Graphical trees: the simultaneous display of risk of fetal death by pregnancy history and maternal age

Many epidemiological studies consider sequences of events over time where t he outcome of interest is determined, at least in part, by similar outcomes that have occurred in the past, e.g. risk of fetal death in successive pre gnancies. In this situation, a probability tree can provide useful insights into the relationship between successive events. However, probability tree s are usually limited to the display of a single variable, whereas the majo rity of studies also include other risk factors. Analyses involving other f actors, particularly those that consider interactions, can produce an abund ance of results that make identification of potentially interesting pattern s difficult. A new graphical approach is described for simultaneously prese nting the effects of two risk factors where one relates to a sequence of ev ents over time. The graphical tree highlights patterns in the raw data and is therefore proposed for use in exploratory analyses and hypothesis genera tion. The approach is introduced and illustrated in the context of risk fac tors for fetal death, and an interaction between pregnancy history and mate rnal age is explored.