 Okay, so just a quick definition of out-of-range values or those values which are outside the expected scope of response. They can be numeric variable values which exceed the anticipated minimum or maximum. They can be response categories that were not previously identified, that you didn't anticipate. You add a new test, you add a new type of test and so that or a new drug comes on the market after the forms were designed, that would be a new category. The reasons for out-of-range values can be due to changes in the environment such as the availability of new drugs or it can be due to data entry errors. If your system is set up to allow some out-of-range values to get into the system then some of it can have to do with with actual data entry errors, not so much the data recording errors. So utilization of range restrictions on numeric fields to prevent entry of erroneous data can help reduce the out-of-range values and all in addition logical checks can conditionally restrict entry of out-of-range values. Now there's a trade-off here. I mean if you put a lot of time in in your data entry system and having conditional logical checks based on other variables, other values that might be in the somewhere else in the system, it can actually slow down data entry. So you need to make choices about how much how much checking you want to do at the point of data entry and how much you want to do post hoc. So we want to try and anticipate the fact that there's going to be out-of-range values just like with missing data. So in the data collection, we want to include a response category of other with a place to specify what other is. So if you're not sure that all of your categorical responses are capturing all possible answers, then you definitely want to include another and maybe a place to write in what that specifically is. It's very very helpful to include version numbers on your data collection forms so that you can track when a response category was added to the actual form. So if you have a new medication come on the market and people and you haven't updated the forms yet and people start checking other and writing in what that new medication is, then hopefully you can get the data entry system set up so that it'll actually create a new category for that system. But before that happens that those that new drug may just go into your a text field in the database or it may not go in at all. So then once you see that this is happening and this new drug is out there and on the market, you can go and modify the data the data collection form. But you want to be sure to you know to note when that modification took place so that and that new medication got added to the form so that you know when that was a specific response and when it stopped being written in as other or maybe not being written in at all and when it was a specific response to that item. So version control of your data collection forms is very important. Version control with and knowing the date that that new version went was created and went into the went into use. And here in Ampath we have a we have a problem because sometimes we get a new version of the form and it only gets just and it gets distributed to different sites at different times or it gets distributed to sites and they continue to use up the old version of their forms before they start using the new version. So in addition to just when you know when the form was finalized, you also need in from to maintain information about when that form was actually that when they actually started to use that form at the sites. OK. At your clinic sites. So with with respect to data data entry, as I've already said, you can utilize range restrictions on numeric fields to prevent entry of erroneous data. And you can tag those out of range values. Now there's I don't know if Ada talked about this when you guys were working on InfoPath. There's a lot of different ways to manage values that are out of range. And I think the way to think about it is an absolute range. Specifically thinking about lab values. This comes up a lot with lab values. So something like CD4 percent, as I said the other day, can only be zero to 100. There's no reason to allow anybody to enter a value greater than 100. Even if it appears on the form, it's not a valid it's not a valid value. If it if someone accidentally writes 100 and 105 on the form, that form should be tagged and sent back to the clinician who filled out the form for clarification. Right. But there are lab values that have normal ranges and and you know, questionable data outside the normal range. For example, hemoglobin, you don't expect hemoglobin to go up above 50. I mean even, you know, 50 is pretty high. So what you could do in the data entry system is you could instead of putting absolute restrictions on that on the range of hemoglobin, because very sick people might have a very high hemoglobin or very low hemoglobin or very low hemoglobin or white blood cell counts. Another one, I mean, if you're sick, you could have a very high white blood cell count, you know, out of the normal range. So what would be ideal is if you could set up in the data entry system an alert or that pops up. So if the data interclerc keys in a number that's outside the normal range for that particular lab value, and in our normal range may not be what the lab says is normal. I mean, it may be, you may have to extend it a little bit, because you do see a lot of values low. This is really questionable. So that that forces the data interclerc to confirm that they've typed in the number right. OK, so it's just an alert to the data interclerc. You know, please this this number is unexpected. Please double check and they can double check and see. Yes, that's actually the value that's written here and then they can go ahead and allow entry of that value. But it's but it's an intermediate check that that says it gives the data interclerc a second chance to confirm what they've entered is correct. This maybe an easier example is to think about weights for children. And then this would be a conditional this would be a conditional check, you know, that you wouldn't expect any baby under the age of two to weigh more than 15. Definitely put a check in there that says, you know, that this weight doesn't agree with this with this calculated age for this particular patient. So that then also forces the data interclerc to look at the date of birth. Did I enter the date of birth right? So so that it's calculating the right age and also reexamine the information they're they're trying to put in for the. So if you allow the entry clerk to enter something that's unexpected, you could have another variable in the system that sort of tags that that value and says. It says this is out of this is out of range, and we know it's out of range and the clerk confirmed that that was what was actually on the form. And then, you know, even to get more elaborate, you can allow a place for the data interclerc to actually make notes about the unexpected values. You know, I mean, sometimes clinicians will write a note, you know, put the put the lab value and write a note next to it saying, yes, this is high, but this is the actual value. From the lab, you know, or something like lab re ran the test and came up with the same result. OK, so if you want to get really elaborate, you can just have some of that information entered into the system. Now, I wouldn't advocate doing this for every variable on your form, certainly. But if there are some specific variables that are problematic or that are very important and you expect there to be out of range values. And you want to know the details about that, then you can you can have a place to record the details. So some so dealing with these out of range values, some of the options are to reject, as I already said, to prevent entry of these out of range values. I mean, this could be problematic because you could be rejecting valid data. So, for example, this labs are notorious for switching assays or switching units of measure. So if you have a lab that's, you know, constantly reporting in one unit of measure and all of a sudden they get a new machine or they get a new assay and they're reporting in another unit of measure that it's 10 times greater, you know, than the previous unit of measure than all of your values. Would would appear to be out of out of range and none of that data would get entered. So you got to be careful about, you know, total rejection of out of range values and just make sure that you're on top of the changes in the environment and the way the data is collected, the way the data is collected. You can accept the out of range values unconditionally and allow entry even though you know the value is wrong. You know, you can let somebody enter 56 kilos for a newborn, you know, and just take it when it's maybe should be 5.6 and just accept the 56 and deal with it later with cleanup queries. You can accept the value conditionally so allow the entry to be made but mark another indicator variable or something that says, we know this is out of range, but we're entering in any way. And then you can also some variables in themselves to correcting the out of range value. So converting an out of range value either to the or converting into the upper or lower limit and a lot of lab values you have lower limits that are undetectable below a certain amount. So when you get a note on there that says undetectable or less than 0.1, you can actually enter 0.1 and as long as it's known that that 0.1 means really means less than 0.1 or that the undetectable for something undetectable you enter the lowest value, then that's an option as well. Although I recommend if lab values are going to come back with a greater than or less than sign that you should have a second qualitative variable for that lab value that holds the greater than or less than sign. So in addition to the to the numeric value of that lab result, you have a separate field that would hold a less than or greater than sign if it appeared on the on the on the summary page on the results page from the lab. So this really so this upper and lower bound value is when you see something like undetectable. So a viral load load is undetectable. A lot of people enter the lower limit of the assay, which is either 50 or 400, depending on which assay you're using, typically. Okay, so that's all that I have on out of range values. Any questions or comments?