Conquering an Empire of Pain: An Optimal Learning Strategy for Identifying the Stage of Opioid Addiction

Abstract

In this thesis, we will implement a sequential decision model that allows us to best choose a medical facility (hospital, doctor's office, etc.) to collect information from. We will use a network of `caregivers,' a category including nurses, physician's assistants, physicians and more, to go about the actual process of collecting information, by asking admitted patients a series of questions to determine a rough category of addiction that they fall into. These categories will range from the least at risk, simply being prescribed pain pills, to the most at risk, exhibiting signs and symptoms of full addiction. We would like to catch people in a middle category, where they are exhibiting some signs of addiction, and where interventions could be most effective, making this problem a classification problem. This is not to say that interventions in people in the first or last category of patients would not be effective, as those also must be explored and implemented, but given limited time and a limited budget, interventions at this critical point could be more valuable. This is a pure learning problem that focuses on the value of information. After we have chosen what facilities to target and collected the category information, the government and other agencies like state Departments of Health can better target interventions as they work to combat the opioid crisis.

After we have gone over the extensive background to this issue, as well as the specific narrative to our problem, we will develop our model. After it has been developed, we will discuss different strategies we can use in evaluation. In later chapters, we will be able to compare and contrast these policies, by analyzing the results. Once we have analyzed these results, we can make some conclusions about which strategies might work best in real life, and actually be implemented in the future, as we will learn that more action must be taken in order to prevent the rising death toll due to opioids.