Health information exchanges — the communications systems that allow patient records to be shared among doctors, insurers and pharmacies — may someday yield the benefits and savings that have been promised for years.

But so far, there’s little evidence of that, according to a new paper published in this month’s Health Affairs journal.

Studies done to date on the purported benefits of health information exchanges “[do] not provide sufficient rigorous evidence for the benefits of HIE,” according to the paper, written by health policy professors from the University of Alabama at Birmingham and Weill Cornell Medical College.

Despite the dearth of evidence, insurers, health systems and the federal government have spent billions of dollars building the systems, on the hope that they will improve outcomes and save money for the U.S. health care system. The health information exchanges will do so, advocates say, by cutting down on duplicative diagnostic tests, allowing doctors to retrieve clinical data more quickly, and spotting fraud by tracking patients who rove from doctor to doctor looking for painkillers.

Advertise with us

Critics say the data exchange systems — and electronic medical records in general — are still an unproven boondoggle that haven't been shown to improve health.

Both perspectives may be true, said Nir Menachemi, public health professor at Indiana University-Purdue University Indianapolis and one of the authors of the paper. (He was at the University of Alabama at Birmingham when the paper was written).

“There’s minimal evidence of the benefit,” he said in an interview. “But that doesn’t necessarily mean there is no benefit.”

Demonstrated benefit or not, it’s full speed ahead for Pennsylvania’s various health information exchanges.

The largest one operating in Pittsburgh, ClinicalConnect, is anchored by UPMC and includes Heritage Valley Health System, Butler Health System, Excella Health, St. Clair Hospital and several other providers (soon, the VA hospitals will be participating, too). Each provider can retrieve patient records from other hospitals systems.

Advertise with us

The information exchange works  — just ask any physician who has used it to look at a patient’s chart at light speed, rather than waiting for a fax or a phone call. But does the exchange actually improve the health of patients?

“ClinicalConnect [recognizes] that it must deliver stakeholder and patient value,” Chris Carmody, president of ClinicalConnect, said in an email. “To show that value, ClinicalConnect has engaged a respected, outside firm to do a formal return-on-investment study for our HIE. Our goal is to complete the study and publish the results this summer.”

As the individual exchanges seek to prove their worth, the state of Pennsylvania is getting closer to connecting all of them by way of its own “network of networks.” The state-level health information exchange would connect regional, commercial exchanges, such as ClinicalConnect, to other exchanges, plus private labs, pharmacies, long-term care institutions, urgent care clinics and insurers, as well as independent practices.

That thin technical layer, called the Pennsylvania Patient and Provider Network, could be fully operational within the next few years. In September 2014, the state began testing the system by connecting with the St. Luke’s University Health Network. As many as five or six other regional HIEs are expected to connect with the state network in 2015, said Roger Baumgarten, spokesman for the Pennsylvania eHealth Partnership Authority, the state office that is building the network and its components.

The Partnership Authority also is working on another system that would give relevant state agencies, such as the Department of Health and the Department of Human Services, a portal into the state Patient and Provider Network.

Advertise with us

Mr. Baumgarten said that the authority recognizes the importance of showing its value, but that could be complicated by the fact that neither the authority nor the statewide exchange will actually be retaining patient health data; the state will merely provide the hub through which the data flows.

“A lot of the outcomes information will have to come form the providers themselves,” he said. But “we want to be able to take what we projected to be the benefits, and see it translated into actual benefits.”

According to a January survey issued by the eHealth authority, only two of the state’s regional exchanges “have developed processes to measure [impact], but neither is able to report measurable return on investment.”

Part of the issue with demonstrating value is that bending the cost curve might take years, and a healthier patient population likewise might not be evident for years. In other words, there’s not enough data to study yet.

But the larger issue is the very nature of a rigorous study. A research team would need lots of money, meaning it might need buy-in from federal health agencies, or at least federal grant-makers. It also would need a control group, which can be difficult to find — once a hospital starts using a health information exchange, how does it meaningfully compare health results to other population cohorts?

Advertise with us

Much of the research done so far on HIEs has been “quick and dirty analysis,” Mr. Menachemi said. But it may also be folly to wait until the systems are “proven” before implementing them.

“That’s the quandary. We as a society have already embarked on a journey to implement these things,” he said.

The U.S. Health Information Technology for Economic and Clinical Health Act, which was folded into the larger American Recovery and Reinvestment Act of 2009, earmarked billions of dollars for IT upgrades within the health care industry. Of that, hundreds of millions of dollars were steered specifically to health information exchanges, and the State Health Information Exchange Cooperative Agreement Program.

For that kind of tax investment, plus all the money hospitals and insurers have committed privately, “we’re expecting a certain type of impact,” Mr. Menachemi said.

Other states are reporting that impact, though perhaps not to the rigorous, “high quality” standards Mr. Menachemi and his colleagues are calling for. A University of Michigan study using Florida and California data found that patients with multiple emergency room visits were 59 percent less likely to have a redundant CT scan when their information was shared across an health information exchange.

Advertise with us

They were also less likely to get a duplicate ultrasound or a chest X-ray.

And a study out of Western New York suggested that nine out of 10 “needless” CT scans were ordered by physicians who never or infrequently used the state’s HEALTHeLINK information exchange network.

Bill Toland: btoland@post-gazette.com or 412-263-2625.