PLOS Digital Health: AI identifies individuals at risk for heart disease complications

Scientists at the University of Utah School of Health have demonstrated for the first time that artificial intelligence can better predict the onset and course of cardiovascular disease

The researchers say this comprehensive approach could help doctors predict, prevent or treat serious heart problems, perhaps even before patients become aware of the underlying disease

Although the study focused only on cardiovascular disease, the researchers believe it may have broader implications

"We can bring artificial intelligence to help refine the risk of virtually any medical diagnosis," said Martin Tristani-Firouzi, corresponding author of the medical study and a pediatric cardiologist at U of Health and Intermountain Children's Hospital, Nora and scientist Ike Harrison Cardiovascular Research and Training Institute

The research was published in the online journal PLOS Digital Health

Current methods for calculating the combined effects of population and medical history of various risk factors for cardiovascular disease are often imprecise and subjective, according to Dr.

To more accurately measure how these interactions, also known as comorbidities, affect health, Tristani-Firouzi, Yandell, and colleagues from U of U health and Intermountain Junior Hospital used machine learning software to Classify more than 1.

These electronic records document everything that happens to patients, including laboratory tests, diagnoses, medication use and medical procedures, helping researchers identify comorbidities that are most likely to exacerbate specific conditions, such as cardiovascular disease

In their current study, the researchers used a type of artificial intelligence called a probabilistic graph network (PGM) to calculate how any combination of these comorbidities might affect a relationship with heart transplant, congenital heart disease, or sinus node dysfunction (SND) associated risks

In adults, the researchers found:

• Individuals with a previous diagnosis of cardiomyopathy (disease of the heart muscle) have an 86-fold higher risk of needing a heart transplant than those without

  
  

• Those with viral myocarditis had a 60-fold higher risk of needing a heart transplant

  
  

• Using milrinone, a vasodilator drug used to treat heart failure, increased the risk of transplantation by 175 times

  
  

In some cases, the risk of a merger is even greater

In children, comorbidities have significantly different effects on transplant risk, Tristani-Firouzi said

The researchers also looked at the impact of the mother's health during pregnancy on the child

Babies who underwent Fontan surgery, a procedure to treat a congenital defect in blood flow to the heart, were 20 times more likely to develop SND arrhythmias than those who did not require surgery

The researchers also found important demographic differences
.
For example, Hispanic patients with atrial fibrillation (rapid heartbeat) had twice the risk of SND as blacks and whites with a similar medical history
.

Josh Bonkowsky, MD, director of the Personalized Medicine Center for Primary Children, who was not an author of the study, believes the research could lead to the development of practical clinical tools for patient care
.

"This new technology shows that we can precisely estimate the risk of medical complications and even determine which drug is better for an individual patient
," Bonkowsky said
.

Going forward, Tristani-Firouzi and Yandell hope their research will also help doctors unravel the growing web of confusing medical information that wraps around them every day
.

"In this day and age, no matter how awake you are, as a medical professional, you can't possibly have all the knowledge you need in your head to treat a patient in the best possible way," Yandell said
.
"We're developing The computers of this era will help physicians make the best possible patient care decisions using all the relevant information available in the electronic age
.
These machines are critical to the future of medicine
.
"

article title

An explainable artificial intelligence approach for predicting cardiovascular outcomes using electronic health records