Chapter 1 Introduction: Domain Problem and Data Characterization

1.1 Domain Problem

For our project we explored data related to opioids, in an effort to better understand and obtain insight into the opioid epidemic. Our domain problem is one for a researcher wanting to explore the connection between prescriber rates of opioid prescriptions and opioid-related deaths both in the country as a whole and drilling down to the state level. The first part of the data we examined was prescriber data. This data would allow the researcher to see the distribution of opioid prescriptions across the US and also find the most commonly prescribed opioids. The second part of the data analysis involved finding information about deaths that occur from opioid overdoses in the United States. This would also allow the researcher to drill down to the state level. Another level of detail that we felt would be important for the researcher is to categorize these deaths into different types of groups such as race and age. This would help identify groups that are suffering from opioid addiction, which would then allow researchers to provide information and determine where attention should be focused to combat the opioid epidemic.

1.2 Data Characterization

Opioids are medicines that are typically prescribed by doctors to treat severe pain. Heroin is also considered an opioid due to its similar chemical makeup and affects. While the benefit is pain relief, opioids are highly addictive and can lead to patients abusing the drugs. The increase in opioid prescriptions over time has led to misuse of prescription drugs; the crisis has gotten so bad, that on October 26, 2017 the Acting HHS Secretary declared a Public Health emergency (What Is the U.S. Opioid Epidemic? 2019).

For data about prescribers, we acquired Medicare Part D Prescriber Data from the Centers of Medicare and Medicaid Services (CMS) for the most recent year available (2016) (Medicare Part D Prescriber Data 2016). Medicare Part D is a prescription drug benefit to help people get at minimum, standard coverage of drug prescriptions for their medical needs (Medicare Part d Prescription Drug Benefit 2018). While this represents a subset of information on opioids, it contains a good sample size to work with. We set the level of detail to be an aggregate of prescribers as a summary at the national and state level. We further limited the scale to only records related to opioid prescriptions. We focused our research on this dataset to indicators we felt would be of most interest to the researcher. These indicators are the number of prescribers, number of claims and total drug cost related to opioids. The second part of our development explored the connection between high prescriber rates and the number of deaths related to opioids. If addiction and overdoses are linked to prescribed medications instead of drugs that were illegally obtained, this could be a good starting place for researchers interested in solving this problem. In addition, researchers may be interested to know the characteristics of the individuals whose deaths occurred due to opioid overdoses in each state. We compared age data, race data and the type of opioids to death rates to see which groups are most affected on a state-by-state basis.

1.3 Target Users

As mentioned in the previous section, this application is being designed for a researcher in the medical field. Our aim was to make this a tool that a researcher can use to find information at both the national and state level. This allows the usability to reach a larger audience where researchers from any state can use this tool to find information about their state of interest and compare what the findings are nationally or compare against other states. With the hope of researchers using this application to find valuable insight that can be used in their research, we provide the ability for users to download our data in Excel and most of our visualization plots can be exported as images.

1.4 Application Information

The application is hosted at https://nocchipi.shinyapps.io/dsba_5122_final_project/. Information about the changes made to the application and how to install it locally can be found on our GitHub repo at https://github.com/nick-occ/dsba_5122_final_project.

References