Modelling the first dose of measles vaccination: the role of maternal immunity, demographic factors, and delivery systems

Abstract

Measles vaccine efficacy is higher at 12 months than 9 months because of maternal immunity, but delaying vaccination exposes the children most vulnerable to measles mortality to infection. We explored how this trade-off changes as a function of regionally varying epidemiological drivers, e.g. demography, transmission seasonality, and vaccination coverage. High birth rates and low coverage both favour early vaccination, and initiating vaccination at 9-11 months, then switching to 12-14 months can reduce case numbers. Overall however, increasing the age-window of vaccination decreases case numbers relative to vaccinating within a narrow age-window (e.g. 9-11 months). The width of the age-window that minimizes mortality varies as a function of birth rate, vaccination coverage and patterns of access to care. Our results suggest that locally age-targeted strategies, at both national and sub-national scales, tuned to local variation in birth rate, seasonality, and access to care may substantially decrease case numbers and fatalities for routine vaccination.

Fig 1

Force of infection over age, based on the pattern observed for Niamey, Niger that peaks at 3 years [24].

Fig. 2

Best strategy (red for 9–11 months and green for 12–14 months) obtained using the ratio of cases weighted by mortality rate over 10 years following initiation of vaccination at 9–11 months, or a t 12–14 months, for varying degrees of seasonality (α = 0, α = 0·2, α = 0·6) across a range of vaccination coverage × axis) and birth rate/1000 per year (y axis). Vaccination at 12–14 months performed poorly at low coverage and high birth rates (top left panel, red), but well over a range of higher vaccination coverage and lower birth rate (bottom right, green). The right-hand panels show time-series plots following the start of vaccination (the timing of which is indicated by a vertical red line) for vaccination at 9–11 months (red) or 12–14 months (green) of total incidence (top) and average age of infection (bottom) for two different locations for the graph indicating intermediate seasonality (symbols : ●, ■). The dashed box in the bottom left panel indicates the range taken in Figure 3 for comparison.

Fig. 3

Optimal time in years, over 10 years, for switching from 9–11 to 12–14 months vaccination as a function of the birth rate/1000 per year (x axis) and vaccination coverage (y axis), identified as the switching time leading to the fewest number of cases weighted by mortality risk; shown for three different levels of seasonality, as in Fig. 1. The time series in the right-hand panels show incidence and average age of infection where α = 0·2, for the three main cases: (i) never switch to 12–14 months [top panel, corresponding to the birth and vaccination level shown by the circle (●)]; (ii) switch at an intermediate level (middle panel, ■); (iii) switch immediately (bottom panel, ▲).

Fig. 4

(a) Pattern of efficacy used in estimating the optimal upper age of vaccination; (b) Patterns of first access to care set to capture a more urban (—) and rural (- - -) community.

Fig. 5

Optimal upper age of vaccination over 10 years for different parameter combinations for different birth rates/1,000 per year (colours corresponding to each birth rate are provided in the legend) with (a) no differences in patterns of access to care over age (vaccination is proportional to the number of susceptible individuals in each age group); (b) with ‘urban-like’ patterns of access to care; and (c) with ‘rural-like’ patterns of access to care (see Fig. 4); horizontal dotted lines (·····) show the range of age where transmission is highest according to data estimated from Niamey, Niger.