Estimating the incubation period of raccoon rabies: a time-space clustering approach.

We used a time-space clustering approach to estimate the incubation period of raccoon rabies in the wild using data from the 1999-2001 invasion of raccoon rabies into eastern Ontario from northern New York State. The time differences and geographical distances between all possible pairs of rabies cases were computed, classified and assembled into a time-space matrix. The rows of that matrix represented differences in cases in weeks and the columns represent distances between cases in kilometers and the values in the cells of the matrix represent the counts of cases at specific time and distance intervals. There was a significant cluster of pairs 5 weeks apart with apparent harmonics at additional 5-week intervals. These results are explained by assuming the incubation period of raccoon rabies had a mode of 5 weeks. The time clusters appeared consistently at distance intervals of 5 km. We discuss the possibility that the spatial intervals were influenced by the 5 km radius of the point infection control depopulation process used in 1999 and the 10-15 km radial areas used in 2000. With the practical limits of those radii, there was an intensive effort to eliminate raccoons. Our procedure is easy to implement and provides an estimate of the shape of the distribution of incubation periods for raccoon rabies.