Salt Lake City, UT -- (ReleaseWire) -- 03/18/2021 --Intermountain Healthcare is using innovative data analytic tools and algorithms to help predict whether COVID-19 patients are at risk for being admitted to the ICU, or if they're at high risk to develop COVID hyper inflammatory syndrome from the virus, or if they may benefit from monoclonal antibody treatment therapy.

Intermountain has done the latter by embedding a risk calculator into its daily report of people who have tested positive for COVID-19, so outreach can be done to connect them with the appropriate care. Currently, within 48 hours of receiving a COVID-19 positive test, caregivers can connect the high-risk patients to receive an infusion of monoclonal antibodies.

Intermountain uses a number of features such as lab criteria to stratify a patient's risk and link them to the best treatment strategies. Patients can also evaluate their own risk using a tool on the state of Utah's coronavirus web page.

"Advanced algorithms help frontline caregivers make clinical decisions based on modern machine learning methods. These artificial intelligence methods can significantly enhance the predictability of patient outcomes and how we support personalized care," said Greg Nelson, Intermountain assistant vice president of analytics services.

"This data can also be used to help caregivers like myself, determine the best clinical care practices for the patient right in front of me at the hospital," said Brandon Webb, MD, an infectious diseases physician and chair of the COVID therapeutics committee at Intermountain.

The analytic tools can also enhance equity of care and address health disparities. The decisions made by clinicians become more data-based and efficient, said Nelson.

Intermountain clinicians provide input to help Intermountain data analysts and medical informaticists design these analytic diagnostic tools. Then clinicians thoughtfully use this data, along with other more traditional diagnostic tools, and combine it with human oversight and insight to help them make the best clinical decisions for their patients.

Since the pandemic began, Intermountain has evaluated 125,000 Utahns with COVID-19. The synthesized results are used across the health system to design the best care possible for communities. The data helped identify treatments that work best early on, and this brings awareness to the treatments available for high-risk patients.

"There are misconceptions that having lots of data can help make better predictions. But what is most important is the quality of the data and the quality of questions being asked," said Dr. Webb.

The analytic tools also enable care managers to proactively reach out by phone to patients who are not in the hospital and do not have COVID-19, but might be at high risk of getting the virus and having complications, due to having other chronic conditions or comorbidities.

To date, 1,500 at-risk patients have been contacted through personal outreach. Care managers teach these individuals the importance of reducing their risk of getting the virus and encourage them to be tested if they have symptoms or have been exposed to COVID-19. They also assure them there are new treatments that can help if they get COVID-19 and have complications.

The proactive outreach intervention models are for patients in value-based care plans. Caregivers from Intermountain's population health subsidiary, Castell, make phone calls to check up on at-risk patients at home.

"The calls are done in a caring and nurturing way. Patients are educated about their risks and how to get tested and encouraged to seek care early. This high-touch method works well in these populations that have a chronic illness or tend to be older," said Nelson.

Nelson said Intermountain uses evidence-based practices and has worked for decades to bring tailored strategy of care to the bedside.

The goal is for more personalized medicine. Prior to the COVID-19 pandemic, Intermountain already had hundreds of algorithms in operation. For example, to help predict a patient's risk for C-Diff or sepsis, he noted.

Intermountain tests all the analytic tools internally and tests predictive models that external parties might have developed as well. The artificial intelligence models used are continuously re-validated and updated to ensure responsible and ethical assurance.

When monoclonal antibodies first showed promise for treating patients, it was in scarce supply. Clinicians needed to get those treatments to the patients with the highest risk for severe illness or death. Intermountain worked with other health systems in Utah to collaboratively develop a simple but accurate prediction tool to determine how likely a patient is of going on to need hospitalization or dying of COVID-19.

The tool was then adopted as part of Utah's state eligibility criteria to ensure that treatments are allocated equitably and to the patients who would benefit the most. The tool was designed to be simple to calculate, even for front line clinicians without advanced informatics. It is currently available on the Utah Coronavirus web page and is available as a pre-print scientific manuscript here.

"To identify patients at risk for COVID hyper inflammation syndrome (CHIS), Intermountain looked at criteria rheumatologists use to identify inflammatory syndromes and adapted them to use COVID-specific biomarkers. Patients were given a CHIS score," said Dr. Webb.

Details were published Dec. 1, 2020, by Dr. Webb and other researchers and physicians in the medical journal, Lancet Rheumatology. A summary of their work can be found in the Intermountain Newsroom.

About Intermountain Healthcare
Intermountain Healthcare is a not-for-profit system of 24 hospitals, 225 clinics, a Medical Group with 2,600 employed physicians and advanced practice clinicians, a health insurance company called SelectHealth, and other health services in Idaho, Utah, and Nevada. Intermountain is widely recognized as a leader in transforming healthcare by using evidence-based best practices to consistently deliver high-quality outcomes and sustainable costs.