How data can shape health insurance
"Big Data" is a term that has generated big interest—and some big promises. Some technology manufacturers have made pretty big claims about the power of Big Data to do things like treat individuals according to their genome and predict disease outbreaks.
Big Data has already revolutionized some industries. Everyday services like the tool that recommends movies and shows you might like on Netflix are driven by data analytics. The mining industry uses Big Data to predict when equipment will need servicing or to anticipate when slight shifts in the rocks present safety challenges.
Many people see big potential for Big Data in the health care industry. But so far the industry has been slower to adopt some Big Data practices.
What makes health data tricky
While providers and practitioners would love to use data to understand health trends and improve outcomes, health care data is challenging to work with.
First and foremost are privacy issue. Personal health care data is protected, and its use often requires a sign-off from patients. Even when patients agree to let researchers use their personal information, health care data is often unstructured. That means somebody or some tool has to scour health records, interpret the notes from doctors and try to assemble matching data sets. Statistics or information present in one record might be left out of another.
The rise of Electronic Health Records (EHR), with its structured codes, has made things a little easier, but it's still hard to standardize or "clean" health data.
Yet, the potential for data to improve treatments and patient outcomes seems huge. The way health data might be used to develop better care breaks into two broad categories: population health and precision health.
Population health—seeing the big pictures
Population health refers to the way diseases and treatments affect large populations. Data in this sense represents anonymous patients and the way diseases present or proceed in a large group of people. These people may be grouped by age, sex or race. Or they may be part of a population based simply on geography. "Americans," for instance, is a population group clinicians may want to study.
Using data in population health focuses on trying to see the trends that emerge when a very large set of statistics is available about a certain condition. How did symptoms develop? How did the disease progress? How often did a certain drug bring a favorable outcome?
Massive data sets about, for instance, people who suffer from diabetes, may lead to better understanding of the disease and how to treat it. A 2017 National Institutes of Health study suggested that analyzing biological, clinical, behavioral, and outcomes data, we may find there are more types of diabetes than we previously thought. Better understanding of diseases can obviously lead to better care.
In certain cases, on a broad scale, data can suggest which treatments provide more value than others—though it remains the role of doctors and researchers to understand why those things might be true. But the potential to scour health data and statistics from millions of people and spot inefficient—or high value—procedures seems great.
Precision health—personalizing medicine
Another general way data can be helpful in medicine is the opposite of looking at huge populations. Precision medicine allows doctors and researchers to apply data to individuals.
Data allows doctors to examine someone and, based on information and statistics—like lifestyle factors, family and medical histories and genomic information—tailor care specifically to that individual.
Precision medicine could make it possible one day for doctors to predict, using lots of data from genetic analysis to social factors, when a person is more likely to develop a certain disease—and proactively begin treatments that could improve that patient's outcome.
How data makes quality clearer
There is another way data can be used in medicine—helping to better understand quality care and how health care spending is linked to medical outcomes. In other posts, we have examined the connection between cost and quality.
Data can help clarify those relationships even more. When researchers use technology to "see," on a broad scale, which treatments are leading to better outcomes, people can start to make stronger assumptions about cost and quality.
Former Surest Chief Clinical Officer Dr. Tara Bishop says that data can paint a clearer picture of the way an insurance plan can be designed. An insurance company might be able to look at data about their members and see what kind of conditions they have, and then compare those to demographic statistics to understand what kind of outcomes are likely.
Designing care pathways
"Can we actually use data to predict if a member is on a certain trajectory?" Bishop asks. "And then can we use our plan design to help drive them to a better trajectory of care?"
What Bishop is referring to is already starting to happen at Surest. Surest assigns prices to health services. Surest lowers prices when the data they analyze in various categories of health services show that treatments and providers meet one or more of these criteria:
- They're more likely to deliver desired health outcomes
- They're less likely to result in unwanted health outcomes, like complications or readmissions
- They're less costly due to more efficient use of health care resources, like fewer tests and surgeries
- They're less costly because the providers charge less for their services.
A lower price may not indicate a lower-quality option. Instead, it is typically an indication that data analysis by Surest shows the service is a higher-value option.
More personalized health plans
There's still a long way to go. Bishop is emphatic in stating that treatment decisions should always be made in concert with clinicians—not by health plan designers. But she says there is a different role data can play in helping insurance companies design better, more personalized plans.
Data can help us understand things like, who are our members are and what are the conditions they’re dealing with? What are the health struggles? What are their previous journeys? Or their potential future journeys? - Tara Bishop, Chief Clinical Officer, Surest
"Data can help us do that in a way that is more personalized to the member, not just the individual condition. If you can actually help members find the best care pathway or the highest value care, that starts to roll up into better population health in general."