The introduction of ‘No jab, No school’ policy and the refinement of measles immunisation strategies in high-income countries

Background In recent years, we witnessed a resurgence of measles even in countries where, according to WHO guidelines, elimination should have already been achieved. In high-income countries, the raise of anti-vaccination movements and parental vaccine hesitancy are posing major challenges for the achievement and maintenance of high coverage during routine programmes. Italy and France approved new regulations, respectively in 2017 and 2018, aimed at raising immunisation rates among children by introducing mandatory vaccination at school entry. Methods We simulated the evolution of measles immunity profiles in seven distinct countries for the period 2018–2050 and evaluated the effect of possible adjustments of immunisation strategies adopted in the past on the overall fraction and age distribution of susceptible individuals in different high-income demographic settings. The proposed model accounts for country-specific demographic components, current immunity gaps and immunisation activities in 2018. Vaccination strategies considered include the enhancement of coverage for routine programmes already in place and the introduction of a compulsory vaccination at primary school entry in countries where universal school enrolment is likely achieved. Results Our model shows that, under current vaccination policies, the susceptible fraction of the population would remain below measles elimination threshold only in Singapore and South Korea. In the UK, Ireland, the USA and Australia either the increase of coverage of routine programmes above 95% or the introduction of a compulsory vaccination at school entry with coverage above 40% are needed to maintain susceptible individuals below 7.5% up to 2050. Although the implementation of mandatory vaccination at school entry would be surely beneficial in Italy, strategies targeting adults would also be required to avoid future outbreaks in this country. Conclusions Current vaccination policies are not sufficient to achieve and maintain measles elimination in most countries. Strategies targeting unvaccinated children before they enter primary school can remarkably enhance the fulfilment of WHO targets. Electronic supplementary material The online version of this article (10.1186/s12916-019-1318-5) contains supplementary material, which is available to authorized users.


The non-stationary age-structured vaccination model
The evolution of susceptibility is simulated through a deterministic non-stationary agestructured model stratified in 85-years age classes. The population is divided into different compartments: individuals protected by maternal antibodies (M), susceptible individuals (S), and individuals who acquired immunity against measles either through vaccination or natural infection occurred before 2018. Among susceptible individuals we kept trace of vaccinated individuals who were not successfully immunized due to first or second dose vaccine failure of (F). Newborn individuals are protected against measles infection for 6 months on average by the passive transfer of maternal immunity after which they become susceptible. In the model individuals are vaccinated with a first and second dose, if any was scheduled in 2017, according with data reported on immunization schedules and coverage levels 1 , by mimicking country-specific vaccination activities performed in 2017. It is worth noting that only individuals who were not successfully immunized with a first dose (F) are considered eligible for a second dose. In contrast, vaccination at school entry and catch up campaigns are designed for children who were not reached by first or second dose vaccination programs and are therefore administered to susceptible individuals except from vaccine failures (S-F). The model takes into account the vital dynamics of the host population, and is informed by country-specific crude birth rate and age-specific mortality rates predicted for the period considered 2 . Epidemiological transitions for each individual's age are described by the following system of ordinary differential equations: where t and a represent time and individuals' chronological age respectively; b(t) and d(a,t) are the crude birth rate and age-specific mortality rate at time t respectively. Coverage levels at time t associated with first dose, second dose programs and school entry vaccination are denoted by ! , ! and !" respectively; ! , ! and !" are respectively the age at first, second and school entry doses; !" is the vaccination coverage characterizing the catch up campaign performed in 2018 and targeting ages between !"! ( ) and !"! ( ); ε(a) represents the vaccine efficacy, which is assumed to be 95% when the vaccine is administered at 15 months or more and 85% otherwise 3,4 . Finally, N(a,t) represents the total population of age a at time t and δa,j is the Dirac delta function, which is equal to 1 for a=j and 0 otherwise.

Demographic data
The model was informed with country specific longitudinal data on crude birth rates (Fig. S1), age specific mortality rates ( Fig. S2 and S3), as predicted by the United Nations World Population Prospect 2 for the period 2018-2050. In particular, according to UN predictions, in countries like Ireland, Italy, Singapore and South Korea, populations are still ageing as a result of the strong decline of fertility rates occurred between 1960-2050. Finally, as shown in Fig.S3 in the UK, US and Australia, the population age structure is expected to remain rather stable between 2018 and 2050.

Initial susceptibility distribution
In a previous modeling work 4 , Trentini et al. estimated the temporal changes in the age-specific measles immunity profiles between 1950-2017 in different countries. Estimates obtained in Trentini et al. 4 for 2015 are here adopted as initial conditions on the age distribution and overall immunity level at the end of 2017, across the 7 high-income countries. The initial age distribution of susceptible individuals here considered is shown in Fig. S4.

Vaccination at school entry
Coherently with laws approved in some European countries 5,6 on compulsory immunization, we simulated a vaccination at primary school entry aimed for all individuals who are not compliant to country-specific programs. Ages at which this new routine dose is administered are countryspecific and depends on the age at primary school entry (see Tab. S1).

Sensitivity analysis on maternal antibodies protection
In our model we assume that all children are protected by maternal antibodies at birth for an avarage period of 6 month. Due to uncertainty of seroprevalence of mothers we decided to conduct a sensitivity analysis to check for the robusteness of our results. In Fig. S5 we show projection of residual susceptibility in 2050 with the conservative assumption that all newborns end up in the susceptible class (absence of maternal protection). Estimates of residual susceptibility vary by only 4% in Australia and US, and by 1% in all other countries with respect to those obtained with our assumption on maternal antibodies. Results seems therefore to be robust to a variation in this particular assumption.

Additional results
In the main text we presented results obtained by simulating current routine programs complemented with both school entry vaccination and a catch up campaign in 2018 targeting individuals between 1 and 15 years of age. In this section, temporal changes in measles susceptibility are investigated when current routine programs are combined with vaccination at school entry alone.
Obtained results show that, under coverage levels above 40%, the introduction of this additional immunization activity in 2018 can reduce the susceptible fraction of individuals below 7·5% of the population within 2050 in UK, Ireland and US, South Korea, Singapore and Australia therefore making possible both the achievement and maintenance of measles elimination in these countries (Fig. S6). In Italy, 100% coverage vaccination at school entry would determine acceptable levels of susceptibility, but that would result still above the elimination threshold of 7·5%.