Healthcare organizations have no shortage of data about their patients, not just from electronic health records (EHRs) but also from remote patient monitoring. But until now, it's been difficult for IT and clinical leaders to turn that information into real-time insights to improve patient care and reduce the cost of care. That could be changing, according to speakers at .conf24, Splunk's annual Las Vegas conference this year.
“The world is living in an era where artificial intelligence and machine learning are playing a vital role in solving healthcare challenges and delivering better care,” said Brett Roberts, global partner technical manager, Splunk.
Alan Peaty, senior partner solutions architect at Amazon Web Services, said Splunk offers “fantastic” out-of-the-box capabilities that allow users to gain insights without needing machine learning expertise.
But for healthcare organizations looking to build their own custom ML models, he says it makes sense to integrate Splunk data with other information in AWS. “This is really about getting data and insights to the people who can act on it,” Peaty says.
Discover: Optimizing clinical workflow improves patient outcomes.
Additional features required by healthcare organizations
For the live demo, Roberts and Peaty imported simulated data from patient wearable devices into AWS from their Splunk environment and integrated it with data from EHRs, customer relationship management platforms, and other sources. They integrated the data in Amazon SageMaker Canvas, AWS's no-code ML interface for business analytics. They then formatted the data using AWS Glue, a serverless data integration tool. Finally, Roberts and Peaty used Amazon Athena, a managed SQL engine, to run queries across 183,000 patient records looking for associations between wearable tracker data and hospital readmission rates.
The results were enlightening and almost instantaneous.
“We can see that age is impacting whether a patient is admitted to hospital,” Peaty says. “That makes sense. But the other feature that Splunk gave us is around uptime on wearables, which also has an impact. So now we can see how, as healthcare providers, these models can help us personalize patient care and intervene proactively.”
Please click on the banner below Learn how modern data analytics programs can optimize care.
Putting patients at the center of data-driven decision making
Hospitals have financial and clinical incentives to reduce patient readmissions. In particular, the Centers for Medicare & Medicaid Services' Hospital Readmission Reduction Program ties payment to a healthcare organization's ability to reduce readmission rates by improving communication and care coordination efforts. Predictive models that help hospitals identify patients at high risk of readmission could enable providers to reach out to patients before they return to the hospital, potentially reducing avoidable readmissions and preventing poor health outcomes.
Peaty also showed how the ML model can generate predictions for specific patients. In the example of an 89-year-old patient who didn't use wearable devices much, the model showed that the patient was more likely to be readmitted to hospital. “So, as a healthcare organization, we're going to reach out to the patient and make sure we have an intervention,” he said.
In an interview Health TechRoberts said the demo represents “the art of the possible” rather than reflecting currently widespread use cases, but he said he hopes it will serve as inspiration for healthcare organizations looking to increase their use of predictive analytics to solve real-world problems.
He emphasized the importance of including clinicians both in the design of ML models and in the decision-making based on the predictions of these models.
“You need to have a high degree of accuracy,” Roberts says. “Training is key. For use cases like these, where it's potentially life or death, you really need to work hard to have both the data scientists and the people with the domain expertise.”
“It's a team sport,” he added. “You have to include the human element.”
