Chapter 5 User evaluation

5.1 Who is the User?

As mentioned in the previous chapter of our report, the application is modeled as a tool for researchers looking for more insight to aid in their research into opioids. The app is designed so that it includes information for all 50 states. We believe application usage could be expanded to be used by different State departments such as Health and Human Services to see what areas need focus when developing opioid action plans. For example, the NC Department of Health and Human Services currently has an ongoing project called the Opioid Action Plan Implementation Initiative (North Carolina’s Opioid Action Plan 2019), they may find an application such as this useful for their research.

5.2 Using the Application

The application is designed so that each section provides insight into different factors of the opioid epidemic and is summarised in the analysis tab. When forming an action plan for the opioid crisis, one question researchers and public policy officials may have is whether the drugs identified as the addictive substance and/or the cause of the overdose are drugs that can be legally prescribed, such as oxycodone or percocet, or if they are illegal drugs such as heroin. We believe the answer to this question would help them choose how to utilize resources and focus their efforts.

We believe a typical scenario for a researcher would be to first get preliminary information about the drugs being prescribed. This could be either to find new information or compare against their findings. For this preliminary information into the drug usage, we provide a high level view of what drugs are being prescribed the most through a word cloud visualization when the application first loads. The map visualization would be the next thing the user can load to get an idea of which states have a high number of drug prescriptions and where they may want to focus their research in the remaining sections of the application.

The two main parts of our research are deaths and prescriber rates and seeing the relationship they have for every state. The user would investigate the deaths tab to find information about deaths by state and examine which characteristics such as race, age groups and opioid type stand out. To see the breakdown of the different groups the user would hover over the state to see a bar chart with the results for that state. The second part that aids in our research is looking at a map of prescription rates. We acquired data at the county level related to prescription rates so the user has the option to drill down further by clicking on the state, to see in which parts of the state the prescription rates have changed.

  • Prescriber Rates

Lastly, the analysis tab is what puts it all together and the user has the option to select a state and see the relationship between prescription rates and death rates over time. In order to truly assess how the state is performing we included the national average for comparison. The second part of the analysis would show which race, age group and type of opioid is most prevalent for the selected state. The application is designed so that any section can be run independently from the other sections, with the realization some users may just be interested in our final analysis result.

In our testing we wanted to find an example of what information we can draw from when using the application. Our results show that Hydrocodone-Acetaminophen was commonly the most prevalent in the US and in most states with a high number of prescribers. We then examined the death data for the different variables and found that there is definitely a common trend of opioid death facing people classified as white non-hispanic. Deaths by age groups and type of opioid varied from state to state, but the state that stuck out at us the most was West Virginia because it was at the top of the list for every year in total deaths. In the prescriber rate analysis, West Virginia was also a state with one of the highest prescriber rates in the country. In the analysis tab we examined West Virginia and see that their prescription rate is very high and is significantly higher than the national average and while we see the prescription rate drop, the deaths are still rising. When we look at the second plot with the death variables we see that in 2015 heroin-related opioid deaths are rising as well, which are not linked to prescriptions. When comparing different states we do see a common trend that most prescription rates appear to decline between 2010 and 2015, but in a lot of the cases we saw the deaths were still either staying the same or rising. This leads us to believe that people may be finding alternate ways to get prescribed opioid medications, or are using illegal drugs classified as opioids such as heroin.

  • Analysis Results of West Virginia

References