‘Big data’ seen as key ingredient for efficient health care
Published 1:30 am Sunday, October 16, 2016
Defense giant Northrop Grumman has a $92 million government contract to build the second phase of a computer system that focuses on reducing Medicare fraud.
Eventually, the system will help anticipate beneficiaries’ medical disorders.
Public and private insurers are spending millions on “big data” — technology to predict people’s future health care needs based on their interactions with doctors, hospitals and pharmacies, as well as information gleaned from sources like social media.
These so-called predictive analytics aim to make health care more efficient by opening the door to addressing medical issues before they become serious problems.
For example: You’ve complained to your doctor that you’re having trouble losing weight. You’re taking a cholesterol medication. You’ve posted on Facebook that you’re stressed over a marital breakup or on LinkedIn that you’re looking for a new job.
An algorithm would connect the dots and alert your doctor that you’re at risk of a heart attack and require medical intervention.
“There are tremendous advantages to big data in health care,” said Gerard Magill, professor of health care ethics at Duquesne University. “It’s about creating a comprehensive approach to using medical information.”
The trade-off: Goodbye privacy.
“Big data requires that information; it’s nonnegotiable,” Magill said. “Individual privacy is gone for the common good.”
Medicare’s contract with Northrop Grumman is one of the largest efforts underway to create a health care crystal ball capable of looking into patients’ futures.
“The use of data in health care is absolutely critical,” said Dr. Shantanu Agrawal, director of Medicare’s Center for Program Integrity, which is tasked with lowering costs. “Having it be predictive of various issues is extremely important.”
Medicare has been criticized for using a “pay and chase” approach to fraud — paying all claims and, afterward, trying to determine which were bogus.
Rep. Peter Roskam, R-Ill., chairman of the House Ways and Means Oversight Subcommittee, said at a hearing on health care fraud last month that the agency needs to move faster in implementing “better data analysis and predictive analytics.”
Medicare says the first phase of its Northrop-designed fraud-detection system produced more than $1 billion in savings over the last two years.
Amy Caro, vice president of Northrop’s health care technology division, told me sophisticated algorithms are clearly the best way to crack down on fraudulent claims.
They’re capable of sifting through millions of submissions and recognizing signs and patterns that indicate a claim may not be on the up and up.
The next step, she said, will be using big data to get ahead of Medicare and Medicaid beneficiaries’ health care needs.
“You have all types of data out there,” Caro said. “You’re able to drill down and look for signs of certain diseases or conditions.”
I shared my own experience:
Type 1 diabetes runs in my family. So, over the years I routinely asked physicians to test me. Even though I was prone to low blood sugar, doctors informed me again and again that I was fine.
Then, a decade ago, I was diagnosed with the chronic condition.
“A perfect example,” said Dr. Sam Shekar, also with Northrop. “The kind of system we’re designing would have seen that you may be in a pre-diabetes stage. It allows you to get ahead of diseases before they’re diagnosed.”
In my case, the episodes of low blood sugar were the red flag, he said. Coupled with my family history — which a national DNA database would know — my doctor could have been more proactive in keeping diabetes at bay.
The thing about big data, though, is that it’s more effective the bigger it gets. Medicare’s Agrawal said the program is eager to share information and algorithms with private insurers. Most private insurers, meanwhile, are busy developing their own predictive analytical systems.
Anthem, one of the country’s largest private insurers, is typical of the industry. Like Medicare, the company wants to find high-tech ways to reduce fraud and improve the well-being of those it covers.
Ariel Bayewitz, Anthem’s vice president of provider analytics, said the goal is to crunch data from multiple sources, including claims for doctor, hospital and drugstore visits, and to be able to alert an individual’s physician that there may be an issue that needs attention.
“We have a unique lens because of the access we have to claims data,” Bayewitz said.
But are people deemed to be at higher risk of a disease or condition in danger of insurance rate hikes? At this point, the answer is no.
Obamacare protects people in the individual insurance market. People in large employer-sponsored plans who submit expensive claims can have an affect on overall rates, but such price increases are spread throughout the coverage group.
Most experts I spoke with said big data for health care is largely uncharted territory, and it’s unclear exactly how insurers will make use of so much information
Existing privacy laws for medical data may be insufficient to address the scope of information-sharing on the horizon.
“There are definitely protections that we need to put in place that haven’t been put in place,” acknowledged Bayewitz.
Northrop’s Caro said “there will have to be a national conversation” about big data’s use of medical info and overhauling regulations to reflect a need for added safeguards.
“We’re not there yet,” she said.
— Los Angeles Times
