Illuminating Your Healthcare AI Journey, Part 4: Taking Action on Your AI Journey

AI is your power source for the future — as an organization dedicated to healthcare, we want you to make it to your best destination. In this series, Lumiata walks you through the AI Journey so your organization can: 

  • Build an AI-enabled strategy (not just acquire new technology)
  • Focus your AI investments
  • Streamline decision-making processes
  • Plot a path forward into the future
  • Generate material impact on your business

This series will prepare you to mature through the AI Journey curve, understanding what’s happened with your data, why it happened, predicting what will happen in the future, and then optimizing your actions and decisions. See the other blogs in this series:

Part 1: Assessing Your Starting Point 

Part 2: Sharpening the Tip of Your Strategy Spear

Part 3: Build vs. Buy in Healthcare AI

Healthcare tends to lag in tech adoption, and this can work in your favor as you progress in your AI Journey. By looking at general AI trends, you can find guidance along your own path to maturity. For example, Deloitte’s State of AI in the Enterprise is a clue to where adoption of automation in healthcare is headed in the near future. 

“In the firm’s previous study, published in 2018, 57% of respondents said the technology would transform their company within the next three years. Now, 75% of adopters say AI is set to transform their company in the same time frame.” 

The translation? AI isn’t a trend. It’s a requirement for any business that wants to keep up with the rest of the field. The benefits include discovering what your data can really do, and, if you’re working with the right partner, the ability to focus more deeply on your core competencies — all while meeting your markets’ most pressing needs. 

Let’s take a look at what your first steps can look like.

How can we access early wins with AI? 

AI holds infinite potential, even in your initial application. 

Most organizations will benefit from a “minimum viable” approach to achieve early wins and ramp up inertia around AI initiatives. Here are a few examples of models we’ve created:

  • Diabetes Care Management: Identifying high-risk, undiagnosed members or patients to funnel them into care management programs. 
  • Avoidable Hospitalizations: Discovering Medicare members who are at risk for hospitalization to inform work with clinical teams and prevent costly and dangerous admissions.
  • COVID-19 Control: Pinpointing members with chronic conditions that would be good candidates for home care to keep them out of facilities and reduce COVID-19 exposure. 
  • High-Risk Care Management: Detecting members who might develop comorbidities and direct them into existing programs. 

Meet your models.

In your new approach to AI, it’s critical to start with the right models. We’re proud to build intelligible, transparent models that generate clear predictions — predictions that can be easily explained to users at all levels of data science proficiency. 

But however you’re applying the technology, models should be built off a large data set. Lumiata’s models, for example, are pre-learned on over 100 million patient records, 35,000 physician curation hours to inform our disease codes and over 50 million published articles. These models are applied to your data and calibrated for your population — all to keep you moving smoothly along your AI Journey. 

Our models outperform large national health plan legacy methods by 20% (which translates to $18-$20 pmpm). In the high-cost claimant space, our models beat Wakely by up to 80% and Cotiviti by up to 60%.

And our cost prediction models outperformed the DHHS hierarchical condition category risk adjustment model by more than 200% (using claims records of over 20 million patients).

Right now, these models are enabling insurance companies and providers across the industry to get in front of their most challenging predictive needs — and because of our exclusive focus on healthcare, they’re doing so in a way that evolves with them over time. 

A Lumiata Case Study

As an example, let’s look at an insurance company in the Northeast that was working on enhancing its pricing strategy in a highly competitive market. Their incumbent vendor wasn’t providing the precision needed to accurately predict individual and group level costs. 

By working with Lumiatas models, they were able to identify twice the number of high-cost members as their previous methods, with accuracy that was almost 40% higher — seeing results after only a little over two months of kicking off the relationship. 

The AI decisions you make now are the first steps toward earlier interventions, confident and decisive leaders that react with increasing precision, and helping the members and patients you impact lead healthier lives through earlier interventions.  We would love the honor of accompanying you on your journey. To get a glimpse of what the path will look like for you, we want to invite you to take a tour of our Lightning Model Builder and Spectrum Model Catalog with a personalized demo.