Last month, researchers from the Brookings Institution, SUNY Buffalo and the University of Connecticut published a paper on how health information exchanges (HIEs) can reduce redundant medical procedures—an important takeaway given that the Congressional Budget Office estimates the cost of procedures that did not improve health outcomes at $700 billion per year.
For the study, Brookings scholar Niam Yaraghi and the paper’s coauthors obtained data from the Center for Medicare and Medicaid Services (CMS) on medical services received by Medicare patients, and compared that data to information on the adoption of HIEs from HEALTHeLINK, the regional health information exchange network for western New York. According to the research, the results showed that an HIE lowers the repetition of therapeutic procedures, those aimed at improving a patient’s condition, but HIE adoption did not affect diagnostic procedures that determine the extent of a patient’s illness.
More specifically, the main variable of interest in the study was the ratio of repeated tests to the unique patients that a physician sees, Yaraghi and his coauthors explained in a brief summary of the findings from Brookings. They wrote, “For example, a physician that performs a specific therapeutic medical procedure 100 times for 73 unique patients repeats the procedure 27 times, an average ratio of repetition of 0.27. Assuming other variables are fixed at their average level, the results show that the average ratio of repetition for that physician will fall to 0.26 after one year of experience in the HIE system. This means that the physician would perform that test 98 times for the same 73 patients, avoiding two repetitions.”
Given these results, the study’s authors estimated that had all the physicians in the U.S. been using HIE, Medicare would have saved $63 million annually for each therapeutic procedure performed at physicians’ offices.
Over an email exchange discussing the study and its impact, Yaraghi said that one of the key takeaways is that health information exchange has never been studied in outpatient settings until this research that analyzed HIEs for outpatient therapeutic procedures. Yaraghi and Healthcare Informatics Managing Editor Rajiv Leventhal further discussed this study, its significance, and the overall HIE landscape. Below are excerpts of that email exchange.
What should be the main takeaways from this study’s findings?
This is the first study that shows HIE is effective in reducing redundancies at physician offices. It is very difficult to appropriately measure the effects on outpatient settings (due to lack of good control) and thus so far, HIE studies have been limited to EDs and hospitals.
We show that HIE affects redundancies differently in inpatient vs. outpatient settings. While at EDs, HIE significantly reduces diagnostic procedures, at inpatient settings, it reduces therapeutic procedures and doesn’t affect diagnostic procedures.
The effect is small, but economically very significant. Per each therapeutic procedure, the potential savings from avoided redundancies adds up to $63 million per year.
I have seen many studies on HIEs’ benefits in the ED, as you mention, but not in outpatient settings. What locations did you look at specifically for this study?
We looked at all physician offices in western New York where the HIE (HEALTHeLINK) operates. You are right, almost all studies look at EDs, because data has been hard to gather at so many doctors’ offices. We used publicly available CMS data and paired it with HIE adoption data.
You estimated, “Had all the physicians in the United States been using HIE, Medicare would have saved $63 million annually for each therapeutic procedure performed at physicians’ offices.” How realistic is it for all U.S. physicians to be connected to HIEs?
I am very optimistic. Although I am not sure if the current technical structure and approaches to HIE is the best way forward, I am pretty sure that regardless of the method, all the U.S. physicians will eventually exchange health information. The move toward open APIs (application program interfaces), patient-mediated exchange methods, new forms of payment and medical organizations, and most importantly, an increased patient demand for access to records, will lead the way to interoperability and exchange sooner than we expect.
What are the main roadblocks to greater HIE adoption and success?
It depends on location. Some places have really good infrastructure and the only roadblock may be the reluctance of doctors; in some other places, HIEs haven't been successful because of technical difficulties and most importantly, failing to establish trust with hospitals and data providers. HIE success is 10 percent technology and 90 percent management, vision, negotiation skills, and trust building. Unless you can bring hospitals on board to send their data, HIEs fail, no matter how much money or how advanced technology you have. I think the most important reason that HIEs have not been successful is the lack of incentives for hospitals to share their data and the failure of HIEs to show their value to hospitals to gain their cooperation.
With many HIEs having been driven by federal funding that has now dried up, how do you envision their future?
I think those that have thrived, such as HEALTHeLINK, will have a great time especially if they continue to innovate and mine their data for other things like precision medicine, finding patients for drug trials, etc. Those that have not become fully operational will have a tough time going forward because other technologies are now available and will compete with them. Also, they have lost the trust of local providers due to their lack of success so far and are thus competing with others, [meaning] they will not have the upper hand.
Talking about other technologies, do you foresee any conflicts with something like blockchain that could get in the way or inhibit the value of HIEs? Or can they work concurrently?
I think blockchain has a long way to go. Banks have started to use it only recently and healthcare is always a couple of years behind. Also, the most important feature of blockchain is its security. While we certainly need better security in healthcare, to be honest, I don't think that is the main concern of providers. The fact that most privacy breaches happen in healthcare just shows the lack of respect for security. Since people unfortunately do not value or pay attention privacy and security in healthcare (as compared to banks), blockchain will take longer than usual to become the norm in healthcare.