Factors Behind Overdose Deaths

Deaths caused by drug overdose have been increasing at an alarming rate. In order to gain more insights about the social and economic factors that lead to a higher risk of overdose in Allegheny County, PA, two IBM statistical tools; SPSS and IBM Watson, were used to explore the data set which contained information of 3551 patients, reported to have died due to the Opioid Crisis in Allegheny County. There has been a great deal of research into the demographic and socio-economic causes of drug overdose deaths, but the scope of the societal problem certainly warrants further research, similar to the need for healthcare dissertation help in exploring these critical issues. We selected a dataset from Allegheny County, Pennsylvania, to gain a microscopic insight about the large epidemic of a drug overdose in the US where the city of Pittsburgh is located so that we could study one area in our own state to highlight both the pressing and personal nature of these issues, as well as to gain a better understanding of one area that may be used as a microcosm for the larger issues at hand. The dataset was derived from www.data.gov, compiled by the Allegheny County Office of the Medical Examiner. It includes fatal accidental overdoses in the county and contains variables for the date of death, the time of death, the manner of death, the age of the decedent, their gender, race, the seven most prevalent drugs found in the overdose in alphabetical order, the zip code of the overdose incident, and the zip code where the decedent lived. The data extends from 2008 to 2017 in separate files which were combined into one master dataset for our analysis.

To gain further insights, we decided to search for other variables that may have a relationship to the overdose rates. These other variables can roughly be broken up into two categories, climate, and economical. The climate variables we used were Monthly Average Temperature and Temperature Departure from Mean 2008-2017 Base Period. These data were retrieved from the National Oceanic and Atmospheric Administration “Climate at a Glance” tool.

The economic indicators were Unemployment Rate for the county, and Income Inequality, which is measured as a ratio of the mean income for the highest quintile of earners divided by the mean income for the lowest quintile of earners in the county. These two datasets were collected from the FRED Economic Data service from the St. Louis Federal Reserve Bank. The final additional variable gathered was the uninsured rate, which was sourced from the Small Area Health Insurance Estimate from the U.S. Census Bureau.

Deaths due to the opioid crisis have reached epidemic levels in the US, with annual numbers more than deaths caused by motor vehicle accidents, gun violence or HIV (Ciccarone, 2017). In 2016, more than 42,000 Americans lost their lives from a drug overdose, including 613 Allegheny county residents, which was a jump of over 44 per cent. This was largely due to the presence of newer, strong drugs such as fentanyl (Lord). Local health officials have found it difficult to keep up with the pace of new drugs being introduced into common usage, especially the presence of fentanyl in most doses of heroin sold on the streets. Fentanyl is also extremely potent and can cause overdose deaths on the first try for some users (Lord). Today, drug abuse is one of the major problems in the US. This can be supported by using the study by the Center for Disease Control (CDC) stating that in 2017, more than 47,000 people died in the country (Lopez, 2019). Further Saloner and et al. (2018) stated that more than 2 million Americans live with addition towards drugs such as opioids.

Looking at every year separately on IBM Watson, a pattern that appeared to be redundant throughout the years captured in the dataset was that months with abnormally high death counts were followed by months with abnormally low death counts, and vice versa. This may be related to the availability of drug supply in the market, reflecting the high addiction level to drugs containing Fentanyl and Heroin. In recent, sometimes, the government and the authorities have a crackdown on the drug ring, but the effects of it are yet to be clearly understood (Mazer-Amirshahi and Nelson, 2015). The authorities have successfully taken down various platforms that were previously used for selling and buying drugs. Not a lot of information has been determined in terms of the effects of such a crackdown (Edlin, 2016). In 2017 this pattern was most visible in the months of August and September. August saw the highest number of deaths and was followed by September, a month with the lowest death rates in that year. The rest of the year ended with relatively low death rates than previous months.

In 2016, and based on our finding above, June was a month of low death rates, which could be related to drug shortage in the market. Subsequently, the following months, starting in July, saw a higher death rate. You can see somewhat of a reversal of this pattern. Overdose deaths decrease to below average levels in June, followed immediately by an increase to above previous levels in July.

This alternating pattern is apparent in essentially every year that there is data for. However, 2013 is a very good example of the potential oscillating trend, in which months of very high overdose deaths are followed by significantly lower overdose deaths.

This statistical analysis study provided an expanded exploration of the overdose drug problem in Allegheny County. Using SPSS, we were able to validate previously known findings of overdose deaths, such as the age range, and general demographics of the affected populations. Much like Lord’s article, we have found that over the last several years, overdose deaths have been on the rise and that the cocktail of drugs involved in this has never been more complex. We found SPSS to be a very useful and capable tool to gain a better understanding of our sample set, as well as using it to generate informative and understandable visuals. Its ability to cluster the dataset based on K-Means clustering is also very powerful and allowed us to gain insight into where the “centres of gravity” are in our dataset, so to speak.

IBM Watson delivered a visual analytic into the overdose drug deaths by month in a unique way, which is based on a cognitive computing system. It was also a very capable tool that allowed us to find new information about our sample, particularly when combined with outside variables added to the base dataset. It was able to generate very practical visuals with clear trends out of a large dataset with many variables. However, our collective skills are limited, and the issue of overdose deaths is so pressing that more research by skilled individuals is needed to gain as much insight into this epidemic so that real solutions can be found. However, we can recommend that IBM SPSS and Watson tools are to be used for further analysis, as they have the capability to reveal more insights into the drug epidemic crisis.

There is an urgent need for the authorities to find a permanent and effective solution for the drug problem in the country. In this regard, it can be recommended that easy access to helpful medicines such as naloxone (Sederer, 2016). It immediately blocks out the spread of deadly respiratory suppression that might have caused due to intakes of drugs like heroin, methadone or any other. It should be available to all the first responders, as currently, it is one of the most effective ways of stopping the spread of drug in the human body. Other than this, the authorities should also spend more time and money on advertisements to promote and educate individuals and families about the problem as well as the ways that can be used to overcome it (Sederer, 2016).

Refrences

Saloner, B. and et al., (2018). A public health strategy for the opioid crisis. Public Health Reports, 133(1_suppl), 24S-34S.

Mazer-Amirshahi, M. and Nelson, L. S. (2015). Drug shortages: implications for medical toxicology. Clinical Toxicology, 53(6), 519-524.

Edlin, B. R., (2016). Access to treatment for hepatitis C virus infection: time to put patients first. The Lancet Infectious Diseases, 16(9), e196-e201.