Analysis of the Incubation Period for Measles in the Epidemic in Greenland in 1951 using a Variance Components Model

This paper presents a model for analysing the incubation period of highly infectious diseases in populations where almost all susceptibles become infected during an epidemic. The model leads to a simple method for estimating the variance of the duration of the incubation period without any distributional assumptions. Further, the influence of covariates on the duration of the incubation period can be analysed. Data from the epidemic of measles in Greenland in 1951 are analysed and it is found that intersymptom times are correlated within households, suggesting that secondary cases are infected almost simultaneously. This result is inconsistent with a variation in the times of infection of secondary cases within a household which is often assumed when analysing data on measles. Prophylactic treatment did not prevent infection in the epidemic in Greenland, but it is found that the incubation period tended to be shorter for persons not receiving prophylactic treatment