How can healthcare organizations use self-service analytics platforms like MDClone’s ADAMS to gain new insights that lead to better patient care?
In a recent webinar, “Empowered Self-Service Insights to Inform Patient Care,” Viet Le, researcher and physician associate in cardiology at Intermountain Healthcare, discusses how organizations can leverage data in their healthcare systems to quickly and easily discover at-risk clinical populations that physicians can target with new resources.
In February 2019, Le attended a National Lipid Association conference where he met with one of Intermountain’s partners from Salt Lake and decided to build an advanced lipid clinic. This clinic would help physicians determine which parents of children with familial hypercholesterolemia (FH) also had the disease. Patients with FH have significantly elevated LDL cholesterol, or “bad cholesterol,” and an increased risk of having coronary artery disease at a younger age. Around the same time, Viet was introduced to MDClone by Peter Calderone, Director of Customer Success at MDClone.
“Having a self-service analytics platform can allow you the flexibility to extend your journey beyond a ticketing queue or a queue for a data analyst to write a SQL query into empowering you with your own self-service strategy and the ability to ask questions at your speed.”
Le said that when navigating an electronic medical record, physicians must look at labs, medications, medication dosages, and other information to understand an individual patient’s journey. However, clinicians may not always have time to do a deep dive into this data.
“I can do this on a per-patient basis if I have time,” Le said. “I can do this if I’m not double- or triple- or quadruple-booked. This requires intention. So why not have everyone who has great thoughts and sees these clinical scenarios ask a question and be able to delve into it and really get some answers?”
Le set out to transform lipid management by leveraging guidelines that were already in place, determining best practices, asking new questions, and discovering patients and opportunities for improvement. Le wanted to understand the magnitude of underlying health risk and the potential for future poor health outcomes. From there, the team could inform the way clinicians find patients that meet those risk factors and then stratify those patients by risk factors.
Currently, half of men with FH that are untreated will have a heart attack, an 8-fold increase from the general population. Additionally, women with FH have a 6-fold increase in risk of heart attack. Le said that if a clinician can identify these individuals, then they may be able to treat them and reduce or normalize that risk. Le and team have been able to hardware guideline therapy into patient-based care and predict risk and probability of adverse events by risk factors, as well as community-based or social determinants and data previously locked away in unstructured text.
Recently Le has used MDClone’s natural language processing platform, NLP Studio, to run an NLP model, pull coronary calcium scores for a population in question, and leverage that from the literature as a risk-stratifying component to the population.
“We were able to look at and prioritize the pieces of information that led us to the greatest amounts of risk and then build some statistical models and categorizations around the patients based on a milieu of clinical characteristics and biomarkers of interest,” Le said. “And from that, we were able to look at the highest pieces of information of predictable value and group those into categories.”
Le emphasized that one thing he enjoys about the MDClone ADAMS Platform is the ability to visualize data easily and quickly.
“You can look at data, unstructured or structured data, and then through the platform, there are visualizations.”
“I’m a very visual person. If you give me SQL, I kind of glaze over. But as these reports come back and the graphs come together, you really get that gestalt, because it’s made for you with these visualizations,” said Le.
When it comes to using MDClone’s ADAMS Platform, Le emphasized that physicians should think about the clinical questions they want to query and start as broad as possible.
“Don’t get tunnel vision, because you’ll interpret the data without the full context,” Le said. “You may be missing some features.”
Today, Le aims to visualize risk for patients with LDL cholesterol greater than or equal to 160 — a main risk factor for young people with FH — and build upon that.
“Perhaps at some point we will develop alerts that are prescriptive as well that will help our primary care providers,” Le said. “When you can monitor and identify a gap, you can do something about it. There are gaps where new therapies are available. This platform allows us to look at specific high-risk populations and when we fix the gaps, see what’s left. Then we can approach these patients and offer them the ability to be enrolled in a clinical trial, for example.”
Learn More
Intermountain Healthcare is a recognized leader in transforming healthcare through innovation and a patient-centric approach.
