Cost-Effectiveness of Nationwide Hepatitis B Catch-up Vaccination Among Children and Adolescents in China

Overview

We considered a strategy of catch-up vaccination for children and compared it to the status quo with current levels of vaccination coverage. The catch-up vaccination strategy provides hepatitis B vaccination immediately for children and adolescents who missed newborn vaccination. For children not participating in the intervention (and for all individuals under the status quo), we assumed that 0.5% would voluntarily seek the vaccine themselves each year in the future.

We developed a model to project costs and health impacts of both acute and chronic HBV infection. We developed a probability tree to evaluate the outcomes of acute HBV infection (new infection that is either asymptomatic or symptomatic) and a Markov model of disease progression to evaluate the health and economic impacts accruing from treated and untreated chronic HBV infection. The Markov model of disease progression captures the long-term effects of chronic HBV infection, which represent a much larger part of the overall disease burden of HBV than the effects of acute infection (12). We modeled costs and health states for a hypothetical cohort of 10,000 children and adolescents age 1-19 throughout their lifetimes. We calculated costs and benefits for the overall cohort, and we examined individual cohorts of age 1-4 years, 5-10 years, 11-14 years, and 15-19 years. We measured costs and health outcomes in discrete time increments of one year, similar to other analyses (13, 14).

According to WHO criteria (15), an intervention can be considered “very cost-effective” if the cost per disability-adjusted life year (DALY) gained is less than the annual per capita GDP, which in China is approximately $2,500 (16-18) [17,000 RMB]. An intervention is considered to be good value for money if it costs less than three times the per capita GDP. Because QALYs are roughly similar to DALYs (19), we use a threshold of 17,000 RMB [$2,500] per QALY gained for the intervention being very cost-effective.

Data and Sources

To gather data needed for our model (Table 1, Appendix Tables 1 and 2), we performed a Medline search to review previous cost-effectiveness analyses of HBV vaccination in both China and the United States (1, 20-22) and, more importantly, published Chinese government reports, national seroepidemiologic surveys, and peer reviewed articles published in Chinese. Values for model parameters were drawn from the existing literature when available. In most instances, Chinese data was used, although parameters that were less specific to China (e.g., those relating to the natural history of disease) were taken from international literature (Appendix Table 2). When no data were available, informed expert judgment was used (personal communication with China CDC, China Foundation for Hepatitis Prevention and Control’s national experts and the WHO’s technical experts in China and the Western Pacific Region). In sensitivity analysis we considered low and high values for all variables, using ranges to reflect values in the literature, where available.

Some parameters varied by age (Appendix Table 1). The initial prevalence of chronic HBV infection and of HBV immunity are different for different age groups. In addition, background mortality and the chance that an acute infection develops into a chronic infection vary by age. For example, an acute infection in a 1-year old has an 86% chance of becoming chronic, whereas an acute infection in a 5-year-old has only a 30% chance of becoming chronic. As the model steps forward in time, these parameters update according to the aging of the children. We assumed that all other parameters relating to disease progression did not vary by age.

Compliance and Vaccine Effectiveness

We assumed that parents who knew of their children’s infection or prior hepatitis B vaccination history would not have their children vaccinated. Of those unaware of their vaccination history or infection, we assumed that 70% complied with the intervention, regardless of the age of the children.

We assumed that the vaccination would consist of a three-dose series, and that the protection rate from three doses of the vaccine was 95% (23-25). If the children complied with the intervention, we assumed that they all received all three doses, based on known high completion rates in the target population (1, 26). We assumed that costs associated with program implementation would be similar to that of a recently completed two-year, province-wide, school-based, pilot catch-up vaccination program that one of the authors (S So) helped to initiate in partnership with the China CDC and Ministry of Health, the China Foundation for Hepatitis Prevention and Control, and the Qinghai provincial CDC and government (27). The program provided free HBV catch-up vaccination and education for more than 500,000 children in kindergarten and primary school in the entire remote province of Qinghai between 2006-08 (26). Vaccination at each school clinic was administered through the existing provincial CDC immunization offices that have branches to serve the immunization needs of the various districts in the province. We did not include program setup costs in the model as such costs could be highly variable depending on the size and scope of a vaccination program. Setup cost for the recent Qinghai province catch-up vaccination program was approximately $0.40 [2.7 RMB] per child; we included this setup cost in sensitivity analysis.

Epidemiologic Parameters

Epidemiological studies of HBV in China over the past 15 years have shown a reduction in the national estimates of HBV prevalence and incidence due to the expanded newborn immunization efforts. Estimates of the prevalence of chronic HBV infection in the general Chinese population dropped from 9.75% in a national survey conducted between 1992-95, to 7.18% in a more recent 2006 national survey (8, 28). The prevalence of chronic HBV infection among young children is lower because children have had higher vaccination coverage than older members of the population. The latest national serological survey of more than 80,000 people conducted in 2006 in China estimated chronic HBV infection prevalence to be 0.96% among the almost 60 million children aged 1-4 years and 2.42% among the almost 190 million children aged 5-14 years (8). Using prevalence data from this serological survey (8), we estimated base case incidence of acute HBV infection to be 100/100,000 per year per unvaccinated person, regardless of age (calculations in Appendix). Because estimates of incidence vary, we varied this parameter between 20/100,000 and 500/100,000 in sensitivity analysis (6, 8, 20, 21, 29, 30). We discuss incidence in terms of acute HBV infection per unvaccinated person since the policy affects unvaccinated persons and we wished to provide a comparable risk of infection across age groups. This incidence number is higher than published incidence numbers that are based on entire populations (for the latter, the denominator is larger and thus the reported incidence statistic is lower).

Acute HBV infections were modeled in a similar manner to those of other published cost-effectiveness studies (31, 32). The likelihood of an acute HBV infection becoming chronic varies by age. We used a data from a meta-analysis (33) to relate this likelihood with age; estimated values are shown in Appendix Table 1. There are differences of opinion about this likelihood, especially at older ages (34), so we performed sensitivity analysis on this probability.

Health Outcomes and Costs

We measured health and economic outcomes for all children and adolescents in the cohort over their lifetimes. Since newborn vaccination coverage is rapidly increasing in China (1), we excluded the effects of infection on future children born to girls currently in the cohort. We took a societal perspective and included all costs and savings, regardless of source or beneficiary, and we discounted all costs and health outcomes to the present at 3% (35).

We included the costs of vaccination, treatment, and all other healthcare costs for HBV-infected and uninfected individuals in the cohort over their lifetimes. We did not count time from work or school due to disease since that morbidity is accounted for in quality-of-life adjustments (35). We gathered treatment cost data from several recent studies of the cost of chronic HBV infection in China (22, 36) and then converted them to US dollars. Since costs vary significantly for different regions of China, we conducted sensitivity analysis on all cost parameters.

We measured several health outcomes including deaths, HBV infections averted, and QALYs gained. Quality adjustments for health states were taken from published literature (13, 36-38).

Base Case

Base case results for the entire cohort of children and adolescents ages 1-19, and for individual age cohorts, are shown in Table 2. In all cases, catch-up vaccination is cost-saving: it increases QALYs and saves costs compared to the status quo.

Sensitivity Analysis

We performed two-way sensitivity analysis on the incidence of acute HBV infection and the age of the cohort (Figure 2). In areas where annual incidence of acute HBV infection is 100/100,000 or higher among susceptible children (base case), it is cost-saving to vaccinate all children aged 1 to 19 years. If annual incidence is 20/100,000 (one-fifth as high as estimated in the base case), catch-up vaccination is still cost-saving for children aged 2 years and younger, and would be highly cost-effective for children aged 3-10 years, costing less than $2,500 [17,000 RMB] per QALY gained. The catch-up vaccination program becomes less cost-effective as acute HBV infection incidence decreases and as the age of the children vaccinated increases.

We performed one-way sensitivity analysis across the ranges of all parameters (Appendix Table 3). In almost all cases, catch-up vaccination was cost-saving. In addition to cohort age and incidence of acute HBV infection (discussed above), the cost-effectiveness ratio was most sensitive to the costs of administering the vaccine. If the total cost of vaccine and administration is at the high end of our range ($8 [55 RMB] per child vaccinated), then the intervention costs $5,800 [40,000 RMB] per QALY gained for a cohort of 10-year-olds.

In our base-case analysis, the chance an acute infection becomes chronic is based on a meta-analysis (33). One previous study, not included in that meta-analysis, found a lower likelihood (34). Reflecting that study, we lowered the chance of an acute infection becoming chronic (calculations in Appendix) and found that vaccination is still cost-saving for children up to age 10 and costs less than $2,500 per QALY gained up until age 15 (Appendix Figure 3).

In the base case, we did not include setup costs since they can be highly variable depending on the size and scope of the program. When we included setup costs similar to those in a recent catch-up vaccination program in Qinghai province (about $0.40 [2.7 RMB] per child), the catch-up vaccination program was still cost-saving.

China has recently experienced rapid economic growth and increased prosperity. This economic growth may lead to higher costs for HBV treatment, and for health care in general. In addition, rural areas in China may have lower healthcare costs and higher HBV incidence (due to lower vaccination coverage) than urban areas. Thus, we conducted two-way sensitivity analysis on healthcare costs and incidence (Figure 3). We varied annual incidence of acute HBV infection between 10/100,000, a rate similar to that among vaccinated individuals (20, 21), and 100/100,000 (the base case incidence). We varied all costs of health care and HBV treatment simultaneously from 50% to 150% of the base case values. If incidence is lower than in the base case, there is less benefit from the vaccination program. Similarly, if the costs of treating HBV and its complications are lower, then the program has less benefit. However, treatment costs must be significantly lower and incidence significantly lower than in the base case before catch-up vaccination is not cost-effective.

We performed probabilistic sensitivity analysis in which we varied all parameters simultaneously in a Monte Carlo simulation (Figure 4). Parameters were drawn from a uniform probability distribution between the low and high values (Table 1) and were assumed to be uncorrelated. With these assumptions, the catch-up vaccination program has a 97% chance of being cost-saving and a 98% chance of having an incremental cost-effectiveness ratio of less than $2,500 [17,000 RMB] per QALY gained.