Acknowledgements

This document was prepared by Edward Mills, Eric Druyt and Sam Keeping, Global Evaluative Sciences, Vancouver, Canada.

Acronyms

AE

adverse event

CPP

cost per patient

CPSVR

cost per sustained viral response

DAA

direct-acting antiviral

DCV

daclatasvir

EMA

European Medicines Agency

FDA

U.S. Food and Drug Administration

HCC

hepatocellular carcinoma

HCV

hepatitis C virus

HIV

human immunodeficiency virus

IF

interferon-free

LDV

ledipasvir

LMIC

low- and middle-income countries

PR

pegylated interferon and ribavirin

R

ribavirin

RNA

ribonucleic acid

SOF

sofosbuvir

SVR

sustained virological response

WHO

World Health Organization

1. Executive summary

Infection caused by the hepatitis C virus (HCV) is a serious public health problem, especially in low- and middle-income countries (LMIC). Access to treatment has generally been poor for patients with hepatitis C in LMIC due to a combination of issues including drug prices, treatment ineffectiveness and complexity, a lack of health system capacity, and insufficient political will. New, highly efficacious, interferon-free regimens provide an opportunity to deliver wider access to treatment for patients with hepatitis C as they have shorter durations of treatment and reduce the need for specialist oversight compared to traditional treatment with pegylated interferon and ribavirin (PR).

This analysis sought to assess the feasibility of delivering new HCV treatments in LMIC by comparing the treatment costs for different groups of diagnosed patients under two different scenarios: treating all patients with PR and treatment with interferon-free regimens.

A model was constructed using Microsoft Excel to estimate HCV treatment costs in a sample of three countries: Brazil, Mongolia and Ukraine. As treatment for HCV is highly patient specific, patients entering into the model were grouped according to a number of different characteristics and treatments allocated to each based on published recommendations. The number of patients in each category was calculated using published epidemiological data.

Costs during active treatment comprised those for drug procurement, laboratory tests for HCV RNA, other laboratory tests, administration and clinical monitoring, and management of adverse events. The duration of treatment for each regimen was split into 4-week cycles, with discontinuation rates applied to determine the number of patients on treatment at each point in time. Costs for each of the categories named above were then applied to these numbers to give the total costs for each patient group.

In the base case analysis, the total number of patients given treatment and the number achieving a sustained virological response (SVR) in Brazil under the PR scenario were 308 584 and 154 354, respectively. The cost of treatment in this scenario was US$ 1.66 billion, split into US$ 1.42 billion for PR and US$ 0.24 billion for other costs. Equivalent numbers under the interferon-free scenario were 313 044 patients, 299 734 SVRs, and US$ 3.39 in total treatment costs (US$ 3.32 billion for antiviral drugs and US$ 0.07 billion for other costs). Compared to treatment with PR, this represented an increase of 145 380 SVRs at an additional cost of US$ 1.73 billion. In the sensitivity analysis, these results were found to be most influenced by changes in drug prices.

Under the PR scenario in Mongolia, the total number of patients given treatment was 58 180, 26 729 of whom were estimated to have achieved an SVR. Total treatment costs in this scenario were US$ 409.38 million (US$ 323.20 million for PR and US$ 86.18 million for other costs). The number of patients given treatment in the DAA scenario was 59 271 patients and the number of SVRs was 299 734, an increase of 31 520 SVRs compared to PR. The cost of treating patients with interferon-free regimens was US$ 101.32 million (US$ 81.42 million for drugs and US$ 19.90 million for other costs), US$ 308.07 million lower than treatment with PR. The sensitivity analysis showed the results to be most sensitive to the cost of PR.

In Ukraine, 404 631 patients were estimated to be treated under the PR scenario, yielding a total of 208 742 SVRs. Costs of treatment were US$ 0.43 billion for drugs and US$ 0.85 billion for other treatment costs, coming to a total of US$ 1.28 billion. When interferon-free regimens were used instead, the respective numbers were 406 171 patients, 387 365 SVRs, US$ 0.97 in drug costs, and US$ 0.23 billion in other costs; an increase of 178 623 SVRs compared to the PR scenario, and of approximately US$ 0.08 billion less in total cost. These base case findings were most sensitive to non-drug costs with PR and drug prices for both PR and DAAs.

The results of the analysis demonstrate that for some countries, expanding access to treatment using interferon-free regimens may entail less cost than if traditional PR regimens are used. The main drivers of differences between Brazil and Mongolia were the ratio of prices for DAAs used in interferon-free regimens and PR, and the genotype distribution among HCV patients. The lower costs overall with interferon-free regimens in Ukraine came from lower non-drug costs in genotypes 1 and 2 patients compared to those with PR. It is also clear from the results that treating patients with interferon-free regimens as opposed to PR is likely to result in substantially higher rates of SVR across all patient groups. The next step for countries will be to quantify what this would mean in terms of reducing the socioeconomic burden of HCV in the long term. This requires data on the costs of treating HCV patients with end-stage liver disease as well as the impact that these conditions have on both the quality of life and life expectancy.

2. Introduction

The hepatitis C virus (HCV) is a serious public health concern, causing substantial morbidity and mortality worldwide among those infected (1). Around 85% of people who contract HCV go on to develop chronic infection, which in time can lead to serious complications, such as liver cirrhosis and hepatocellular carcinoma (HCC) (2,3). The total number of people who are chronically infected with HCV is around 80 million, the majority of whom reside in low- and middle-income countries (LMIC) (4,5). In 2013, an estimated 700 000 deaths were attributable to HCV (6). Epidemiological models suggest that unless access to screening and treatment becomes more widespread, the number of deaths can be expected to rise in the coming years (7,8).

Obstacles to more comprehensive treatment for HCV-infected individuals in LMIC include cost, a lack of specialist staff, limited numbers of facilities able to carry out diagnostic testing, and the perceived complexity and ineffectiveness of treatment with pegylated-interferon and ribavirin (PR) (9–13). Nonetheless, some countries have been able to successfully implement treatment programmes, albeit on a limited scale, with outcomes broadly in line with those seen in high-income countries (14). The emergence of interferon-free antiviral regimens has been regarded as a major step towards facilitating wider access to treatment for HCV patients in LMIC. Interferon-free regimens are highly efficacious across a range of patient characteristics and are generally well tolerated, meaning that adherence is likely to be high. As they are also orally administered, they provide the opportunity for task-shifting away from specialist facilities towards community and primary care (10).

When seeking funding for new HCV treatment programmes using interferon-free regimens, it will be necessary for planners to generate estimates of their expected health and economic benefits. A key step in this process will be for countries to quantify the current burden of HCV within their respective populations. This is a complicated exercise, as around 15% of those with acute HCV will spontaneously clear the infection and therefore need to be followed up in order to confirm whether they have progressed to chronic infection (2,3). Population-level epidemiological studies are limited largely because most LMIC do not have the resources to carry out this type of study at the required scale. In those countries where robust data on prevalence are available, these estimates can be used to calculate the direct and indirect costs that HCV places on both individuals and society, so that the relative cost–effectiveness of new treatment programmes can be assessed.

A major influence on the investment case for new treatment programmes will be the cost of delivering interferon-free regimens (9). The overall cost of providing treatment for HCV has hitherto been largely determined by the price of pegylated interferon, which unlike ribavirin, is subject to patent protection in a number of LMIC (9). Based on data from a sample of such countries, a full 48-week course of pegylated interferon has been estimated to cost between US$ 10 000 and US$ 20 000 per patient (15). These high costs are one of the major reasons that PR has not seen extensive use outside of those countries where large price reductions have been achieved. For example, in Egypt, where generic competition has reduced the cost per dose of pegylated interferon to roughly US$ 40, treatment has been offered to over 200 000 people since 2006 (16).

Prices for pegylated interferon are low in comparison to those currently being charged in high-income countries for some direct-acting antivirals (DAAs) used within interferon-free regimens. A 12-week course of sofosbuvir plus ledipasvir (SOF + LDV), a fixed-dose, all-oral combination that has demonstrated high efficacy against HCV genotypes 1, 2 and 4, currently costs around US$ 113 400 (17). The price for the equivalent regimen using the ombitasvir plus paritaprevir/ritonavir and dasabuvir oral combination is US$ 99 983 (17). These high prices have been heavily criticized by some and have led to concerns that these drugs will not be made available in LMIC (18,19). Progress has already been made towards lowering these prices for LMIC; at least one manufacturer has signed licensing agreements with counterparts in India so that they can produce their own generic versions of some medicines, with other companies set to follow suit (20,21). Tiered pricing will also allow for the originator products to be made available (22). It is anticipated that these agreements will largely determine which treatments are made available in different countries.

Beyond drug prices, a host of additional factors will need to be taken into consideration when evaluating the costs of delivering interferon-free regimens in resource-limited settings. These include prioritization of groups for treatment, e.g. those at the highest risk of serious complications, the genotype distribution among cases, access to laboratory testing, and non-drug-related treatment costs. The content of this report details a model that has been developed to assist decision-makers in bringing together information on some of these elements to produce estimates of the costs of different treatment strategies for HCV. A series of case studies have been carried out for countries where epidemiological information is available. This has been combined with data on the efficacy, safety, and the costs associated with different treatment regimens for HCV, to calculate the short-term budget impact associated with switching from treatment with PR to interferon-free regimens, where indicated.

Although some of the data used within the model is of a low quality, requiring the need for a number of strong assumptions based on limited evidence, the view was taken by the core research team and expert advisory committee that this should not act as an impediment to efforts to quantify the costs of providing treatment to HCV patients. This analysis should therefore be seen as a guide for those working in LMIC, for whom access to data is not confined to the published literature, who wish to carry out their own more comprehensive economic analyses in support of funding applications.

3. Objective

To assess the feasibility of delivering new HCV treatments in LMIC by comparing the treatment costs for different groups of diagnosed patients under two different scenarios: treating all patients with PR and treating with interferon-free regimens.

4. Methods

4.3. Treatable populations

As changes in screening programmes were not included within the model, the potentially treatable population was taken to be those patients who had previously received a positive diagnosis of chronic HCV infection. These populations were calculated by first applying the prevalence of HCV viraemia in each country to their total population in 2015, giving the total number of chronic HCV infections. Prevalence estimates for Brazil and Mongolia were taken from a recent systematic review of the global epidemiology of HCV (4). No data were available for Ukraine so an estimate of the regional prevalence for Eastern Europe from the same study was used. The number of those diagnosed in each country was then derived by applying diagnosis rates from a pair of modelling studies looking at the future burden of HCV across a number of different countries (7,8).

The efficacy of treatment regimens for HCV varies depending on patient characteristics, so it was necessary to split the HCV population in each of the countries of interest into groups. The diagnosed population was first split into treatment-naive and treatment-experienced patients. The probability of any patient currently diagnosed with chronic HCV having failed previous treatment was taken to be the product of the proportion of patients offered treatment in the past within each country and 1 minus the historical average sustained virological response (SVR) for that country. The probability of being treatment naive was 1 minus the probability of being treatment experienced. Country values for the proportions given treatment previously and the historical SVRs used to calculate these probabilities were taken from the same modelling studies that supplied the diagnosis rates (7,8). The authors of the study noted that as no data were available on the exact treatments provided to patients in each country, the average SVRs were based on the expert opinion of treating physicians.

Severity of liver disease is another relevant factor in both the choice and efficacy of treatment for HCV. Treatment-naive and treatment-experienced patients were therefore further subdivided into four states based on the extent of their liver disease, assuming that the underlying distribution of severity was the same for the diagnosed and undiagnosed populations. These categories were patients with no cirrhosis (METAVIR score F0-F3), those with compensated cirrhosis (F4), those with decompensated cirrhosis (F4), and those with HCC. Because of additional complexities, including additional costs involved in treatment for patients with HCC, it was decided that this subpopulation would not be evaluated in this model. Tables 1–3 show the size of each of these patient populations for each country.

Table 1

Epidemiological input parameters and estimated patient group sizes for Brazil.

Table 3

Epidemiological input parameters and estimated patient group sizes for Ukraine.

Table 2

Epidemiological input parameters and estimated patient group sizes for Mongolia.

Finally, the numbers in each patient group were multiplied by the genotype distributions (Figure 1) to give the final structure of the treatable populations used within the model. Differences in efficacy have been observed between sub-genotypes, in particular 1a and 1b, and it was considered whether this should be accounted for within the model. However, as the facilities needed to carry out genotyping, let alone sub-genotyping, are limited in most LMIC, it was decided that this was not necessary. This left us with 2 × 3 × 6 = 36 treatment groups included within the model.

Figure 1

Distribution of HCV infections by genotype in Brazil, Mongolia, and Ukraine (4).

4.6. Intervention parameters

4.6.1. Interventions

The model was used to compare two different treatment scenarios. The first was treating all diagnosed patients without a contraindication with PR. In the second scenario, each patient group was assumed to be treated with an interferon-free regimen.

4.6.2. Allocations

As pegylated interferon is contraindicated in patients with decompensated cirrhosis, only those patients without cirrhosis or with compensated cirrhosis were assumed to be offered treatment in the PR scenario (24). The duration of treatment with PR was set to 48 weeks for genotypes 1, 4, 5, and 6, and 24 weeks for genotypes 2 and 3. Combinations included within the interferon-free scenario were sofosbuvir plus ribavirin (SOF + R), sofosbuvir plus ledipasvir (SOF + LDV), and sofosbuvir plus daclatasvir (SOF + DCV), with or without ribavirin. The treatment allocations in the interferon-free scenario were based on expert input from the World Health Organization (WHO) regarding the supply of treatments to LMIC, and are presented in Table 4.

Table 4

Interferon-free treatment allocations for patients with/without cirrhosis, by HCV genotype, used within the model.

4.6.3. Efficacy

Pooled SVR proportions for each of the regimens were taken from a systematic literature review of safety and efficacy data that was carried out alongside this analysis (see report entitled: Systematic review and network meta-analysis to support the development of WHO Guidelines for the treatment of persons with the hepatitis C virus). These estimates contained data from both randomized controlled trials and non-randomized controlled trials. In the event of missing data, assumptions were made using information from comparable regimens in patient populations of interest. Only one trial included an arm that assessed the efficacy of SOF + LDV for 12 weeks in genotype 6, the results showing an SVR at 12 weeks of 96% (25). As this was in line with estimates for genotypes 1 and 4, we used the pooled data from these genotypes as our estimate for genotypes 5 and 6. The same assumption was applied to the efficacy of PR for these two genotypes.

Clinical trials of PR for 24 weeks in treatment-experienced patients with genotypes 2 and 3 were not captured in the accompanying systematic literature review, as PR was not a primary intervention of interest. It was therefore assumed that efficacy would be equivalent to efficacy in the treatment-naive population. Although the data to support this assumption are minimal, because numbers of treatment-experienced patients were low in all countries, its likely influence on the final results was assessed to be low.

There has also been only one trial of the SOF + R combination for 16 weeks in treatment-experienced, genotype 2 patients (26). Again, the estimated SVR in this trial was similar to the pooled estimate for the same regimen over 12 weeks; therefore, the latter was used as it was based on data from considerably more patients (346 and 141 for treatment-naive and treatment-experienced patients, respectively). Data were only available for SOF + DCV + R for 24 weeks in a combined population of genotypes 2 and 3 treatment-naive patients (27). Efficacy in treatment-experienced patients was assumed to be equivalent to this group, as a similar pattern was observed in genotype 1 patients (27). All the efficacy estimates used within the model are presented in Table 5.

Table 5

Pooled SVR proportions by regimen, prior treatment experience and genotype.

4.6.4. Discontinuation due to adverse events

Pooled discontinuation rates due adverse events were also taken from the systematic literature review (SLR) and are presented in Table 6. A similar set of assumptions to the ones used to fill in missing data for efficacy was applied to discontinuation. First, discontinuation among patients with genotypes 5 and 6 treated with PR or SOF + LDV was assumed to be the same as the rates for genotypes 1 and 4 for the reasons stated above. Only a single study of 9 treatment-naive patients included data on discontinuation with PR for genotype 2 (28). It was decided that the larger, combined genotypes 2 and 3 population would provide a better approximation of discontinuation as, compared to efficacy, rates of adverse events leading to discontinuation are less sensitive to genotype than they are to sample size. Data on discontinuation for SOF + R were also available only for a combined naive and experienced population so the same estimate was applied to both groups. The same values were used for the 16-week regimen. The likely impact of these two assumptions was assessed to be minimal due to the limited extension in treatment duration and small numbers of experienced patients. The only trial of SOF + DCV + R for 24 weeks in genotype 3, treatment-naive patients also included genotype 2 patients within the safety analysis. Again, discontinuations were assumed to be equivalent between treatment-naive and treatment-experienced groups based on data for genotype 1.

Table 6

Pooled proportions of patients discontinuing treatment due to adverse events by regimen, prior treatment experience and genotype.

4.6.5. Cost

The price of PR in Brazil and Ukraine were obtained from the literature (15,18,29). No price for units of pegylated interferon or ribavirin in Mongolia could be identified. Prices from Brazil were used instead, as total treatment costs with PR are similar between the two countries (18,30). Antiviral drug prices were taken from an unpublished survey of national experts carried out by Médecins Sans Frontières. A price for DCV has not yet been agreed upon in Mongolia or Ukraine, so the average of prices available for LMIC was used. Ukraine, along with Brazil, also does not have a price for SOF + LDV. Rather than use the average of SOF + LDV prices, which would be based on data from only two countries, we assumed that this combination would be priced at the same level as the combined cost of SOF and DCV or SOF plus simeprevir, depending on availability. The final drug prices used within the model are presented in Table 7.

Table 7

Cost of drug per four-week cycle in Brazil, Mongolia and Ukraine (2015 US$).

4.7. Other parameters

4.7.1. Non-drug costs

Costs during active treatment were broken down into those for drug procurement, laboratory tests for HCV RNA, other laboratory tests (e.g. full blood count, creatinine, and alanine transaminase), administration and clinical monitoring, and management of adverse events. Country-specific data on non-drug costs were identified only for Brazil. The cost per sample tested for HCV RNA was taken from a study of different screening algorithms for HCV (31). The costs of other laboratory tests, clinical monitoring, and management of adverse events were taken from a micro-costing study carried out in 2012, which focused on patients treated with PR. The different costs within the study were mapped to the cost categories as seen in the first two columns of Table 8. Data on non-drug costs when treating with interferon-free regimens, with or without ribavirin, in LMIC could not be identified, as these treatments have only recently been introduced. As an alternative, safety information and WHO treatment guidelines (32) were used to determine a cost breakdown for regimens containing ribavirin without pegylated interferon, and those with neither ribavirin nor pegylated interferon (all DAA regimens). These are the final three columns of Table 8.

Table 8

Components of each cost category by treatment regimen.

All non-drug costs were inflated to 2015 prices using the Brazilian National Consumer Price Index (33) and converted to US dollars (34). To account for variation in prices between Brazil and the other countries in the study, we multiplied all costs by the ratio of unit costs for inpatient or outpatient facilities from the WHO-CHOICE unit's study of service delivery costs (35). The final costs used within the model can be found in Tables 9–11.

Table 9

Non-drug treatment costs (2015 US$) per four-week cycle in Brazil, by regimen.

Table 11

Non-drug treatment costs (2015 US$) per four-week cycle in Ukraine, by regimen.

Table 10

Non-drug treatment costs (2015 US$) per four-week cycle in Mongolia, by regimen.

5. Results

5.1. Brazil

5.1.1. Base case

The results of the base case analysis for Brazil can be seen in Table 12. In Scenario 1, the total cost of treating all patients currently diagnosed with chronic HCV using PR would be just over US$ 1.67 billion, US$ 1.42 billion of which would be drug costs and the remaining non-drug costs. As a result of receiving this treatment, an estimated 154 354 patients would achieve an SVR. The average cost per patient and SVR in this scenario would be US$ 5368 and US$ 10 733, respectively, across all patient categories. Restricting treatment to only patients with evidence of cirrhosis would cost approximately around US$ 217.22 million, leading to 20 239 SVRs.

Table 12

Base case results for Brazil (all costs in 2015 US$).

Alternatively, treating all patients with interferon-free regimens would result in an additional 145 380 SVRs at the cost of an extra US$ 1.73 billion (US$ 10 831 per patient and US$ 11 312 per SVR). These additional costs resulted from higher drug costs (US$ 1.90 billion), although there was reduction of US$ 169.10 million in expenditure on elements of treatment other than for the antiviral drugs themselves. Equivalent figures for just cirrhotic patients (both compensated and decompensated) were 22 148 additional SVRs at an additional cost of US$ 623 million. Figure 2 shows the distribution of total treatment costs among genotypes, and drug and non-drug costs. The pattern of higher drug but lower non-drug costs in both scenarios is consistent across genotypes.

Figure 2

Total costs (2015 US$) of treating all diagnosed HCV patients using interferon-free regimens compared to PR, by genotype, in Brazil. PR, peginterferon + ribavirin; IF, interferon free

5.1.2. Sensitivity analysis

The results of the analysis were most sensitive to changes in drug prices. Figure 3 shows the variation in th difference in the average cost per patient in the interferon-free scenario compared to the scenario where PR was used. This difference ranged from US$ 4055 and US$ 6870 with the cost of SOF + LDV, from US$ 4541 to US$ 6384 with the cost of PR, from US$ 4746 to US$ 6179 with the cost of SOF + DCV + R, and from US$ 4920 to US$ 6004 with the cost of SOF + R.

Figure 3

Sensitivity of cost (2015 US$) per patient to changes in cost parameters (Brazil). DCV, daclatasvir; LDV, ledipasvir; PR, pegylated interferon + ribavirin; SOF, sofosbuvir; R, ribavirin

5.2. Mongolia

5.2.1. Base case

Table 13 shows the base case results for Mongolia. The total cost of treating all patients diagnosed with chronic HCV is estimated to be US$ 409.38 million using PR, and this would result in 26 725 SVRs. This equates to an average cost per patient of US$ 7 036 and cost per SVR of US$ 15 316. SVR numbers and treatment costs were 3 552 and US$ 54.40 million, respectively, for patients with cirrhosis.

Table 13

Base case results for Mongolia (all costs in 2015 US$).

The estimated number of additional SVRs when using interferon-free regimens was 31 520. This was associated with a reduction in costs compared to use of PR of around US$ 308 million (cost per patient US$ 2649 and cost per SVR US$ 3450), split into US$ 241.70 million less spent on antiviral drugs and US$ 66.28 million less on other elements of treatment. These reductions were primarily driven by the lower unit costs of SOF + LDV compared to PR, and the much shorter time spent on treatment, as the majority of patients in Mongolia are treatment-naive, genotype 1 patients (Figure 4). Savings when treating only those patients with cirrhosis were US$ 31.45 million, even after accounting for the additional costs of expanding treatment to those with decompensated cirrhosis, and a total of 8604 SVRs would be expected to be achieved within this population.

Figure 4

Total costs (2015 US$) of treating all diagnosed HCV patients using interferon-free regimens compared to PR, by genotype, in Mongolia. PR, peginterferon + ribavirin; IF, interferon free

5.2.2. Sensitivity analysis

The majority of patients in Mongolia are infected with genotype 1. As a result, the differences in average cost per patient were most sensitive to costs relating to treatments for this group in the two scenarios: PR for 48 weeks and SOF + LDV for 12 or 24 weeks. Varying the cost of PR by 20% resulted in the difference in average cost per patient of between −US$ 6438 and −US$ 4216. The results were far less sensitive to changes in the cost of SOF + LDV (−US$ 5599 to −US$ 5055), non-drug costs with PR (−US$ 5623 to −US$ 5031) and non-drug costs with all interferon-free regimens (−US$ 5393 to US$ 5261).

Figure 5

Sensitivity of cost (2015 US$) per patient to changes in cost parameters (Mongolia). DCV, daclatasvir; LDV, ledipasvir; PR, pegylated interferon + ribavirin; SOF, sofosbuvir; R, ribavirin

5.3. Ukraine

5.3.1. Base case

Base case results for Ukraine are presented in Table 14. Total treatment costs for patients diagnosed with HCV in the PR scenario were estimated to be US$ 1.28 billion (cost per patient US$ 3173) and this would lead to 208 742 SVRs (cost per SVR of US$ 6150). Unlike Brazil and Mongolia, non-drug costs made up a majority of total costs, US$ 852.27 million compared to US$ 431.60 million for the drugs themselves. Costs for treating patients with cirrhosis were US$ 41.44 million and an estimated 6737 SVRs would be achieved in this population.

Table 14

Base case results for Ukraine (all costs in 2015 US$).

Interferon-free regimens were estimated to deliver 178 623 additional SVRs compared to PR alone, at a total cost of US$ 1.20 billion (cost per patient US$ 2953 and cost per SVR of US$ 3097), US$ 80.30 million lower than using PR. These lower overall costs were mainly the result of a large reduction in non-drug costs when using SOF + LDV to treat genotype 1 patients, as shown in Figure 6. The costs of treating patients with compensated and decompensated cirrhosis using interferon-free regimens were US$ 82.19 million.

Figure 6

Total costs (2015 US$) of treating all diagnosed HCV patients using interferon-free regimens compared to PR, by genotype, in Ukraine. PR, peginterferon + ribavirin; IF, interferon-free

5.3.2. Sensitivity analysis

The differences in cost per patient in Ukraine were most sensitive to non-drug costs with PR, varying from −US$ 641 when these costs were higher to US$ 202 when they were lower. The difference in cost also became positive (US$ 64) when the costs of SOF + LDV were high. Reductions in the average cost per patient ranged from −US$ 433 and −US$ 6 with the cost of PR, from −US$ 415 to −US$ 25 with the cost of DCV + SOF + R, from −US$ 317 to −US$ 122 with the non-drug costs with all DAA regimens, and from −US$ 302 to −US$ 137 with the cost of SOF + R.

Figure 7

Sensitivity of cost (2015 US$) per patient to changes in cost parameters (Ukraine). DCV, daclatasvir; LDV, ledipasvir; PR, pegylated interferon + ribavirin; SOF, sofosbuvir; R, ribavirin

6. Summary of findings

The results of the analysis demonstrate that for some countries, expanding access to treatment using interferon-free regimens may entail less cost than if traditional PR regimens were used. The main drivers of differences between Brazil and Mongolia were the ratio of prices for DAAs used in interferon-free regimens and PR, and the genotype distribution among HCV patients. Genotypes 2 and 3 patients, for whom the duration of treatment with PR is equal or close to that for interferon-free regimens, are found in larger numbers in Brazil compared to Mongolia. Mongolia also has a lower ratio of DAA to PR price per month. These two factors combined meant that the cost of providing interferon-free regimens was lower than for PR in Mongolia, while in Brazil there were significant additional costs. This pattern was consistent across all genotypes. In contrast, for Ukraine, the interferon-free scenario was less expensive than PR in genotypes 1 and 2 but more expensive for genotype 3. The additional costs for genotype 3 were mainly due to the high cost of the SOF + DCV + R combination used in this population. The lower costs overall with interferon-free regimens came from lower non-drug costs in genotypes 1 and 2 patients compared to costs with PR. This reflects the lower ratio of drug to non-drug costs compared to the other two countries.

7. Strengths and limitations

In making the case for expanded access to treatment for HCV patients, it is critical that planners are able to generate estimates of both the investment required to deliver new programmes and what the return on this investment will be, in both health and economic terms (10). Although this analysis only deals with the drug and other costs related to the period while on treatment, it is often these short-term costs that domestic budget holders are most concerned with (36). It also helps to highlight the data that are needed if more comprehensive analyses are to be carried out, and also how the relationship between the prices of different drugs and the epidemiology of HCV might influence the shape of future HCV treatment programmes.

Better local data on the epidemiology of HCV would have made it possible to more accurately quantify the costs of new treatment programmes. Progression from chronic HCV infection to cirrhosis is highly variable across countries, but generally occurs in between 10% and 15% of patients (37,38). This analysis was originally going to include a category for the costs of treating all diagnosed patients with a METAVIR score of either F3 or F4. These are the groups at the highest risk of going on to develop more serious, potentially life-threatening complications and some have suggested that in resource-limited settings, they should be granted priority access to treatment (39). Outside of a small number of case studies, however, data on fibrosis scores for patients in LMIC were not available from public sources. The analysis therefore had to use data from a modelling study, which itself utilized expert opinion for some key inputs that provided information only on the proportion of patients with a score of F4 or under. The lack of data on the severity of liver disease was also a major reason why it was not possible to include an African country within the analysis. Data on fibrosis scores may be available from HCV treatment centres within countries, and those involved in building the case for new funding for HCV programmes should look to incorporate these within their own analyses.

Mortality was also not factored into the analysis due to a lack of information on the characteristics of the population with chronic HCV, such as age, alcohol consumption, injection drug use and the prevalence of HIV coinfection. This approach was deemed acceptable as the model only covered a short time horizon, and clinical trials have shown that mortality is <1% while on treatment with interferon-free regimens (see accompanying systematic literature review). For analyses with a longer time horizon, increases in life expectancy as a result of treatment will be a large component of overall health benefits and will have to be accounted for.

The high costs of purchasing and delivering PR have been perhaps the biggest historical barrier to access to treatment for patients with HCV (15). This analysis shows that, at the prices currently being agreed for DAAs in LMIC, interferon-free regimens will in some cases reduce the overall costs of treatment. To maximize the benefits of the simplified administration, tolerability and efficacy of interferon-free regimens, countries should focus on how to deliver them in the most efficient manner, potentially through task-shifting or the integration of HCV treatment with other services, such as those for HIV patients (10,12). Understanding what level and type of monitoring is required to achieve high levels of adherence will be central to this.

The analysis would have benefited from the inclusion of scenarios looking at different methods of delivery and their associated costs, along with the costs of transitioning to them. However, as very limited data exist on the costs of current treatment programmes, it was possible to model only one scenario of the reductions in cost that interferon-free regimens might deliver. A key priority in all LMIC should be the collection of better data on the costs of new and existing HCV programmes. Some LMIC such as Egypt and Ukraine have already begun delivering interferon-free regimens to patients (40,41). These programmes can provide useful data on delivery costs, which other countries might use in analyses such as this one to help inform their own new programmes or in the scaling up of others. In the longer term, it will be necessary to produce more detailed evidence to support wider investment in HCV treatment services.

Recommended treatment for HCV is now highly patient specific, and this was reflected in the different interferon-free regimens that were applied to the patient categories within the model. Only one combination of interferon-free regimens was considered, however. In reality, the choice of which regimens are used will be dependent on the price and availability of different drugs within countries (10). The analysis would have benefited from a more in-depth exploration of the costs of different combinations of interferon-free regimens. Another consequence of the growth of available treatments is that there is now the real possibility of a patient being retreated with an alternative regimen if they fail to achieve an SVR with their primary treatment. This was not included within the analysis due to the short time horizon but should be considered in future models.

Although the cost per SVR was a secondary outcome, the analysis did make use of all available information relating to the efficacy and safety of interferon-free regimens when calculating the additional SVRs that they are estimated deliver. These data were primarily from trials carried out in high-income countries, however. The data for PR regimens was less robust as only trials with DAAs were included within the systematic literature review carried out in conjunction with this analysis. Nonetheless, it is clear from the results that treating patients with interferon-free regimens as opposed to PR is likely to result in substantially higher rates of SVR across all patient groups. The next step for countries will be to quantify what this would mean in terms of reducing the socioeconomic burden of HCV in the long term. This requires data on the costs of treating HCV patients with end-stage liver disease as well as the impact that these conditions have on both quality of life and life expectancy. Reductions in indirect costs, for example through increased productivity, should also be included.

8. References

1.

Lavanchy D. Evolving epidemiology of hepatitis C virus. Clin Microbiol Infect. 2011;17(2):107–15. [PubMed: 21091831]

2.

Thomas DL, Seeff LB. Natural history of hepatitis C. Clin Liver Dis. 2005;9(3):383–98. [PubMed: 16023972]

3.

Alberti A, Chemello L, Benvegnù L. Natural history of hepatitis C. J Hepatol. 1999;31(Suppl 1):17–24. [PubMed: 10622555]

4.

Gower E, Estes C, Blach S, Razavi-Shearer K, Razavi H. Global epidemiology and genotype distribution of the hepatitis C virus infection. J Hepatol. 2014;61(1 Suppl):S45–57. Epub 2014 Jul 1030. [PubMed: 25086286] [CrossRef]

5.

Mohd Hanafiah K, Groeger J, Flaxman AD, Wiersma ST. Global epidemiology of hepatitis C virus infection: new estimates of age-specific antibody to HCV seroprevalence. Hepatology. 2013;57(4):1333–42. [PubMed: 23172780]

6.

Global, regional, and national age–sex specific all-cause and cause-specific mortality for 240 causes of death, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet. 2015;385(9963):117–71. [PMC free article: PMC4340604] [PubMed: 25530442]

7.

Razavi H, Waked I, Sarrazin C, Myers RP, Idilman R, Calinas F, et al. The present and future disease burden of hepatitis C virus (HCV) infection with today's treatment paradigm. J Viral Hepat. 2014;21:34–59. [PubMed: 24713005]

8.

Hatzakis A, Chulanov V, Gadano AC, Bergin C, Ben-Ari Z, Mossong J, et al. The present and future disease burden of hepatitis C virus (HCV) infections with today's treatment paradigm – volume 2. J Viral Hepat. 2015;22:26–45. [PubMed: 25560840]

9.

Ford N, Singh K, Cooke GS, Mills EJ, von Schoen-Angerer T, Kamarulzaman A, et al. Expanding access to treatment for hepatitis C in resource-limited settings: lessons from HIV/AIDS. Clin Infect Dis. 2012;54(10):1465–72. [PubMed: 22431808]

10.

Suthar AB, Harries AD. A public health approach to hepatitis C control in low- and middle-income countries. PLoS Med. 2015;12(3):e1001795. [PMC free article: PMC4355406] [PubMed: 25757228]

11.

Lemoine M, Nayagam S, Thursz M. Viral hepatitis in resource-limited countries and access to antiviral therapies: current and future challenges. Future Virol. 2013;8(4):371–80. [PMC free article: PMC3646239] [PubMed: 23662157]

12.

Ford N, Swan T, Beyer P, Hirnschall G, Easterbrook P, Wiktor S. Simplification of antiviral hepatitis C virus therapy to support expanded access in resource-limited settings. J Hepatol. 2014;61(1, Supplement):S132–S138. [PubMed: 25443341]

13.

Manns MP, Wedemeyer H, Cornberg M. Treating viral hepatitis C: efficacy, side effects, and complications. Gut. 2006;55(9):1350–9. [PMC free article: PMC1860034] [PubMed: 16905701]

14.

Ford N, Kirby C, Singh K, Mills EJ, Cooke G, Kamarulzaman A, et al. Chronic hepatitis C treatment outcomes in low- and middle-income countries: a systematic review and meta-analysis. Bull World Health Organ. 2012;90:540–50. [PMC free article: PMC3397704] [PubMed: 22807600]

15.

Milani B, Gaspani S. Pathway to affordable, quality-assured sources of pegylated interferon alpha for treating hepatitis C. GaBI Journal. 2013;2(4):194–203.

16.

Ministry of Health and Population. Plan of action for the prevention, care & treatment of viral hepatitis, Egypt 2014–2018. Cairo: Ministry of Health and Population; 2014. [16 November 2015]. http://www​.emro.who.int​/images/stories/egypt​/VH_Plan_of_Action_FINAL_PRINT1.pdf.

17.

Gritsenko D, Hughes G. Ledipasvir/sofosbuvir (Harvoni): improving options for hepatitis C virus infection. P T. 2015;40(4):256–76. [PMC free article: PMC4378517] [PubMed: 25859119]

18.

Blatt CR, Bernardo NLMdC, Rosa JA, Bagatini F, Alexandre RF, Neto GB, et al. An estimate of the cost of hepatitis C treatment for the Brazilian health system. Value Health Reg Issues. 2012;1(2):129–35. [PubMed: 29702891]

19.

Steinbrook R, Redberg RF. The high price of the new hepatitis C virus drugs. JAMA Intern Med. 2014;174(7):1172. [PubMed: 24798183]

20.

Gilead announces generic licensing agreements to increase access to hepatitis C treatments in developing countries. Sep 15, 2014. [16 November 2015]. [press release] http://www​.gilead.com​/news/press-releases​/2014/9/gilead-announces-generic-licensing-agreements-to-increase-access-to-hepatitis-c-treatments-in-developing-countries.

21.

Bristol-Myers Squibb. HCV developing world strategy. 2015. [20 August 2015].

22.

Gilead Sciences. Tiered pricing. 2015. [25 August 2015].

23.

World population prospects: the 2012 revision. New York: United Nations; 2012. [16 November 2015].

24.

Dusheiko G. Side effects of α interferon in chronic hepatitis C. Hepatology. 1997;26(S3):112S–121S. [PubMed: 9305675]

25.

Gane EJ, Hyland HR, An D, Svarovskaia E, Pang PS, Brainard D, et al. Ledipasvir plus sofosbuvir, with or without ribavirin, for 12 weeks in patients with hepatitis C genotype 3 or 6 infection. Gastroenterology. 2015;149(6):1454–61. [PubMed: 26261007]

26.

Jacobson IM, Gordon SC, Kowdley KV, Yoshida EM, Rodriguez-Torres M, Sulkowski MS, et al. Sofosbuvir for hepatitis C genotype 2 or 3 in patients without treatment options. N Engl J Med. 2013;368(20):1867–77. [PubMed: 23607593]

27.

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):211–21. [PubMed: 24428467]

28.

Foster GR, Hézode C, Bronowicki JP, Carosi G, Weiland O, Verlinden L, et al. Telaprevir alone or with peginterferon and ribavirin reduces HCV RNA in patients with chronic genotype 2 but not genotype 3 infections. Gastroenterology. 2011;141(3):881–9. [PubMed: 21699786]

29.

Hepatitis C in Eastern Europe and Central Asia. Civil society response to the epidemic. International HIV/AIDS Alliance in Ukraine, International Treatment Preparedness Coalition in Eastern Europe and Central Asia; Oct, 2015. (http://www​.aidsalliance​.org.ua/ru/news/pdf/28​.10.2015/EECA%20HCV%20EN.pdf)

30.

Jazag A, Puntsagdulam N, Chinburen J. Status quo of chronic liver diseases, including hepatocellular carcinoma, in Mongolia. Korean J Intern Med. 2012;27(2):121–7. [PMC free article: PMC3372793] [PubMed: 22707881]

31.

Barreto AMEC, Takei K, E CS, Bellesa MA, Salles NA, Barreto CC, et al. Cost-effective analysis of different algorithms for the diagnosis of hepatitis C virus infection. Braz J Med Biol Res. 2008;41:126–34. [PubMed: 18235967]

32.

Gudielines for the screening, care and treatment of persons with hepatitis c infection. Geneva, Switzerland: World Health Organization; 2014. [PubMed: 25535634]

33.

Banco Central do Brasil. Economic indicators. Nov 11, 2015. [17 July 2015]. [website]

34.

Board of Governors of the Federal Reserve System. Country data: historical rates for the Brazilian Real. Nov 09, 2015. [17 July 2015]. [website]

35.

Cost–effectiveness and strategic planning (WHO-CHOICE). Health service delivery costs. 2015. [20 July 2015]. [web page]

36.

Hauck K, Smith PC, Goddard M. The economics of priority setting for health care : a literature review. Washington DC: World Bank; 2004.

37.

Chen SL, Morgan TR. The natural history of hepatitis C virus (HCV) infection. Int J Med Sci. 2006;3(2):47–52. [PMC free article: PMC1415841] [PubMed: 16614742]

38.

National Institutes of Health Consensus Development Conference Statement: Management of hepatitis C: 2002. June 10-12, 2002. Hepatology. 2002;36(5B):s3–s20. [PubMed: 12407572]

39.

Obach D, Yazdanpanah Y, Esmat G, Avihingsanon A, Dewedar S, Durier N, et al. How to optimize hepatitis C virus treatment impact on life years saved in resource-constrained countries. Hepatology. 2015;62(1):31–9. [PubMed: 25581111]

40.

International HIV/AIDS Alliance in Ukraine. The cutting edge drug to treat patients with hepatitis C is finally in Ukraine. May 26, 2015. [16 November 2015]. [press release] http://www​.aidsalliance​.org.ua/ru/news/pdf​/01-12_2015/05/Sofosbuvir​%20in%20Ukraine_en.pdf.

41.

Doss W, Esmat G, Serafy ME, El Sayed MH, Hassany M, Youssry A, et al. Interim analysis for sofosbuvir national treatment program in Egypt. J Viral Hepat. 2015;22(Suppl S2):94. [16 November 2015]; [P149] http:​//onlinelibrary​.wiley.com/store/10.1111/jvh​.137_12425/asset/jvh137_12425​.pdf?v​=1&t=ih24z3qs&s​=3af350230add4fe381f86531368260b7d7ed3d06.

Appendix A. Cost inflation calculations