their drinking increased, decreased or remained stable over a given period of time 13. In addition, there is no clear agreement concerning the appropriate means of classifying drinkers who cease consumption prior to baseline measurement. A number of proposals have been put forward, including an intention‐to‐treat approach, which assigns former drinkers with a current drinking value predicted to be most representative of their prior consumption 36, or the use of a retrospective life grid as a means of soliciting participants to estimate their past consumption 37. Although the best approach for dealing with both issues will differ to some extent according to the aim of the study and the data available, the incorporation and treatment of longitudinal data for the analysis of alcohol‐related risks represents an important area for future debate.
