The risk of infection that is calculable from the prevalence of infection is thus an average annual risk. It is a proxy for the incidence of tuberculous infection. It differs from the true incidence of infection that led up to the observed prevalence of infection in that it is constant while the true incidence is likely to have varied over time.
The prevalence of infection is measured at calendar time b+a. The calculated risk of infection assumes a constant incidence of infection over the entire lifespan of the individuals in the survey. This is, however, the least likely course over time. The true incidence of infection that led up to the measured prevalence may have declined (incidence 1) or increased (incidence 2) or may have taken any shape.
This schematic presentation shows that the average annual risk that was calculated existed only at one particular time between calendar time b and calendar time b+a. This point in time b+x cannot be known unless there are serial surveys. It is therefore often approximated to lie in the middle of the period, i.e. at b+(a/2).
