The Informonster Podcast
Episode 10: The Architecture of Intolerance: Discussing How Healthcare IT Documents Substance Intolerances and Allergies
September 29, 2020
In this episode of the Informonster Podcast, Charlie Harp talks about the origins and importance of allergies and tolerances in Healthcare IT, as well as what makes them unique from other terms within the system. He then discusses how different systems handle allergies and intolerance as well as the issues that come up with them and how he thinks they could be solved.
Have a question or topic idea?
Get our News and Updates
Get notified about new podcast episodes, upcoming events and webinars, and more!
I’m Charlie Harp and this is the Informonster Podcast. Today’s Informonster Podcast is going to be about the architecture of intolerance. And by that, I mean how we document substance intolerances and allergies in healthcare, which is kind of a pet peeve of mine so I might go off on a rant as we get into this. So let’s start out with what is an allergy and what is an intolerance. So typically what we’re trying to document in healthcare are things that you want the patient to avoid, either for dietary reasons, environmental reasons, or medication reasons. We tend to lump them together. So whether they don’t like fried eggs and shrimp, or mayonnaise hits them in the gut, or if you give him penicillin they’ll going to anaphylactic shock, those things are all things you can document for different purposes. And typically you document the intolerance and you document some other things about the intolerance and the idea is that the people that are caring for the patient will see that, or the software will see that, and it will stop you from doing something that puts the patient in an uncomfortable or life-threatening situation. It’s not super complex, but it’s interesting the way it all works.
Sometimes there’s a debate on allergies and intolerances because an allergy is something kind of specific. It’s an immune response to some substance that can cause the patient to ultimately go into anaphylactic shock or have other reactions that could be life-threatening. Whereas an intolerance could just be, “Well, it makes me feel queasy,” or, “It causes me to be dizzy.” It could be a side effect of the drug that you might be susceptible to. But from my perspective, it really doesn’t matter whether it’s specifically an allergy or specifically an intolerance because really what you’re saying is if you give the person this substance, they are going to have this reaction. And if you spend a lot of time noodling over whether it’s an allergy or an intolerance, I hate to tell you this, but at the end of the day, it’s not going to matter. Either the human or the software that’s trying to avoid the reaction is not going to do it differently, necessarily. At least not in any clinical decision support module I’ve ever seen is not going to process it differently if it’s an actual allergy, versus something that is not an allergy but just an intolerance. And there are other things too, that when you think about these intolerances, you also have things like foods and substances and environmental. And I’m going to get into that a little bit later as to why I think those things are pretty different, but we tend to kind of group them together nowadays.
So let’s go back in time and talk a little bit about the history of allergies in healthcare and EHRs from a coded perspective. And so originally when the doctor was taking their notes from your encounter, they would ask you the question, “Do you have any allergies?” and they would write it down in the chart. And that way, when they’re going to prescribe medications or do other things, or if you’re in the hospital and they’re going to feed you, they want to know that you can’t have certain things, or if you’re taking a particular drug, (if you have) an intolerance or an allergy to a particular drug, they don’t want to give it to you because obviously they don’t want to put you in harm’s way. And that’s sometimes moderated by what your reaction is. So if you say, “Well, my reaction is mild. I get hives.” They might decide, if they need to give you a drug or a drug in that class, to give it to you even though it might give you hives.
But if you were to say, “Well, if I take that, I go into anaphylactic shock,” then obviously they’re going to avoid it. They might even avoid it if it gives you hives, just because they don’t want to have a worse allergic response than you had before. So initially they would write that down so that other humans that were doing things like prescribing meds and giving you food, could look at your chart and say, “Oh, we can’t give them that because they have an allergy to it.” But then in 1990, the Omnibus Budget Reconciliation Act, or OBRA ’90, created a requirement that outpatient pharmacy subscriptions had to be checked. There was a number of things in OBRA ’90, but one of the big ticket items was this drug utilization review, or DUR, and the things included in the drug utilization review, when you were giving a patient a medication, was duplicate therapy checking, drug interaction, checking, and, you guessed it, drug allergy checking.
As we’re doing more and more electronically, and we’re doing things in the outpatient world, we wanted to make sure that in a less attended environment, you know, they weren’t giving drugs that were gonna harm the patient. And so what happened when they released OBRA ’90 was it was kind of the genesis for drug terminologies and clinical decision support in the US. And what it did was it forced the evolution of several of the medication formulary publishers, like first DataBank, and Medi-Span, and Multum to turn into something more sophisticated, and that was decision support vendors because before OBRA ’90, and someone can check me if I’m wrong, a lot of these companies were really all about drug pricing. But they had the pharmacist, and they were doing the research on formulary and pricing, and they said, “You know what? We can do the research on interactions and allergies.” And so they started providing that data because, you know, if something’s mandated by the government, it’s a good market to be in. That was kind of the evolution of these decision support vendors. And allergies are not as simple as you might think. That has to do with the fact that how allergies are typically described (sic). So when you think about a substance that you’re allergic to, when you go to the doctor and they say, “Do you have any drug allergies?” – let’s stick with drug allergies for a second – you have a couple of options of what you can say. You can say, “I’m allergic to penicillins or quinolones.” So if you were told by your provider that you shouldn’t take a particular class of drugs, you might remember that. And you might say, “I’m allergic to this class of drugs,” which is a very broad “something” that incorporates and covers a lot of specific products in the marketplace. If you say, “Penicillins, send into anaphylactic shock,” that’s going to take a pretty decent chunk of products off the table as to what the provider might give you, both in terms of a direct allergic reaction and something called a cross-sensitivity reaction. It might not be a penicillin, but it’s known to elicit the same allergic response that a penicillin might. So there’s a checking when it goes to the class level. When you say a class that’s a big, broad bucket of products. You could also say that, “I’m allergic to codeine.” So you could say a very specific ingredient, even when that’s not part of a drug. There’s no single ingredient codeine on the market. Codeine usually comes in a multi-ingredient formulation. But you might say, “Well, I can take acetaminophen, but I just can’t take codeine.” And that’s a specific ingredient. That’s a very narrow item. The narrower, the better, frankly, when it comes to allergies because it gives people options. You could also say a formulation because you don’t know. I’m allergic to Vicodin, and that’s a multi-ingredient formulation, and no one ever told me which of the ingredients I’m allergic to. And you can even go broader. You can say, “I’m allergic to Pepto-Bismol.” What the ingredients in Pepto-Bismol (are) can change over time. So that’s going at it from a brand name perspective. It’s a much fuzzier thing than if you were to go in and say, “I’m allergic to acetaminophen,” for example. So unlike some other domains, when you’re interacting with allergies, you’re throwing out these concepts from multiple terminologies, potentially, and multiple code systems. And when you lay on top of that food, and environmental allergies, and the things that I call the “no-knowns” – in SNOMED, there’s no known drug allergy, no known food allergy, I’m going to vent about that later – unlike some other areas where you give an ICD-10 code, or a SNOMED code, or an RxNorm code, allergies is kind of a wider field of things. There’s different animals that live in that jungle, in terms of what you can say and not say.
Now, when you think about what goes beyond that, and when you’re talking about an allergy, when you say, “I’m allergic to X,” the question that follows is, “How bad is it? What happens when you do it? When did it first happen?” There’s this information model around intolerance, and it typically is: “What type of thing is it,” which is kind of implied. If it’s shrimp, we know it’s a food. If it’s Coumadin, we know it’s a drug. “
What is the substance itself,” which in most modern systems, it is a code system followed by a code like an RxNorm, RXCUI, or an FTB GCN sequence number, Medi-Span GPI, but it could also be free text because sometimes people get lazy and they just type in whatever they want. “All cillins.” I mean, I’ve seen people say they’re allergic to lions in free text allergies. Um, I think I might’ve even seen somebody be allergic to their mother-in-law, but probably was – wasn’t a true allergy.
The next thing is, we talked about it earlier, “type.” If you look at the fire resource for intolerance, it’s got a type: allergy, or intolerance. And once again, I would suggest it doesn’t matter unless you’re getting into some nuanced clinical decision support. For example, you may not go to a cross-sensitivity class if somebody says it’s an intolerance, but I would argue that if there’s a cross-sensitivity class, the person that said it’s an intolerance may not actually know it’s an intolerance. I just tend to fail over to worst case scenario. And I would probably always kind of lean that way unless I had a really good information that said it couldn’t possibly be an allergy. So don’t worry about looking across sensitivity classes, but I’m not a doctor. So, you know, you got to make your own decisions.
The next thing is, “What is the severity of the allergy?” Typically what I’ve seen is mild, moderate, severe, and life-threatening. And that’s important because if it’s a mild allergy, you might say, “Well, they really need this drug. If it was mild, let’s go ahead and try it and see what happens.” Whereas if it’s life threatening, you’re not going to do that because that, in all the clinical decision support, whether or not to surface something to a provider because alert fatigue is a real thing. And some people would say, “Well, I want to suppress certain things.” With allergy, the severity is a big deal because even if it’s giant hives, if somebody went to the trouble to say this is life-threatening, that should always put the brakes on something that’s happening relative to that thing that they cannot tolerate. I also noticed in FHIR, they have something called an “allergy criticality,” and it’s low, high, or unable to assess. That’s kind of a new thing to me, and I’m not really sure how that would be used unless it’s kind of like a life-threatening thing, where they just basically simplified it into a low, high, and “I don’t know.”
The next thing is the onset. “When did this first happen?” People like to document that if they’re being thorough. And then the other thing is really the “reaction.” I mentioned one before: You know, giant hives, or anaphylaxis, or gastroenteritis, nausea. So you basically saying that, “When I take this substance, or when I’m around this substance, I have a reaction,” and that can also be really important because, you know, if you start sneezing, that’s one thing. If you’re going to anaphylaxis, that’s another.
With an information model, and this idea that I’m collecting things in that way, when we talk about the substance itself, we have this weird anomaly in allergies that has been inherited through time. It kind of goes back to this idea of healthcare being very episodic, where, you know, we process one visit chunk, and then we throw the visit away. That’s historically how healthcare works. So this thing that I’m talking about is the “no-known” concepts for allergies. So we have a concept of “no-known” drug allergy, “no-known” allergy, “no-known” food allergy, “no-known” substance allergy. And if you’re living in an episodic world where you’re in that moment, and I say, “The patient has no-known allergies,” the reason for something like that is there’s a difference between no information and knowing that there’s nothing to talk about because if I just said nothing in the patient’s allergy list, it could be that I never asked them and they actually have a life-threatening allergy to penicillin, whereas if I document “no-known allergy,” I’m letting somebody know that I did ask the question and they said nothing that they know of, and so I’ve documented that. In an episodic universe – like kind of, you imagine the two-dimensional versus three-dimensional line universe – in an episodic universe, that works just fine because I asked you a question, you said none, and I basically am saying, when I document that, that you are allergic to no known allergies. Because when I put something in there, it’s like, if I ask what drugs you’re taking and you say, “I’m not taking any drugs,” I typically would not go into a list and add a known drug item to your medication list. I would leave it blank. In this case, we’re putting kind of a negative concept in there to say, “Hey, I checked.” That’s really what we’re doing, but in a longitudinal universe, in a universe where I’m pushing data around, the “no-knowns” cause a problem and I’m going to talk about that in a minute.
All right. So what I’m going to do now now is I’m going to talk a little bit about challenges we face when it comes to patient intolerance and tolerances in Healthcare IT. Now, I tend to lump challenges, when it comes to software, into four buckets: People, process, structured knowledge, and systems design. So the people aren’t doing what they’re supposed to do, we don’t have a good process in place to make sure things are happening, we need some kind of structured artificial experience at the point of care and it doesn’t exist or we don’t have it, or the systems aren’t well-designed. And so those are the buckets I tend to lump things into when I’m trying to figure out where the problems are in Healthcare IT.
And like with many things, the biggest challenge, when it comes to documenting and managing and tolerances, happens at the beginning. And this is when the patient is sitting in front of somebody in the process, and they’re tasked with documenting their intolerances. How they ask the questions, what they know, what the patient knows, that whole beginning can stop everything else from happening the way it’s supposed to. There’s no system on earth that’s sophisticated enough, at least that I know of, that can perform well if it’s got bad, missing, malformed, wrong data. So really the biggest challenge (is) with, when it comes to patient and tolerances, a lack of quality data capture. So when you’re getting the data, you’ve got a good source of things to document, the patient knows things, the person documenting knows how to characterize and knows what to document. They do all that, then that information is going to flow into the system and it’s going to be very powerful.
Now beyond the human challenge, when it comes to allergies, the next big challenge is that, historically, and you know, I, for 10 years, I was in the drug decision support world. That was my world. And I’ve been out of that for the most part for the last 13 years, but I know that it was still a big issue and it may still be a big issue, and that is free text. So people that don’t want to put in a documented allergy, and sometimes it’s easier to just type in whatever you’re thinking, and sometimes it’s you type, “Bee sting, bananas, Coumadin,” and then you might even type in “anaphylaxis” because you’re typing free texts. It’s stream of consciousness. You’re going to put in whatever you’re thinking about, and that is a problem. And we’ve done projects at Clinical Architecture where we’ve taken that free text apart and tried to convert it to codify data. A lot of times there are codes for all those things. It’s just the person doing the documentation didn’t want to go through the process of searching. And I know there are probably a number of really good systems that make that super simple and easy to do because I also know that people sometimes struggle with, “Well, what lists should I use? If I have a list it’s only ingredients and a patient tells me a brand name, what am I going to do? Look it up? I guess I have to look it up and find out what the ingredients are and type in all those ingredients.” So there’s also, sometimes, this conceptual disconnect between how the patient is articulating the intolerance and what I’m able to document. And that’s another big challenge, is there’s this conceptual misalignment where whenever you’re documenting something, the first place where you start to transform the data is that chair-to-keyboard interface of a human being deciding what they’re going to put in the system. In the case of allergies, if I were to say, “I’m allergic to penicillin,” and somebody types in penicillin as an ingredient, that’s fine. If they say, “Oh, well, they must be allergic to penicillins,” or I say, “I’m allergic to Vicodin,” and they decide that they’re going to put down that that’s acetaminophen, I might’ve actually been allergic to the other ingredient. There’s that first step of them interpreting what you said. And it might be that you say “I’m allergic to,” well, how about NyQuil? “I’m allergic to NyQuil,” and they go to look up NyQuil and NyQuil is not there because their system only allows for ingredients. Well, there might be like 38 ingredients in Nyquil. I don’t, I don’t know what the formulation is today, but somebody’s got to make a call at that point. And so they might say, “Well, I can’t find all the ingredients in NyQuil. So what do I do? Ah, go to free text and I type NyQuil.” And as soon as I go to that free text bucket, there’s a very good chance that that allergy is now slipping through the cracks because we’ve got so used to relying on decision support, I don’t know how well people look at the free text allergies. I would assume they’re looking at them and doing their due diligence, but the whole point of computerized decision support is things theoretically don’t slip through the cracks as easily. And if you document it electronically with codes, the system can do its processing and work through it.
I’ve seen a number of different systems that deal with looking and evaluating allergic reactions. Here’s typically the way it works. You put in something, whether it’s a class, or a brand name, an NDC code, or a drug, or an ingredient, and typically what happens is that gets turned into a list of ingredients. So if it’s a class, we say, “What are all the ingredients in that class?” If it’s a drug, we say, “What are all the ingredients in that drug?” If it’s a formulation or an NDC code, we do the same thing. We break it down into an ingredient list, and then when someone goes to prescribe something, they do the same thing with that. They look at the drug, they break it down into its ingredients and they do a comparison. They say, “Are any of the ingredients in the class they’re allergic to? Or is there an ingredient overlap? Because if they said they could be allergic to an ingredient and I’m about to give an ingredient, I’ve got to stop and evaluate that, depending upon the severity of the reaction.” And in some cases, if certain ingredients have cross-sensitivity classes, I’m going to check the ingredients in the cross sensitivity classes too, just to make sure that I’m not going to put the patient in harm’s way. So it’s a very straightforward collision pattern, where I’m just looking to see if the thing on list A is anywhere in list B. And if it is, I’m going to throw up a warning and say, “Hey, watch out. These things could be a problem.”
So this goes to the next point, and that is most systems today, when you’re documenting an allergy, are one of those lists that come from a compendia. So if you’re using Medi-Span, or first DataBank, or MULTUM, or whatever you’re using, and you go to use those, typically those things all come with like an allergy picklist which is a combination of brand names and ingredients and multi-formulation products, and technically you can put anything that is anywhere in the drug ontology for any of those guys because they all break down into ingredients that can be handled by the clinical decision support.
But what there isn’t really today in healthcare is a good standard for interoperating allergies. Before you get upset, let me explain what I mean. If you document an allergy as a drug, there’s a very good chance you can crosswalk into RxNorm. If it’s a class, you used to be able to maybe crosswalk into NDF-RT, but since that’s been deprecated, now you can maybe crosswalk into MED-RT, and you might also be able to crosswalk in a SNOMED CT. But unlike other domains in healthcare, there isn’t a, at least not what I would consider to be, a comprehensive list of things that you’re allowed to use for interoperability for allergies. I mean, you can always use RxNorm for allergies. You can probably use SNOMED, but you’re going to be operating with a list. I mean, at Clinical Architecture, we produced a list that we call the Common Allergy Target. And we’re kind of going through it again now because HL7’s come out with something; FHIR’s got something. We basically did exactly that. We took a bunch of things and munched them together for food, drug, class, brand name, substance, the “no-knowns,” and we merged them into a multi-code system value set, and we make that available to our clients so they have something that they can pivot on.
But that brings me to a funny story. Now, there might be people that were there at the time, so forgive me for the story that I’m about to tell. Several years ago, right after I started Clinical Architecture, there was a movement to use UNII codes for allergy. Now, for those of you that don’t know what UNII codes are, UNII codes are substance codes from the FDA that are basically ingredients for products. There was a movement to say, “Let’s use UNII codes for allergy,” and that’s not uncommon because we have this issue where, sometimes, when we need a new terminology, we kind of go through the drawer and see what we can find is that, “Oh, wait, we’ve this thing, UNII. Let’s use that. It’s got a bunch of ingredients in it.” But the problem with that is that most allergies that are documented are not ingredients. Most people don’t know the ingredient they’re allergic to. They know the product they can’t take, or the class maybe, or the brand name. When I found out about that, it was probably the very first time I got involved in a working group. So I joined the organization that was working on the working group for that, and I joined with one purpose in mind. I had to stop it from happening. I had to stop UNII codes from being the standard because UNII codes, whenever you take a terminology, and I’ve talked about this before, whenever you take a terminology that is purpose-built, in this case ingredients for products, and you try to use it for something else, you immediately start to create this broken thing. And in the case of UNII, it’s an ingredient list. What happens when we try to put brand names in there, since people document brand names? What happens when we try to put classes in there? All of a sudden we’ve taken this thing that’s relatively pure and good, and broken it when we start to introduce these other things. And by the way, all the medication ingredients are already in RxNorm. You know, you can document food allergies until the cows come home, but you’re probably not going to see that come up in a clinical decision support module. There might be a few, but if I don’t like broccoli, I’m allergic to broccoli, I’m not going to get an interaction alert or a drug allergy alert for broccoli. Well, I don’t know, I’m not a pharmacist. I might, but I don’t think I’m going to.
So I just wanted to stop it, and one of the points that I tried to make when we were having the discussions – and I successfully stopped it, and then I quietly receded into the background – but one of the things I pointed out is shrimp. So let’s say you’re allergic to shrimp. In UNII, there were 72 varieties of shrimp. So what are you allergic to? Are you allergic to a Micronesian Tiger Shrimp? Are you allergic to, you know, the Galapagos galloping green shrimp? I don’t think that’s a real thing. I think I just made that up. The bottom line is, it was a great exercise in what not to do. I don’t blame the people that were trying to do it. They’re doing what we always try to do. We say, “We’ve already made an investment in this terminology. Can we use it?” But when you do that, you kind of have to ask yourself the question of, “what is the consequences of doing that?” And it’s like I said, my objective was to stop it and it was stopped. And so you might blame me for the fact that there isn’t a standard, but you know, I’ll take it because if we’d gone forward with UNII, I don’t think it would have been a very good standard. And I think what we really need to do in healthcare is probably get a good handle on medication classes in the public domain. And that’s been a debate in the RxNorm world, and with MED-RT, I think that’s evolving. And I also think that with SNOMED, there’s a good opportunity We could do things there. I think my perspective is if we don’t have something, and we’re trying to think of where to put it, and it’s a healthcare thing, that’s what SNOMED is supposed to be, as far as I can tell. I would say let’s put those things in SNOMED, because it’s something we can all get. They do a good job. But anyways, I’m pontificating.
So I talked about this a second ago, but one of the things I think we might consider doing, going forward in healthcare, is separating out intolerances, as to whether it’s an environmental, or a medication, or a food intolerance, into their own bucket, or environmental, for that matter. The reason I feel that is if we combine those into a list, you still have to identify what it is if you want to be able to route it for decision support purposes or for nutritional purposes. So if you’re in the inpatient environment and a patient’s allergic to a food, you’re routing that to the people that are controlling their diet, ideally. They’re already kind of separate, and the people asking the questions, it’s just as easy for you to say, instead of, “Do you have any allergies,” to go through the list and say, “Do you have any food allergies?” Document it. “Do you have any environmental allergies?” Document it. “Do you have any substance allergies that I should be aware of, like latex?” Document it. “And do you have any medication allergies?” Having those as distinct resources as distinct lists of things, If I want to cross-check things, I can always do that as long as I have code systems that support it. But it also gives everything a nice clean lane to operate in. So that’s one thing that I think would be useful.
The other thing, when you’re dealing with allergies, is to make sure you’re dealing with the reaction. And this is another problem I think we have with allergies. Allergies are documented in a kind of a weird way, If you think about it. When I document that John Smith has an allergy to codeine, I put codeine as an allergy. I say, “Allergy: Codeine.” And then I put in a reaction. So we’re kind of documenting the substance as the primary thing, when really what we’re documenting is a medical condition, that they can’t tolerate something. So they really have a substance intolerance, or a drug intolerance, or a food intolerance. The substance itself is really a payload of that intolerance. This is kind of a holdover. It’s an anachronism from the way we looked at things. It’s like, “Acquired absence of thumb.” You know, instead of documenting the physical manifest of anatomic locations, we document the acquired absence of something, which is a little bit weird, if you think about it. With allergies, we do the same thing. We document it as the thing, not the intolerance of the thing, but we worked through it. Don’t get me wrong. I just struggle with it, um, from a data pattern purist perspective, but it is a little weird that we document Coumadin, and then we, we say, “Oh, here’s the reaction and here’s this. And here’s that.” And then later, if you have a different reaction to Coumadin, we document Coumadin again. And we say, “Well, now it used to be giant hives, and now it’s anaphylaxis.” Instead of saying, “Well, we have this intolerance and it’s the same thing, and it has changed over time,” we treat them as separate entities.
So the last thing I’m going to talk about, when it comes to allergies, and you’re probably very grateful to know that this is coming to a close, is that when you think about the classic pattern of allergies, it goes back to that episodic documentation pattern we have with allergies. Because one of the things that happens from time to time is I’m aggregating data for John Smith, and I’m pulling data from multiple places, and one place says no known allergies, and one place as they’re allergic to Coumadin. And of course, that’s one of those things that either someone didn’t bother to ask, or the no-known allergy encounter happened before they had their onset of allergic reaction to Coumadin. And so you have this whole temporal misalignment of, “Oh, well, do they, or do they not have an allergy?” Now, of course, anybody who’s paying attention would just assume that, if there an allergy, they can take the no-known and they can chuck it out the window. But it is still a piece of data that’s moving around, and it’s one of those things where, even when you look at FHIR resources, there’s almost this notion of I can pass you a code that is basically a negative. I can pass you a code that says, “No, there’s nothing.” And then when you go to reconcile that, somebody has to have logic or thought process that says, “Oh, wait, this means nothing, and this means something. Something wins over nothing, unless something is wrong, and pitch one of them out.” Now, what I would think we need to do, going forward, is find a better way to do that. To really have something that documents that I did check because, really, what you could say are there are no allergies. If there are no allergies, you leave the allergy bucket empty. But what you need is another thing at the patient level that says, “I asked the question about allergies on this date.” So what you’re really doing is you’re confirming that I did ask and there were no allergies. That’s why there’s nothing on the list. Because what that does, it allows you to approach the alternative to that, or the opposite of that, or the reciprocal of that, which is it doesn’t say whether they asked them about allergies or not. Whereas, if I have a record of allergies or intolerances and I have three choices, it’s empty, it says no known allergy, or it’s got stuff on it. When I do that, I have to make a decision with those. If I have this other thing that maybe sits at a patient level and says, “On this date and time, I asked them if they had any allergies. I confirm their allergy status.” And that should be enough to take off our plate this concept of no-known allergies, because I don’t like anti-terminology. So matter,anti-matter; I see the no-knowns as anti-terminology, and I think they’re an acronystic holdover from a simpler age when we just had these simple pre-coordinated lists of things, and we were dealing with them in that way. And we still do a lot of things in how we store data that way. That would take a long time to dig into all those.
So ladies and gentlemen, this has been, uh, the Informonster Podcast talking about architecture of allergies and intolerance. I hope you enjoyed it, and I look forward to talking to you again. Please, we have got a few people with some, some requests for topics, love to get those. So if you ideas or things you want to hear me or me and my colleagues talk about, please do let us know. And if you ever want to be on the Informonster Podcast and talk about something, I love to do that. So let us know that as well. Thanks.