Cost-Effectiveness and Cost Thresholds of Generic and Brand Drugs in a National Chronic Hepatitis B Treatment Program in China

Chronic liver disease and liver cancer associated with chronic hepatitis B (CHB) are leading causes of death among adults in China. Although newborn hepatitis B immunization has successfully reduced the prevalence of CHB in children, about 100 million Chinese adults remain chronically infected. If left unmanaged, 15–25% will die from liver cancer or liver cirrhosis. Antiviral treatment is not necessary for all patients with CHB, but when it is indicated, good response to treatment would prevent disease progression and reduce disease mortality and morbidity, and costly complications. The aim of this study is to analyze the cost-effectiveness of generic and brand antiviral drugs for CHB treatment in China, and assessing various thresholds at which a highly potent, low resistance antiviral drug would be cost-saving and/or cost-effective to introduce in a national treatment program. We developed a Markov simulation model of disease progression using effectiveness and cost data from the medical literature. We measured life-time costs, quality adjusted life years (QALYs), incremental cost-effectiveness ratios (ICERs), and clinical outcomes. The no treatment strategy incurred the highest health care costs ($12,932-$25,293) per patient, and the worst health outcomes, compared to the antiviral treatment strategies. Monotherapy with either entecavir or tenofovir yielded the most QALYs (14.10–19.02) for both HBeAg-positive and negative patients, with or without cirrhosis. Threshold analysis showed entercavir or tenofovir treatment would be cost saving if the drug price is $32–75 (195–460 RMB) per month, highly cost-effective at $62–110 (379–670 RMB) per month and cost-effective at $63–120 (384–734 RMB) per month. This study can support policy decisions regarding the implementation of a national health program for chronic hepatitis B treatment in China at the population level.

Introduction previous study on CHB in Shanghai, China [6]. Treatment-naïve, chronic HBV, HBeAgpositive or HBeAg-negative patients eligible for treatment under international treatment guidelines enter the model either in the cirrhotic or non-cirrhotic health state (Fig 1). Patients can progress to compensated cirrhosis, decompensated cirrhosis, hepatocellular carcinoma, and would be eligible to receive a liver transplantation. All patients face age-specific mortality plus increased mortality if they have cirrhosis, HCC, or a liver transplant. If patients receive treatment, they can develop drug resistance or sustained virologic response. Strategies Eight different strategies were analysed in this study. "No antiviral treatment.". Chronic hepatitis B patients progress according to the natural history, following annual disease progression estimates ( Table 1). The disease progression estimates were derived from recent age-specific cohort studies on inactive and active CHB in Asia [7][8][9][10][11][12][13][14][15]. We assumed that patients received best supportive care, except for drug treatment. Patients followed the natural history according to their HBeAg and disease status (with or without cirrhosis). Spontaneous virologic response was defined as seroconversion to anti-HBe in HBeAg-positive patients, and as persistent HBV DNA suppression and ALT normalization in HBeAg-negative patients. We assumed that a proportion of patients with decompensated cirrhosis and HCC became eligible for liver transplantation.
"Lamivudine mono-therapy" (LAM). Patients receive 100mg orally once daily of the first licensed antiviral HBV drug. This drug is known to be associated with a high incidence of resistance [16]. Such mono-therapy is still commonly prescribed in China [17]. We assigned different rates of virologic response under long-term therapy between resistant and non-resistant patients (Table 2) [16,[18][19][20].
"Entecavir monotherapy" (ETV). Patients in this strategy received 0.5mg entecavir once daily. The treatment related probability estimates for responding and resistant patients are shown in Table 2 [24][25][26]. Since entecavir and lamivudine share cross-resistance, it is not a recommended salvage therapy for lamivudine resistant patients.
"Pegylated Interferon" (PEG-IFN). Patients receive 180mcg of pegylated interferon administered subcutaneously once a week for 48 weeks. If the patient does not respond or relapses in the second year after treatment, they follow the transitions in the natural history (no treatment) strategy.

Assumptions
Causes of death not related to liver disease associated with CHB were included in the model, based on age-specific mortality rates from the National Bureau of Statistics China [38]. The annual probabilities of receiving a liver transplant for decompensated cirrhosis and HCC (12% and 4.7%, respectively) in China were calculated based on data from China Liver Transplant Registry [39].
If progression rates were reported, these were transformed into annual probabilities using a standard formula (P = 1-e -r×t ), where P is the probability, e is the base of the natural logarithm, r is the event rate, and t is the time interval [40]. We assumed that it was possible to develop cirrhosis and hepatocellular carcinoma while on treatment, but with a 50% reduction in the rate decrease from the natural history [41]. To HCC All ages 0.8 (0.5-1.0) 9 To HBV-related death All ages 0.6 (0.2-0.9) 9 From active CHB, HBeAg-negative To inactive CHB, HBsAg-positive All ages 1.6 (0.0-11) 8 To cirrhosis All ages 2.4 (1.3-3.4) 9 To HCC All ages 0.8 (0.5-1.0) 9 To HBV-related death All ages 0.6 (0.2-0.9) 9 From seroconversion

From cirrhosis
To decompensated cirrhosis All ages 3.9 (3.2-4.6) 9 To HCC All ages 5.0 Treatment guidelines often recommend a finite period of therapy with oral nucleoside analogs for patients with HBeAg-positive CHB who undergo HBeAg seroconversion. However, prolonged therapy is to be considered for patients with evolving HBeAg-negative CHB and active HBV DNA replication (> log 4 IU/ml) and for patients with cirrhosis [42][43][44][45][46]. A more recent publication [47] suggests continuation of long-term nucleos(t)ide analogue treatment, irrespective of the occurrence of HBeAg seroconversion in HBeAg-positive patients. Following these recent findings, our model assumes continued antiviral therapy for HBeAg-positive patients even if seroconversion occurs. Also our model assumes that the resistance rate for ETV and TDF remains as low as recent studies report [33,48]. The scenario for PEG-IFN assumes that non-responders and relapsers continue with long-term ETV treatment both in HBeAg-positive and negative patients.

Cost and utility estimates
We used generic antiviral drug costs for the base case analysis and examined brand drug costs in the sensitivity analysis. Generic drugs are widely available in China, and whether they are prescribed often depends on the physician's preference and the choice of the patient based on their health insurance coverage and out of pocket costs. The generic and brand drug costs were obtained from official prices approved by the Shanghai Municipal Bureau of Pricing. Prices were available for all generic antiviral drugs except for PEG-IFN and TDF which are not yet available in generic form in China. The medical management costs for CHB and other related costs were derived from the retrospective analysis of the medical records of patients with CHB by Zhiqiang et al. [49].
All costs were inflated to 2014 prices using China National Healthcare Index from the National Bureau of Statistics of China and converted to US dollars using the exchange rate as of 2014 (1 USD = 6.23 RMB). The age-specific utilities were obtained from a multinational study on hepatitis B [50] (Table 3). Following the World Health Organization guidelines for cost-effectiveness estimates [51], we regarded the incremental ratio of less than one times the gross domestic product (GDP) per capita for each QALY gained as an cost-effective intervention, and an ICER between one and three times GDP per capita per QALY as a potential cost-effective intervention, in which the GDP per capita in China in 2013 was $6,800 (41,775 RMB) [52].

Sensitivity analysis
A Monte Carlo simulation was conducted, assuming that all variables followed a triangular distribution, with base case, low and high range values. We simulated 10,000 trials and plotted the results on willingness to pay threshold acceptability curves. We also analyzed at which threshold To HBV-related death All ages 15.0 (9.9-20.0) 9

Base case
A cost-effectiveness plot of the various scenarios according to HBeAg status with or without cirrhosis is shown in Fig 2. For the base-case analysis, the lowest available prices were taken for  To  Table 4 and for the branded price outcomes in S2 Table. The health outcomes for each subgroup in terms of morbidity (decompensated cirrhosis, hepatocellular carcinoma, and liver transplantation) and mortality are shown in Fig 3. Non-cirrhotic HBeAg-positive patients in the no CHB treatment strategy would likely incur the highest health care costs, and the worst health outcomes, compared to the other strategies. With no antiviral treatment, each HBeAg-positive patient had $12,932 in total healthcare costs, 11.8 QALYs and 65% of them would die of hepatitis-B related liver disease, and 32% will develop HCC over their lifetime. When comparing the different therapies, some therapies appeared better than others. LAM!ADV salvage gave patients 18.0 QALYs at a cost of only $844 per QALY gained. Both ETV and TDF monotherapy offered the highest benefit, 19.0 QALYs, and 7.2 QALYs gained over the no treatment strategy. LAM monotherapy, PEG-IFN monotherapy, and the PEG-IFN!ETV salvage therapy did not appear cost-effective compared to the other therapies, so were all dominated (which means the intervention costs more and is less effective than the comparator). LAM monotherapy provides fewer QALYs at a higher costeffectiveness ratio than LAM!ADV salvage (known as "dominated by extended dominance" in health-economics). PEG-IFN!ETV salvage therapy had fewer QALYs at a higher cost than ETV alone (known as "absolutely dominated" in health economics). If a low resistance, highly potent drug such as ETV or TDF is used for treatment of HBeAg positive patients who met the treatment criteria, approximately 60% of HBV-related mortality and 31% of HCC can be prevented. Non-cirrhotic HBeAg-negative patients in the no treatment strategy had $11,735 in costs, 12.2 QALYs and 59% of them would die of hepatitis-B related liver disease, and 30% will develop HCC over their lifetime. When comparing the various therapies, both ETV and TDF offered the highest benefits, 17.7 QALYs and 5.5 QALY gained over no treatment. With no generic TDF and the high cost of brand TDF in the Chinese market, the ICER for generic ETV is $4,066 while the ICER for TDF is $1.5 million per QALY in China. Approximately 49% of HBV-related deaths, and 26% of HCC can be prevented if a low resistance, highly potent drug such as ETV or TDF is used.  If CHB cirrhotic patients are left untreated, 95% would die of HBV related liver disease, and 39% would develop HCC over their lifetime.
For cirrhotic HBeAg-positive cirrhotic patients, LAM!ADV salvage was the lowest cost alternative, it dominated no treatment and LAM therapy alone. LAM!TDF salvage was the next-best option, which added 1.9 QALYs over LAM!ADV salvage therapy at a cost of $15 per QALY. PEG-IFN monotherapy and PEG-IFN!ETV was dominated by ETV monotherapy. Nearly 79% of HBV-related mortality and 33% of HCC in HBeAg-positive patients can be prevented if they are treated with a low resistance, highly potent drug such as ETV or TDF.
For cirrhotic HBeAg-negative cirrhotic patients, LAM!ADV salvage would provide 12.67 QALYs and has the lowest ICER, $79 per QALY. ETV monotherapy or TDF monotherapy had the best health outcomes with 14.0 QALYs, at an additional cost of $2,583, and $1.8 million per QALY, respectively. Approximately 61% of HBV-related mortality and 26% of HCC in HBeAg-negative cirrhotic patients can be prevented if they are treated with a low resistance, highly potent antiviral drug such as ETV or TDF.

Sensitivity analysis
The results of the probabilistic sensitivity analysis (Fig 4) indicated that unless the cost of TDF drops, among all the therapies, generic ETV is likely the most cost effective therapy for both HBeAg-positive and negative patients with or without cirrhosis in China. For non-cirrhotic HBeAg-positive patients, ETV was 24% likely to be cost-effective at a willingness to pay threshold of $6,800 (1xGDP), and increased to 42% likely to be cost-effective at $20,400 (3xGDP). For the non-cirrhotic HBeAg-negative patients, ETV was 43% likely to be cost-effective at a threshold of $6,800 and further increased to 54% likely to be cost-effective, when compared to other treatments, at a threshold of $20,400. For cirrhotic HBeAg-positive patients, ETV was 32% likely to be cost-effective at a threshold of $6,800 and 55% likely to be cost-effective at a threshold of $20,400. For cirrhotic HBeAg-negative patients, ETV was 45% likely to be cost-effective at a threshold of $6,800 and 48% likely to be cost-effective at a threshold of $20,400. This shows the ETV is most likely to be cost-effective at commonly cited thresholds for cost-effectiveness. We did an additional analysis on the costs of the two highly potent and low resistance profile drugs, ETV and TDF, to determine the thresholds at which treatment becomes cost-saving (with a willingness to pay (WTP) threshold of $0), highly-cost-effective (WTP $6,800, 1xGDP) and cost-effective (WTP $20,400, 3xGDP) for a national procurement and treatment program in China, compared to no treatment, for all sub-groups (Table 5). We chose ETV and TDF since they were the most effective strategies and had the highest health outcomes in all subgroups. In non-cirrhotic HBeAg-positive patients, ETV and TDF are cost-saving if the drug cost is below $523/year ($44/month), and considered highly cost-effective at $1,008/year ($84/ month), and cost-effective at $1,173/year ($98/month). In non-cirrhotic HBeAg-negative patients, ETV and TDF are cost-saving if the drug cost is below $381/year ($32/month), and considered highly cost-effective at $1,313 year ($109/month) and cost-effective at $1,442 ($120/month). For cirrhotic HBeAg-positive patients, ETV and TDF are cost-saving below $899 ($75/month), and considered highly cost-effective at $1,027/year ($85/month), and costeffective at $1,284/year ($107/month). For cirrhotic HBeAg-negative patients, ETV and TDF are cost-saving below $886 ($74/month), and considered highly cost-effective at $1,120/year ($93/month), and cost-effective at $1,065/year ($89/month).

Discussion
We examined the cost-effectiveness of treatment for four sub-groups of CHB patients; HBeAgpositive and HBeAg-negative patients with or without cirrhosis. We found neither pegylated interferon with or without salvage therapy were cost effective. Lamivudine monotherapy was  also not cost effective. For patients who were previously on lamivudine treatment and developed drug resistance, salvage therapy by switching to generic adefovir was cost saving in cirrhotic, HBeAg-positive patients, and cost effective in cirrhotic, HBeAg-negative patients and in non-cirrhotic patients. Salvage therapy by switching to tenofovir was cost effective in cirrhotic patients and in non-cirrhotic, HBeAg-positive patients. In our study we found entecavir or tenofovir monotherapy treatment in non-cirrhotic patients would prevent 49-69% of liver related deaths and 26-31% of HCC, and would prevent 61-79% of liver related deaths and 26-33% of HCC in cirrhotic patients. The QALYs gained from entercavir or tenofovir treatment compared to no treatment ranged from 5.5-7.3 for non-cirrhotic patients, and 9.3 to 12.1 for cirrhotic patients. These QALY gains are impressive, since few treatments for chronic non-communicable diseases or chronic infectious diseases result in such a gain in healthy life years. The QALYs gained by treatment is similar to that reported in HIV.
Among the two highly potent, low resistance drugs, entecavir is available as a branded drug at about $196/month (1,204 RMB) or as the less costly generic drug at about $102/month (626 RMB) for CHB treatment. Branded tenofovir is available to the public health system in China for the treatment of HIV/AIDS at the cost of less than $30/month (184 RMB) [53]. Even lower-cost generic tenofovir priced at as little as $5/month (30 RMB) for HIV treatment is available for low-income countries through the Global Fund [53,54]. Recently, several lowincome countries in Asia also have access to generic tenofovir for CHB treatment. Unfortunately, following approval of tenofovir for CHB treatment in China, there is no generic tenofovir available and the price of branded tenofovir for CHB treatment in China is high at $240/ month (1,474 RMB), making it out of reach for most infected persons and the public health treatment programs in China.
Our sensitivity analysis indicated that at current prices of the available CHB treatment in China, entecavir is highly likely to be the preferred therapy. According to our threshold analysis, in order for entecavir to be cost-saving, meaning that using this strategy will have the lowest total costs with the best health outcomes, the drug price needs to drop to $671/year ($56/month (344 RMB)) from its current price of $1,229/year ($102/month (626 RMB)) in China. If the price of branded tenofovir for CHB treatment drop to the same level as the price afforded for HIV treatment in China (at less than $30/month), tenofovir treatment would be cost saving.
According to Li et al., a majority of physicians in China still prescribe lamivudine as first line CHB therapy and was prescribed by 54% of physicians in urban areas and 37% of physicians in rural areas, whereas, entecavir was prescribed by 5.4% of physicians in urban and 10% of those in rural areas [55]. Although several studies on the cost effectiveness of CHB treatments in China have been published, our study is the first to include tenofovir, the first to compare nucleoside, nucleotide and interferon therapies and include all four treatment subgroups, and the first to calculate cost thresholds for the most effective treatments. Another strength of our study is that we use age-specific transition estimates taken specifically from recent Chinese cohort studies to capture disease progression in this population as closely as possible. One limitation of our study is that we were unable to find literature on the rate of progression to cirrhosis and HCC for individuals who experience sustained virological response, so we assumed there is a 50% decrease in the rate of disease progression compare with the estimate of patients on treatment without sustained viral response [41,56].
For a treatment program to be cost-saving, the price of the two highly potent low resistance drugs, entecavir and tenofovir would need to cost between $32-75/month . For it to be highly cost-effective or cost-effective, the drug price would need to be lowered to $62-120/month (380-737 RMB). Our study was undertaken to potentially support policy decisions regarding implementation of a national public health treatment program for CHB in China at the population level. This study not only compares the health outcomes but the impact of a reduction in the price of antiviral therapy. Our study could also be adapted to provide similar cost benefit estimates for decision makers in other endemic countries in developing their national viral hepatitis action plan.
Supporting Information S1