126 ESTADO DE LA NACION EQUIDAD E INTEGRACIÓN SOCIAL CAPÍTULO

Contribution

The thesis aimed at following the entire data science process from collecting raw data to generating models to infer information. The entire process workflow is documented to help further research on this front.

The logistic regression resulted in selecting the influential set of variables in determining if a patient is opioid addict or not. These features include the final set of variables that helped in building the regression model are – Age, race, public health insurance coverage, depressed for more than half of the days, pain affected their normal functioning and joint pain. The factors like education, total expenses did not have an impact on the final model.

Implication

The logistic regression model devised above has identified some independent variables that affect the prescription opioid analgesic use. This information can be useful when we are targeting initiatives to control the spread of opioid use and to identify the people group. The exploratory analysis revealed that most of the overuse belonged to non- Hispanic white, female population. Even though this study produces some interesting insights, the results should be handled with caution for future research due to the limitations of MEPS data.

Limitation

MEPS data is restricted to non-institutionalized Americans, this does not include military population where the problems of opioid abuse are wide spread after retirement from active service. In fact, a part of the MEPS survey is self-reporting of the information by the patient. The patient cannot be trusted in the self-reporting model especially when we are dealing with overdoes problems.

Future work

Our model does not include the factors of cost, form and dosage of consumption of the opioids. If this information along with the deeper understanding of the health variables present the consolidated data file for each patient can give more data points for future research on opioid addiction. Increasing the accuracy of the models is always a problem of data science. With this model being an above average accurate one, future models with better prediction accuracy can be path-breaking.

Bibliography

[1] Neuman, S. (2018, January 17). Opioid Crisis Blamed for Sharp Increase in Accidental Deaths in U.S. Retrieved from:

https://www.wabe.org/opioid-crisisblamed-for-sharp-increase-in-accidental-

deaths-in-u-s/

[2] Drug Facts: What are Common Prescription opioids? National Institute on Drug Abuse. (June 2018). Retrieved from:

https://www.drugabuse.gov/publications/drugfacts/prescription-opioids

[3] Opioid Overdose Crisis, National Institute on Drug Abuse. Retrieved from:

https://www.drugabuse.gov/drugs-abuse/opioids/opioid-overdose-crisis

[4] Vowles KE, McEntee ML, Julnes PS, Frohe T, Ney JP, van der Goes DN. Rates of opioid misuse, abuse, and addiction in chronic pain: a systematic review and data synthesis. Pain. 2015;156(4):569-576.

doi:10.1097/01.j.pain.0000460357.01998.f1.

[5] Muhuri PK, Gfroerer JC, Davies MC. Associations of Nonmedical Pain Reliever Use and Initiation of Heroin Use in the United States. CBHSQ Data Rev. August 2013.

[6] Stein, R. (December 8, 2016). Life Expectancy in U.S. Drops for First Time in Decades Report Finds. North Carolina Public Radio. Retrieved from:

http://www.npr.org/sections/health-shots/2016/12/08/504667607/life- expectancyin-u-s-drops-for-first-time-in-decades-report-finds

[7] Stein, R. (December 8, 2016). Life Expectancy in U.S. Drops for First Time in Decades Report Finds. North Carolina Public Radio. Retrieved from:

http://www.npr.org/sections/health-shots/2016/12/08/504667607/life- expectancyin-u-s-drops-for-first-time-in-decades-report-finds

[8] Preventing Opioid Misuse and Overdose: Data Sources and Tools to Inform Assessment and Planning Efforts. Retrieved from:

https://www.samhsa.gov/capt/sites/default/files/resources/national-data- sources-opioids-related-needs-assessment.pdf

[9] Understanding the Epidemic: Opioid Overdose, Division of Unintentional Injury Prevention, Division of Unintentional Injury Prevention. Retrieved from:

https://www.cdc.gov/drugoverdose/epidemic/

[10] Sites, B. D., Beach, M. L., & Davis, M. A. (2014). Increases in the use of prescription opioid analgesics and the lack of improvement in disability metrics among users. Regional anesthesia and pain medicine, 39(1), 6-12. [11] (March, 2017). International Narcotics Board Report – 2016. United

Nations – Vienna, Austria. Retrieved from:

http://www.incb.org/documents/Publications/AnnualReports/AR2016/English /AR2016_E_ebook.pdf

[12] (July 22, 2016). Comprehensive Addiction and Recovery Act (CARA) Public Law 114–198. Retrieved

from:https://www.congress.gov/114/plaws/publ198/PLAW-114publ198.pdf

[13] (August 2, 2017). Attorney General Sessions Announces Opioid Fraud and Abuse Detection Unit. Department of Justice, Office of Public Affairs. Retrieved from:

https://www.justice.gov/opa/pr/attorney-general-sessions-announces- opioidfraud-and-abuse-detection-unit

[14] Dunham, J. & Kearney, S. Jr. (June, 2016). Data and Analytics to Combat the Opioid Epidemic. Research Brief, International Institute for Analytics – SAS. Retrieved from:

https://www.sas.com/content/dam/SAS/ja_jp/doc/whitepaper1/wp-iia- dataanalytics-combat-opioid-epidemic-108369.pdf

[15] Bresnick, J. (September 6, 2017). CDC Awards $28.6M for Big Data Analytics to Track Opioid Abuse. HeathITAnalytics. Retrieved from:

https://healthitanalytics.com/news/cdc-awards-28.6m-for-big-data-analytics-

totrack-opioid-abuse

[16] Greengard, S. (October 10, 2018). Predictive Analytics in Health Care: Opportunity or Risk? Retrieved from:

https://www.physicianleaders.org/news/predictive-analytics-health-care- opportunity-risk

[17] Bresnick, J. (September 4, 2018). 10 High-Value Use Cases for Predictive Analytics in Healthcare. Retrieved from:

https://healthitanalytics.com/news/10-high-value-use-cases-for-predictive- analytics-in-healthcare

[18] Bresnick, J. (May 25, 2017). UPenn Uses Machine Learning, EHRs to Target Severe Sepsis. Retrieved from:

https://healthitanalytics.com/news/upenn-uses-machine-learning-ehrs-to- target-severe-sepsis

[19] LaPoint, J. (February 16,2017). 78% of Hospital Staff Still Face Manual Supply Chain Management. Retrieved from:

https://revcycleintelligence.com/news/78-of-hospital-staff-still-face-manual- supply-chain-management

[20] Charumilind, S. Escobar, Elena. Latkovic, T. (June, 2018) Ten insights on the US opioid crisis from claims data analysis. Retrieved from:

https://www.mckinsey.com/industries/healthcare-systems-and-services/our- insights/ten-insights-on-the-us-opioid-crisis-from-claims-data-analysis

[21] Medical Expenditure Panel Survey, MEPS Topics. Retrieved from:

https://meps.ahrq.gov/mepsweb/

[22] Medical Expenditure Panel Survey, Prescribed Medicines file. Retrieved from:

https://meps.ahrq.gov/data_stats/download_data_files_detail.jsp?cboPufNumb er=HC-188A

[23] Pryor, Alan. (2017) U.S. Opiate Prescriptions/Overdoses. Retrieved from:

https://www.kaggle.com/apryor6/us-opiate-prescriptions

[24] Jupyter Notebook. Retrieved from:

https://jupyter.org/

[25] Patil, Prasad. (March, 2018) What is Exploratory Data Analysis? Retrieved from:

https://towardsdatascience.com/exploratory-data-analysis-8fc1cb20fd15

[26] Changes in Opioid Prescribing Practices. Retrieved from: https://www.cdc.gov/drugoverdose/data/prescribing.html

[27] Shmueli, G., & Koppius, O. (2011). Predictive Analytics in Information Systems Research. MIS Quarterly, 35(3), 553-572. doi:10.2307/23042796 [28] Fabian Pedregosa, Gaël Varoquaux, Alexandre Gramfort, Vincent Michel,

Bertrand Thirion, Olivier Grisel, Mathieu Blondel, Peter Prettenhofer, Ron Weiss, Vincent Dubourg, Jake Vanderplas, Alexandre Passos, David

Cournapeau, Matthieu Brucher, Matthieu Perrot, Édouard Duchesnay; Scikit- learn: Machine Learning in Python 12(Oct):2825−2830, 2011.

[29] Navlani, Avinash.(December, 2018). Decision Tree Classification in Python. Retrieved from:

https://www.datacamp.com/community/tutorials/decision-tree-classification- python

[30] Logistic Regression. Retrieved from:

https://www.statisticssolutions.com/what-is-logistic-regression/

[31] Chawla, NV, Lazarevic, A, Hall, LO. SMOTEBoost: improving prediction of the minority class in boosting. In: Proceedings of European conference on principles of data mining and knowledge discovery, Cavtat, Croatia, 22–26 September 2003, pp.107–119. Berlin, Heidelberg: Springer.

[32] Li, Susan. (September 2018). Building A Logistic Regression in Python, Step by Step. Retrieved from:

https://towardsdatascience.com/building-a-logistic-regression-in-python-step- by-step-becd4d56c9c8

[33] sklearn.feature_selection.RFE. Retrieved from:

https://scikitlearn.org/stable/modules/generated/sklearn.feature_selection.RFE .html

[34] Narkhede, Sarang. (June, 2018). Understanding AUC - ROC Curve. Retrieved from:

https://towardsdatascience.com/understanding-auc-roc-curve-68b2303cc9c5

[35] Thomas, Scott. (March, 2019). The Big List of Narcotic Drugs. Retrieved from:

Appendix