Zafar Chaudry, MD, has been involved in health IT for more than three decades, working in both the United States and United Kingdom. So he has a unique perspective on how patient data exchange operates in two totally different healthcare systems.
Becker's recently talked to Dr. Chaudry, now senior vice president and chief digital and information officer of Seattle Children's, as part of its series examining what it will take to make electronic health data fully interoperable in the U.S. (Read parts one, two and three here.)
Note: The interview has been condensed and lightly edited for clarity.
Question: What do you think it's going to take to get to full data interoperability in the healthcare system, or at least close to it?
Dr. Zafar Chaudry: Before you get into the technology piece, it is a cultural challenge to do a lot of the work to get organizations to think about data in a particular way. Lots of data-sharing agreements make this stuff prohibitive. It's not as easy as, technologically, we just turn a few things on and information is shared. Every organization has protocols, data-sharing agreements. Legality is a big issue. So there are lots of stakeholders in this, right? You've got the healthcare providers, you've got the clinics, you've got the insurance companies, you've got the software providers themselves. So it's pretty siloed.
In my experience, the legal components take the longest time versus the technological components. Most government-certified EHR-type systems that we use across the country have certain technical and functional capabilities. The government has defined how systems should talk to each other. Now, that doesn't mean they all do it the same way and you can connect them. And that doesn't mean the software vendors have made it easy. When you're buying an interface from one system to another, you would think it would be rinse and repeat. And it actually isn't. You actually have to pay — there are fees involved, there's professional services involved. So that causes delays in time.
When you look at interoperability, there are probably three ways you look at it. There is, can one EHR system receive data from another? Yeah, pretty much that can happen. Is it able to interpret that data? Not always, because every system has different tables, fields and how it recognizes that data. And there's the semantic piece — can two or more systems exchange the information, and can the exchanged information actually be used?
Technologically, I can send information from my system to another. We just moved to Epic and we can send information to other Epic systems and Cerner systems. And can we receive information? Yeah, we can. It takes some configuration and some interface costs.
But then what do you do with that information? Is it in a format that makes sense to the clinician? That's when the cultural pieces come into it. And how long does it take for us to agree between organizations whether that can or cannot happen?
So there's more than just one issue. Many people will say, let's just blame the software companies. Yeah, I think they have a hand in it and they could make it easier for all these things to happen. But then at the same time, if we truly believe that health information should be exchanged, why is it so complex to agree to that from a data-sharing perspective?
In terms of definitions of standard datasets, that's been done. There are different formats to receive data: HL7, C-CDA, etc. Many systems have robust APIs [application programming interfaces].
The other question you have to think about is, So even if I send you the data, will you look at it? The average clinic visit in our organization is, what, eight to 12 minutes? So if you're looking at data on a patient that has, let's say, a long-standing chronic illness, do you have enough time to look through that? Or are you going to do all that prep work the night before? I'm not sure if you have 15 patients you're going to do that every night the night before to figure out what the patient has. Hence, the concept of when you go to see the doctor, the doctor says, "So why are you here?" Even though they might be looking at the data right in front of them.
And so the bigger question in my head is always, What information do people need to provide care for the patient? During this pandemic, we went from Cerner to Epic, and I asked our physicians, "Well, what data do you need from the old system (i.e., Cerner)?" and they said, "Well, we need everything."
I said, "OK, let's leave Cerner on while we move to Epic." And we did, for a six-month period of time. We did that to try and understand how many docs are actually logging into that system to see that historical data that's so critical that we have to bring it across. And it wasn't a huge uptake. So then we decided to not portal the data but move all the data to an archive, and then connect the archive to Epic. That way you can still see it. But it wasn't like it was being used at 80, 90 percent. Less than 10 percent of people were looking at that historical data on a daily basis.
So one could argue, alright, so if you do have all the information flowing into the system, do you need it all, or do you need it from a legal perspective? And do you only need to highlight certain things like previous conditions, medications that they've been on, lab tests that they've had the last six months? Maybe that's enough to make future decisions for the patient. Because if you had a lab test 12 years ago, I'm not sure that really is going to help you provide care moving forward.
So we have been able to exchange data and most organizations can. I'm not sure anybody's done the study that says, Well, if you did exchange all of your data, what impact would that have on the patient outcome itself? And isn't that why we're here, to exchange the data so that we can improve patient outcomes?
Of course, we should stop transacting data on fax machines. I'm totally 100 percent on board with that. But at the same time, you're pretty much exchanging data only with organizations that you predominantly do business with or get referrals from. We get referrals from all over, but typically Washington, Alaska, Montana and Idaho. And most of those health systems already have Epic, so it's quite easy to exchange the data. Prior to that, we were doing a lot of faxing.
It's quite interesting because when I worked in the U.K., they have a national [health information exchange called] Spine. So all the patient data, if you're a patient in the U.K. — it's one single health system — goes into the NHS [National Health Service] national Spine. And the national Spine receives data from multiple systems. But the only way that happened was the government said, "If you want to do business in this country, your system must be able to send data to the national Spine, so that when a patient shows up in one city or another, the doctor can log into the Spine and see the information on you. And if your system can't send that data, or receive that data, then don't let the door hit you on the way out."
And that's obviously something you can do in a socialized environment. But it also proves the simple fact that technologically, if vendors want, they can send the data and they can receive the data. But what's the incentive here? ONC can have a set of guidelines, which they do. But then what's the level of enforcement, right? How do you enforce that on a particular vendor without them saying, "Yeah, this is gonna take time, it's gonna cost you money." And the number of times you have to pay for an ADT [admit, discharge and transfer] interface is just mind-boggling with vendors. It's a good money-making model.
Q: Was the system in the U.K. effective? Was it easy for providers to use?
ZC: Yeah, they pretty much got rid of fax machines, and the last set of fax machines that were left the government set a target to shut that down too. And you have to comply because your health system is paid by the government. So, yes, sometimes the big stick does help. But at the same time, it's more to force vendors to have standards, and then have that information interoperate.
Has it improved patient outcomes? Well, you can see the patient information. I'm not sure they did a study that said it improved patient outcomes. But, yeah, you could see the information for that patient. And the same vendors that provide systems there provide systems here. There are not a lot of people in the space, right? So they can do it.
It's IT's responsibility to make sure that the technology interoperates, and if not, discuss that with the vendor to try and get that to happen. But it's the organization's responsibility to culturally adopt that information and figure out what to do with that information and how that information will impact patient outcomes, and then it's a legal responsibility to make sure you've dotted all the I's and you've crossed all the T's so that information is being used for the right purpose, that it's not being monetized without the patient's permission.
Because that's the other thing that people worry about — my information is my information. I want you to use it, but I don't necessarily want you to sell it. And we've made information more available to the patient because, as part of our project to shift systems, we launched MyChart, but at the same time we launched Open Notes. So our parents and patients can see everything that goes into their record, and that's been well received. Because people want to know everything about their child. Not every organization does that, but you can and it's possible to do.
Q: Why did you switch to Epic?
ZC: We effectively used Cerner for 15 years. And it was time to reevaluate. And so we did ask our clinical teams to reevaluate — at that time it was 2020 — what would work best for our organization. And the clinical staff decided that they wanted to move to Epic. My job is to make that happen. It wasn't like people hated a particular system. It's "what do you think you need now?" Everybody should reevaluate. Fifteen years is a long time to have a system.
Q: If we got to interoperability, what do you think would be the ideal uses for it?
ZC: As people roll out social determinants of health, it would also help with that. So the equity issue is also big. Are people getting all the things they need to help improve their health when they move from one health system to another?
Information, in effect, is power. But once you've collected it, it still goes back to, So what do you do with it? And how does that impact the patient outcome? Of course, all of it needs to be done in collaboration with the patient. Because every patient is different. Some patients want you to use all their information. Some patients don't want you to use all the information. The whole privacy component still comes in.
But, yeah, it's life-saving events. Certainly at the top of my list, if I'm on holiday somewhere and I go to the emergency room, they can see key things about me and therefore they won't give me penicillin, which would cause me a major allergic reaction, and I'm unconscious so I can't tell them. I think it's already used for those types of purposes where people have access to the information.
But there isn't a national way to, if you're an unconscious patient, look you up by name. If you're not from that state, they might not be able to look you up. The difference in the national Spine in the U.K. is everybody's identified by a single identifier, which is the National Health Service number. It's similar to a Social Security number. And so they start there to look you up and if they can't find you, then they look you up based on date of birth or address. And if you don't have a national identifier number, because you may be an illegal immigrant, then they'll struggle to find you too. There has to be something where they can track you down. And it's similar here: It starts with your name, date of birth. If you happen to be in the local health information exchange, or in the Epic system you might already be a patient.
Q: But to be in the health information exchange, you'd have to be from that state or in that state's exchange?
ZC: Yeah, you would be, right? And if you're unconscious, how would I know? Only your ID will tell me you're from out of state and then maybe I would know which state and then I could look at that state if I have access to that state's health information exchange. But if you're not from that state or you have no ID, then it's still the same scenario of I have to use my clinical judgment till the patient or family member or someone taking care of that patient can tell me.
And we don't carry the equivalent of a mini health record on a credit card. Ideally speaking, if you could embed key medical data into a standard document — let's say like your driver's license — then they could scan that and have that basic information at least.
Basic information to me is, are you on any medications? Do you have any chronic illnesses? Do you have any allergies? And what's your blood type? Now, of course, your driver's license, they can look up whether you're a donor or not. They added that functionality a while ago. But they can't determine anything else.
Is interoperability about, Can I see the information that you have with your care providers? Or is it really emergent care, and can you opt in to have that information placed on a document that you may carry at all times on a scannable chip?
Q: Have you ever seen anything like that discussed?
ZC: I haven't seen that. It's quite interesting because when you look at other parts of our lives, we carry that information with us. So if you're traveling and you carry your passport, your passport has information in it, right? They have a chip, they scan that chip, and it tells them a lot about you: your name, your date of birth, your address, when you were naturalized, if you were born in the United States — that's all on that little RFID [radio frequency identification] chip.
You could carry these Google records or these Apple records and you could input some of that key information, which could then be used in emergent care, but there's still voluntary opt in, right? And it's still linked to your smartphone, and whether the smartphone is with you when you have an injury. Of course, the smartphone could be locked and they'd have to use your face to unlock it to see that. Or you might not have power on your cellphone for them to look at that information.
It's not as common as everybody's got to carry a state ID.
Q: ONC has put out a 2030 goal for improved interoperability. Do you think that's reachable, or is that just an arbitrary date?
ZC: Oh, I think we will reach it. I think we're already seeing progress in the interoperability space. I don't think it's as dire as it used to be.
The question will really be, will they then enforce it? So there comes a time when somebody says, "Hey, I'm ONC and I'm going to give you this guidance level: By this year, you should achieve X." But it's sort of like the price transparency thing, right? Nobody really actually enforces it.
What gets us going when we wake up is, "Oh my god, am I gonna get fined? Then I better do it." And so I'm not seeing that level of enforcement yet. Maybe it's coming. But if they're giving us another five to 10 years to do something, you would think that we should be able to get that done.
And the concept of, is technology going to hold us back, or are we going to information block or any of those things — we should be working them out now. But if we get to that date and people still haven't done it, what's the scenario? What would happen?
Ideally, you would say it's the right thing to do, outside of what anybody tells you. If you work in healthcare, I would argue it is the right thing to do for me to provide information that could improve the care of the person that I'm trying to serve. That's how I look at it. But it's always more than that. When money's involved, it's always more than that.
Q: It might be costly to make systems interoperable, but wouldn't it save money in the long run by cutting down on unnecessary tests and other care?
ZC: People will argue the money concept. I would argue the safety concept. For us as a healthcare system, when we went down this journey and said, "We really need to interchange this data. Who do we need to interchange it with, based on who sends us patients?" — the focus we had was purely on patient safety.
And yes, it costs money. Interfaces between systems cost anything from $5,000 to $50,000 at a time. But is it the right thing to do? Yes. And does it improve patient safety? Yes. And do we have data that we have now interchanged hundreds of thousands of pieces of information? Yes, absolutely.
Can we talk to clinicians in our emergency room and say, "Were we able to pull the data from that patient who was a previous patient of another health system in this state?" Yes, we can. That's great. It's very simple. And we're able to do that. We weren't able to do that before. We would have to wait for a fax to come in from that facility. That took a long time. Now it's pretty much instantaneous.
It saves money because I don't need to buy hundreds of fax machines. I don't need to buy toner for fax machines, which you can't find besides on eBay half the time. And there's also the administrative burden, because at the end of the day if you're receiving reams of paper, you have to have people to process that paper and send it to the right place. So we don't need those folks anymore.
So we've seen that financial benefit, but most importantly, if you're making split-second decisions, and having that information almost instantaneously, it's pretty much a no-brainer for why you would interact with this information.
Lots of admin time is saved from moving from paper to electronics. But the flip side of this would be as we move more electronic, systems must always be on, right? The EHR just has to be on all the time. Because when it's not on, then there's a crisis or hell to pay because people are so reliant on the technology.
So I've seen a higher demand on technology services to make sure that the system is always on. There was a time when 99 percent uptime was good. Now, for our system, I have to guarantee five nines to my clinicians because they may only accept a few minutes a year of downtime, not an hour to three days. It's not acceptable. Because you can't now practice that way.
Q: What else could get the healthcare system to interoperability? Or is it just too complicated to boil it down to just a few things?
ZC: I still center my comments around the three concepts: There's the technology piece, there's the cultural piece, and there is the "what happens to my data?" piece.
We're all striving to get that 100 percent electronic exchange. I don't think technology is the limiting factor here. It's more around: What does an organization truly need? How do they negotiate that with the next organization? And then how do you make it happen in a legal way and where you have patient input into that decision making?
Otherwise, people are sitting on mountains of data, and patients get really, really upset when their data is used without their consent, right? So we do have to bring patients into the conversation as well, which happens and then it doesn't happen and then it happens. And so that's where you get those, "This X health system has been partnering with some technology company using data," and then people get upset.