“Vaccinating children against flu may be an effective way of protecting the rest of the population”, The Independent reported. Many newspapers covered the same story, with The Daily Telegraph saying that, “flu could be virtually wiped out if all under 16s were vaccinated against the disease”, and the Daily Mail saying that vaccinating children under five could cut infection rates by 70%. Most newspapers also said that the Joint Committee on Vaccinations and Immunisation (JCVI), considered and rejected the idea in 2006, but was keeping the issue under review.
These reports are based on a study that used mathematical techniques to model the potential effects of vaccinating children of different ages in England and Wales against flu. Although this model provides a useful forecasting tool to look at the possible effects of vaccination, further real-life studies on efficacy and safety, and modelling of cost-effectiveness of a flu vaccination programme in children, would be needed before being considered to become national policy. Carrying out a vaccination programme of all children in the UK on an annual basis would be a massive undertaking and there are also ethical issues to be considered.
Dr Emilia Vynnycky and colleagues from the Health Protection Agency Centre for Infections carried out the research. One of the authors was funded by a grant from the UK Department of Health; the Department did not have a vested interest in the study’s outcome. No other sources of funding are reported. The study that was appraised was a pre-publication version awaiting publication in the peer-reviewed medical journal: Vaccine.
In this mathematical modelling study, the researchers used complex mathematical techniques to predict the long-term effects of vaccinating children of different ages in England and Wales against flu. As it had not been modelled before, the researchers were particularly interested in the effects of only vaccinating preschool children, and looking at how the children’s contact patterns (i.e. who they came into contact with) affected transmission rates.
The researchers based their model on existing models of the effects of global flu epidemics. To develop the model for the UK population, they had to make assumptions about major contributing factors in the occurrence of influenza. These assumptions were based on real life observations as much as possible. Their model was adapted to take three main factors into account. First, that there are annual flu epidemics in the winter in the UK, with influenza type A epidemics occurring annually, and type B epidemics occurring every 2-3 years. Second, that children have a higher rate of infection than adults as they have not been exposed to as many flu viruses in the past. And third, that re-infection with the same strain of flu can occur, because the virus gradually changes (mutates).
The model took into account the various flu strains and their reoccurrence over time, based on what has been observed in the past. The model also took into account: percentage of people covered by vaccination, efficacy of the vaccine, immunity based on previous exposure, mutation in the influenza virus, period of infectiousness (set at an average of two days), and the percentage of people who would show symptoms if they were infected (64%). The study assumed that asymptomatic individuals would not be infectious.
The model’s birth and mortality rates from flu were taken from England and Wales in 2003. Children were assumed to have different levels of contact with people of different ages. Five different sets of probabilities of contact by age were used; these were based on assumptions used in previous models and have been shown to give good estimates of what happens in real life. These were adjusted further to better fit observed UK data. The definition of an effective contact was given as contact between an infected and an uninfected susceptible person that was sufficient for transmission of infection. Efficiency of contact was presumed to be greatest in the winter, when flu epidemics occur. Vaccination was assumed to be finished before the influenza season began. Vaccinations were assumed not to be given to children aged under six months. The model assumed that 60% of children within the targeted age groups could be effectively vaccinated (i.e. would receive the vaccination, and would develop immunity against flu).
The researchers found that the effect of a vaccination programme depended on the age of children vaccinated. Vaccinating children aged under one year had little effect on overall cases of flu in the entire population. Vaccination in other age groups led to reduction of clinical flu cases not only in their own age group but also in the overall population. Vaccination in these groups led to an initial reduction in cases of flu (a “honeymoon” period), followed by an increase, and finally settling at a rate lower than the pre-vaccination rate.
The effect on influenza B was greater than the effect on influenza A. Vaccinating children aged under two years could reduce clinical cases of influenza A by 11-22% in the overall population, and influenza B by 25-35%. Vaccination in children under five years of age was estimated to reduce clinical cases of influenza A by 22-38%, and influenza B by 44-69%, and in children aged under 16 years the estimates were 65-97% and 85-96% respectively.
These predictions were sensitive to certain assumptions used to make the model, mainly the patterns of contact the children had; this means the results change if you change these assumptions.
The researchers concluded that achieving a high level of effective flu vaccination among preschool children could bring benefits both to the children themselves, and to the wider community. They suggest that further population based studies on the effects of child vaccination programmes are needed to confirm their results.
This modelling study provides predictions for the effects of flu vaccines in different age groups of children. As with any model, the accuracy of the predictions depends on how accurate the assumptions are. The authors note that although some of their assumptions may be overly simplistic compared with real life, they are based on real data where possible. They say the model improves on other existing models by including the indirect effects of vaccination (known as “herd immunity” – where immunity at a certain level in the community protects other, non-immune members of the community).
As the authors suggest, actual population based studies will be needed to confirm the accuracy of their model, and to refine it if necessary. Also needed are further studies that assess the costs of different vaccination programmes, and balance them against the costs saved by avoiding illnesses, to see which strategy is most likely to be cost effective. These studies would need to be carried out before any nationwide programmes were considered.
As there are different strains of influenza virus circulating during different seasons, the vaccinations are prepared before the flu season starts and are designed to protect against the strains that are expected to be predominant. Even then, it is not always possible to get this 100% correct and the vaccination is more effective in seasons where it matches well with the strains of virus that are causing infection. Carrying out a vaccination programme of all children in the UK on an annual basis would be a massive undertaking. Even if immunisation reduced transmission rates to vulnerable and elderly population groups, the ethical issues of giving yearly injections to children who would normally suffer only an uncomplicated and self-limiting illness, must be considered.
Children bring joy into our life ... and viruses.