Blue Cross Blue Shield: Blue Cross Blue Shield predicts individual propensity for opioid abuse with 85% accuracy to modify insurance pricing and support appropriate interventions using machine learning analytics | Case Study

Blue Cross Blue Shield predicts individual propensity for opioid abuse with 85% accuracy to modify insurance pricing and support appropriate interventions using machine learning analytics

Blue Cross Blue Shield maps risk factors associated with the potential of opioid abuse among patients. It is using data such as demographic, location and frequency of pharmacy claims data to build models to predict individual propensity of abuse. They can then stage interventions or use nudges to prevent the behaviour. It is using machine learning predictive analytics.

Context

"More than 60 percent of drug overdoses in the United States involve opioids, and healthcare costs associated with abuse exceed $25 billion a year. "

The Project

"Using Machine Learning, Fuzzy Logix is developing an actionable predictive model that identifies likely opioid abuse before it occurs. The health plan is using DB Lytix analytics software on Teradata to develop a holistic view of a member’s health based on past medical and pharmacy utilization, location data and demographic information. Findings help guide outreach efforts more sensitively and effectively. To build the models, Fuzzy Logix consultants worked with the plan’s analysts and data scientists to perform predictive and advanced statistical analysis on Teradata using fast and scalable in-database analytics."

Data

Past medical and pharmacy utilization, location data and demographic information.

Results

The company claims ~85% accuracy in predicting risk.

Related Use Cases

Use Case

Predict personalised health outcomes to recommend individual treatment approach

→

Use Case

Predict prescription addiction and abuse

→

Use Case

Personalise messaging, incentives and media used to improve wellness and prescription adherence

→

←Back to Case Studies