|The cost-effectiveness of introducing hepatitis B vaccine into infant immunization services in Mozambique
|Griffiths U K, Hutton G, Dores Pascoal E D
This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.
The use of a hepatitis B vaccine in routine infant immunisation.
The study population comprised a hypothetical 2001 birth cohort in Mozambique, based on United Nations Population Division estimates. The future population numbers for the 2001 birth cohort were adjusted for co-morbidities using gender-specific life tables from the World Health Organization (WHO).
The setting was community care (infant immunisation services in Mozambique). The economic study was carried out in Mozambique.
Dates to which data relate
The effectiveness data were derived from published literature dating from 1977 to 2003. The price year was 2001.
Source of effectiveness data
The effectiveness evidence was derived from a review of published studies, supplemented with assumptions made by the authors based on clinical expert opinion.
A linear model was constructed to estimate the cost per number of deaths and disability-adjusted life-years (DALYs) averted. The model was constructed on the basis of two linear relationships, the probability of dying from acute disease (age and gender-specific) and the probability of being a hepatitis B surface antigen (HBsAg) carrier. A risk of death from carcinoma or cirrhosis was associated with hepatitis B carriers. The model assumed that immunisation costs will be incurred, on average, 40 years into the future. Both the costs and benefits were discounted at a rate of 3%.
Outcomes assessed in the review
The main outcomes estimated from the literature were:
the annual HBsAg carrier rate (based on 7 published studies and expert opinion);
the probability of death from acute disease (based on sub-Saharan Africa estimates); and
the annual relative risk of carriers dying from hepatocellular carcinoma or cirrhosis (based on published estimates for Gambia, Taiwan and Alaska).
Study designs and other criteria for inclusion in the review
Sources searched to identify primary studies
The authors reviewed published literature reviews covering economic evaluations of immunisation against hepatitis B. These reviews comprised studies from both developed and developing countries.
Criteria used to ensure the validity of primary studies
Methods used to judge relevance and validity, and for extracting data
Number of primary studies included
Fourteen primary studies were included as sources of effectiveness evidence. Estimates on the coverage rate of the DTP-hepatitis vaccine and vaccine efficacy were derived from WHO official country estimates and guidelines.
Methods of combining primary studies
Investigation of differences between primary studies
Results of the review
The probability values of the most important outcomes drivers were as follows:
the annual HBsAg carrier rate was 14% nationwide (95% confidence interval, CI: 12 - 16, normal distribution);
the probability of death from acute disease was 10% (range: 8 - 11, triangular distribution);
the annual relative risk of carriers dying from hepatocellular carcinoma or cirrhosis was 1.05% (range: 0.88 - 1.22, triangular distribution) compared with other death causes.
The 2001 birth cohort in Mozambique was estimated to be 794,650 children.
Measure of benefits used in the economic analysis
Two summary benefit measures were used in the economic analysis, the deaths averted and DALYs averted. Both were derived from the decision model. For the estimation of DALY valuations of health states, the authors reported only that it was based on the formula published in the 1996 Global Burden of Disease series. The basic method of valuation of health states was not reported.
The health service and patient costs were evaluated. The quantities and the costs were analysed separately. The authors identified all resource items used for vaccine delivery using registry data (Expanded Programme on Immunization) and assumptions on staff time and vehicles. Medical treatment associated with hepatitis B infection was also collected: the costs of treatment comprised hospital overhead costs as well as patient-specific costs for drugs and laboratory tests. For patients with cirrhosis and hepatocellular carcinoma, there was essentially no treatment available in Mozambique. The cost data for inpatient care were collected at Maputo Central Hospital (national referral hospital). The authors disregarded costs due to adverse events based on evidence that they are minor events. All of the costs were discounted at a rate of 3% per annum. The price year was 2001.
Out-of-pocket costs to traditional healers were also included in the analysis. In particular, the proportion of patients seeking traditional healer assistance and the frequency of visits (both based on assumptions) and the cost of treatment based on estimates from the Traditional Medicines Unit, National Institute of Health, Mozambique.
Statistical analysis of costs
The costs were analysed stochastically. Standard deviations and probability distributions were provided in table 2 of the original paper.
No indirect costs were included in the analysis.
US dollars ($). The exchange rate used when calculating the values was MTS 24,000 = $1.
A sensitivity analysis was carried out to investigate the impact of different patient management strategies (monovalent vaccine or DTP combination vaccine) on the incremental cost-effectiveness ratios (ICERs). Cost-effectiveness ratios were calculated in a deterministic way, with and without discounting future health benefits at 3% per annum (see 'Synthesis of Costs and Benefits' section for details).
Estimated benefits used in the economic analysis
In the base-case scenario, there were 4,383 deaths averted and 53,375 DALYs averted by the introduction of hepatitis B vaccination. After discounting health benefits, these results were reduced to the equivalent of 1,043 deaths and 40,752 DALYs per year. The authors did not include adverse events in the economic analysis.
Crystal Ball predicted a total of 5,767 deaths from acute and chronic disease in the base-case scenario (undiscounted health benefits), with the majority occurring in the age group 35 to 56 years. With a birth cohort of 794,650 children, it was predicted that 0.73% of the cohort would die from hepatitis B-related disease in the absence of vaccination.
The cost results were hard to understand. The table of results appeared inconsistent with the text. The text did not quote the total costs for either of the alternatives, although it did state that the incremental programme costs were $2.1million. It was unclear, but it appears that there was a potential $125,930 per year saving from future costs due to hepatitis B vaccination.
Synthesis of costs and benefits
Cost-effectiveness ratios were calculated probabilistically in order to incorporate parameter uncertainty and translate it into overall cost-effectiveness results, but only for the combination vaccine. The mean discounted results were $2,034 per death averted (95% CI: 1,404 - 3,067) and $52 per DALYs averted (95% CI: 36 - 78). When the health benefits were not discounted, the mean cost per death averted was reduced to $496 (95% CI: 329 - 777) and $40 per DALY averted (95% CI: 27 - 59).
Results for the monovalent vaccine were only presented deterministically, that is, $749 per death averted and $19 per DALY averted when discounting future health benefits at 3% per year. When the health benefits were not discounted, the ICERs decreased to $178 per death averted and $15 per DALY averted.
The results of the sensitivity analyses suggested that only the health benefit discount rate has an important impact on the cost per death averted ratio, especially for the combination vaccine scenario (ICER $436 undiscounted compared with $1,833 discounted). Changes in the cost per DALY were minimal in both combination and monovalent vaccination scenarios.
The cost-effectiveness of infant hepatitis B vaccination was shown to be in the range of other primary health care interventions for which similar analyses have been undertaken. The (discounted) cost-effectiveness results for the monovalent vaccine ($749 per death averted and $19 per DALY averted) are considerably more favourable than those for the combined vaccination strategy ($1,833 per death averted and $47 per DALY averted).
CRD COMMENTARY - Selection of comparators
The authors did not provide any justification for the choice of the comparator. The option of reduced coverage instead of the do nothing option could have been more realistic and/or provide important information for decision-making. You should decide whether it represents a valid comparator in your own setting.
Validity of estimate of measure of effectiveness
The analysis of effectiveness used data derived mainly from other sub-Saharan African countries and expert opinion. It was unclear whether the authors had conducted a systematic review. They seemed to have opted for obtaining the studies from published reviews covering economic evaluations of immunisation against hepatitis B. The method of data synthesis was not described. These issues tend to limit the internal validity of the analysis. The authors made assumptions to provide additional inputs for the model, and these seem to have been reasonable. However, two key parameters in the model, the annual HBsAg carrier rate and the probability of acute death, required further justification, such as how the information from 7 studies and expert opinion was combined and why the probability of acute death was assigned a triangular distribution. It was unclear whether other potential relevant variables (e.g. vaccine efficacy and vaccine wastage) were incorporated in the model. The section on the sensitivity analysis and figure 2 and table 5 were confusing, and it was unclear whether this section reported the results of a probabilistic analysis or the results of a univariate sensitivity analysis. The "sensitivity chart" (figure 2) and the scale of the graph require further explanation.
Validity of estimate of measure of benefit
Both benefit measures (death and DALYs averted) were obtained from the decision model, and are comparable with the benefits of other health care interventions. There was no difference in the benefits of both vaccine strategies; they differed only in terms of the additional costs of the DTP-hepatitis combination vaccine.
Validity of estimate of costs
The authors explicitly stated the perspective adopted in the study. All the relevant cost categories were included in the analysis. Mean costs by treatment were provided. The costs were treated deterministically and no sensitivity analysis seems to have been carried out. All the costs were discounted at a rate of 3% in all scenarios. The cost results were not reported clearly and table 4 of the original paper showed inconsistencies with the text. Consequently, it is hard to have confidence in the results.
The base-case scenario was unclear, the authors implying that the results of the undiscounted health benefit scenario represented the base-case. The authors did not explore any difference in efficacy between the hepatitis vaccine strategies. The results reported seem to have been deterministic, with table 5 of the original paper reporting probabilistic results only for the DTP-hepatitis combination vaccine. However, this was not explicitly stated in the report and table 5 was rather confusing. There was lack of clarity in the reporting of the results for both vaccine strategies. A sensitivity analysis using different coverage rates could have provided further insight into the decision-problem.
Implications of the study
The study results suggested that the introduction of a hepatitis B vaccine could increase the annual budget for immunisation services by approximately 25 to 56%. The model predicts that more than 4,000 future deaths could be averted annually by routine hepatitis B infant immunisation. The monovalent hepatitis B vaccine is considerably more cost-effective than the combined hepatitis B and DTP vaccine. No specific recommendations for further research were made.
Griffiths U K, Hutton G, Dores Pascoal E D. The cost-effectiveness of introducing hepatitis B vaccine into infant immunization services in Mozambique. Health Policy and Planning 2005; 20(1): 50-59
Subject indexing assigned by NLM
Carrier State; Child Health Services /economics; Cohort Studies; Cost of Illness; Cost-Benefit Analysis; Health Care Costs; Health Services Research; Hepatitis B /complications /economics /epidemiology /prevention & Hepatitis B Vaccines /administration & Humans; Immunization Programs /economics /utilization; Infant; Models, Econometric; Mozambique /epidemiology; Prevalence; Quality-Adjusted Life Years; control; dosage /economics
Date bibliographic record published
Date abstract record published