Lifestyle-based analytics hold promise for proactive care
Lifestyle-based analytics may be an "emerging" predictive health model, but experts note that it's "simply taking data that we already have at our fingertips" and analyzing it in ways that weren't possible before.
The benefit? “Moving from a reactive mode to a proactive mode” in healthcare, says Chris Stehno, senior manager at Deloitte Consulting.
In the past, predictive healthcare modeling has used claims data, but the majority of the population doesn’t have good data – making predictions about life events and diseases difficult.
Stehno says their model uses consumer spending data, which is “chock-full“ of information on how individuals lead their lives. This data also provides high correlations for lifestyle-based diseases, which account for 75 percent of the total medical dollars spent in the U.S.
As Bill Preston, a principal at Deloitte, points out, when a person changes addresses, he or she starts getting bombarded with offerings for new siding – because consumer data indicates they're a new home buyer. Preston says their model takes that same data and it mines it for “specific variables that will be indicative of a particular situation or disease.”
For instance, Preston and Stehno note studies that have shown that individuals with a commute of 90 minutes or longer round-trip are 20 percent more likely to become diabetic or obese.
“Where in claims data you know for sure that they have a condition,” in the lifestyle based analytics, “this information doesn’t tell you that [this individual] has diabetes it tells you that they have an elevated risk for diabetes," says Stehno.
Preston notes that this can allow health insurers and providers to be proactive and not wait to do something until they are sick, which lowers overall healthcare spending.
But the model can also indicate an individual's “willingness to change.” For example, says Stehno, if an individual is seen to have purchased weight loss training products, it suggests that he or she has “change behavior” in mind. This can help identify people at risk when the chances of helping them can be maxmized, he says.
With the diagnoses codes changing due to healthcare reform and ICD-10, predictive modeling will have to be rebuilt, Preston notes.
"Lifestyle based analytics gives them a way to bridge that gap while this work is being done,” he explains.
For health insurers and providers, “the biggest hurdle,” says Stehno, is changing their “mindset and strategy” about how they approach people, and how they will use their existing programs to talk to people and reach out to them.