Using Patient Identifiers in HIEsHow Matching Patients to EHRs Could Improve Care
Whoever is taking care of a patient "needs to be able to have the whole picture of what's going on with the person so that care can be managed appropriately," says Laura Landry, the HIE's executive director, in an interview with HealthcareInfoSecurity.com's Howard Anderson (transcript below).
When matching patients to all their electronic health records, the California HIE now relies primarily on a master patient index, which uses algorithms to match demographic information so all appropriate records about a patient from multiple organizations can be located, Landry says. But that approach is far from fool-proof, she explains, because, for example, multiple patients in the same city can have the same name and share other characteristics. And mismatches can lead to privacy violations.
By using the patient identifier, the HIE hopes to help advance its goal of creating "a virtual medical record across multiple organizations to best support the patient," Landry says. The pilot, funded through a grant from the Robert Wood Johnson Foundation (see: HIE Tests Patient Identifiers), runs through mid-2012.
If the identifiers work well, Landry says, "We free up the physician from worrying about 'do I have everything I need to know about this patient' to focusing on 'how do I improve their health and well-being?'"
In the interview, Landry:
- Describes the goals of the project. "We just need to know, does this solve the problem or doesn't it? If it does, we can put more resources into it and refine it some more. If it doesn't solve the problem ... we need to know it's a dead end and ... look for another solution."
- Outlines how the identifiers could eventually enable patients to designate who they want to access their records.
- Explains that when the test concludes, the HIE will prepare a detailed report on the lessons it learned. "This is truly a demonstration project ... focused on what works, what doesn't work, what did we learn, and what do we think the next steps need to be in order to make [identifiers] useful ubiquitously across populations."
Landry is the founding executive director of Western Health Information Network, formerly known as Long Beach Network for Health. She has 24 years experience in developing technology solutions for government and healthcare.
Health Information ExchangeHOWARD ANDERSON: For starters, why don't you tell us a little bit about Western Health Information Network, including the organizations it supports and the transactions it accommodates?
LAURA LANDRY: Western Health Information Network is a non-profit organization that was founded in 2007, originally named Long Beach Network for Health. It was a project in the public health department focused on getting data shared to the appropriate location, including for public health purposes. We became a member of the Nationwide Health Information Network File Implementation Contract Group, and we developed our health information exchange along with the Nationwide Health Information Network Connect Gateway. Now we are actually supporting organizations in both Las Angeles and Orange Counties and we are supporting two health systems, Providence Health and Services California and Citrus Valley Health Partners, as well as several community clinics.
What we do is make information available through the health information exchange ... transactions like demographics for patient identification and lab results, transcribed notes, medications and radiology results. The idea is to create an aggregated, virtual patient record across multiple organizations to best support the patient and what it is that they need in order to be taken care of properly.
Patient IdentifierANDERSON: You recently announced plans to test the use of the Voluntary Universal Healthcare Identifier System from Global Patient Identifiers Inc. What was the motivation behind this pilot project, and why did you choose this particular technology?
LANDRY: One of the things that is key to making cross-organizational health information exchange work is good identification of the patients in multiple organizations. We have a really robust master patient index; however, that is only part of the equation. It gives an algorithm and it determines whether or not it thinks that these patient demographics and those patient demographics are the same patient. In order for us to use it at the community level, we need to have a really high level of assurance that the patient is the same patient. We have a very low threshold for matching, so it has to be a very high percentage of match-ability.
For instance, I'm somebody who has three other people with my name. One of them has the same middle initial as I do, lives in the same city as I do, is the same age as I am and is also born in November. We've got a lot of similar demographics, and she also has a brother named Mark, interestingly enough. That's a really hard thing for one of these master patient indexes to [sort out], and we have a lot of [cases] that. We want to be more certain, and the way we get more certain is we include the patient, or the consumer, in helping to identify themselves at these different locations. So GPII [Global Patient Identifiers Inc.] presented itself as an opportunity because we know that we have this problem and we have to solve it somehow.
There are lots of different options we could have tried, including a volunteer one where the patient chooses their identification. But what I liked about GPII is their job is to make sure that an identifier is active, to know what community the identifier is active in and to connect the organizations that have the demographics for that patient so that they can say, "Yes, this really is the same patient." They've done a lot of work on standards. ... It was a natural opportunity for us to go to the Robert Wood Johnson Foundation and say: There is a lot of concern about whether this would work. Let's make it voluntary. ... And if it works and the patient says "here is my identifier at this location" and then goes to a hospital and comes through the identifier, that gives us one more piece of data and we can weigh it pretty heavily. We can say this is a very high matching probability. If I take this to each of these locations, that is a pretty clear indicator that Laura Landry with this number is probably the same Laura Landry in another organization as long as she has that number over there too.
We are giving it our best shot. We are going to see if it really works. The answer isn't necessarily that it has to work. We just need to know does this solve the problem or doesn't it. If it does, we can put more resources into it and refine it some more. If it doesn't solve the problem, and if it ends up not being the solution, we need to know that it is a dead-end and we need to look for another solution.
Privacy IssuesANDERSON: Over the years, some privacy advocates have expressed some concern that the use of any kind of patient identifier could jeopardize privacy. But you argue that the identifiers give patients more control over the privacy of their information. Can you explain how that is the case?
LANDRY: I can't talk about what the privacy advocates believe. I can only talk about what I believe. I'm somebody who has been in several jobs over my 20-year career. My records are scattered among many healthcare payers. And I had a pretty significant health event back in 2000 when I was working with another organization, and most of that information is actually not available to me, plus I wouldn't know how to go back and get it. I have no way of tracking whether those people are using my information appropriately or not. They have no accountability to me, because I'm no longer their member.
The way I look at it is, if I have an identifier, it's my identifier, which I manage and make people accountable to. Eventually we can do things like report against who has been accessing my data in the health information exchange. Who has claimed that they are one of my physicians? Over time, we'll be able to build applications that support the consumer auditing their own access of records and allowing for their own access of records by the people they want to access it. Unfortunately, the way things have been built now, with all the constraints around data moving, it doesn't just protect it from when we don't want it to move, it also protects it from when we do want it to be shared. That is what we are trying to solve. We are trying to solve how we take out of the equation the difficulty of making it shared when we are the ones who want it to be shared.
Quality of RecordsANDERSON: And how could the identifier play a role in improving the quality of records? Is it by making sure that all records associated with the patient are available when they are needed?
LANDRY: Somebody with even a moderately complex disease, who has multiple providers, case managers or even physicians for different issues, needs all of their data to be viewed by at least one person, whether it is a care manager or whether it is their primary care physician. Whoever it is needs to be able to have the whole picture of what is going on with the person so that it [care] can be managed appropriately.
I am currently taking care of my elderly parents. They are in their eighties and they both have a regime of drugs, doctors they have to see and tests they have to take on a regular basis. And their primary care physician manages all of those things. But there are people who don't have that. What I see as a benefit we get out of it is that when we know this is truly all the information, we free up the physician from worrying about, "Do I have everything I need to know about this patient," to focusing on, "Okay now that I know what I know about the patient, how do I improve their health and their well being in their daily life?"
By aggregating the data, making it available to the people who are accountable for taking care of the patient and making sure that the patient has access to what they need access to, they can see what it is that their doctor recommended they do. I think all of those things play a part. And at the root of solving that problem on a societal level is being able to identify the patient on the one hand. On the other hand, allow the patient to say, "Here are the things that I want people to know about me. Here are the things I don't want people to know about me." You can do that through this mechanism because they have different levels of identification. There is a private identifier and there is an open identifier.
Demonstration ProjectANDERSON: Finally, once your pilot, which runs through mid-2012, concludes, will you be issuing a report on lessons learned and making it available to others?
LANDRY: Absolutely. That is a core part of what we're doing. This is truly a demonstration project. It is a pioneer project focused on what works, what doesn't work, what did we learn, and what do we think the next steps need to be in order to make [identifiers] useful ubiquitously across populations.
This is a very small, focused project that will give us some answers. Things as simple as: Will a patient want to carry a card around? Once they have a card, will they actually use it at multiple locations? Will they remember? Things like that are so important. We will be publishing this. It is part of our grant agreement. We're absolutely going to be making sure that everybody knows what we learned in this project.