Following the lead of nearly a dozen other states, Connecticut officials are developing a repository of health insurance claims data that’s intended to provide a better view of health care costs and the health of the population.
David Newman runs a similar data-gathering operation, and on Thursday, he offered Connecticut officials some advice.
“It’s easy to do it wrong and it’s tough to do it right, but it’s worth doing,” said Newman, executive director of the Health Care Cost Institute, which is building a national database of health insurance claims information.
All-payer claims databases, as they’re known, grew out of a push for more transparency in health care costs and spending. They typically include information on medical, pharmacy and, in some cases, dental claims paid by commercial insurance companies, Medicare and Medicaid. Ten states, including Massachusetts, Vermont, Massachusetts and Maine, have all-payer claims databases, and several other states, including New York and Rhode Island, are developing them.
Newman and other experts on all-payer claims databases spoke during a forum sponsored by the Connecticut Health Policy Project, the Connecticut Center for Patient Safety, and Access Health Analytics, the organization developing Connecticut’s database.
There’s “something for everyone” in the databases, said Jo Porter, deputy director for the New Hampshire Institute for Health Policy and Practice and co-chairwoman of the All-Payer Claims Database Council.
Public health agencies, insurance departments and policymakers can use the information to better understand the health of the population, what problems or gaps exist, and where the cost and use of medical services is highest. Health care providers can use it to adapt to payment models that increasingly tie compensation to patient outcomes. And consumers, who are increasingly covered by health plans that require paying a portion of medical costs, can use the databases to find out what different hospitals charge for the same procedure.
But the plans for an all-payer claims database in Connecticut have also drawn questions about how the data will be used and whether it’s possible to adequately protect the privacy of patients whose records are used.
Crude measurement without data
State Comptroller Kevin Lembo’s office handles the health care benefits of state employees and retirees. The health plan has undergone significant changes aimed at improving health and controlling costs. But Lembo said his office must rely on a crude measurement to tell if those efforts are working: Are costs going up or down? Is the trend of cost increases rising or decreasing?
Without more detailed data, Lembo said, even if things are moving in a positive direction, officials can’t tell why. And it’s hard to make nuanced decisions about the health plan, such as whether to set lower cost-sharing for certain services they want more people to get.
While Lembo spoke of the database’s promise, he also raised questions about who should administer it. The all-payer claims database originated last year as part of the state’s Office of Health Reform and Innovation, which has since been eliminated. It’s now run by Access Health CT, the state’s health insurance exchange. Access Health is a quasi-public agency, and Lembo said it was worth considering whether the database should be housed in government instead.
Tamim Ahmed, executive director of Access Health Analytics, said his organization has already started gathering Medicaid and Medicare data, and hopes to get claims data from private insurers in the next few months. The goal is to have enough information by next August to publish tools to help consumers choosing insurance plans. One could be information on what an individual or family could expect to spend on care, based on their health.
Ahmed said there will be two levels of information: Raw data that gets submitted to the organization, and data that would be used for research and for the public, which would be “de-identified” so those seeing it would not be able to tell who it represented.
But Dr. Susan Israel, who said she had been following patients’ “privacy rights, or lack thereof,” said it was important to know who would be handling patient data that hadn’t been de-identified.
“The public really needs to understand, first of all, that their data’s being taken without their consent, and that there may be a number of people who will be seeing their identified data,” she said.
Israel, who described the database as an “unprecedented threat to our medical privacy,” said the goal of transparency in medical costs could be achieved by instead gathering payment information from insurers and health care providers without patient-specific identification.
Ahmed said Access Health Analytics had not yet completed its rules for how data can be used or governed, but it’s aiming to apply rules that emulate federal medical privacy standards.
Other states take a range of approaches to handling data privacy, Porter said. Minnesota only allows its data to be used by one state agency and doesn’t release it to the public. Other states that are less restrictive typically have boards that review requests for data and exactly what information gets released, she said.
Newman said his firm only licenses data to academics through their universities. The data is de-identified so there’s a 1 in 2 million chance that an individual could be identified, he said.
But Newman noted that one concern is that researchers can merge de-identified data with other data sets, producing information that can identify specific patients. His organization prohibits researchers using its data from doing any merges on their own. Instead, they must let his institute do it and be certified by a statistician that no information will be identifiable.
When giving data to researchers, Newman said he prefers a “data enclave” approach, in which people must go into his organization’s system to use it, rather than receiving their own copy of the information. When their license to use it expires, the researchers’ access can simply be turned off. If they received a copy, he noted, people could make a copy before turning it in.
Of privacy issues, he added, “These are real, legitimate concerns.”
Why are costs rising?
Despite the concerns, Newman thinks that all-payer claims databases are worth creating. He noted that federal policy has been driven largely by information drawn from Medicare claims, which don’t reflect trends involving kids or people under 65 particularly well.
Having the information isn’t enough; it must also be understood, he said. Newman’s database, which has information from across the country, found that spending on mental health and substance abuse services and psychotropic drugs for kids was increasing. But he noted that that alone doesn’t provide enough information to take action. It’s possible that spending is rising because more people are getting better care.
But if policymakers want to reduce the growth in health care spending, he added, they must first understand where the costs are rising, and why.