[MUSIC] With this tackle information, moving up our little pyramid here. So let me show you some examples of information. And I'm going to be making the point that basically, information is organized or aggregated data. We've seen this front sheet, the electronic health record, several times. What I want to point out today is that it's organized. So once you know where to find different types of data, you don't really think too much. You know that the upper left is active problems, that below that is medications and the middle is allergies, and the right hand side are alerts of one sort of another. So you're no longer thinking about it, but you know what the data mean. So the organization helps you with the interpretation. I want to point out something subtle, if you look in the middle top you see that under allergies is the drug ibuprofen. So ibuprofen is a medication, that piece of data could live in the medication field and/or it could live in the allergy field. So the meaning of the same data changes, depending on which box it's in, which is a classic use of information. Another type of information is range. In this case, electronic health record called OpenMRS, an open source medical record system from around the world over like 150 countries are the most prevalent. Each are used in the developing world. And here you see I'm looking at temperature. The number is 44. The machine reports that the maximum is 43. Now 44 degrees celsius is 111 degrees. I don't think anybody can live with a temperature of 111. I think the highest temperature I've ever seen was a child with 106.5, and that was a child who had pneumonia and was dehydrated. But the machine is telling you that it doesn't know whether or not the 44 is plausible, it's just telling you that it's larger than its set limit at 43. Information could depend on knowledge of norms and the guidelines. This is a bit more complicated. You see here a growth chart. And on this growth chart, I've plotted a child's height and that's the red dot in the middle. But I've also drawn a vertical line showing that that dot is in between certain limits. The limits on the left hand side are 52 inches and 60 inches and if I follow the curves, I can see that we're talking about the top of 75th percentile and 25th percentile. So rather than just telling me what the limits of normal are, what this graph is telling me is that this child for this age, really fit compared to other children. It's not telling me what to do with that information. It's just giving me that information of putting the data of that child's height into a context of other heights. Maps are another way of organizing data and therefore providing information. So this map of the United States, I am not telling you what the colors represent but I think you can see that the left-hand side, the darker green places are kind of similar and different from the lower right-hand side which is brown. Now if I tell you that these are mortality rates and green is lowest and the dark brown is the highest. I think you would say that if in fact this represents your risk, you would prefer to live in the left-hand side of the United States than the lower right-hand side of the United States. This information, because I'm not telling you what to do with the data, I'm simply telling you where the data for each kind of this lives compared to everybody else. If you take this to mean that you should investigate further, that is an action that you're taking, but that the machine is not telling you the take. There might be other things that you would do with this data, the size, investigate, but the machine is just telling you what the data mean. When it's talking about data, I distinguish between data about Individuals and data about more than just the individual. Mortality rate like what we just saw on the map is something I calculate by kind of adding up experience of individuals. So I can get the mortality rate by looking at the number of people I have, count the number of people in the population. I can count the number of people who've died. I get a ratio, I have a rate. On the other hand, there are things like herd immunity of a population that doesn't belong to any one individual. So that's kind of a nonlinear property, and that's what I intend to show by that curve. Sometimes these are called emergent properties. I can't say where it is but it comes out of the totality. Classic examples of picture, so that I show you a computer graphic of the Mona Lisa portrait by Leonardo da Vinci. I could look at any individual pixel and you cannot tell me by adding or multiplying, or something, you can't tell me what the Mona Lisa is. But clearly you step back, you see the Mona Lisa in the totality. So that's an emergent property. So which I mean is up here is that when you get to numbers like rates in the population, it's sometimes hard to know whether you're talking about data or information. So, information that I calculate from individuals might be information, right? So if I come up with a mortality rate that may be an information. On the other hand, if it's emergent property that belongs to the entire population, like herd immunity that I mentioned. Then it's data of the population. My favorite is that one way of detecting whether or not a community has a level of opiate abuse is by measuring the opiate metabolites in the sewage. So you cannot calculate that concentration from knowing anything about individuals in the population, it's the property of a sewage, which you then apply to the population. The last type of information is even more subtle. This is what we call metadata. Remember we talked about the drug ibuprofen being either in one field or another field, therefore getting its meaning from that. Similarly, the metadata tell us how to interpret something about the data we're looking at. So it could be a sodium level of 135. It could be diagnosis like E11, but who has that sodium level? Is that 135 a sodium level or is it something else that's measured? And when was it measured is clearly really important. If you have a low blood count before the transfusion, that's okay. If you have a low blood count after the transfusion, I'm not so happy. And then the method by which it was made is a subtle thing. Even when it comes to blood you can have arterial, you can have veins. You can have different types of questionnaires. It could be different values. So in summary, when you talk about data versus information, one man's data can be another man's information. So I have Marie Curie's portrait there on purpose to make a comment on the word, man's data. And she took what was garbage to coal miners, which is pitchblende, and discovered uranium and polonium in there, and change the course of the 20th century. So the distinction between data information can be life changing.
