Image: Anadmist, Getty Images/iStockphoto

An artificial intelligence company has pivoted from finding the right treatment plan for patients to helping them find health insurance to pay for that care. At the start of the pandemic, Jvion used its Eigen Sphere engine to build a map that shows which residents in a community are at highest risk for complications if they contract the coronavirus. Now the company is helping hospitals protect the bottom line by identifying which patients are most likely to have lost insurance and helping them find a new policy.

Many people who have lost their jobs due to the coronavirus pandemic also have lost their health insurance. The Commonwealth Fund found that 41% of people who lost their jobs had insurance through that job. The Kaiser Family Foundation estimates that four out of five Americans who lost health insurance when they lost their jobs can get coverage through the Affordable Care Act or Medicaid, but many do not know about it.

Dr. John Showalter, chief product officer at Jvion, said that the company’s CORE analytics engine has always been able to understand how people use or don’t use the healthcare system. The company originally focused on matching the right care with the right patient based on age, illness, and other factors. When the COVID pandemic started and people delayed healthcare visits and lost insurance, the company shifted its focus to understanding who is most in need of insurance coverage.

The Jvion CORE generates two types of recommendations: How to get insurance coverage and how to get preventative care and avoid hospitalizations. Coverage recommendations focus on identifying uninsured patients and helping them enroll in the most effective health insurance plan for them.

“Based on estimates in our clients’ data, over 25% of patients with uncompensated care qualify for Medicaid and an additional 20% or more qualify for ACA marketplace plans,” he said.

SEE: Robotic process automation: A cheat sheet (free PDF) (TechRepublic)

Showalter said that Jvion data sets have always included the patient’s insurance information, but insurance coverage has not been an important data point previously.

While this helps individuals manage their diabetes and other ongoing health problems, it also helps hospitals improve the bottom line because when more patients have insurance, hospitals provide less care for free.

The Affordable Care Act significantly reduced these costs for hospitals. According to the Center on Budget and Policy Priorities, the nationwide uninsured rate fell 35% from 2013 to 2015, and nationwide hospital uncompensated care costs fell by about 30% as a share of hospital budgets — a $12 billion drop in 2015 dollars. In states that expanded Medicaid, the number of uninsured people dropped the most and uncompensated care dropped by 57% on average.

Showalter said that the company’s research suggests that up to two-thirds of hospital visits can be prevented by incorporating Jvion’s insights into preventive care management programs for asthma, cancer, diabetes, and heart disease. This kind of care involves preventative care, patient education, and addressing the social determinants of health.

The Centers for Medicare and Medicaid Services has recently started reimbursing doctors and nurses for providing these check-up services. The idea is that receiving advice from a healthcare professional on a regular basis can keep a person out of the hospital or nursing home, which is better for the individual and saves money on healthcare costs at the same time.

The Jvion CORE does not use reimbursement rates to identify patients most in need of care but the algorithm does consider coverage and patient costs when identifying patients who may be underinsured.

“If a patient is highly likely to need a lot of care, a high deductible plan is probably not a good option for them financially,” Showalter said.

Can AI replace insurance navigators?

In addition to making health insurance somewhat easier to get, the Affordable Care Act funded navigators who helped individuals choose the right insurance plan. The Trump administration cut funding for the navigators from $63 million in 2016 to $10 million in 2018. During the 2019 open enrollment period for the federal ACA health insurance marketplace, overall enrollment dropped by 306,000 people.

“While that may not seem like a lot, the average annual medical expense is around $3,000 per person, and a shortfall of covered patients could represent over $900,000,000 of medical expenses will not be paid by health insurance,” Showalter said.

When states banned elective medical procedures temporarily during the early months of the pandemic, this cut off an important revenue stream for hospitals and many laid off workers. Some of these layoffs included patient navigators who helped patients enroll in health insurance, particularly Medicaid.

Showalter said that all Jvion customers have had at least a few navigators on staff but not enough to reach every patient in need of assistance.

“To bridge the gap, we are deploying advanced technologies, like chatbots, to support our clients and we will also be deploying a full-service option in the fall,” he said. “Our clients will simply provide us with their data and we will do the rest—identify, engage, and enroll.”

Mapping the risk of contracting COVID-19

Jvion also built a map to help businesses and public health officials understand the varying COVID-19 risks based on location. This map estimates the risk of contracting the virus based on vulnerability calculations that include access to jobs, access to healthcare, race, length of commute, and environmental risks.

The analysis identifies specific census tracts where the population is at greater risk of hospitalization and mortality due to COVID-19. After studying research from the US Centers for Disease Control and Prevention on environmental factors that led to certain outcomes with respiratory infections, Jvion took anonymized data on 30 million Americans and combined it with models of patients with comparable respiratory infections, as well as with virus and geolocation data.

The service provides individual employees information about their risks so they can request appropriate protections at work. It also alerts employers to areas where a high number of employees are vulnerable so they can implement appropriate safeguards.