How Geneia data scientists and clinicians collaborate to create and hone predictive models
Data Science

Geneia Conversations: Health Plans, AI and Chronic Disease

May 5, 2021
A data scientist and clinician discuss the hypertension complications model
Vice President, Marketing


At Geneia, our data scientists and clinicians collaborate to create and hone models to predict chronic disease complications. In this episode of the Geneia podcast, Jasmine McCammon, lead principal data scientists, and Natalie Benner, principal clinical transformation consultant, discuss their work on the Geneia Data Intelligence Lab’s hypertension complications model.

In Jasmine’s words,

“When I first thought about complications related to hypertension, my mind immediately jumped to things like strokes, heart attacks and kidney failure. But the reward of working with Natalie and other clinicians is that they pointed out that there are a lot of other signs and symptoms of exacerbating hypertension before you get to these devastating consequences…The clinical team helped us identify different stages of severity across different target organs.”

As the direct result of Geneia’s robust data science – clinical collaboration, the hypertension complications model predicts, for people with hypertension, the probability of a complication event in the next 12 months across three stages. The complications steadily increase in severity. Stage 1 is the least severe, e.g. protein in the urine, retinopathy or an enlarged heart. As Jasmine notes, the stage 1 complications ‘are not often associated with hypertension and, absent clinical input, could have been initially overlooked by the data science team.’ For stage 3, the complications are very severe, e.g. end stage renal disease, heart failure, stroke and more.

Hypertension complication stages

The bottom line, says Jasmine, is:

“It’s a two-way street. Natalie’s team helps ensure the clinical validity of our models through variety of strategies, and my team helps her understand how the models are working under the hood, which is important for their buy-in. If clinicians don’t trust the models, why would they use them? That trust feeds into them making the models actionable through population health management strategies. Designing AI models with the end goal in mind is what makes our products work.

Chronic Disease Care: Essential AI for Health Plans

Jasmine and Natalie also talk about Geneia’s new white paper, Chronic Disease Care: Essential AI for Health Plans. They answer questions such as:

  • How can a health plan use the hypertension complications model?
  • How big of a problem is chronic illness for health plans?
  • How has the COVID-19 pandemic exacerbated the challenge of chronic disease?
  • Ten years from now, how do you think the management of chronic disease will have changed?

I encourage you to listen now.