Using such data as a source of social determinants of health information can boost the accuracy of population risk predictions.
It’s no surprise when your mailbox is full of credit card offers or coupons and catalogues from your favorite stores. After all, it’s fairly common knowledge that credit reporting agencies like Experian and Equifax (which recently made headlines for a massive data breach) provide your purchasing data to companies for direct marketing purposes.
But what if healthcare providers could tap into that data too? How might it be added to claims and clinical data to better predict population health risk?
That’s exactly what Mission Health Partners, Mission Health’s Medicare ACO, has been doing for the past six months or so, augmenting patients’ clinical records with other data that can reveal lots of insights about their lifestyle. It’s working with value-based care solutions company Lumeris, which gathers data from those agencies, populates it into Mission Health’s patient records, and uses it to create a score that evaluates the risk—from low, medium to high—for ACO patients to end up in the hospital or ED.
Instead of using that data to sell products, “we’re using it to help improve health,” says Michael Cousins, PhD, chief analytics officer at Lumeris.
The idea that such data could help improve health and reduce healthcare utilization relies on the notion that social determinants of health are some of the most critical factors in patient outcomes.
Just a name and address—without even being seen by a provider—can reveal a lot: Someone’s transportation and housing situation, vacations, marital status, lifestyle habits, and whether they pay bills on time. When combined with claims and clinical data, this information can help determine whether a patient might need extra clinical support.
Rob Fields, MD, medical director at Mission Health Partners, says when the ACO first started in 2015, they were collecting this information the “hard way.”
“It’s an exceedingly painful process,” he says. “The only way to get at this data is through patient interviews and assessments.”
He describes long, sensitive conversations between patients and a relatively small team of care managers (Mission Health Partners serves about 90,000 patients across 18 counties in western North Carolina).
“It’s often not until the first or second, or maybe even third, phone call when you start to get at a housing need or a food insecurity need. It’s not the sort of thing people generally willingly share,” Fields says. “We might believe and know in our heart of hearts that this is the driver of bad outcomes, but having to gather that data in a very manual way is challenging.”
Although case managers might have this information at their fingertips, broaching the conversation with patients is still a delicate matter, so Fields says care managers start by telling a patient that they work with their doctor, they know they have a certain health condition, and that often, people with that condition have trouble affording their medications. He says doing so “unlocks the conversation.”
“The data helps inform who we call, but at the time of the call we need to let the patient give us that information back,” Fields says.
In fact, the patients don’t actually know that the case managers have this information.
“This is an incredibly powerful tool, and people have perhaps, knowingly or unknowingly, come to expect it from Walmart or Target,” Fields says. “But they don’t necessarily expect healthcare providers to have this same information, and it would seem likely uncomfortable to say we have all this information about you, about your housing or transportation…that’s not likely to go over well.”
Cousins says Mission Health Partners is a pioneer in using this information and that they’re currently in “version one.”
Being at the early stages also means that it’s early from a legal perspective.
HIPAA does not govern the use of direct marketing information, only medical information. Both Lumeris and Mission Health Partners are taking “the conservative route,” says Cousins, because they anticipate a time when HIPAA will govern the use of this data.
They treat the data as if it is personal health information, and abide by the same rules that they follow when dealing with medical diagnoses or medications.
Although the data collection process and predictive analytics is cutting edge, Fields says what they do with the information is decidedly old-school and rooted in building relationships. He points to one case of an elderly man who had been labeled belligerent and non-adherent, but who was plagued by frequent COPD exacerbations.
After several nurse phone calls and community paramedic visits (provided at no charge), the man finally let people into his home, which revealed hoarding, a rain catchment system instead of running water, and a broken air conditioning and heating unit. They helped clean his home and get him other help he needed.
“It was a terrible situation that wasn’t going to improve no matter how many medications or prescriptions you write,” Fields says. “None of that would’ve happened without that relationship piece and the element of trust”
Mission Health Partners’ early work in using social determinants of health to guide care has paid off. CMS’s 2016 Quality and Financial Performance Results showed that Mission Health Partners achieved a Quality Score of more than 97% and saved Medicare more than $11 million beyond its target, despite having one of the lowest spending benchmarks in the nation.
In addition, Mission Health Partners has boosted the accuracy of its population risk predictions over the past six months by 25% since incorporating Lumeris’ social determinants of health data into its EHR and analytics reporting tools. Cousins notes that the costs of their solution are on par with predictive analytics tools that don’t include this data.
Although much of Mission Health Partners’ early work in this area has focused on readmissions, Fields says they’d next like to think more “upstream” and use this data to keep people out of the hospital or emergency department in the first place, identifying which high-risk people would benefit most from intervention.
“We’re trying to make that intervention before they end up there, and that’s obviously the more powerful piece of that,” he says.
Alexandra Wilson Pecci is an editor for HealthLeaders.
