EPtalk by Dr. Jayne 10/15/15

Both the American Hospital Association and the American Academy of Family Physicians have publicly weighed in on ONC’s revised Interoperability Roadmap. Both organizations believe it doesn’t go far enough or fast enough.

Although quite a few stakeholders accuse the EHR vendors of being the villains with regards to data blocking, my personal experience has been that health systems and hospital-controlled physician networks are really behind it. Not to mention that ACOs naturally create barriers to interoperability as they encourage physicians to keep patients within a well-defined (often narrowly so) network.

I had the adventure of standing up a private HIE seven or eight years ago. The sponsoring health system’s express purpose was to share data only within our referral network. Providers affiliated with other hospitals were not invited to join.

At least in my metropolitan area, I haven’t seen that philosophy change across the years. While the AAFP calls for “increased accountability on industry and decreased accountability on those who are using their inadequate products,” I think we need to focus the microscope on some of the other interested parties.

Let’s hold health systems accountable. When their vendors provide interoperable products, they need to be using them in an interoperable fashion. They need to force their owned hospitals and employed physicians to accept and integrate electronic health data brought by patients. And they need to release the data electronically when patients request it. Mr. H’s story of his inability to get his records electronically is just one of thousands out there.

Let’s also bring the non-provider technology stakeholders into the business of being regulated in the same way that providers and hospitals are. For example, reference laboratories. The fact that there are still numerous laboratories who cannot (or will not) send LOINC codes with their results is appalling.

Let’s also start regulating the pharmacies, getting their information systems under control. One of my local pharmacies sends scads of duplicate prescription refill requests all the time. This leads our practice to have to spend time determining if they’re really duplicates or not. When we call the pharmacy to complain, we’re told that it happens due to limitations in their computer systems and the inability to match refill responses to the original requests in an accurate manner.

Since all patients and physicians depend on pharmacies and labs, why shouldn’t this be part of the solution? What about standardizing all the other systems we have to interact with, such as home health systems and those used by nursing homes and other extended care facilities? Many patients that we used to hospitalize for care are instead receiving services through home health agencies. It doesn’t do me much good to interoperate with the hospital if the patients aren’t in it.

I started reading the Roadmap essentially to have a break from reading the Meaningful Use final rules. I’ve been jumping around in it rather than reading it through, however. I was pleased that they called out “Authentication and Identity Proofing” early in the paper (page 11 if you’re interested). They admit that lack of standard approaches has hampered information sharing. They mention that Direct was intended to “work much like email and lower the barrier to exchange for providers and hospitals.”

In my market, however, Direct has added all kinds of confusion, especially for providers like me who have multiple practice situations. I have different addresses for the different urgent care groups I work with on different EHR platforms. If a PCP is looking for me to send a follow-up, how do they know where to send it? Other health organizations have created additional complexity. One organization created separate addresses for each location for each provider, so a cardiologist who sees patients at six physical locations has six addresses. Although I understand their original reasoning for doing it, it’s untenable in reality.

Page 36 addresses individual data matching. They point out that although HIPAA required creation of national identifiers for patients in 1996 and Congress blocked appropriations, there is no prohibition against private or non-HHS governmental agencies creating their own. They note that the Department of Veterans Affairs and the Department of Defense assign unique identifiers for service members. Although I understand the arguments against it, I’d volunteer to test drive a national ID myself even if it’s private and voluntary.

I’ve had too many bad experiences with using existing primary data elements (name, DOB, sex, phone number), both personally and professionally. Even existing non-healthcare governmental systems don’t always handle the data in the same manner.

Here’s an example. Due to my southern roots, my actual legal name has a compound first name. My legal middle name is actually a family name, which looks like a last name to most people. Then I have my actual last name. Let’s work with “Peggy Sue Herrington Mitchell” as our example. My original birth certificate was completed on a typewriter, with “Sue” landing squarely on the line between the caption for “first” and “middle” names.

When I was in medical school, I had to get a certified copy of my birth certificate to prove identity for licensure. My birth county had computerized, and the person keying the data had made “Peggy” my first name and “Sue Herrington” as my middle name. It also truncated the “Herrington” by a few letters since there weren’t enough characters in the middle name field. With the layout and truncation issues, the computerized copy didn’t match my college transcript or my medical school transcript or my passport, which had my legal name listed correctly. The authorities refused to accept my identity proofing. I had to petition the county to pull a microfiche copy of the original and certify it.

When I complained about the truncation issue as the reason for needing the original, they said that they were limited by their systems. I pointed out that it was good that George Herbert Walker Bush wasn’t born in my county because his middle names wouldn’t have fit either. The county clerk didn’t find that amusing.

Fast forward a number of years to last year when I applied for TSA PreCheck status. Guess what? TSA in 2014 had surprising character issues. I was told that if I wanted to continue through the application process, I needed to formally shift Sue into my middle name or they couldn’t process my application since “space” was considered a special character and wasn’t allowable in the first name field and they couldn’t run PeggySue all together because it wouldn’t match my passport. This meant changing my name on all my frequent flyer profiles.

Additionally I’ve had hundreds of issues with patient matching during system migrations and data cleanups, not to mention the HIE project. How many people with either longer names or those that are more complex than mine will look like different people depending on how the data was entered? Of course there will be other matching criteria, but it’s going to take a lot of work to meet their stated goals of reducing system duplicate record rates from 2 percent in 2017 to 0.5 percent in 2020 and 0.01 percent in 2024. It should be interesting to see where already strapped (or cheap, as the case may be) organizations find the resources to get it all done.

I still have more to read, but at least it’s more interesting than MU rules. What’s your take on the Interoperability Roadmap? Email me.

Email Dr. Jayne.