CALMA, an algorithm framework for mobile just in time interventions

Advancements in ubiquitous computing are rapidly changing preventative health care. These quick changes allow not only to track in real time the heath of an individual, but also to react to any anomalies that may indicate the need of help. This new health care paradigm (i.e., Just-in-time interventions) allows to support treatments and deliver help anytime and anywhere. In this work, we take a first step in modeling a flexible mechanism for choosing the most effective intervention from a set of available ones. We model this optimization problem as a Markov Decision Process (MDP) and we solve it by using Value Iteration in an online fashion. We show that the policy found by our algorithms for selecting interventions, at least doubles the efficiency achieved by a random policy, in minimizing the average number of interventions to relieve a patient from a condition such as stress.