Throughout the last year, in addition to the millions of individuals who contracted the COVID-19 virus in the U.S. and needed medical support, patient care leaders were facing another, parallel issue: the number of people who were delaying other care, not related to the virus, was also increasing at alarming rates.
According to the CDC, at least 41 percent of U.S. adults have delayed or avoided necessary medical care to prevent infection from COVID-19. Without proactive interventions, many patients who deferred care will deteriorate to more advanced disease states, end up admitted to the hospital, or develop new diseases that go undetected in their early stages, demanding more costly care in the future. All told, deferred care could increase annual U.S. healthcare costs by $30 to $65 billion, according to a McKinsey analysis from last fall.
For healthcare organizations on the leading edge, being proactive with identifying patients who are deferring important visits with physicians as a result of the pandemic, and then intervening when necessary, will go a long way. To that end, Community Health Network, an integrated health system with over 200 locations throughout Central Indiana, is using artificial intelligence (AI) to target outreach to accountable care organization (ACO) members who are likely to suffer negative health outcomes—such as ER visits and avoidable hospitalizations—from putting off care.
Patrick McGill, M.D., executive vice president and chief analytics officer at Community Health Network, says that early on in the pandemic, health system leaders realized that different populations of patients are affected differently, and there’s a certain group of folks who were very hesitant to come back into the healthcare system. “We needed a way to not only identify the patients, but also better understand what methods we should leverage to connect with those patients to let them know that it’s safe to come back, and also specifically provide outreach to try to engage them back with the health system,” he says, adding that “knowing which patients are at risk of deteriorating isn’t always easy.”
That’s when Community Health Network turned to Jvion, an artificial intelligence solutions company, whose CORE AI technology goes beyond predictive analytics and machine learning to identify patients on a trajectory to becoming high-risk by accounting for social determinants of health (SDOH) and lifestyle factors to reveal hidden patient risk, according to officials. They add that CORE also determines whether that risk can be modified, and recommends the interventions and outreach channels most likely to be effective in reducing risk.
The initiative, which McGill says is still in the early phases and should ramp up in the second quarter of 2021, will entail leveraging Jvion’s AI platform to analyze the medical records of Community Health Network’s members. John Frownfelter, M.D., chief medical officer at Jvion, notes that the CORE AI platform applies machine learning algorithms to both structured and unstructured data that comes from healthcare organizations themselves, but also from other publicly available sources that Jvion collects.
“We start at the individual level with a picture of who that patient is from a socioeconomic standpoint, while also getting the clinical data from Community Health Network,” Frownfelter explains. “Those social elements include everything from food insecurity to access to transportation, to how digitally fluent [someone is]. Access to the internet is one of those social determinants, for example, that is driving health inequities today. The combination of those things is really what drive risk for an individual patient,” he says.
Frownfelter adds that understanding the context within those SDOH elements is also very important. For example, he offers, you might have a COPD patient who has moved to a city to be closer to his or her grandkids, but that city also has poor air quality. That’s a very different scenario than another person who just moved somewhere because he or she is retiring, is flush with finances, and is going to a retirement community. “On a scale, if you just use an index around social determinants, short-lived housing equals housing instability, but you also need to understand the interactions around those features to comprehend what's driving risk,” he notes.
McGill adds that one of the things Community Health Network is going to specifically focus on will be patients who were overdue for screening gaps in care. For some conditions, such as diabetes, hypertension, and certain other chronic medical conditions, Community Health Network already does have tools in place in its Epic EHR in to address this, though it’s not at the same level as what Jvion provides. But leveraging Jvion’s AI platform can take that next step, and help provider leaders figure out how to engage with patients who might be overdue for other preventative screenings such as mammograms or colonoscopies.
For example, if Mrs. Smith is overdue for a mammogram, a provider organization might send her a letter reminding her to make an appointment. “That's not really doing enough,” McGill admits. But from the Jvion platform, he notes, “We can [learn] that Mrs. Smith might have trouble with transportation and might not be able to afford a mammogram. Yes, she’s overdue, but she struggles with these other things, so getting a mammogram is not at the top of her priority list. So we may need to engage her with a resource coordinator or provide transportation for her to get the mammogram. Essentially, if we can address the food insecurity, she might be more likely to get her mammogram.”
What’s more, McGill says that while there have been a lot of conversations around the incidence of cancer, he believes people need to understand that the incidence of cancer isn't going to change because of the pandemic. “People will develop cancer at the same rate that they that they historically would, but what's probably going to happen—or what we're trying to prevent happening—is people showing up with cancer at more advanced or later stages because they've delayed either getting their screening tests or delayed coming in at all. “So that's really what our initial focus is, and we hope to close those gaps so that we don't see an increase in late-stage cancers appearing.”
