WEBVTT 1 00:00:09.433 --> 00:00:15.433 We'll now transition to a discussion of the validation and verification essentials, 2 00:00:15.666 --> 00:00:21.666 and apply that to the automated urinals methods. 3 00:00:22.166 --> 00:00:29.566 First, we're going to take a pit stop and talk about what method validation is to clear. 4 00:00:30.900 --> 00:00:33.500 And let's go over a couple of definitions. 5 00:00:33.500 --> 00:00:36.833 Even before we get to that, 6 00:00:36.833 --> 00:00:43.066 validation refers to the process of establishing the performance specifications 7 00:00:43.066 --> 00:00:49.500 of a new diagnostic tool, such as a new test, laboratory, developed test, or modified method. 8 00:00:50.633 --> 00:00:58.800 In contrast, method verification is a one time process to determine the performance characteristics of a test before use. 9 00:00:58.800 --> 00:01:02.966 In patient testing, with verification. 10 00:01:02.966 --> 00:01:09.900 You're essentially verifying that the diagnostic tool meets or exceeds the labeling. 11 00:01:11.466 --> 00:01:15.666 We're going to use these two terms interchangeably during this webinar. 12 00:01:15.666 --> 00:01:19.800 And that's okay. Going back to the clear concepts 13 00:01:21.466 --> 00:01:26.366 method validation and in the clear regulations 14 00:01:26.366 --> 00:01:34.666 is described depending on the complexity of the test or the instrument that you're trying to document. 15 00:01:35.866 --> 00:01:41.166 For example, moderate complexity tests clear specifies 16 00:01:41.166 --> 00:01:47.166 that you must document method performance in the form of these four studies. 17 00:01:47.366 --> 00:01:53.366 Accuracy, precision, reference interval, and reportable range 18 00:01:54.200 --> 00:02:03.166 for a high complexity method or test, or maybe a moderate complexity test that you've modified in some meaningful way. 19 00:02:03.833 --> 00:02:09.766 Clear specifies that you must also assess analytical sensitivity 20 00:02:09.766 --> 00:02:15.733 and analytical specificity. 21 00:02:15.733 --> 00:02:21.766 These are the essential steps of any validation or verification process. 22 00:02:22.766 --> 00:02:26.133 First, you want to gather your resources. 23 00:02:26.133 --> 00:02:28.633 This is a very important step. 24 00:02:28.633 --> 00:02:34.333 At this point in the process, it behooves you to familiarize yourself 25 00:02:34.333 --> 00:02:40.100 with the manufacturer's requirements to document method performance 26 00:02:40.100 --> 00:02:49.066 and also spend some time studying what others have done and applying some of the best practices that you can find to the method. 27 00:02:49.066 --> 00:02:52.133 Validation and challenge that you have ahead of you. 28 00:02:53.900 --> 00:02:58.800 Your next step is going to be to write a validation plan, 29 00:02:58.800 --> 00:03:03.533 followed by execution of the validation studies. 30 00:03:03.533 --> 00:03:11.866 I'd like to point out that before you actually perform the studies, it's incredibly important to train 31 00:03:12.166 --> 00:03:20.533 any of the individuals who will be executing the studies to the new instrument before getting started. 32 00:03:20.933 --> 00:03:27.000 You want to be sure that the method performance, the results that you see are truly reflective 33 00:03:27.433 --> 00:03:33.433 of the the new test or the new instrument that you're working with. 34 00:03:34.133 --> 00:03:42.566 Once the studies are done, you will analyze the data and compare the results to predefined acceptability criteria. 35 00:03:43.700 --> 00:03:48.166 And then that last step listed on the slide is optional. 36 00:03:48.166 --> 00:03:54.166 And we will talk about that later in this webinar. 37 00:03:56.100 --> 00:04:00.433 Let's spend another minute or so talking about gathering resources. 38 00:04:00.433 --> 00:04:03.466 As I mentioned this is a really important step. 39 00:04:04.800 --> 00:04:07.166 There's a lot of sources of information 40 00:04:07.166 --> 00:04:13.166 that I think are commonly overlooked in terms of contributing value to this planning process. 41 00:04:13.333 --> 00:04:19.333 Make sure to thoroughly review any package inserts that come with the new instrument. 42 00:04:19.366 --> 00:04:21.800 Instructions for use. 43 00:04:21.800 --> 00:04:23.933 Do a quick PubMed search. 44 00:04:23.933 --> 00:04:29.233 Look at the literature. See if others have published their validations. 45 00:04:29.233 --> 00:04:36.100 Look at best laboratory practices, perhaps from Clinical Laboratory Standards Institute, for example, 46 00:04:37.200 --> 00:04:47.200 and then also understanding the components of the automated system that you're trying to validate, as well as how the methods work behind each. 47 00:04:47.633 --> 00:04:54.133 You may find that the manufacturers requirements to document method performance are pretty sparse. 48 00:04:54.833 --> 00:05:00.833 They will typically recommend accuracy, precision reportable range, and carry over. 49 00:05:02.133 --> 00:05:08.800 And one important concept here is that the foundation of all method validation is clear. 50 00:05:09.433 --> 00:05:14.733 But I would also add to that meeting the manufacturer's minimum requirements. 51 00:05:14.733 --> 00:05:19.033 For example, clear doesn't specify that carryover must be done. 52 00:05:19.033 --> 00:05:22.833 But a lot of manufacturers mandate that it be so. 53 00:05:22.833 --> 00:05:29.400 So you could think of the clear requirements and the manufacturer's recommendation as being the foundation of any thorough 54 00:05:30.266 --> 00:05:33.133 method validation. 55 00:05:33.133 --> 00:05:38.866 And then really above that, there's layers of regulatory hierarchy that may inform your approach 56 00:05:38.866 --> 00:05:45.133 to the validation, such as checklist requirements with the College of American Pathologists, 57 00:05:46.066 --> 00:05:49.933 Best Practice from Clinical Laboratory Standards Institute 58 00:05:49.933 --> 00:05:55.933 and other professional societies who publish recommendations in this area. 59 00:05:56.833 --> 00:05:59.000 Let's talk about writing a plan. 60 00:05:59.000 --> 00:06:05.000 This is actually a very difficult and time consuming component of the process. 61 00:06:05.733 --> 00:06:10.833 It's really important that you go into this knowing what you're going to report, 62 00:06:10.833 --> 00:06:16.666 what parameters are you going to report, and what are the characteristics of your patient population. 63 00:06:16.666 --> 00:06:25.833 That information needs to be factored in to each of the items you see on this slide, particularly the components of your plan. 64 00:06:25.933 --> 00:06:30.300 And we talked about the four right that are required by clear. 65 00:06:30.300 --> 00:06:38.800 So your plan will include the each of the component studies typically accuracy precision reportable range and reference interval. 66 00:06:39.633 --> 00:06:45.633 But you must include sufficient detail on the plan such as where your sample is going to come from. 67 00:06:45.966 --> 00:06:47.533 What type will they be? 68 00:06:47.533 --> 00:06:53.533 How many, how many of those samples will be abnormal versus normal? 69 00:06:53.900 --> 00:06:59.900 Will you include important patient demographics in the collection of those samples? 70 00:07:00.033 --> 00:07:02.466 Fresh versus preserved samples. 71 00:07:02.466 --> 00:07:08.366 Not all laboratories accept preserved samples for testing, but if you do, you want to include 72 00:07:08.366 --> 00:07:11.166 a sufficient number of preserved samples in your plan. 73 00:07:12.333 --> 00:07:18.266 And then things like minimum volumes, exclusion criteria 74 00:07:18.266 --> 00:07:25.700 and even the procedure and the procedure, you want to think about pre analytic variables and analytic variables. 75 00:07:25.833 --> 00:07:32.033 So storage process sourcing mixing decanting and the testing aspects. 76 00:07:33.266 --> 00:07:41.533 Lastly don't forget to find acceptability criteria for each of the studies and a data analysis plan. 77 00:07:44.200 --> 00:07:48.566 We're going to briefly discuss an overview of the component studies. 78 00:07:48.566 --> 00:07:53.133 And then we're going to apply it to automated urinalysis methods. 79 00:07:53.133 --> 00:07:58.200 Starting with the accuracy study or method comparison study, 80 00:07:58.200 --> 00:08:07.333 an accuracy study in concept determines bias to an existing method or a gold standard reference method, 81 00:08:07.833 --> 00:08:14.333 and that's done by analyzing samples and duplicate on the new method and the reference method. 82 00:08:15.300 --> 00:08:24.033 An accuracy study should be done for results that are quantitative, semi quantitative, and qualitative. 83 00:08:24.700 --> 00:08:28.300 So in the context of our urinalysis validation, 84 00:08:28.300 --> 00:08:35.966 it would apply to the reportable parameters from the chemical UA as well as the quantified formed elements of the urine. 85 00:08:39.300 --> 00:08:44.100 The precision study applies to the quantitative results. 86 00:08:44.100 --> 00:08:51.300 And there's typically two types of precision within run and between run or total precision, 87 00:08:51.900 --> 00:08:57.900 which represent repeatability and reproducibility of the results, respectively. 88 00:08:58.933 --> 00:09:05.966 Precision studies are conducted by analyzing materials repeatedly within a run and over several days. 89 00:09:08.833 --> 00:09:11.700 The reportable Range study 90 00:09:11.700 --> 00:09:20.533 verifies the lowest and highest results that can be directly reported from the analyzer by conducting a linearity experiment. 91 00:09:21.566 --> 00:09:29.966 This is typically required for components that report numerically, such as red blood cells, white blood cells, and bacteria. 92 00:09:29.966 --> 00:09:34.033 In the context of the automated UA methods, 93 00:09:34.033 --> 00:09:38.000 and this study is very helpful for, 94 00:09:38.000 --> 00:09:44.000 the laboratory to determine the reportable limits of the new instrument or test. 95 00:09:46.900 --> 00:09:50.166 Reference intervals are critical 96 00:09:50.166 --> 00:09:59.533 when transitioning from a manual to an automated method, because sometimes the reference intervals are different between methods. 97 00:09:59.533 --> 00:10:04.900 Keeping in mind reference intervals are method specific in all of laboratory medicine, 98 00:10:06.333 --> 00:10:08.533 you typically have three options. 99 00:10:08.533 --> 00:10:14.533 One is to verify the current reference interval in use by the lab, 100 00:10:14.700 --> 00:10:21.766 or you can verify the manufacturer's proposed reference interval or something else entirely. 101 00:10:21.766 --> 00:10:25.233 Maybe there's one in the literature that you like, 102 00:10:25.233 --> 00:10:29.933 but it needs to be considered for all reportable parameters. 103 00:10:29.933 --> 00:10:37.566 There's a pretty big difference between reference interval verification versus establishing a new reference interval. 104 00:10:38.300 --> 00:10:43.733 The minimum number of samples for the study is vastly different. 105 00:10:43.733 --> 00:10:48.866 And if your reference interval is partitioned by sex or age, 106 00:10:48.866 --> 00:10:54.866 that minimum number of samples applies to each partition in the reference interval. 107 00:10:55.233 --> 00:10:58.266 Carryover is an important assessment. 108 00:10:58.266 --> 00:11:07.400 Carryover determines whether samples with a high concentration of a substance impact subsequent samples with low concentration. 109 00:11:08.200 --> 00:11:15.933 And these studies are typically done by analyzing material with a high concentration, followed by samples with a low concentration. 110 00:11:19.233 --> 00:11:20.533 Method validation is 111 00:11:20.533 --> 00:11:26.800 incredibly challenging, and I've listed here on the slide some of the most common challenges. 112 00:11:27.733 --> 00:11:35.633 Even things as simple as determining what materials to select for each component study isn't always clear. 113 00:11:36.500 --> 00:11:42.500 Then sample selection, collection, and stability can be a challenge. 114 00:11:42.700 --> 00:11:50.133 Curating abnormal samples and samples from diverse patient populations is a frequent barrier. 115 00:11:51.600 --> 00:11:55.800 You may encounter differing reporting formats between methods. 116 00:11:55.800 --> 00:11:59.900 Different sample processing requirements, 117 00:11:59.900 --> 00:12:07.566 and even something that you would think is relatively simple, such as determining the acceptability criteria for each component. 118 00:12:07.566 --> 00:12:16.133 Study can actually be quite difficult, and there isn't always an obvious answer for what the acceptability criteria may be. 119 00:12:16.800 --> 00:12:24.100 That's where the laboratory director or medical director clinical and technical expertise shines, 120 00:12:25.466 --> 00:12:29.700 writing a plan that fulfills the requirements for an automated system. 121 00:12:29.700 --> 00:12:36.033 When your automated system has multiple components like most of them do, also makes it 122 00:12:36.333 --> 00:12:42.133 a layer more complex than your average validation of an instrument. 123 00:12:43.933 --> 00:12:48.000 We'll now transition to applying all of these concepts 124 00:12:48.000 --> 00:12:54.533 to a sample validation plan for an automated urinalysis instrument. 125 00:12:57.100 --> 00:13:03.100 And we'll organize that by component of the automated urinalysis system. 126 00:13:03.566 --> 00:13:09.133 We'll begin with describing studies for the chemical portion of the UA 127 00:13:09.133 --> 00:13:16.333 as sample plan for this could include a mixture of normal and abnormal samples. 128 00:13:16.333 --> 00:13:22.333 And I've listed some minimum numbers here on preserve I've specified in this plan. 129 00:13:22.466 --> 00:13:29.766 But as I mentioned earlier, if you accept preserved samples for analysis in your lab, just add in some of those. 130 00:13:30.833 --> 00:13:41.033 What you'll do is you'll analyze these samples as soon as possible and within stability of the samples on both the new and reference method. 131 00:13:41.366 --> 00:13:43.966 So in duplicate 132 00:13:43.966 --> 00:13:50.733 for data analysis you'll look at percent agreement for the qualitative and semi quantitative results. 133 00:13:51.533 --> 00:13:57.533 And for the quantitative results you'll look at slope intercept R value and bias 134 00:13:57.566 --> 00:14:03.533 pH and specific gravity sometimes are able to be quantitative with the chemical UA. 135 00:14:03.533 --> 00:14:10.566 So those are examples of a couple analytes where you might take that approach to data analysis for the chemical UA. 136 00:14:12.866 --> 00:14:17.533 For the precision study for the chemical UA, 137 00:14:17.533 --> 00:14:22.366 you could take normal and abnormal control material 138 00:14:22.366 --> 00:14:27.966 as your sample to select for this particular study 139 00:14:27.966 --> 00:14:33.966 and analyze it ten plus times consecutively for within run precision. 140 00:14:34.500 --> 00:14:39.200 And then for total precision you have two options. 141 00:14:39.200 --> 00:14:45.200 Option one is to do a full 20 day total precision study. 142 00:14:46.233 --> 00:14:52.500 Option two is to conduct what's called a precision verification study. 143 00:14:52.833 --> 00:14:58.833 I have abbreviated here as a five by five study where you take two samples. 144 00:14:58.833 --> 00:15:03.533 In this case, we're taking a normal and an abnormal control. 145 00:15:03.533 --> 00:15:09.533 And you would run five replicates one run per day for a total of five days. 146 00:15:11.933 --> 00:15:20.533 According to clear, you're only required to do the method verification precision study, which is the five by five I just described. 147 00:15:20.866 --> 00:15:26.333 But if you wanted to do something more extensive and embark on a full 20 day 148 00:15:26.333 --> 00:15:36.900 total precision study, that would be completely acceptable and exceed the minimum requirements for data analysis, you would look at agreement 149 00:15:36.900 --> 00:15:43.300 for the qualitative and semi quantitative parameters and percent CV for a quantitative results. 150 00:15:47.200 --> 00:15:51.633 For the reportable range study of a chemical UA, 151 00:15:51.633 --> 00:15:57.633 you're typically only applying this method performance study to specific gravity 152 00:15:59.166 --> 00:16:01.066 and sample selection. 153 00:16:01.066 --> 00:16:03.566 Here is a bit variable. 154 00:16:03.566 --> 00:16:13.433 You could either use patient samples or you could take control material or you could make a stock solution. 155 00:16:14.366 --> 00:16:24.033 For example, you could dilute 10% saline with denied water to span the analytical measurement range for your instrument. 156 00:16:25.200 --> 00:16:30.500 Regardless of what you do, whether it's preparing serial dilutions or diluting 157 00:16:30.500 --> 00:16:36.500 from a stock solution or diluting patient samples, you're going to do a linearity study. 158 00:16:37.333 --> 00:16:43.333 And once you've generated your set of samples, you will analyze them in triplicate 159 00:16:44.100 --> 00:16:50.100 and conduct a linear regression analysis and percent recovery calculation. 160 00:16:50.733 --> 00:16:57.866 You're just proving the linear relationship between measured and target concentration of each sample in the set. 161 00:16:59.100 --> 00:17:04.133 And the statistics will help prove that there is that linear relationship, 162 00:17:04.133 --> 00:17:11.566 and also reassure you that the expected recovery of specific gravity, which you're trying to measure, is acceptable. 163 00:17:14.233 --> 00:17:22.400 Now let's talk about method performance studies and a sample validation plan for the urine particle analyzer component. 164 00:17:23.166 --> 00:17:29.800 We mentioned previously that it will vary depending on what automated urinalysis system you're validating, 165 00:17:30.300 --> 00:17:36.300 whether the technology is flow cytometry or digital imaging. 166 00:17:36.900 --> 00:17:44.266 But regardless, this slide shows one potential approach to documenting accuracy method performance. 167 00:17:45.266 --> 00:17:50.533 Again, you're going to take some normal and abnormal samples, 168 00:17:50.533 --> 00:17:56.233 and then you're going to analyze them and 169 00:17:56.233 --> 00:18:03.433 in duplicate right between your your new instrument and your reference instrument. 170 00:18:04.566 --> 00:18:07.466 And that reference instrument could be a gold standard. 171 00:18:07.466 --> 00:18:13.900 Or it could be the instrument that is currently in use in the laboratory. 172 00:18:15.333 --> 00:18:19.266 If you're transitioning from one automated, 173 00:18:19.266 --> 00:18:25.133 microscopic UA to another, it's relatively straightforward. 174 00:18:25.133 --> 00:18:34.900 But some laboratories may be validating an automated urine particle analyzer for the microscopic and comparing that to a convention 175 00:18:34.933 --> 00:18:40.933 on a microscopic method using a brightfield microscope. 176 00:18:40.966 --> 00:18:45.000 And there are some challenges with that that we'll discuss a little bit later. 177 00:18:45.000 --> 00:18:47.966 Regardless, it's all possible. 178 00:18:47.966 --> 00:18:51.766 The concept is you're going to take some normal and abnormal samples. 179 00:18:51.766 --> 00:18:57.933 You're going to test them and duplicate once on each new and you're comparison method. 180 00:18:59.166 --> 00:19:03.266 One thing I've called out on this slide is that 181 00:19:03.266 --> 00:19:12.533 if you're if you're validating a new automated method and comparing it to an existing automated method in use, 182 00:19:13.133 --> 00:19:21.366 I think it's very helpful to perform a conventional microscopic analysis on a subset of the abnormal samples. 183 00:19:22.333 --> 00:19:24.133 The reason for that is that 184 00:19:24.133 --> 00:19:31.766 if you do see discrepancies in the quantitative results between the two automated methods, 185 00:19:32.133 --> 00:19:39.733 you can use the data from the conventional microscope spec analysis to help adjudicate any of those differences. 186 00:19:40.200 --> 00:19:48.733 Plus, it's kind of a nice reassurance that, both of the automated methods are appropriately sensitive 187 00:19:48.733 --> 00:19:55.566 to detecting abnormal formed elements in the selection of samples 188 00:19:55.566 --> 00:20:01.566 that's used for this, validation study. 189 00:20:01.833 --> 00:20:05.300 One challenge with accuracy studies for the urine 190 00:20:05.300 --> 00:20:11.300 particle analyzer is that unit conversion may be necessary to compare results. 191 00:20:11.800 --> 00:20:17.433 Your approach to data analysis, as well as your acceptance criteria, will vary depending on 192 00:20:17.433 --> 00:20:23.433 how the two methods measure analytes and report them. 193 00:20:23.433 --> 00:20:28.933 One word of caution is that turbid, abnormally colored, 194 00:20:28.933 --> 00:20:34.633 and specimens that have been stored may more frequently show disagreement, 195 00:20:35.866 --> 00:20:37.333 and sometimes 196 00:20:37.333 --> 00:20:43.500 methods have different sample processing requirements, so you could see some bias between methods 197 00:20:43.766 --> 00:20:49.766 that's solely attributable to those differences. 198 00:20:51.133 --> 00:20:57.133 Here is a potential approach to the precision studies for a urine particle analyzer, 199 00:20:58.466 --> 00:21:02.633 you may select normal and abnormal materials 200 00:21:02.633 --> 00:21:06.833 prepared from reagents or collaborators, 201 00:21:06.833 --> 00:21:15.866 and conduct a precision study for quantification of bacteria casts, epithelial cells, red blood cells, white blood cells, etc.. 202 00:21:16.733 --> 00:21:22.166 Now, you may be wondering why not use patient samples for precision study? 203 00:21:22.166 --> 00:21:26.600 And that's a great question and I will tell you why I would not recommend doing that. 204 00:21:26.600 --> 00:21:29.266 And it has to do with stability. 205 00:21:29.266 --> 00:21:35.266 Patient urine samples are typically only stable for a couple hours 206 00:21:35.400 --> 00:21:38.133 longer if you're using a preservative. 207 00:21:38.133 --> 00:21:45.533 But regardless, when we're thinking about our precision study, our precision study isn't typically complete in one day. 208 00:21:45.933 --> 00:21:51.166 So you need to source material for this that will be stable for the duration of the precision study. 209 00:21:52.666 --> 00:21:54.433 The precision studies is 210 00:21:54.433 --> 00:22:00.433 conducted by analyzing control material over a period of several days. 211 00:22:00.733 --> 00:22:07.266 You could do a full 20 day precision study for the urine particle analyzer, or you could conduct a five 212 00:22:07.266 --> 00:22:11.266 by five verification precision study, 213 00:22:11.266 --> 00:22:14.700 and which I described on a previous slide. 214 00:22:14.700 --> 00:22:20.700 You'll compare the results to the manufacturer's specification. 215 00:22:20.966 --> 00:22:27.966 And we're typically comparing the coefficient of variation, the CV, the coefficient of variation 216 00:22:27.966 --> 00:22:33.966 to that specification. 217 00:22:34.300 --> 00:22:36.533 The reportable range study for a urine 218 00:22:36.533 --> 00:22:42.533 particle analyzer can be conducted by 219 00:22:42.700 --> 00:22:47.966 serial dilutions spanning their range from a prepared stock solution. 220 00:22:47.966 --> 00:22:52.566 So it's the same concept as what I described on the previous slide. 221 00:22:52.566 --> 00:22:55.466 When we talked about reportable range studies. 222 00:22:55.466 --> 00:23:01.466 You're going to generate a series of samples and then 223 00:23:01.700 --> 00:23:06.133 analyze each of those samples in the set, at least in duplicate. 224 00:23:07.833 --> 00:23:09.933 It's helpful to know if the 225 00:23:09.933 --> 00:23:17.533 instrument manufacturer has some kind of ready made material that you could use to to generate the samples 226 00:23:17.533 --> 00:23:23.533 for this study, but the concept is a linearity study, like we talked about a few minutes ago. 227 00:23:24.200 --> 00:23:30.200 And your data analysis will be by linear regression statistics. 228 00:23:32.233 --> 00:23:35.233 Moving on to the last component of an automated 229 00:23:35.233 --> 00:23:41.233 urinalysis system, which is often a digital imaging device, 230 00:23:42.000 --> 00:23:43.366 strictly speaking. 231 00:23:43.366 --> 00:23:47.566 And accuracy study for this component is optional. 232 00:23:47.566 --> 00:23:50.466 But I would argue that it's best practice. 233 00:23:50.466 --> 00:23:57.133 It's nice to know if the device accurately classifies the particles based on images. 234 00:23:58.500 --> 00:24:03.733 You could conduct an accuracy study using a limited number of abnormal samples. 235 00:24:03.733 --> 00:24:14.233 Let's say a minimum of ten, and then just seeing how well the device classifies them on its own, keeping in mind that in the real world, 236 00:24:14.233 --> 00:24:21.866 outside of the validation environment, tech confirmation of classification of all the images will happen in production. 237 00:24:21.866 --> 00:24:27.866 We wouldn't just result, something out without it being reviewed by an expert. 238 00:24:30.133 --> 00:24:34.866 The precision study for the digital imaging device is also optional, 239 00:24:34.866 --> 00:24:42.833 but if you want to do that, what you could do is you could assess the precision of the sensitivity of the non 240 00:24:42.833 --> 00:24:50.766 reportable parameters of the device, usually by using commercial QC materials specific to the instrument manufacturer. 241 00:24:53.200 --> 00:24:57.666 The last study that is worth mentioning. 242 00:24:57.666 --> 00:25:03.733 Again you know carryover is not necessarily implicit in the clear regulations. 243 00:25:03.733 --> 00:25:09.733 But I mentioned is very important overall for the automated urinalysis system 244 00:25:10.333 --> 00:25:13.200 is you could take 245 00:25:13.200 --> 00:25:19.200 for the chemical portion of the UA system, some abnormal material, 246 00:25:19.900 --> 00:25:25.900 for example, control material and analyze that first, followed by normal control material. 247 00:25:26.366 --> 00:25:30.633 For the urine particle analyzer component, you could take a high sample. 248 00:25:30.633 --> 00:25:40.466 Let's say it's high for the cell components or for bacteria, and analyze it multiple times followed by a low sample analyzed multiple times. 249 00:25:41.333 --> 00:25:47.033 And for a digital imaging device, you could perform consecutive analysis of a high solution 250 00:25:47.033 --> 00:25:52.033 followed by a blank, and then calculate percent carryover using a standard equation. 251 00:25:53.466 --> 00:25:59.100 Often, in my experience, the clinical application specialists of the system 252 00:25:59.100 --> 00:26:05.133 that you're evaluating will perform the carryover studies on site for you using a standard protocol. 253 00:26:08.200 --> 00:26:11.633 We're going to spend the next few minutes talking about the reference interval studies, 254 00:26:11.633 --> 00:26:15.733 which if you recall, is one of the required evaluations. 255 00:26:15.733 --> 00:26:18.233 According to Clea. 256 00:26:18.233 --> 00:26:24.233 And it's also very, very important clinically, 257 00:26:24.500 --> 00:26:29.300 the manufacturer for the urinalysis system that you're validating 258 00:26:29.300 --> 00:26:35.733 may have recommendations for how many samples to include in your reference interval verification study. 259 00:26:37.133 --> 00:26:44.833 And typically, this type of study is applicable to the chemical portion of the system as well as the urine particle analyzer. 260 00:26:46.366 --> 00:26:52.366 And so we're going to talk through a few caveats of designing an effective reference interval study. 261 00:26:52.633 --> 00:27:00.433 If a preservative is routinely used you must also include samples containing preservative in your reference interval study. 262 00:27:01.500 --> 00:27:02.233 The samples 263 00:27:02.233 --> 00:27:08.233 must be stored, processed, and tested the same as patient samples will be. 264 00:27:08.233 --> 00:27:14.233 It's recommended that 40 fresh urine samples be collected for such a study, 265 00:27:15.166 --> 00:27:24.733 and you must collect urine samples from normal individuals who have not been under treatment 266 00:27:25.233 --> 00:27:33.433 in the past six weeks for any illness they're not under treatment for a yeast infection or a bladder infection, 267 00:27:34.033 --> 00:27:40.033 do not have a diagnosis of hypertension, and are not menstruating or pregnant. 268 00:27:41.400 --> 00:27:46.233 Also, the the normal samples should reflect 269 00:27:46.233 --> 00:27:52.033 the laboratories patient population and include both male and females 270 00:27:52.033 --> 00:27:58.033 and a variety of age ranges if at all possible. 271 00:27:58.500 --> 00:28:02.266 If you analyze 40 urine samples 272 00:28:02.266 --> 00:28:09.866 and find that you can't verify the proposed reference interval, what are you going to do? 273 00:28:11.000 --> 00:28:13.300 You have a couple of options. 274 00:28:13.300 --> 00:28:18.133 One is to continue collecting normal samples. 275 00:28:18.133 --> 00:28:24.100 Perhaps you'll collect another 20 or 40 and see if you can get the proposed reference interval to verify. 276 00:28:25.133 --> 00:28:27.866 But if you can't get it to verify, 277 00:28:27.866 --> 00:28:36.733 you're in a position of having to conduct a full reference interval study, which would typically include a minimum of 120 samples. 278 00:28:37.300 --> 00:28:42.866 And keep in mind that if your reference intervals are partitioned, that minimum number would apply 279 00:28:42.866 --> 00:28:48.866 to any partitions that you have. 280 00:28:49.400 --> 00:28:51.766 I mentioned on an earlier slide 281 00:28:51.766 --> 00:29:01.600 that an optional last step of this process is a post go Live assessment, or a clinical validation of the new SAS system. 282 00:29:02.833 --> 00:29:08.833 Another term for that might be a clinical validation, but the reasons why you might do it, 283 00:29:10.000 --> 00:29:14.900 is because you may see differences in clinical sensitivity depending 284 00:29:14.900 --> 00:29:20.900 on how urinalysis testing is commonly utilized in your health system. 285 00:29:21.566 --> 00:29:28.300 What you could do is perform a clinical sensitivity study for your unique patient population 286 00:29:29.233 --> 00:29:31.700 or user defined parameters. 287 00:29:31.700 --> 00:29:37.700 For example, if you want to customize your instrument to discriminate normal from abnormal samples, 288 00:29:38.100 --> 00:29:45.400 it would be useful to do a clinical validation to make sure that, those criteria have been optimized 289 00:29:45.733 --> 00:29:53.433 and that patient management is flowing appropriately from the laboratory results that clinicians are interpreting. 290 00:29:55.033 --> 00:29:59.566 This is especially important if you have local clinical protocols 291 00:29:59.566 --> 00:30:05.800 that could be impacted by integrating automated urinalysis as methods in your laboratory.