Epidemiological modelling and health economic evaluation of vaccination programmes against varicella and herpes zoster in Norway
Report
|Published
The Norwegian Institute of Public Health has utilised a mathematical infectious disease transmission model to simulate what would happen if a vaccination programme were introduced against either (1) varicella (chickenpox) or (2) herpes zoster (shingles) in Norway.
Summary
We utilise a mathematical infectious disease transmission model to simulate what would happen if a vaccination programme were introduced against either (1) varicella (chickenpox) or (2) herpes zoster (shingles) in Norway. The model is calibrated to the current situation without widespread vaccination, where varicella circulates and infects most children, and many of those infected develop shingles later in life. We use the model to simulate different vaccination scenarios and compare epidemiology, health loss, lost quality-adjusted life years (QALY), and costs in vaccination scenarios with the current situation. Based on this, we calculate whether the different vaccination programmes could be considered cost-effective.
Varicella
The model shows that the introduction of vaccination against varicella in the childhood vaccination programme (BVP) will lead to near eradication of varicella in Norway within a few years, provided the vaccination coverage achieves similarly high levels as the rest of the BVP.
In our main scenario, the vaccine is offered in two doses, first at 15 months of age and then at 7 years of age. When vaccination begins, community transmission will drop quickly, resulting in children older than 15 months at the implementation time possibly lacking immunity—they are neither infected nor vaccinated. Since varicella is more dangerous to contract as an adult, this may pose a health risk. The model also indicates that there will be a temporary resurgence with increased transmission five to ten years after programme initiation due to changed transmission dynamics in the population when the group lacking immunity ages. Therefore, we also simulate a catch-up programme that ensures immunisation of all age cohorts up to adolescence.
The catch-up programme can be structured in various ways. In our main scenario, all children aged 7 years or younger at the programme’s start are offered two doses of the vaccine within four years. Additionally, all 15-year-olds without known previous varicella infection are offered two doses. The model finds this approach to be almost as effective as a (logistically unrealistic) scenario where all age cohorts from 2–15 years are offered two doses at programme start. On the other hand, a catch-up programme that only offers the vaccine to 15-year-olds is insufficient to avoid the temporary resurgence five to ten years later.
In our main scenario, we have a vaccination coverage of 96 percent, in line with the high coverage seen in other BVP components. In alternative scenarios, we examine the effects of lower coverage, from 80 percent and upward. Since varicella is a highly contagious disease, the model finds that coverage below 90 percent is insufficient to achieve herd immunity. Lower coverage will nevertheless reduce the overall incidence of the disease, but the unvaccinated may have an increased risk of contracting the disease later in life.
We present health economic analyses based on simulations from the healthcare, extended healthcare, and societal economic perspectives. The analyses indicate that a vaccination programme is cost-effective from a healthcare perspective and an extended healthcare perspective and leads to better health and lower costs from a societal perspective. In healthcare and extended healthcare perspectives, vaccination has a low cost per QALY, or incremental cost-effectiveness ratio (ICER), at NOK 131,946 and 43,904 respectively over a hundred-year period, assuming vaccines are purchased at list price. If vaccines are purchased at half of the list price, the ICER is reduced to NOK 34,253 from a healthcare perspective, and vaccination becomes dominant in the extended healthcare perspective (i.e. leading to better health and lower costs). There is marginal economic difference between scenarios with and without the catch-up programme.
The virus causing varicella disease, the varicella zoster virus, can also lead to herpes zoster later in life. This occurs when the virus, latent in the body after previous varicella disease, is reactivated. The varicella vaccine significantly reduces the risk of herpes zoster compared to contracting the disease—in the model, we assume a 90 percent risk reduction. There are hypotheses that new exposure to varicella infection helps prevent flare-ups of herpes zoster, so-called exogenous boosting. Thus, one might think that when varicella transmission disappears due to vaccination, it leads to increased incidence of herpes zoster. Recent research from countries that have had varicella vaccination for up to 30 years suggests this does not happen to a significant extent. Our model takes this re-exposure mechanism into account to keep herpes zoster in check, albeit with limited effect in line with new studies.
Herpes Zoster
Herpes zoster (HZ) is not an infectious disease, hence there is a more direct linear relationship between the number of vaccinated individuals and the amount of health gained. Still, there are a few complicating factors, notably that the HZ risk is age-dependent and that the vaccine wanes over time. To facilitate such aspects, the analysis has been carried out using the same epidemiological model as for varicella.
The model is calibrated to age-specific general practitioner (GP–in Norwegian fastlege) and emergency department consultations and, as for varicella, assumes a demographically stable population. Over a 100-year horizon, HZ vaccination produces a clear and lasting reduction in HZ incidence compared with no vaccination: incidence declines over the first 10–25 years after introduction and then stabilizes at a lower equilibrium. In the base case scenario, we assume 95% vaccine effectiveness (VE), 2.5% waning per year, 75% coverage at age 65 and limited catch-up during the first 5 years. In this scenario, long-term incidence is about 2.75 with vaccination, compared to 3.75 cases per 1000 person-years without vaccination. This corresponds to roughly 27% fewer cumulative HZ cases.
Catch-up vaccination mainly accelerates short- and medium-term gains. Broad catch-up in older age groups produces a sharp initial drop in HZ incidence, whereas more restricted catch-up yields smaller short-term benefits. After around two decades, however, all catch-up strategies converge to similar long-term incidence, close to that achieved with routine vaccination alone at the same target age.
Long-term HZ incidence levels depend on vaccine effectiveness, waning, coverage, and age at first dose. Programmes starting at 70 years with catch-up at 75 years achieve smaller reductions than vaccinating at 65 years, because fewer individuals spend a substantial part of their remaining lifetime protected. The model suggests that, under 95% VE, 75% coverage, and 2.5% waning per year, lowering the first-dose age below 65 years only marginally impacts overall incidence, reflecting the low baseline HZ risk at younger ages combined with waning dynamics.
We also performed health economic analyses of herpes zoster vaccination from healthcare, extended healthcare and societal perspectives. In the main scenario (vaccination at 65 years with catch-up at 70 and 75 years, 75% coverage, 95% VE, 2.5% waning per year, 100-year horizon and list-price vaccine), vaccination leads to more QALYs and higher costs, with ICERs of approximately NOK 474 000, 574 000 and 518 000 per QALY gained, respectively for the three perspectives. Using the extended healthcare perspective recommended by the Norwegian Medical Products Agency and an assumed threshold of NOK 275 000 per QALY, the programme is therefore unlikely to be considered cost-effective at current list price. From a societal perspective, more of the upfront vaccination costs are offset by reduced healthcare use and productivity losses, but since the relevant willingness-to-pay threshold is uncertain we cannot make a firm conclusion.
One-way sensitivity analyses show that the results are particularly sensitive to assumptions about vaccine price, discounting, herpes zoster incidence and the proportion of patients with severe pain. For example, a 50% reduction in vaccine price lowers the ICER to NOK 212 680, 313 416 and 256 559 per QALY respectively from the three perspectives, improving the cost-effectiveness profile, whereas assuming fewer patients with severe pain substantially increases the ICER. When we vary the age at vaccination in scenarios without catch-up, the lowest ICER is obtained for vaccination at 70 years, reflecting a balance between increasing disease risk and remaining life expectancy. In these analyses the extended healthcare perspective gives the highest ICERs, because it includes time costs of vaccination but not all productivity gains, while the societal perspective yields similar or lower ICERs, especially in the younger age groups where work absenteeism is relevant.