Prescription medication abuse is a major healthcare problem and can lead to addiction syndrome, higher healthcare cost, and serious harm to patients. Mobile health can play a major role in addressing prescription medication abuse. This is due to the ability to (a) monitor patient's health conditions anywhere anytime, (b) monitor patient's medication consumption, and (c) connect with healthcare professionals and utilize suitable interventions in time. More specifically, medication behavior can be monitored using smart medication systems, specialized wearable sensors or mobile devices with patient-entered consumption data. This data can then be analyzed for certain patterns to detect medication abuse. The goal is to design and develop an advance warning system based on the patterns of medication use to alert healthcare professionals and/or family members. Such system will utilize additional contextual knowledge of patient's condition and past history, current use, and information on abuse and addictive potential of medications. In this paper, we present medication related challenges and a preliminary design of a system to monitor and analyze the patterns of medication use, and utilize an analytical model for performance evaluation. The known patterns are utilized to estimate probability of near-future addiction. Our results show that medication adherence can be estimated and probabilities of multi-dosing and super adherence (>100% medication adherence) can be computed based on thresholds supplied by healthcare professionals. The work applies to m-health analytics and decision support systems.