Assessing individual treatment effectiveness in the presence of structurally missing measurement occasions

In this paper, we propose an exploratory method for assessing individual treatment effectiveness over repeated measures, when the number of measurement occasions depends on whether subjects fulfil or not certain conditions. We address this situation with the concept of “structurally missing occasions”. Our method involves three main steps. Firstly, we apply several strategies to overcome the problem of different numbers of occasions available for subjects. Secondly, given that data we consider are of “subjects-by-variablesby- occasions” type, we apply a specific multiway data analysis technique, using an iterative schema, for setting up indicators and assessing treatment effectiveness. In this context, we will focus on three-way multidimensional scaling, in that we adopt a subjectoriented approach to analyses. Finally, indicator trends can be studied over the different occasions by means of suitable graphical tools. To demonstrate the potential of our method, we consider a case study, based on data pertaining to personalized treatments for obesity and provided by the International Center for the Assessment of Nutritional Status (ICANS), University of Milan.