Leveraging AI for COVID-19 Outreach, Population Health Management

March 24, 2020 - In the time since the COVID-19 outbreak has escalated to a full-blown global pandemic, buzzwords like artificial intelligence, population health management, and the social determinants of health have become vital defenses in the fight against infection.

For more coronavirus updates, visit our resource page, updated twice daily by Xtelligent Healthcare Media.

Across the healthcare industry, organizations are using their analytics and big data resources to track the spread of the virus, monitor the use of resources, and identify individuals who may be particularly vulnerable in the midst of a health crisis.

Patients who experience food insecurity, a lack of transportation access, or social isolation often suffer from worse health outcomes, but when faced with a threat like coronavirus, it becomes even more critical to mitigate the downstream effects of these social factors.

At Medical Home Network (MHN), leaders are aiming to prioritize care for vulnerable patients. The Medicaid accountable care organization is using an artificial intelligence platform to identify individuals who have a heightened risk of developing severe complications from COVID-19.

READ MORE: Is Artificial Intelligence Mature Enough to Combat COVID-19?

“MHN is composed of ten federally qualified health centers in Chicago and three health systems. Altogether, we’re responsible for the care management of 122,000 Medicaid beneficiaries,” Art Jones, MD, chief medical officer of MHN, told HealthITAnalytics.com.

“We've been using the AI platform for a couple of years to identify who our high-risk members are for case management, as well as anticipate which of our hospitalized members are at risk for rehospitalization.”

In the case of COVID-19, MHN is leveraging AI to identify patients at high risk of experiencing severe respiratory infections or respiratory failure, a particularly vulnerable group of people.

“We used machine learning to identify which of our patients had a high risk of admission for COVID, or for unrelated complications from respiratory ailments,” Jones said.

“The results showed that 4.4 percent of our patients would represent about half of those patients at risk. So instead of calling all 122,000 of our members, we can focus our initial outreach on that 4.4 percent of our adult population. That’s what we’re doing now.”

In addition to respiratory issues, MHN care managers are basing their outreach efforts on a less obvious, non-clinical risk factor: Social isolation.

“With COVID-19, there's a lot of talk about social distancing, which is to protect people from getting the virus. But there's also the concern that if socially isolated patients get sick, do they have nearby family or friends to support them?” Jones said.

To determine which patients are experiencing social isolation, MHN care managers ask their members if they are homeless, if they live alone, and if they have people to turn to for help if they get sick.  

“After we cross those two lists of patients – those patients who are at risk for being admitted for respiratory failure or COVID-19, and those patients who are socially isolated – we know who to reach out to first,” Jones said.

“And once we get our care managers reaching out to those patients, then we would reach out to the next group of patients, which is composed of individuals who don't necessarily have those social risk factors but are at high risk for hospitalization or respiratory infection.”

While care managers are principally aiming to check on people’s well-being, these health assessments will also involve a good amount of patient education about coronavirus, Jones noted.

“We want to make sure people know what they can do to reduce their risk of getting an infection in the first place. We help them identify what symptoms they should be looking for, as well as what they should do if they develop those symptoms,” he said.

With this approach, the MHN team is working to get ahead of the care access concerns that have threatened the US healthcare system since the outbreak of COVID-19. The pandemic has raised serious questions about hospital staffing, bed capacity, and levels of preparedness across the nation.

By educating patients about where and when they should seek medical attention, providers and care managers can help lessen the anticipated strain on healthcare organizations.

“What we don't want is for everyone to go to the emergency room as soon as they develop symptoms. Because while 80 percent of people can deal with the virus at home, about 20 percent of people that get the virus may end up having to be hospitalized,” Jones said.

“We want to help patients know that when they develop these symptoms, they’re able to call their care manager or their primary care practice. They have a place to turn.”

Using AI and machine learning, MHN has been able to identify specific groups of high-risk patients, which will lead to better long-term health outcomes.

“There are broad categories of people that are at risk for respiratory complications from COVID-19, including the elderly, people with diabetes, and people with heart and lung problems. The problem is that these individuals make up most of the regular Medicare population, and we have limited resources. In care management, you want to be able to focus your outreach on those people that will be most impacted,” Jones said.

“What AI allows you to do is to be able to evaluate multiple items, and multiple contributing factors. It allows you to see patients who have a history of consistent hospitalizations, or those who have a history of not taking their medication. With this information, you can target specific groups instead of having to reach out to everyone with diabetes or everyone with heart problems.”

However, MHN’s technology platforms aren’t the only tools that will help patients during the COVID-19 pandemic, Jones stated.

“Typically, within managed care, the care management is done by the health plan instead of the practice. Because our care managers run the practice, we’ve found that they have been able to build trusting relationships with our high-risk members. It also means that patients are much more likely to answer the phone, because the call is coming from the care managers,” he said.

“That contact has been important, as well as the fact that our system screens for individuals’ social determinants of health. Because we’ve been systematically collecting that data, we were really prepared for something like this.”