How health insurers profit by hoarding your data

A comprehensive All-Payer Claims Databases would help the health care system become more effective and efficient. The Department of Labor could do something about that, but hasn’t.

Merrill at GoozNews is depicting why providers and software vendors closely guard patient information, maintaining exclusivity so as patients have to rely on them. Insurance companies follow suit so as to maintain the frequency and particularity of care for disease, etc. to set payments. Reporting requirements would resolve much of the difficulty in healthcare.

Originally Published at GoozNews by Merrill Goozner

Last week, I reviewed how health care providers and their computer software vendors refuse to share patients’ electronic health records (EHRs). This issue isn’t just for health care wonks. Allowing patients to control and share their data with other providers would improve care coordination and help eliminate duplication and waste. If the data (stripped of personal information, of course) were housed in a comprehensive health information exchange, it would become a valuable research tool for improving health outcomes.

The main reason providers and their software vendors block the free flow of information is that they are primarily interested in serving their economic self-interest, not the health of their patients or public health. Neither group is interested in facing heightened competition, which would be enabled by the easy transfer of individual EHRs.

The other side of health care transactions are the insurance companies — the payers, who have comprehensive data on how much they paid to whom and for what. The private firms process claims for more than 180 million Americans with employer-provided health care coverage. They also manage the policies of more than 21 million individuals and family members whose plans were purchased on the Obamacare exchanges.

Like providers, these insurers engage in information blocking. Self-interest is at play here, too. They make a lot of money from selling the data.

Insurer data differs in fundamental ways from the data controlled by providers. An EHR contains an individual’s medical history. It typically includes every interaction between a patient and one provider. If the patient receives all their care from an integrated delivery system (a hospital system or large physician practice may have many providers feeding into its EHR), the record may be complete. But that’s rarely the case.

That’s why many states and regions created health information exchanges (HIEs) to facilitate EHR portability. Unfortunately, they are severely underfunded and rely largely on user fees. A recent recent GAO study found only half of hospitals “often” share their EHRs with an HIE.

A payer claim, on the other hand, whether it is paid by public agencies like Medicare and Medicaid or private insurance companies, contains much less information than an EHR. Each claim is a discreet record of what an individual payer paid an individual provider (hospital, doctor, clinic, pharmacy, drug company) for a single episode of care. That ranges from a bundled payment (replacing a knee, for instance) to an individual test, product, or service. In addition to the provider bill and the amount paid, the claim may include demographic data about the patient. It usually has minimal information about the patient’s medical condition that triggered the claim.

Using claims

Though claims databases are limited in scope, they do provide enough information to compare prices, resource use, quality of care, and outcomes. They can also be mined to measure providers on their use of high-value and low-value care, use of high-priced branded drugs instead of generics, and similar issues that make the U.S. the highest priced health care system in the world.

Medicare, which covers about 65 million seniors and disabled Americans, maintains the most comprehensive claims database. Its de-identified records have been publicly available at least since the mid-1990s when Dr. Jack Wennberg and colleagues used Medicare claims to create the Dartmouth Atlas of Health Care, which documented why clinicians in states like Florida, which cost Medicare $13,652 per beneficiary in 2020, spent so much more than Vermont, which cost Medicare $8,726 per beneficiary.

Providers in high utilization states fought back by claiming the Dartmouth Atlas failed to consider patient differences or the higher cost of living in some states. It also failed to include data from private payers, which cover nearly two-thirds of all patients and pay much higher rates than Medicare. The Atlas also didn’t include claims from state Medicaid agencies, which cover about the same number of beneficiaries as Medicare and pay somewhat lower rates.

In response to those criticisms, patient advocacy groups, cost control advocates, employers and researchers began pushing for the creation of state-based all-payer claims databases (APCDs). As its name implies, APCDs hoped to collect data from every payer, whether governmental or private.

They’ve had some success. Beginning with Maine in 2003, about half the states have either created or are in the process of building an APCD. The most recent states to launch efforts included Georgia, Texas and Indiana. The demand for data transparency is bipartisan.

Insurers’ self-interest blocks the way

Unfortunately, insurer reluctance to share data with an APCD is just as strong as provider resistance to sharing data with an HIE. The major insurers who control most of the claims are in the business of selling their de-identified data to drug and device firms and hospital equipment makers; marketing firms and data brokers; major universities and independent researchers; and anyone else who can afford the tab.

Optum, a wholly-owned subsidiary of UnitedHealth Group, the nation’s largest insurer with the largest repository of claims data, is a major data purveyor. So is the non-profit Health Care Cost Institute, which is controlled by and receives claims from Aetna, Humana and Kaiser Permanente (UnitedHealth dropped out of the consortium in 2019). Market research firms like Definitive Health and IQVIA have become major purchasers of claims data, which they repackage for a clientele that is willing to spend generously for insights into where they can best market their products.

The financialization of claims data was given a boost by a 2016 Supreme Court decision in Gobeille v. Liberty Mutual Insurance Co., which gave insurers the right to deny requests for claims data from state-run APCDs. The 6-2 vote (only Justices Ginsburg and Sotomayor in dissent) said the Employment Retirement Income Security Act (ERISA), which governs most employer-provided health insurance plans, preempts any state law requiring multi-state plans to submit their data to a state-based APCD. The majority argued that meeting the demands of 50 different state agencies would present an unwieldy administrative burden on multi-state employers.

But in a concurring opinion, Justice Stephen Breyer offered a roadmap for meeting the needs of state-based APCDs, or, for that matter, a national APCD, should the Labor Department, which oversees ERISA, want to pursue that option. He wrote:

“I see no reason why the Secretary of Labor could not develop reporting requirements that satisfy the States’ needs, including some State-specific requirements, as appropriate. Nor do I see why the Department could not delegate to a particular State the authority to obtain data related to that State, while also providing the data to the Federal Secretary for use by other States or at the federal level.”

In other words, the Labor Department could standardize claims reporting to a national database by establishing a common set of variables with a common reporting format. This would eliminate most of the administrative burden, establish a national APCD, and serve as the conduit for data flowing into state-run APCDs.

Labor Department missing in action

Unfortunately, the DOL under both the Trump and Biden administrations has not pursued that option. Spokespersons for the department failed to respond to repeated phone calls and emails seeking comment.

Advocates for greater data transparency weren’t willing to wait. The No Surprises Act, bipartisan legislation passed in the waning days of the Trump administration, originally included a section that would have required the DOL to follow up on Breyer’s suggestion. “We pushed hard for a national APCD,” said Frederick Isasi, the former head of Families U.S.A., “but it was killed at the last minute.”

Instead, the legislation created an advisory committee to come up with recommendations. It also earmarked $125 million to strengthen existing APCDs and encourage more states to set one up.

The recommendations in the advisory committee’s report, which arrived on the DOL’s doorstep in late 2021, followed Justice Breyer’s script. The 15-member committee, chaired by Maureen Mustard, then director of health analytics in New Hampshire’s insurance department, called for a standardized format for submitting data; inclusion of a broad array of relevant data fields including non-claims based payments to providers like capitated payments, bundled payments, and performance-based incentives; and “other variables needed to enable cost and utilization analyses, including plan benefit design and cost-sharing features, and the actuarial value of plans.”

Mustard, now an independent consultant, told me in an interview earlier this month that “we offered to help the DOL to work through these issues, but that never came to fruition … I have not heard from anyone at the department.”

The report did not shy away from assigning blame for the glacial pace of using claims data to improve health care system performance. “Employers’ reliance on third-party administrators for their health benefit plan management and administration places an entity that may have an economic or business interest in not (emphasis in original) submitting data,” the report said.

“Data is valuable. Data is money,” Mustard told me. “Companies have realized if they keep their data locked away, it is their information and they can monetize it. They can learn from it and get a competitive edge. They don’t want the government getting its hands on it.”

Or, for that matter, do they want to give it to you, the patient public, whose health care needs generated the data in the first place.