Pinpointing Root Causes and Trends in Opioid Abuse
25 Jan

Pinpointing Root Causes and Trends in Opioid Abuse

– Jon Hoehne, Vice President Quality Assurance

In response to the national opioid crisis, the Department of Health and Human Services (HHS) tried a new approach to determine better ways to use the large amounts of data available to predict and prevent opioid abuse. HHS’s Opioid Symposium and Code-a-Thon brought together people from federal and state government, academia, and the private sector to examine ways that better tools for data sharing can help public health efforts around opioid usage, treatment for opioid use disorder, and supplying prevention resources.

The event was a first-of-its-kind for the agency bringing more than 300 citizen coders from around the country physically into the agency. For the event, more than 100 datasets from HHS, other federal agencies, state and local partners, and the private sector were opened for use.

Dovel was proud to be among the 50 teams working on the 24-hour event. Our solution was based on Amazon Web Services (AWS) SageMaker. It integrated datasets from Indiana on opioid abuse, census.gov population stats, and data from the CDC on opioid prescribing rates.

AWS SageMaker is a new service that has the potential to greatly simplify and reduce the cost of machine learning. SageMaker provides all the components needed to build, train, and deploy machine learning models. The role of SageMaker in the Dovel Code-a-Thon prototype was to provide the backend data analysis that could be simplified and displayed to state and local officials researching ways to address opioid problems in their community. For example, what specific factors may be playing a greater role in my community? Poverty levels? Median age? The figure below illustrates the Dovel solution of simplifying complex analysis to a mobile friendly output.

Simplifying complex analysis to a mobile friendly output

 

Our objective is to re-use these components for upcoming opportunities and continue to provide solutions for our clients with machine learning.

This is the kind of innovation experience we seek to provide more of across the company, whether it is participating in a broad government-led initiative, supporting a specific Dovel customer, or focused on internal business processes.