Vaccination Strategies Against Hepatitis A in Travelers Older Than 40 Years: An Economic Evaluation
This paper has been presented at the 2nd International Conference of the Journal of Travel Medicine and Infectious Disease, 12th September 12, 2007, as a poster.
Laura Costas, MD, MPH, Preventive Medicine and Epidemiology Department, Hospital Clinic, UASP, Villarroel 170, 08036 Barcelona, Spain. E-mail: email@example.com
Background In recent years, the number of travelers aged >40 years who acquire hepatitis A while traveling has increased. Therefore, there is a need to review hepatitis A vaccination protocols in travelers. The aims of the study were to assess immunity levels to hepatitis A virus (HAV) in international travelers >40 years and to determine the least costly immunization strategy.
Methods A serological examination of HAV antibodies in 427 international travelers aged >40 years traveling endemic zones was carried out. The prevalence of antibodies in each age group was assessed. The costs of two preventive strategies, direct vaccination of all subjects (independent of the immune status) or screening and subsequent vaccination of susceptible subjects were compared. The critical value of prevalence (CVP) (the value at which the costs for the two strategies are equal) was calculated.
Results Total prevalence of HAV antibodies was 78.9% [95% confidence interval (CI): 74.8–82.5] and was 80.0% (95% CI: 73.8–85.2) in men and 77.9% (95% CI: 71.9–83.2) in women. There was a positive association with age. In the 40 to 49, 50 to 59, 60 to 69, and 70 to 95 years age groups, the prevalence rates were 62.6 (95% CI: 53.8–71.5), 76.8 (95% CI: 70.0–82.7), 91.7 (95% CI: 85.2–95.6), and 97.5% (95% CI: 87.4–99.6), respectively. The CVP was 58.4% using two doses of vaccine.
Conclusions The CVP was lower than the prevalence rate found in our international travelers. Therefore, we recommend systematic screening for HAV antibodies before selective vaccination of international travelers aged >40 years traveling to hepatitis A endemic zones.
Hepatitis A has a worldwide distribution, although prevalence rates vary according to geographic regions.1–4 Spain is now a low endemicity country due to a marked decrease in prevalence rates over recent decades.2,5,6
The presence and severity of hepatitis A symptoms are related to age. The disease is often mild and asymptomatic in children but may be more severe in patients aged >40 years.7–9
Two effective immunization strategies can be developed to ensure immunity in all travelers: (1) vaccination of all travelers and (2) screening for hepatitis A virus (HAV) antibodies in all travelers and vaccination of only susceptible individuals. Both strategies are cost effective.10,11 Vaccination is highly recommended as a preventive measure, particularly in high-risk groups.6,8 The hepatitis A vaccine is safe and immunogenic.7–9 If a universal strategy of vaccination of all travelers is considered, the same level of protection against hepatitis A will be achieved; however, screening and vaccination of susceptible individuals is a least costly immunization strategy in population groups with a high prevalence of HAV antibodies.
In Catalonia (Spain), official guidelines include universal hepatitis A vaccination of all individuals born after 1960 and recommend determination of HAV antibodies prior to vaccination only in subjects born before 1960 undertaking high-risk travel (long-term cooperation in endemic countries and adventure travel or long stays in endemic areas).12
In our hospital, in recent years, there has been an increase in international travelers aged >40 years becoming infected by hepatitis A. Between January 2006 and December 2007, nine cases of hepatitis A were recorded in international travelers aged >40 years, including one fatal case and another case of fulminant hepatitis A, which required a successful liver transplant.13
We decided to add a screening strategy to assess the prevalence of HAV antibodies in international travelers aged >40 years, and we determined in this study whether this strategy was the most efficient.
A total of 427 travelers aged >40 years traveling to international endemic areas were considered for hepatitis A vaccination. All of them received medical care at the Travelers’ Clinic of the Hospital Clinic Barcelona, from December 2006 to April 2007. Travelers with either a previous history of hepatitis A or who have been vaccinated in the past were excluded from the study. We investigated these data by surveying the travelers and consulting our databases.
Sample size for estimating the prevalence of antibodies against hepatitis A was determined considering a population prevalence rate of 0.5, accepting an alpha risk of 0.05 for a +/−0.05 percentage units precision, with a 0.1 reposition rate.
The prevalence of hepatitis A antibodies according to gender and age group was determined using four age groups: 40 to 49, 50 to 59, 60 to 69, and 70 to 95 years of age.
Informed consent was not required since the screening was introduced as a standard proceeding when we detected an increase in international travelers aged >40 years infected by hepatitis A. The project obtained the approval from the Hospital Ethics Committee.
Hepatitis A antibodies were determined in the Hospital Clinic Microbiology laboratory using an ADVIA Centaur assay, an automated competitive immunoassay that uses direct chemioluminescence.
Efficiency of Immunization Strategies
A cost analysis was performed for two hepatitis A prevention strategies: (1) direct vaccination of all travelers aged >40 years and (2) screening for HAV antibodies and vaccination of subjects without significant antibody titers (prevaccine detection).
To identify the least costly immunization strategy, we first determined the critical value of prevalence (CVP). This is defined as the prevalence of HAV antibodies that makes cost-effectiveness of vaccinating all individuals equal to that for screening and vaccination of susceptible individuals. When the prevalence of antibodies is lower than the critical value, vaccinating all travelers is more efficient than screening and vaccination of susceptible individuals. On the other hand, when the prevalence of antibodies is higher than the critical value, screening and vaccination of susceptible individuals is the more efficient strategy.
The CVP was obtained by using the Plans Rubio method,13,14 which takes into account clinical, epidemiological, and economical information. The CVP depends on the relation between screening costs (S) and vaccination costs (V) minus averted disease (Ad) costs not reduced in individuals with a false-positive screening result:
Vaccination (V) costs are calculated from the cost of one-dose vaccine (Cd), vaccine administration costs (Va), and vaccination compliance (C):
If two doses of vaccine are considered, the vaccination cost is:
Disease costs not averted in travelers with a false-positive screening result (Ad) are determined considering the predictive positive value of a positive screening test (Pv), mean disease costs (D), attack rates (A), vaccine efficacy (E), and vaccination compliance (C):
Vaccination and Screening Costs
The mean cost of hepatitis A was calculated using a decision tree model.14,15 The best available estimates were used. We considered that 70% of hepatitis A cases in patients aged >40 years are symptomatic, of which 47.9% are mild, 30% moderate, 19.9% severe, and 2.1% fulminant. The estimated case-fatality rate of fulminant cases was 70%.8,14,15
A mild case of hepatitis A was estimated to require a mean of one physician visit and three follow-up outpatient visits. A moderate case of hepatitis A requires a mean of one physician visit and four follow-up outpatient visits. A case of severe hepatitis A requires a mean of one physician visit, 11 days of hospital stay, and two additional follow-up visits. A case of fulminant hepatitis A requires seven ICU days (if fatal) and five ICU days plus 11 days of hospital stay and four follow-up visits (if nonfatal).14,15 The relapse rate (9% for mild cases, 7% for moderate, and 2% for severe cases) was also included in the model.14,15
The mean costs for all episodes and physician visits were determined using the financial accounting system of the Hospital Clinic. The first visit was assumed to occur in the emergency department and follow-up visits in the outpatient’s clinic.
The predictive positive value of the test was measured using the total prevalence rate of HAV antibodies together with a sensitivity and specificity rate of 99.0%. The attack rate in travelers (3 cases/1000 persons-month), vaccine protection (95%), and compliance rate figures (60%) were obtained from the best published estimates.2,8,16,17
The cost of screening was obtained from the Hospital Clinic 2006 tariffs, and the cost of a vaccine was its acquisition cost.
Administrative costs were determined by dividing the mean salary of a registered nurse by the mean time taken for a single vaccination.
The mean cost per person of preventing hepatitis A was assessed by taking into account the percentage of travelers aged >40 years who would be protected in each age group.
A sensitivity analysis was carried out to assess consistency of results obtained considering 10% higher and lower vaccination and screening costs; 45, 75, and 90% vaccination compliance; and 10% higher and lower disease costs avoided.
Between December 2006 and April 2007, 427 travelers aged >40 years who met the inclusion and exclusion criteria were included in the study.
There were 205 men (48%) and 222 women (52%). The mean age was 56 years (range 40–95 years), with no significant differences between genders. A total of 111 travelers were aged 40 to 49 years, 165 were aged 50 to 59 years, 110 were aged 60 to 69 years, and 41 were aged 70 to 95 years.
Prevalence of HVA Antibodies
Total HAV antibodies prevalence rate was 78.9% (95%CI: 74.8- 82.5): 80% (95% CI: 73.8–85.2) in men and 77.9% (95%CI: 71.9–83.2) in women (nostatistically significant differences). A positive association between HAV antibodies prevalence rates and age was found (test for trends: p value < 0.001) (Table 1).
Table 1. Prevalence of hepatitis A virus (HAV) antibodies and least costly immunization strategy in different age groups
|40–49||111||62.6||(53.8–71.5)||Vaccination||Screening + selective vaccination||7,218€||5,518€||8,455€||8,829€|
|50–59||165||76.8||(70.0–82.7)||Vaccination||Screening + selective vaccination||6,433€||5,518€||7,201€||8,829€|
|60–69||110||91.7||(85.2–95.6)||Vaccination||Screening + selective vaccination||5,611€||5,518€||5,884€||8,829€|
|70–95||41||97.6||(87.4–99.6)||Screening + selective vaccination||Screening + selective vaccination||5,293€||5,518€||5,376€||8,829€|
Vaccination and Screening Costs
The cost of screening was 20.62€ (blood sample plus antibody test). The cost of one dose of vaccine was 19.16€. The administration costs were 2.91€. The total cost of disease episodes avoided was 686.7€. The positive predictive value of the screening test was 99.7%.
Least Costly Immunization Strategy in Different Age Groups
The CVP obtained by Plans Rubio method was 93.4% for one vaccine dose and 58.4% for two vaccine doses.
The CVP for one vaccine dose (93.4%) was greater than the real prevalence rate in the 40 to 49, 50 to 59, and 60 to 69 years age groups (62.6, 76.8, and 91.7%, respectively) and lower than the real prevalence rate in the 70 to 95 years age group (97.6%). For two vaccine doses, the CVP (58.4%) was lower than the real prevalence rate for all age groups.
Screening and vaccination of susceptible individuals is the more efficient strategy when the critical value is lower than the real prevalence of antibodies. This occurs in the 70 to 95 years age group for one vaccine dose and in all age groups for two vaccine doses.
A sensitivity analysis was performed to test likely changes in the CVP according to variations in the values of the main variables in the cost analysis (Table 2). The CVP was sensitive to changes in compliance rate, vaccination costs, and screening costs but not to changes in the attack rate, the cost of disease, and vaccine protection.
Table 2. Sensitivity analysis: critical value of prevalence (CVP) variations
|Vaccination costs (%)|
|Screening costs (%)|
|Compliance rate (%)|
|Vaccine protection (%)|
|Disease costs avoided (%)|
Recent decades have seen a reduction in the prevalence rate of HAV antibodies in developed countries.2,18–20 The current situation warrants a reassessment of preventive strategies against hepatitis A in international travelers aged >40 years, particularly considering the potential severity of the disease.1,7,8,17
Recent hepatitis A seroprevalence studies in Spain5,21 have shown a protection rate of around 94% in people aged 40 to 54 years higher than the levels found in this study in travelers aged >40 years. The mean socioeconomic status of international travelers may be higher than those of the general population.
Immunization against hepatitis A usually consists of a two-dose vaccination strategy, even though 40% of subjects are lost to follow-up. Therefore, given that Plans Rubio method takes into account these losses to follow-up, we preferred this strategy as the standard for the analysis. Other studies have performed similar analyses with one or three vaccine doses.14,22,23
As expected, our results showed different values according to the number of doses considered. Using a single dose strategy, vaccination of all international travelers aged 40 to 69 years and screening tests for HAV antibodies followed by selective vaccination of susceptible individuals in all other age groups are recommended. If a two-dose strategy is accepted, screening followed by selective vaccination is the most cost-effective approach for all age groups older than 40 years, as shown by various reports.14,22–24
The use of a screening strategy was supported by the potential savings in terms of vaccination administration in settings with higher prevalence rates in older populations. It also offers the advantage of monitoring possible changes in prevalence rates over time.
In spite of a steady recent decline in prevalence rates of HAV antibodies, the situation in the near future cannot be foreseen as immigration to Spain has increased and most immigrants come from countries with higher endemic rates of hepatitis A.20 The time elapsed if the screening strategy is adopted threatens its feasibility in busy, last minute consultations, a common situation in travelers’ clinics.25 Therefore, screening may only be feasible in a proportion of international travelers aged >40 years, although we believe the proportion could be substantial. Screening should be performed in international travelers older than 40 years of age whenever time before departure allows it.
The Plans Rubio method has been used in previous studies22,23 but has several drawbacks, notably due to lack of information and comparisons between its results and those achieved by other methods of economic analysis. Our study shows that the method can be used to assess the least costly immunization strategy without developing a cost-effectiveness analysis. Application of this method can increase the efficiency of immunization programs.
Cost-effectiveness analyses are suitable for conducting economic evaluations of preventive strategies against hepatitis A but require detailed clinical and epidemiological data that are not always easily available. Furthermore, they are expensive, time consuming, and sometimes difficult to interpret in each age group.14,15 When the main question to be answered is to vaccinate with or without a screening strategy, the Plans Rubio method is very useful.
Studies have shown critical values of prevalence of 22% (considering three doses of vaccine),14 37.5% (one dose of vaccine),22 and 22.5% to 23% (two doses of vaccine).15,23 Our study found higher critical values of prevalence, mainly because we used the pharmacy wholesale price rather than the pharmacy retail price. In addition, we did not apply discount rates.
Our analysis had several limitations. Some travelers may reject screening due to the need for blood tests, or because one more visit is needed for this strategy, but only 2.3% (n= 10) of our study subjects rejected the procedure. The inclusion of all international travelers aged >40 years prevented any major selection bias, although travelers’ samples were collected only from December 2006 to April 2007.
We investigated some exclusion criteria, ie, travelers with either a previous history of hepatitis A or who have been vaccinated in the past, by surveying them and consulting our databases. Due to unapparent infections, record bias or the existence of records in other centers, there could be an excess of positive results for HAV antibodies. However, this is the situation we observe in practice and it estimates the proportion of unknown immunized people older than 40 years old visited in our clinic.
Information on some costs is less sensitive and could be a limitation of our study because the positive predictive value of screening is closer to 1.
The cost analysis established a vaccine protection rate of 95%, but some recent studies suggest that people aged >40 years could be less responsive to hepatitis A vaccination.26,27 However, our results were the same when vaccine protection rates of 90, 80, and even 70% were considered.
We have used the attack rate of disease considered by WHO,7 although a recent study28 calculated lower incidences for nonimmune travelers, varying between 6.0 and 28.0 per 100,000 person-months of stay. Nevertheless, our results showed the same CVP value considering these figures.
In summary, taking into account the advantages and drawbacks of the screening strategy, the use of this approach (screening followed by selective vaccination) in international travelers aged >40 years seems to be the best option. The strategy had a very good cost-effectiveness ratio in this age group and would allow us to monitor time trends in seroprevalence, which is useful because our international travelers showed lower prevalence rates of antibodies against HAV than the general population, and it is uncertain how this may change in the future due to immigration patterns.
We thank Sarah Lafuente, Teresa Mejías, David Buss, Sonia Paytubi, Luis Izquierdo, Francesc Cots, Joan Sánchez, and Dolors Heras for their contribution to this study and Anna Llupià and Maria Jesús Bertran for manuscript review.
Declaration of Interests
The authors state that they have no conflicts of interest.