The Controlled Substances Act of 1970 declared that certain opioid formulations are considered a Schedule II Controlled Substance, which subjects the drugs to a strict regulation by the U.S. Drug Enforcement Agency. Opioids have since then been classified as (a) having a high potential for abuse, (b) potentially harmful and may cause severe psychological or physical dependency if abused, and (c) accepted medical use in treatment within the United States. Over the past decade opioid prescriptions have since increased, causing major misuse and abuse of the opioids.

Throughout this course of observational study, my partner and I were able to obtain quarterly data over a ten year span from 2011 to 2021. This data was then broken down by county, substance types, and a number of other categories within the state of Pennsylvania. Specifically, we looked to assess the impact of ruralness and education level on opioid dispensation within the year 2020 by looking at the obtained county-level data. To analyze these findings, we then decided to use "R Programming" to make use of different ANOVA methods, in particular we used a one-way ANCOVA model. 

After exploring the ANCOVA model which required there to be a linear relationship between the response and covariate within our data, we began to look more in depth at the assumptions that needed to be met within our observational study. The assumptions that had to be met were the Gaussian Residuals assumption, Homoscedasticity, and Homogeneity. Our Gaussian Residuals assumption was checked by creating a Q-Q plot of the residuals with a 90% envelope to assess the assumption even further. Approximately 10.4% of the residuals fell outside of the envelope, which was not too much cause for concern. To check for homoscedasticity a Tukey-Anscombe plot was created and we wanted to assess whether there was a distinct pattern that fitted to the related value of the residuals. Homogeneity slopes assumption assumed the constant rate of change of the response with respect to the covariate. Here we used a Type III Sums of Squares interaction, which resulted in a F-statistic value of 0.2654 and a significant p-value of 0.899. From the significant value of both of these terms, we could determine there was no significant interaction between the density quantile (opioid dispensation) and proportion of the population with a Bachelor's degree (education level). 

The overall issue of prescription opioid usage and abuse presents and overall concern for the health and safety of the public. By understanding the trends and patterns in the dispensing, it may help healthcare workers mitigate the usage and educate both patients and others in relevance to their geographical location. It may also help future research on opioid dispensation, and could potentially operationalize the economic wellbeing through per capita income for individuals who live below the federal established poverty line. 

Take a look at the link below to see the entire project report!