e24. 33. Freeman A, Hamid S, Morris L, Vowler S, Rushbrook S, Wight DGD, et al. Improved detection of hepatocyte proliferation using antibody to the pre-replication complex: an association with hepatic fibrosis and viral replication in chronic hepatitis C virus infection. J Viral Hepat 2003;10(5):345e50. 34. Mohsen AH, Murad S, Easterbrook PJ. Prevalence of hepatitis C in an ethnically diverse HIV-1-infected cohort in south London. HIV Med 2005;6(3):206e15. 35. Ryder SD. Progression of hepatic fibrosis in patients with hepatitis C: a prospective repeat liver biopsy study. Gut 2004;53(3):451e5. 36. Wong V, Caronia S, Wight D, Palmer CR, Petrik J, Britton P, et al. Importance of age in chronic hepatitis C virus infection. J Viral Hepat 1997;4(4):255e64. 37. Patel K, Norris S, Lebeck L, Feng A, Clare M, Pianko S, et al. HLA class I allelic diversity and progression of fibrosis in patients with chronic hepatitis C. Hepatology 2006;43(2):241e9. 38. Micallef JM, Kaldor JM, Dore GJ. Spontaneous viral clearance following acute hepatitis C infection: a systematic review of longitudinal studies. J Viral Hepat 2006;13(1):34e41. 39. Shepherd J, Jones J, Hartwell D, Davidson P, Price A, Waugh N. Interferon alpha (pegylated and non-pegylated) and ribavirin for the treatment of mild chronic hepatitis C: a systematic review and economic evaluation. Health Technol Assess Winch Engl 2007;11(11). 1e205, iii. 40. Personal Social Services Research Unit. Unit costs of health & social care 2013. In: Section 16.2: the hospital & community health services (HCHS) index; 2013. 41. Linas BP, Barter DM, Leff JA, Assoumou SA, Salomon JA, Weinstein MC, et al. The hepatitis C cascade of care: identifying priorities to improve clinical outcomes. PloS One 2014;9(5):e97317. 42. Martin NK, Vickerman P, Foster GR, Hutchinson SJ, Goldberg DJ, Hickman M. Can antiviral therapy for hepatitis C
43.
44.
45.
46.
47.
48.
49.
51
reduce the prevalence of HCV among injecting drug user populations? A modeling analysis of its prevention utility. J Hepatol 2011;54(6):1137e44. Martin NK, Vickerman P, Grebely J, Hellard M, Hutchinson SJ, Lima VD, et al. Hepatitis C virus treatment for prevention among people who inject drugs: modeling treatment scale-up in the age of direct-acting antivirals. Hepatology 2013;58(5):1598e609. Martin NK, Hickman M, Hutchinson SJ, Goldberg DJ, Vickerman P. Combination interventions to prevent HCV transmission among people who inject drugs: modeling the impact of antiviral treatment, needle and syringe programs, and opiate substitution therapy. Clin Infect Dis 2013;57(Suppl. 2):S39e45. Bennett H, McEwan P, Sugrue D, Kalsekar A, Yuan Y. Assessing the long-term impact of treating hepatitis C virus (HCV)-Infected people who inject drugs in the UK and the relationship between treatment uptake and efficacy on future infections. PLoS One 2015;10(5):e0125846. Harris M, Rhodes T. Hepatitis C treatment access and uptake for people who inject drugs: a review mapping the role of social factors. Harm Reduct J 2013;10:7. World Health Organization. Barriers and facilitators to hepatitis C treatment for people who inject drugs. A Qual study. 2012. Available from: http://www.euro.who.int/__data/ assets/pdf_file/0011/179750/Barriers-and-facilitators-tohepatitis-C-treatment-for-PWID-A-qualitative-study-June2012-rev-5.pdf. (last accessed 2 December 2015). Townsend R, McEwan P, Kim R, Yuan Y. Structural frameworks and key model parameters in cost-effectiveness analyses for current and future treatments of chronic hepatitis C. Value Health 2011;14(8):1068e77. R, Shiffman ML, Suter F, Bacon BR, Nelson D, Harley H, Sola et al. Peginterferon alfa-2a and ribavirin for 16 or 24 weeks in HCV genotype 2 or 3. N Engl J Med 2007;357(2):124e34. http:// dx.doi.org/10.1056/NEJMoa066403.
Appendix A. Supplementary data Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.puhe.2016.08.002.