Abstract

The statistical analysis of infectious disease data is complicated by the fact that infections in different individuals are not independent, especially when disease is transmitted directly from person to person (Andersson and Britton, 2000; Becker, 1989). This problem was recognized long ago; Sir Ronald Ross called it “dependent happenings” (Ross, 1916). In this chapter, we show how dependent happenings can be handled using methods adapted from survival analysis. Instead of considering failure times in individuals, we consider failure times in ordered pairs ij that consist of an infectious individual i and a susceptible individual j who is at risk of infection from i. We call this approach pairwise survival analysis, and these methods are being implemented in the transtat package for R (available at github.com/ekenah/transtat).