Presentation Title

Predictive Diagnosis of Physician Overdose Relating to the US Opioid Epidemic

Faculty Mentor

Rahul Bhaskar

Start Date

17-11-2018 9:00 AM

End Date

17-11-2018 9:15 AM

Location

C301

Session

Oral 1

Type of Presentation

Oral Talk

Subject Area

interdisciplinary

Abstract

Sales of prescription Opioids in US nearly quadrupled from 1999 to 2014 but there was no change in the amount of pain Americans report. In general, prescription Opioid abusers tend to get pills from family and friends for free. Those at the greatest risk of overdosing, however, are more likely to get the drugs through their own doctor's prescription, according to the Centre for Disease control and Prevention. Identifying the doctors who have a tendency to prescribe these drugs, examining prescribing practices of physicians and proper training on addiction medicines is the key for substance abuse disorders. We find that females in Southern United States have a higher tendency to overdose and typically Orthopedic Surgeons have 13 times higher odds of prescribing Opioids than any other specialty. The insights in this research can be used to fight this public health crisis.

A Minnesota doctor was accused of illegally prescribing an Opioid painkiller to musician Prince after Prince died of Fentanyl overdose and it is a part of the painkillers driving America’s overdose and addiction epidemic. The over prescribing of pain medication at the physician level has led to an unusually large number of addiction cases. Trump administration led a national “Take Back Day” which collected 456 tons of expired and unneeded prescription drug and FDA is considering new ways to facilitate appropriate prescribing practices to cut nationwide prescription fills by one third within three years. This study is directed towards aiding these goals.

From our initial data exploration, we found that only half of medical professionals within the Medicare medical providers prescribe Opioids to patients. However, the distribution of Opioid prescriptions is not normally distributed, in fact, 2499 Doctors (10%) prescribe 70% of Opioids, and 5% of Opioids prescribed originate from 20 medical professionals (~0.08%).

Summary of research results to be presented

  • Highest risk region of Opioid use is Southern region of US and females have a high tendency to overdose.
  • All other things being equal, Orthopedic Surgeons have 13 times the odds to prescribe a high quantity of Opioids. Other practices also have higher odds include: Family Practice, Cardiology, Dentist, Psychiatry and General Surgery.
  • Based on medications linked to high dosage of Opioids, possible causation of high use of Opioids include medical issues such as, infections, blood and heart, cholesterol, nausea, and bone health.
  • The Receiver Operating Characteristic (ROC) curve is an indication of the diagnostic accuracy of our binary logistic regression model. Our training model, Selected Cases, has high a specificity of 98.5% and sensitivity of 71.7%, while our validation model has almost equally strong results with the same high specificity of 98.5% and slightly lower sensitivity of 68.3%.
  • American lives are being cut short by Opioid overdoses and there are a few states where yearly Opioid prescriptions outnumber the population. Using the data and the insights generated in the study it could play a major role in managing this public health crisis. We have used Descriptive Analytics to provide a snapshot of what is happening currently and to gain insights such as the major regions and the major specialty who overdose. Through this predictive modeling analysis, identified the physicians using a nature of drugs having the overdose risk patterns thus preventing the easy access to drugs in the first place and keep an eye on the red flags for the physician. These predictive solutions help to deliver automated warnings in the healthcare management systems which can be used for training the physicians and provide the right kind of interventions to prevent abuse.

This document is currently not available here.

Share

COinS
 
Nov 17th, 9:00 AM Nov 17th, 9:15 AM

Predictive Diagnosis of Physician Overdose Relating to the US Opioid Epidemic

C301

Sales of prescription Opioids in US nearly quadrupled from 1999 to 2014 but there was no change in the amount of pain Americans report. In general, prescription Opioid abusers tend to get pills from family and friends for free. Those at the greatest risk of overdosing, however, are more likely to get the drugs through their own doctor's prescription, according to the Centre for Disease control and Prevention. Identifying the doctors who have a tendency to prescribe these drugs, examining prescribing practices of physicians and proper training on addiction medicines is the key for substance abuse disorders. We find that females in Southern United States have a higher tendency to overdose and typically Orthopedic Surgeons have 13 times higher odds of prescribing Opioids than any other specialty. The insights in this research can be used to fight this public health crisis.

A Minnesota doctor was accused of illegally prescribing an Opioid painkiller to musician Prince after Prince died of Fentanyl overdose and it is a part of the painkillers driving America’s overdose and addiction epidemic. The over prescribing of pain medication at the physician level has led to an unusually large number of addiction cases. Trump administration led a national “Take Back Day” which collected 456 tons of expired and unneeded prescription drug and FDA is considering new ways to facilitate appropriate prescribing practices to cut nationwide prescription fills by one third within three years. This study is directed towards aiding these goals.

From our initial data exploration, we found that only half of medical professionals within the Medicare medical providers prescribe Opioids to patients. However, the distribution of Opioid prescriptions is not normally distributed, in fact, 2499 Doctors (10%) prescribe 70% of Opioids, and 5% of Opioids prescribed originate from 20 medical professionals (~0.08%).