Intermountain Clinicians Benefit From Self-Serve Analytics Software

Spur-of-the-moment care and research is enabled by posing questions in plain English.

Clinicians and researchers are benefitting from new self-serve analytics software in use at Intermountain Healthcare.

The Utah-based system, which operates 24 hospitals and 215 clinics in Utah, Idaho, and Nevada, has been using software from MDClone to investigate everything from cardiology to genetics to hyperbaric medicine, according to officials at the health system.

The software taps Intermountain's electronic health record (EHR) system to perform English-like queries on the 3 million lives of current patients, but also on an enterprise data warehouse on Intermountain patients dating back 25 years, says Jeffrey L. Anderson, MD, a clinician and research physician at Intermountain, and immediate past director of cardiovascular research at the Intermountain Medical Center Heart Institute.

The MDClone software tools are also being applied to Intermountain's HerediGene™ population study, which the healthcare system says is the largest single-population genome study in the world, Anderson says.

The population study is being undertaken with deCODE genetics, a biopharmaceutical company based in Ireland.

"We're being courted by many companies, to see if we have the patients they would like to test with their new therapeutics," Anderson says. "So we need to answer questionnaires. And what we need to do is find out really who we have in the Intermountain system and within a certain distance that can come in and potentially participate in these intervention trials."

The prior way to perform these queries relied on the expertise of Intermountain data analysts, whose very expertise made them so popular, researchers had to "get in line, and wait weeks, even months to have an analysis done," Anderson says.

The new analytics tool allows untrained investigators to pose sophisticated queries without knowledge of the structure of the underlying databases and tables where the data is stored.

Previously, such queries would require sophisticated knowledge about table structures, plus a working knowledge of Structured Query Language (SQL), says Viet Le, a researcher and physician assistant in cardiology at Intermountain, who studies mobile health technologies and cardiovascular genetics.

"MDClone is democratizing our ability to query quickly, to allow us as clinicians to ask very important questions at the point of care," Le says.

In certain cases, in the previous arrangement, researchers could forget the original impetus for the question they posted to expert analysts in the gap of time between submitting the question and receiving an answer, Le says.

This level of ad-hoc query is particularly useful to formerly frustrated researchers at Intermountain, such as the hyperbaric medicine department.

Kayla Deru, who started in the department as a secretary, became a department researcher, and developed research skills to the extent that she recently joined the Intermountain Connect Care team, Intermountain's telehealth division.

The hyperbaric department "never had analytical support, ever," Deru says. "I'm not a statistician, or a computer programmer. These intellectual curiosity questions, or even patient care questions, tend to be fairly time-sensitive. And you don't want to wait for six months, a year, or never, to find out the answer to your question."

The technology works by creating a history of each patient's interaction with the health system, not limited to EHR data, but also including admissions, discharges, transfers, genomics data, claims data, and administrative data. This data is organized as events on patient timelines, making it easier for would-be analysts to query. Natural language processing (NLP) brings in unstructured notes from physicians.

This timeline approach makes it much easier for researchers to search for and find data related to readmissions, queries that would otherwise have required intimate knowledge of underlying data structures, according to Jon D. Morrow, MD, senior vice president and physician executive at MDClone.

One area of inquiry pursued by Deru allowed clinicians to look at connections between carbon monoxide poisoning and blood clots, and on-the-fly results from queries help accelerate the presentation of research papers, Deru says.

In another example, Le used the ad-hoc queries to investigate histories of elevated lipids on relatively young patients coming through Intermountain's cath lab. "I would prefer we catch these individuals before they have their first [adverse] event," Le says.

This ability to quickly ask and get answers back also is helping Intermountain to rapidly respond to opportunities to host clinical trials, and thus to become one of the early places where a particular breakthrough treatment becomes available.

"It seriously has helped our department and other departments in our system enormously well" in this regard, says Lindell Weaver, MD, medical director of the Intermountain Hyperbaric Medical Center at Intermountain's LDS Hospital.

Another area where the queries of Deru and other researchers aided clinicians was focusing on the importance of blood pressure control and renal function to patients with diabetes and a risk of foot ulcers and amputations, Weaver says.

At the Heart Institute, another set of ad-hoc queries zeroed in on patients with hypertriglyceridemia to examine how many had experienced pancreatitis and found that half the patients had done so. "Those are numbers we could not have gotten otherwise," Anderson says.

Finding this connection in the traditional way that would have allowed the Heart Institute to qualify as a candidate site to offer a new therapeutic would have taken "weeks or months to do," Anderson says.

"These trials come along all the time in premier medical centers and medical systems," Anderson says. "That's just part of our everyday existence to try to be on the cutting edge when new therapeutics come along."

Ultimately, tools like this will remove the kind of monopoly on data analysis that more highly skilled but short-staffed data analyst staffs have enjoyed until now, according to officials at MDClone.