Health economic evaluation of immunization strategies of hepatitis E vaccine for elderly population

ABSTRACT

Objective This study was conducted to assess the cost-effectiveness of hepatitis E vaccination of elderly population in the sporadic regions in China.

Methods We used a decision tree-Markov model to evaluate the cost-effectiveness of 3 kinds of hepatitis E virus vaccination strategies from societal perspectives. Parameter estimates were obtained from published researches and experts' opinion. The time horizon was 16 years, and the discounted rate was 3% annually. Costs are expressed in 2016 US dollars.

Results The universal vaccination strategy had an incremental cost-effectiveness ratio (ICER) of US$ 8475.90 per QALY gained versus no vaccination. The implementation of screening and vaccination strategy would have an ICER of US$ 4044.28, compared with no vaccination. The vaccination was cost-effective (ICER< 3 times China's per capital gross domestic product/quality-adjusted life years). The QALY of asymptomatic infection, vaccine coverage and vaccine protection are the important parameters impacting the ICER in one-way sensitivity analysis and screening and vaccination being the dominant strategy in probabilistic sensitivity analysis.

Conclusion This analysis indicates that screening and vaccination is the most cost-effective hepatitis E intervention strategy of elderly population in sporadic region in China.

Introduction

Hepatitis E infection caused by the hepatitis E virus (HEV) is a communicable disease. It is transmitted mainly through fecal-oral route and it remains a public health concern in the developing countries of Asia, Africa, and Latin America. The hepatitis E virus (HEV) is a major etiological agent of acute viral hepatitis identified in 1983.¹

There are 4 genotypes of HEV that infect humans and one serotype.² HEV genotype 4 is endemic or sporadic in 63 countries.³ The HEV genotype 4 caused most of the sporadic cases in China.^4-7 Moreover, HEV genotype 4 is associated with high morbidity and mortality among the aged population.^8-10 There are no globally-recognized guidelines for treatment of hepatitis E. As such evaluating the cost effectiveness of different vaccination strategies could serve as a powerful measure in the prevention and control of hepatitis E infection.¹¹ There are currently 2 vaccines available based on human-HEV sequences, but the HEV239 vaccine is the only vaccine that has been commercialized in China for aged more than 16 y old.^{12, 13} A large-scale randomized double-blind placebo-controlled phase 3 clinical trial, showed vaccine efficacy of 3-doses hepatitis E vaccine was 100% (95% CI 72·1–100%).¹⁴ A recent study published in the New England Journal of Medicine reported that after 4.5 years, the efficacy of the vaccine was 93.3% (95% CI 78.6%–97.9%). Furthermore, most adverse events were mild.¹⁵ As such recommendations concerning vaccination strategies urgently need to be explored.¹¹

To assess immunization strategy, a model has to be established since it is difficult to obtain long-term follow-up data of vaccination. The decision tree model is simple in structure and can be used to compare cost and effectiveness of different programs in the same period as well as time dependence of parameters. It also takes into account the mutual influence that exists between events.¹⁶ The progress of disease is described by Markov model via interconnected transition probability.¹⁷ Moreover, the economic analysis of elderly population had been constructed using Markov model.^18,19 Therefore, the present study intended to use a decision tree -Markov model to analyze the health economic benefit of immunization strategies of elderly population living in HE sporadic regions in China.

Results

Base-case analysis

Table 1 demonstrated the ICER values of 3 immunization strategies evaluated by the decision tree -Markov model. The results revealed that the strategy of screening and vaccination was the dominant strategy for elderly population (ICER less than 3 times GDP per capita) compared with no intervention strategy. Compared with no vaccination, universal vaccination strategy would reduce HEV infection by 5019 and dead by 106 and screening and vaccination strategy would reduce HEV infection by 6370 and dead by 138. Moreover, universal vaccination and screening and vaccination strategy would save 12825.09 and 15968.47 QALYs, respectively. The incremental cost-effectiveness ratio (ICER) of universal vaccination and screening and vaccination strategies were US$ 8475.90 per QALY gained and US$ 4044.28 per QALY gained respectively compared with no intervention strategy. Compared with universal vaccination strategy, the screening and vaccination strategy was cost-saving and produced more QALYs gained with an ICER<0.

Table 1.

Health economic evaluation of different strategies of immunization against hepatitis E virus in elderly population.

Variable	universal vaccination	Screen and vaccination	no vaccination
Total cost($)	128749816	84626522.40	20045567
Incremental cost($)	108704249	64580955.40	—
Total QALYs	14256773.74	14259917.12	14243948.65
Incremental QALYs	12825.09	15968.47	—
ICER	8475.90	4044.28	—

Note: QALYs, quality adjusted life years; ICER, incremental cost- utility ratio.

When calculating ICER, take all no vaccination strategies was used as control.

Sensitivity analysis

To determine the impact of parameter uncertainty on the ICERs, we conducted both one-way sensitivity analysis and probabilistic sensitivity analysis using a range of scenarios. For screening and vaccination strategy compared with no vaccination, the results of a one-way sensitivity analysis showed that no parameter had a significant impact on the ICERs. But the vaccine coverage, vaccine protection and losses of vaccine protection were the most influential variables on the ICERs, compared with universal vaccination strategies. The results of probabilistic sensitivity analysis based on Monte Carlo simulations confirmed that, compared with no vaccination, the most point of ICERs of vaccination strategies were located below WTP(willingness-to-pay) thresholds, it manifested that the vaccination strategies was the dominant scenario in elderly population. Compared with universal vaccination, most points of ICERs of the screening and vaccination strategy were located in the third and fourth quadrant. However, less than half of points of ICERs were located below WTP (willingness-to-pay) thresholds. It indicated that the probability of absolute dominant strategy is about less than 50% for screening and vaccination strategies (Fig. 1).

Figure 1.

Result of probabilistic sensitivity analysis (The slash represents the WTP threshold) (a) ICER of universal vaccination VS no vaccination (b) ICER of screening and vaccination VS no vaccination (c) ICER of screening and vaccination VS universal vaccination.

Discussion

According to the World Health Organization, an estimated 20 million people get infected with HEV, 3 million cases and 56600 deaths occur annually worldwide.²⁰ The report of notifiable infectious diseases in China shows that the incidence of hepatitis A has been decreasing gradually, however the incidence of hepatitis E has been on the rise since 2004.²¹ To date, there is no specific drug for hepatitis E. Vaccination is considered one of the most important means of prevention of infectious diseases. Development and application of hepatitis E vaccine has recently become the center of focus. It is difficult to develop inactivated HE vaccine and attenuated live HE vaccine because the HEV cell culture conditions are not perfect.²² Therefore, research of HEV vaccine is genetic engineering vaccine which mainly aimed at ORF2 gene. At present, there are only 2 hepatitis E vaccines: 56 K_D vaccine and HEV239 vaccine. Result of phase 2 clinical trial showed that the vaccine protection rate of 56KD was 95.5% (95% CI: 85.6%–98.6%), the phase 3 clinical trial revealed that the vaccine protection rate of HEV239 vaccine was 100% (95% CI: 72.1%–100.0%) and 93.3% (95% C1:78.6%–97.9%) 4.5 y after vaccination. There have been no serious adverse vaccine events during these phases.^14,15,23 Currently HEV 239 vaccine has been approved for prevention of hepatitis E in China.

The incidence and severity of the elderly is significantly higher than other age groups during the case in China caused by HEV genotype 4. From the societal system perspective and parameters at baseline, our health economic analysis of sporadic regions of elderly population demonstrated that vaccination is a cost-effective intervention. Furthermore, compared with universal vaccination strategy, the screening and vaccination strategy was less costly but more effective.

The probabilistic sensitivity analysis indicated that compared with no intervention, most of the ICER scatter in the first quadrant below WTP threshold, the vaccination strategy was the more economically strategy. We could conclude that vaccination could be a cost-effective strategy for elderly, conform to the argument that the elderly population was the first choice for hepatitis E vaccine. The vaccine protection, vaccine coverage, QALY of asymptomatic infection and vaccine protection are the important parameters which influences the vaccination strategies.

There are some limitations in the present study: (1) efficacy of HE vaccine, decline rate of vaccine efficacy, and the quality of life of asymptomatic were important parameters of immunization strategies. But due to lack of decline rate of vaccine efficacy and the quality of life of asymptomatic data, no verifications were made to the model. So we need more in-depth research regarding disease burden of hepatitis E and more long-term observation about the changes of efficacy of HE vaccine. It is only through this way we can get the model parameters more accurately as well as the model results to be more reliable. (2) It does not take into account the herd immunity, secondary attack rate, vaccine compliance and efficacy of different doses. And in this study, the cost and the utility parameters were both from a highly prevalent area in China,²⁴ so the immunization strategies in middle and low endemic areas should be further explored.

In conclusion, the present study demonstrated that screening and vaccination is the dominant strategy for the elderly population living in hepatitis E sporadic regions and provided valuable information to establish a more comprehensive evaluation of vaccination for policy makers in China.

Materials and methods

Model structure

In this study, we assumed a single 60-year-old cohort of 1,000,000 individuals enrolled in 2016 using the Hepatitis E vaccine. The model involved a 16-year time horizon (the average life expectancy in China) with 1-year cycles for elderly population. We evaluated 3 intervention strategies: no vaccination, universal vaccination, and screening and vaccination. The definition of screening and vaccination is all subjects were initially tested for the presence of HEV antibody, and if susceptible, then they would be vaccinated. Among the vaccinated population, we estimated 100% of cohort members received 3 doses of the vaccine. In addition, the cost-effectiveness of vaccination was evaluated from the societal perspectives.

The economic analysis was conducted using a decision tree-Markov model (Fig. 2). The model included 5 kinds of disease states: susceptible to infection, HE infection, nature immune, vaccine-induced immunity, and death. Based on the efficacy and safety of the HEV239 vaccine, Persons with vaccine-induced immunity may have 2 outcomes: immunization success or immunization failure. Populations of immunization failure and unvaccinated population were referred to as susceptible people. Asymptomatic infection and symptomatic infection can occur after exposure to HEV in susceptible population. According to the severity of the disease, the people in symptomatic infection could be treated in hospital as outpatients or inpatients. The hospitalized patients had 2 kinds of outcomes: survival or death due to HEV infection.

Figure 2.

Framework of decision tree-Markov model for elderly population.

Several assumptions are used in the model: (1) it is assumed that 100% of cohort members would receive the HE vaccine at 0, 1 and 6 months via intramuscular injection, with 100% vaccine coverage;¹⁴ (2) the herd immunity, adverse effect and secondary attack rate are not considered; (3) The patients with symptomatic infection had the behavior of hospital visits, and no untreated cases; (4) Assuming Markov model cycle of one year, the time horizon of the analysis was based on the average life expectancy, death is considered an absorbing state.

Model parameters

The parameters of model were summarized in Table 2. The baseline value of probability parameter was determined based on the searches of published literature and experts' opinion: the infection rate of HE was 0.140,^25-28 symptomatic infection rate was 0.026,^29,30 natural immune rate was 0.506,^{25-28, 31-36} antibody negative-conversion rate was 0.081,^25,27,28 hospitalization rate was 0.500^14,15 and the baseline of vaccine protection rate was 0.933.^14,15 The average life expectancy from the national census in 2010 is 74.83 y and population-based mortality rate was 0.007, the mortality rate of hepatitis E is 0.056–0.063 according to investigation of China's hepatitis E disease burden in 2010.³⁷

Table 2.

Model parameters.

Parameter	Baseline	Range	distribution	Source
life expectancy	74.83	—	—	NBSC
probability
attack rate	0.140	0.069–0.158	Triangular	25-28
Symptomatic rate	0.026	± 50%	Triangular	29,30
Natural immunity rate	0.506	0.329–0.593	Triangular	25-28, 31-36
IgG antibody negative rate	0.081	± 50%	Triangular	25,27,28
Hospitalization rate	0.500	0.328–0.731	Triangular	14, 15
Mortality rate				37
60–69 y	0.056	± 20%	Triangular
≥ 70 y	0.063	± 20%	Triangular
Vaccine protection	0.933	0.721–1.000	Triangular	14, 15
loss of vaccine protection	0.015	0–0.030	Triangular	15
discount rate	3%	0–0.05	Triangular
Cost ($)
Vaccine price (per dose)	36.64	5.75–71.97	Triangular	44-46
Administration cost (per dose)	1.72	0.45–5.44	Triangular	38
Screening test	7.24	5.02–8.72	Triangular	PBJP
Outpatient	126.90	4.26–448.15	Triangular	38, 40
Hospitalizations	2381.53	1029.87–22330.13	Triangular	39
Deaths	11643.59	1913.78–22402.97	Triangular
Per capita GDP	7292.68	—	—	NBSC
Utilities (QALY)				37-39
Healthy	1	—	—
Asymptomatic	0.95	0.90–1.00	Triangular
Outpatient	0.77	0.45–0.95	Triangular
Hospitalizations	0.65	0.11–0.77	Triangular
Deaths	0	—	—

Note: NBSC, National Bureau of Statistics of China. PBJB, Price Bureau of Jiangsu Province. QALYs, quality-adjusted life years.

Cost inputs

The costs of the immunization program included cost of vaccine, and administration. The cost of HE vaccine was US$ 36.64 /per dose based on INNOVAX Biotechn-ology Company (Xiamen, China), and the preliminary results had shown that the cost of administration per dose was estimated at US$ 1.72,³⁸ which covers costs for nurse labor, storage, transportation and injection of vaccine. Costs of adverse events concerning hepatitis A vaccination were not considered. The medical costs of infected elderly people were acquired from field survey and published studies.^38-40 The discount rate is 3%. The per capita gross domestic product and natural mortality of the population was determined according to the national statistical data. In China, per capital GDP in 2015 was US$ 7292.68, and the study assumed that GDP/capita in 2016 was equal to per capital GDP in 2015.⁴¹ All costs were expressed in 2016 US dollars (1 dollar = 6.7672 RMB in 2016).

Utility inputs

The QALYs (quality adjusted life years) was used to synthetically evaluate health outcomes in the cost-effectiveness analysis. The QALYs of infected cases were assessed using field investigation and published literatures,^37,38,42 the state lived without hepatitis A was assigned a maximum utility score 1, and death a minimum 0; score of non-hospitalized patients was 0.77; hospitalized patients had a score of 0.65. There hasn't being life quality investigation for asymptomatic infection by HEV, so we selected the loss of asymptomatic infection's QALYs from hepatitis A (0.01 QALY in children in Suwantika et al.⁴³) as one from hepatitis E, due to the unavailability of the latter. However, we assumed the lost for asymptomatic infection as 0.05 QALYs, considering elder people with weak immunity and low resistance in our study.

Model analysis

The tree-Markov model was developed in TreeAge Pro 2013 software. To compare 3 immunization strategies, the ICER was used in the analysis. All parameters within the range of values shown in Table 2 were analyzed using one-way sensitivity analysis. To determine the impact of parameter uncertainty on the ICER, we conducted a probabilistic sensitivity analysis based on their parameter distribution. Due to limited published data, the probabilities of all parameters were assigned triangle distributions.

Abbreviations

(HE)

Hepatitis E

(HEV)

Hepatitis E virus

(ICER)

Incremental cost-effectiveness ratio

(NBSC)

National Bureau of Statistics of China

(PBJB)

Price Bureau of Jiangsu Province

(QALYs)

Quality-adjusted life years

Disclosure of potential conflicts of interest

All authors report no conflict of interests.

Funding

This study was funded by the Chinese National Natural Fund (Code 81573258) and the Social Development Fund of Jiangsu Province (Code BE2013723).