This blog post was co-authored by Slawek Kierner, SVP of Enterprise Data & Analytics, Humana, and Tie-Yan Liu, Assistant Managing Director, Microsoft Research China.
Using AI models to influence the real world impact
There are trips to the hospital. And while everyone in the industry strives to provide world-class care for inpatient experiences, everyone – patients and care teams alike – would prefer to avoid these hospitalizations. Humana teams believed they had enough data to investigate the ability to proactively identify when patients were heading for a high-risk event, and they determined Microsoft cloud for healthcare and AI technology to the test.
Humana’s questions were simple: How do we take the data we have today and use it proactively? How do we use AI to identify signals in our existing ecosystem that tell us that someone may be experiencing a scenario that will endanger them? And most importantly, how do we proactively get involved to meet our members in their own environment before they land in the emergency room?
The first approach to chronic patient monitoring often focuses on remote patient and IoT device monitoring, but to address this challenge we wanted to take a different and much broader approach with AI. By combining clinical data, key event triggers that could indicate a patient’s health condition deteriorating, and a combination of predictive models, the data science teams at Microsoft Research and Humana worked together on research to investigate whether it was a system that would identify potential gaps in care between patients and high-risk patients, engage with care teams who could reach out and offer support.
The power of AI model refinement
The result of the research was a look into the future of AI in healthcare. Health organizations like Humana have spent the past few years developing powerful predictive models with a single focus. Humana had existing models that predicted the likelihood of hospitalization in the near future for its 4.9 million Humana Medicare Advantage members, as well as additional models that predicted the cost of treatment and the likelihood of readmissions. Microsoft Research and Humana data science teams merged these models with structured data to create and test a combination of neural networks and tree-based models using Microsoft cloud technologies.
Cloud-scale tools were critical in developing the multivariable model as well as the technology in the Microsoft Cloud for Healthcare to unify the diversity of patient data streams. In addition, Microsoft Research has developed an advanced, deep learning-based sequential modeling approach to capture the dynamics of health status, which is critical to accurately predicting the likelihood of resumption. To further increase the robustness of the research model it has learned, Microsoft Research has developed resampling techniques at its own pace to address the sample imbalance challenge in this resumption prediction scenario. Research showed that integrating all of these technologies improved the precision of the model by over 20 percent. Most importantly, the advanced models with anonymized data are designed to protect patient data.
Empower care teams to help patients when they need it most
“Model accuracy is critical here in identifying members at risk,” said Mike Hardbarger, director of data science at Humana and a research contributor on the project. “Our members deserve individual, proactive support. By using this model in conjunction with others, not only can we help them avoid hospital readmission, but the nursing teams can have the data they need to follow a custom plan. ”From effective prescription management to Combating food insecurity, a care manager can then work directly with the member to initiate the next best action.
Proactive problem solving like this relies on collaboration and innovation. Deep learning enabled research teams, including Sean Ma, Lead Data Scientist at Humana, to have a comprehensive framework for both scientific and industrial considerations. “Working directly with the algorithm authors has accelerated progress considerably. I’m excited about what’s to come, ”says Ma.
Get more out of your data with Microsoft Cloud for Healthcare
This research project is only one step in the evolution of the Humana Analytics Engine. The improvements will continue over time as additional research is carried out and the model will continue to be validated.
Learn more about Microsoft cloud for healthcare.