Moving from Issue Management to Prevention with Predictive Analytics

Nov 8, 2018

You’ve probably heard of the term “big data,” which really is a euphemism for the amount of data available to us in the world today, especially healthcare. At Express Scripts, we have access to 22 petabytes of patient, physician, pharmacy, and medical data. But it’s more than just volume. It’s also velocity – receiving data at unprecedented speeds and needing to address it in a timely manner, and variety – data that comes in all types of formats.

As data scientists, we are constantly listening, interrogating that data, and asking questions to solve the problems that are impacting your health and your wallet. The way that we do that is through advanced analytic methods like: natural language processing, text mining, machine learning, linear programming, deep learning and artificial intelligence. These tools allow us to tackle three challenges:

  • Learn from previous adverse events by building advanced insights to uncover patterns that predict the likelihood of those same events happening again in the future (predictive analytics)
  • Provide those predictive, advanced insights to the right decision makers at the right time to help mitigate emerging issues and minimize or eliminate risk.
  • Combine predictive insights with prescriptive, actionable data to determine what should be done to reduce risk and improve health

By leveraging the most advanced technical platforms and methods in data science we are building a system that creates better value, better health, and better service for patients.

In order to create better value, we identify hidden relationships between data points, using advanced analytics and machine learning. An example of that is how we work to eliminate fraud, waste, and abuse from the care continuum. Through patterns in behavior, like a patient visiting multiple pharmacies or prescribers for the same medication, we’re able to intervene with a targeted outreach to mitigate negative health and financial outcomes. Prior to deploying these advanced insights and our ability to integrate data from disparate, siloed sources like prescriber and pharmacy data, this wouldn’t have been possible.

At the heart of better health is getting patients to take their medications through predictive models that allow for early detection of patients at risk of non-adherence. This leads to prevention by arming health providers with what they need to educate patients on how to stay adherent and healthy, or intervening directly with a patient before adherence becomes a problem. Machine learning also helps us identify the risk of disease diagnosis, which means that we’re able to mitigate behaviors that could lead to an illness before a patient ever steps foot down that path, thus enabling them to live healthier lives.

And finally, better service, which allows us to make something complicated like healthcare, simpler and manageable for patients. We use advanced analytic methods and tools to help  patients understand how to make optimal decisions related to their health and find trends in feedback data to continually improve member experience in a changing healthcare industry.

In all, our strategy is to dive into the complexities of data to make healthcare simpler and more effective for everyone involved in a patient’s healthcare journey. It’s through advanced analytic tools and forecasting of drug trends and costs that we’re able to create better and more affordable outcomes for our members.

comments powered by Disqus