So, you basically gone through the stack, we are going to say a little bit about technology today and later. But I want to revisit interoperability because now that you so sophisticated about the entire stack, I think you'll understand issues of interoperability that we couldn't address especially about data information and knowledge. Let's start off with text data. In order to think about that, you can see where we are, we are at a high level of the stack than we were in talking about exchange of data. I want to start off by talking about envelopes in the US Postal Service. Now, you may remember letters, these are things that you wrote on a piece of paper with a pen. This is the device that deposits a pigment on some surface. You fold it up your letter, you put it into this thing called an envelope and I did not know until preparing for this type of discussion that the Postal Service has a standard for where things should be on the envelope. So, now you can see that the Optical Character Recognition software used by the Postal Service is expecting to see the name and address in a certain area. It's expecting to see the stamp in a certain area. If there's bar-codes, expecting to see that the area. If it is a return address, it's supposed to see that in a certain area. Probably this answers the question that, the cheaper the return address on the front of the back. I think put it on the front. So, an information exchange standard is very similar. It's not telling you about the data inside the system, inside the package. It's telling you about how the data are packaged. The most famous example of this is called HL7. In what's called the HL7 messaging standard, we often call version 2.x there's 2.2, 2.3, 2.5, 2.7. I think you get the idea. Different environments use different standards and they claim to be conforming with that particular standard. Another is a Clinical Document Architecture. This is a structure for documents that was dictated by the meaningful use legislation. Turns out not to have been as successful as we've wanted for the purpose of communicating documents, and in fact, the new golden child for communicating more complicated data then the message is fire. We mentioned fire earlier in a previous lecture and we're not going to be saying much about it today. Why is it called HL7? This gives me a little opportunity to bring in technology. The core technology in informatics today is actually the network and the exchange of data occurs at the technology level, and when we talk about network connectivity, we actually talk about this stack of protocols. But the reason why I call it the informatic stack is actually informed by this very picture. In brief, when you send an email from you to your buddy, your work in the application layer right whether it's Gmail or Outlook or Thunderbird or whatever it is that you work in, [inaudible] application. You're typing with a certain character set. It could be Unicode, It could be Ascii. So, that's called a presentation layer. You may be attaching documents in certain formats. You may be having images in certain types of formats. All that would be called the presentation layer. Below that is the session layer where one machine connects to another machine, and below that is what's called a transport layer. That layer takes responsibility for your email getting from you to your buddy. But you don't know about is that deeper than that, the next level down at the network layer is that the machine chops up your email and your files into little pieces, it packages each piece into a packet and then that's what gets sent down the network. So, the next layer down is connecting that packet of information to the physical medium that's going to be sending it down. That's the data link level and finally you get to the physical layer or you sending it by a cell phone in which case you're using radio waves. Are you sending it at work and you're connect by internet with copper or using fiber optics or using satellite? Well, that is the physical layer. So, and then what happens is when your buddy gets your packets, it actually waits for the packets to all arrive, then assembles them at its transport layer into one email. Then it makes the connection to your computer. It ensures that's in the standards that the image and file and character set standards that you've were determined and it presents it to you in the application. So since, the health level seven is exchange of EHR data of people using the EHR application or the EHR modules, that's at the application layer. So that's why it's called HL7. The standard for exchange of data at the application level. Even though it's exchanged at the application level, the way it exchanges it's pretty dramatic. This here are five chunks of information about a patient's admission to the hospital and I think you can intuit that there's a bits of pieces of information between these vertical bars, and probably the position and the bars mean something. You see this ADT or a O4 MSH, these are probably codes for something. So, to parse this format. So you see that between the vertical bars are fields. That whole thing that's called MSH is a segment and multiple segments make up a message, right? So you can see that this is a message about an event called an AO4 event, which has better admissions about that particular patient, and it's informing the hospital who the next of kin are in case a patient I suppose dies or more importantly, who's responsible for the bill, and finally in this case, what visit prompted the admission. So, this format that you see in front of you plus the vocabulary for the different segments is what comprises that 2.x standard, and so if one hospital says we support 2.5 or 2.7, they are referring to the types of messages that they can deal with. I talked about admission, that's a type of domain, that would be patient administration. But there are all these other domains that HL7 deals within each one of these bullets is actually a book, that you can look up to read one of the types of messages what are their formats? What the rules? What's mandatory? What's not mandatory? It's quite a big deal and managing this message is again the messages to mean, what the conditions want these messages to mean is a huge chunk of the work of any IT operation in any hospital or a patient-care system. The ONC produces a report every year or so about the adoption of different standards and you can see that HL7 has the highest adoption of anything. So, if you saw this report for fire, you wouldn't see so many bullets. The big problem is that even though it's supposed to be plug and play, it isn't plug and play. The idea is supposed to be that if I'm in one hospital and you're in another hospital, you speak 2.5, I speak 2.5, we should be done. I connect to the network, you connect to network, I send you data everybody's happy. Unfortunately, it doesn't work that easy because if two different modules even within the same hospital, that they both support 2.5. They both comply with 2.5. They still one may be missing fields at the other expects. One may put data what expects from one field to put another. One they have different formats for the exact same thing and you wouldn't be surprised to hear that sodium levels might be named differently. Almost use the link codes, but something like a timestamp, even that could be different between different modules than on different information systems. Two point five, 2.3 different versions, even the mandatory fields can be missing up to five percent of the time. So, making sure that you know who this message is about is not a trivial task. Some people may not put the fields together. I feel really sorry for the person has to count those vertical bars and say, "Okay, I need to put something between a 10th and the 11th bar." How many times do they have to account those bars to make sure they get to the right place? This is a subtle issue, but that even though the data on a network it's really liked this puzzle system, I am sending data to you. Now, that sounds fine. But let's say on the lab system, right? I want to inform many different people about the lab value because so many different people taking care of this patient. That is not point-to-point. That's point to many. It's a broadcast. So, HL7 has to be dealt with to give that level of functionality. So in summary, HL7 is the text information exchange standard in place today, and it looks like it's going to be around for a while. It's not perfect. Its complicated and does require attention if you're going to be in that field.
