Predictive Analytics Accurately Screens for Pulmonary Hypertension

Researchers from Michigan Medicine have developed a predictive analytics tool that can help identify pulmonary arterial hypertension, a condition that causes blocked or destroyed blood vessels in the lungs.

Pulmonary arterial hypertension occurs among people with systemic sclerosis, or scleroderma, a rare autoimmune disease. Pulmonary arterial hypertension is marked by tightening of the skin that can damage internal organs, leading to heart failure and in some cases, death.

Scleroderma affects about 70,000 people in the US each year, and around ten percent of these individuals develop pulmonary hypertension. Current guidelines recommend that providers screen scleroderma patients for the condition by observing an annual echocardiogram, or ultrasound.

Although the ultrasounds are an effective tool for symptomatic patients, they don’t always accurately predict the condition in asymptomatic people or early in the disease.

“These ultrasounds miss around one in three patients who may have pulmonary arterial hypertension. And by the time we diagnose a patient so late, the story is over – the patient will likely die in the next two or three years,” said Dinesh Khanna, MBBS, MSc, senior author of the study and director of Michigan Medicine’s Scleroderma Program.

Researchers developed a two-step algorithm, called DETECT, that uses six different clinical variables to determine whether a patient requires an ultrasound of the heart. The second step then informs whether the patient should be referred for a right heart catheterization.

The results showed that the predictive analytics algorithm accurately identified all ten patients with pulmonary arterial hypertension in a study of 68 subjects.

“It didn’t miss a single patient; it can’t get better than that,” Khanna said. “This is a highly sensitive screening tool and can be very useful.”

Of the times that DETECT identified signs of pulmonary hypertension during the study, however, just 20 percent of patients who had right heart catheterizations actually suffered from the condition. But the team stated that it’s better to err on the side of caution.

“That’s the trade-off of having such a sensitive test,” Khanna said. “The right heart catheterization is invasive, but because the mortality of pulmonary arterial hypertension is so high, and the prevalence is so high, the benefits outweigh the risks.”

The algorithm outperformed standard methods used to identify the form of high blood pressure in the lungs that causes the heart to weaken and fail, potentially leading to earlier detection of the condition.

“We’ve been advocating for a long time that every scleroderma patient should be screened on an annual basis using DETECT, and this data supports that,” said Khanna. “Pulmonary arterial hypertension is a leading cause of death for these patients, and we want to diagnose them early.”

This study was the first to compare the predictive analytics algorithm to echocardiogram guidelines published in 2015. The team expects that more providers will consider using DETECT, enabling them to treat the condition earlier. Additionally, researchers anticipate that more studies will result in similar recommendations.

“I’m sure people around the globe will be doing this work and validating it,” Khanna noted. “Early diagnosis and treatment of pulmonary arterial hypertension will lead to better outcomes, including improved quality of life and survival in people with scleroderma.”

Predictive analytics tools are increasingly helping providers stay ahead of poor outcomes. A team from the Michigan Center for Integrative Research in Clinical Care (MCIRCC) and Fifth Eye recently developed a tool that can alert providers which patients are at high risk of deteriorating.

The predictive tool, called the Analytic for Hemodynamic Instability, uses real-time streaming data to help clinicians stay ahead of adverse events.

“Hemodynamic instability can happen irrespective of a patient’s condition. Whether a patient is hemorrhaging because of an injury, or a patient is septic or has some sort of infection, these kinds of things can lead to hemodynamic instability,” Ashwin Belle, PhD, former analytics architect at MCIRCC and chief analytics officer and cofounder at Fifth Eye, told HealthITAnalytics.

“Today, clinicians have very few tools in their hand to be able to see this coming. It can come on really quickly for some patients, while it takes much longer for others. To be able to see that this is happening at the onset is extremely important.”