Q&A: UHIN pilot program targets rural vets
 |
“From get-go, UHIN had a vision of not just administrative, but clinical data sharing, too,” said Jan Root, PhD, president of UHIN, during a recent interview with CMIO. Now, with technology partner Axolotl, federated databases, and secure information exchange, UHIN is poised to pilot data exchanges with the largest healthcare organization in the country.
What will this program do?
The VA has a major data center [in Utah], and a lot of their IT people here. They were interested in the concept of trying to connect the dots [for veterans to get care in a rural area]. The [Community HIE, or cHIE] connects the dots.
For billing reasons, veterans don’t go to [their local] clinics in Moab, they go to Grand Junction VA hospital [in Colorado,] because it’s just easier. That, of course, creates a hardship, particularly if you’re in any kind of distress.
We wanted to connect to the VA data. If we can connect the VA data to UHIN’s Axolotl hub, you connect everybody.
When is this pilot program expected to launch?
We want to start a pilot by March, to go live in the summer. [We’re] starting in Moab, where there’s one small hospital, Allen Memorial Hospital. Once we’ve completed the pilot, the program is immediately extensible to the entire state.
How will it work?
The VA spent years creating essentially a single data warehouse. If a vet shows up in the ED, the doctor queries the system and can see when [he or she] had care at other VA facilities. That’s the advantage of having a single warehouse. … The nice thing from UHIN’s perspective is, although the physicians who will be involved in this are the physicians at Grand Junction VA facility and physicians at Allen Memorial. If physicians at Allen Memorial do a query, they’ll actually be querying the entire VA system.
During the pilot, the Grand Junction VA physicians will only be querying data from Allen Memorial. But once we get past the pilot, then we can make that available to any physician who’s on the cHIE, who’s caring for a veteran.
Approximately how many patients will be included in this effort?
We estimate the maximum is around 700. The VA has an opt-in policy. I don’t know how many will ultimately decide to join. [In a separate HIE pilot program] in San Diego, they had a pretty high percentage, somewhere in the high 80s.
What data will be exchanged?
For the purposes of this program, the VA wants to limit it to a CCD. This turned out to be a challenge because the VA is a single entity [and keeps] data in a single warehouse. UHIN, through the work of Axolotl, is creating an aggregated CCD that ultimately will come from many data sources that are connected to the cHIE. So if a veteran receives care at Allen and they live in Salt Lake City, [and get care] all outside the VA, when VA queries for that person’s information, Axolotl has created an aggregate [and can] pull data from 17 CCD buckets—patient demographics, insurance, history, labs, problems, allergies [and others]. They would query all the sources [at which] this person has received care. Then instead of sending the VA just the lab work or the medication list, they would put it all into the aggregated CCD.
Can the HIE accommodate images?
The cHIE can handle many things. We are looking at [Elysium Exchange, a centralized image storage system] as well as some alternatives. Axolotl is working with a partner and our community is looking to see is there enough pain here that hospitals and clinics would elect to participate in such a system and financially support it.
What precautions will be taken to ensure privacy and data security?
Data are encrypted, only authorized users have access, and we have a system for determining if a person is a real physician or not. The feds are [also] weighing in on some standards for that.
UHIN’s policy is that all data both at rest and in motion is encrypted to federal standards. Although hackers catch the headlines, the usual culprits are authorized users. So we have developed a couple of approaches and we’re not done yet. We’re also developing a system of audit flags. How do you know that Dr. X is inappropriately accessing someone and they’re not a valid patient? That’s a kind of nut that IT is not good at cracking.
We’re trying to work with the knowledgeable members of our community to see if they will share with us what they do. For example, one of the flags that one of our members will throw is if [a doctor] accesses information about a patient [with the same last name]. However, I don’t know what they do about names like “Smith” or “Johnson.”