Reagent analyzer and method for calibrating the reagent analyzer

Abstract

A method is described in which patient tests are executed with reagent test devices on a first reagent analyzer to obtain a series of reagent test device measurements, the first reagent analyzer having a plurality of first error sources and a plurality of first performance requirements. Control tests for control test devices are executed on the first reagent analyzer to obtain control test device measurements. Control element test acceptance criteria are generated for control test device measurements of control tests to be run by the one or more first reagent analyzer on control test devices. The acceptance criteria is calculated with the reagent test device measurements and the control test device measurements and allow for analyte measurement precision that meets the plurality of first performance requirements. The acceptance criteria is stored in a non-transitory computer readable medium coupled to a processor of a second reagent analyzer.

Description

REAGENT ANALYZER AND METHOD FOR CALIBRATING THE REAGENT ANALYZER

[0001] This application claims benefit under 35 USC § 119(e) of U.S.Provisional Application No.63 / 636,351, filed April 19, 2024. The entire contents of the above-referenced patent application are hereby expressly incorporated herein by reference. BACKGROUND

[0002] To satisfy the needs of the medical profession as well as otherexpanding technologies, such as the brewing industry, chemical manufacturing, etc., a myriad of analytical procedures, compositions, and tools have been developed, including lateral flow immunoassays, and the so-called "dip-and-read" type reagent test devices. Regardless of whether lateral flow immunoassays, or dip-and-read test devices are used for the analysis of a biological fluid or tissue, or for the analysis of a commercial or industrial fluid or substance, the general procedure involves a test device coming in contact with the sample or specimen to be tested, and manually or instrumentally analyzing the test device.

[0003] A lateral flow immunoassay is a diagnostic device used to confirm thepresence or absence of a target analyte. Lateral flow immunoassays typically contain a flow path which conveys a sample past a control line position and a test line position. A control line at the control line position confirms the test is working properly, and a test line at the test line position provides the result of the lateral flow immunoassay. Lateral flow immunoassays are developed to be used in a dipstick format or in a housed test format. Both dipsticks and housed tests work in a similar way, and generally fall within one of two categories: sandwich assays – a positive test is represented by the presence of a coloured line at the test line position; and competitive assays – a positive test is represented by the absence of a coloured line at the test line position.

[0004] Dip-and-read reagent test devices enjoy wide use in many analyticalapplications, especially in the chemical analysis of biological fluids, because of their relatively low cost, ease of usability, and speed in obtaining results. In medicine, for example, numerous physiological functions can be monitored merely by dipping a dip-and-read reagent test device into a sample of body fluid or tissue, such as urine or blood, and observing a detectable response, such as a change in color or a change in the amount of light reflected from, or absorbed by the test device.

[0005] Many of the dip-and-read reagent test devices for detecting body fluidcomponents are capable of making quantitative, or at least semi-quantitative, measurements. Thus, by measuring the detectable response after a predetermined time, a user can obtain not only a positive indication of the presence of a particular constituent in a test sample, but also an estimate of how much of the constituent is present. Such dip-and-read reagent test devices provide physicians and laboratory technicians with a facile diagnostic tool, as well as with the ability to gauge the extent of disease or bodily malfunction.

[0006] Illustrative of dip-and-read reagent test devices currently in use areproducts available from Siemens Healthcare Diagnostics Inc., under the trademark MULTISTIX, and others. Immunochemical, diagnostic, or serological test devices, such as these usually include one or more carrier matrix, such as absorbent paper, having incorporated therein a particular reagent or reactant system which manifests a detectable response (e.g., a color change in the visible or ultraviolet spectrum) in the presence of a specific test sample component or constituent. Depending on the reactant system incorporated with a particular matrix, these test devices can detect the presence of glucose, ketone bodies, bilirubin, urobilinogen, occult blood, nitrite, and other substances. A specific change in the intensity of color observed within a specific time range after contacting the dip-and-read reagent test device with a sample is indicative of the presence of a particular constituent and / or its concentration in the sample. Some other examples of dip-and-read reagent test devices and their reagent systems may be found in U.S. Patents Nos.3,123,443; 3,212,855; and 3,814,668, the entire disclosures of which are hereby incorporated herein by reference.

[0007] However, dip-and-read reagent test devices suffer from somelimitations. For example, dip-and-read reagent test devices typically require a technician to manually dip the test device into a sample, wait for a prescribed amount of time, and visually compare the color of the test device to a color chart provided with the test device. This process is slow and the resulting reading is highly skill-dependent (e.g., exact timing, appropriate comparison to the color chart, ambient lighting conditions, and technician vision) and may be inconsistent between two different technicians performing the same test. Finally, the act of manually dipping the testdevice into the sample may introduce cross-contamination or improper deposition of the test sample on the test device, such as via incomplete insertion of the test device into the sample, insufficient time for the sample to be deposited onto the test device, or having too much sample on the test device which may drip, leak, or splash on the technician’s work area, person, or clothing.

[0008] Testing tools and methods have been sought in the art for economicallyand rapidly conducting multiple tests, especially via using automated processing. Automated reagent analyzer systems have an advantage over manual testing with respect to cost per test, test handling volumes, and / or speed of obtaining test results or other information.

[0009] Automated instruments which are currently available for instrumentallyreading individual reagent test devices, such as lateral flow immunoassays, or dip- and-read reagent test devices, or reagent strips, (e.g., CLINITEK STATUS reflectance photometer, manufactured and sold by Siemens Healthcare Diagnostics, Inc.) require each test device to be manually loaded into the automated instrument after contacting the test device with specimen or sample to be tested. Manual loading requires that the reagent test device be properly positioned in the automated instrument within a limited period of time after contacting the solution or substance to be tested. At the end of the analysis, used reagent test devices are removed from the instrument and disposed of in accordance with applicable laws and regulations.

[0010] Another development is the introduction of multiple-profile reagent cardsand multiple-profile reagent card automated reagent analyzers. Multiple-profile reagent cards are essentially card-shaped test devices which include multiple reagent- impregnated matrices or pads for simultaneously or sequentially performing multiple analyses of analytes, such as the one described in U.S. Pat. No. 4,526,753, for example, the entire disclosure of which is hereby incorporated herein by reference. The reagent pads on the multiple-profile reagent card are typically arranged in a grid- like arrangement and spaced at a distance from one another so as to define several rows and columns of reagent pads. Adjacent reagent pads in the same row may be referred to as a test strip, and may include reagents for a preset combination of tests that is ran for each sample, for example.

[0011] Multiple-profile reagent cards result in an efficient, economical, rapid,and convenient way of performing automated analyses. An automated reagent analyzer configured to use multiple-profile reagent cards typically takes a multiple-profile reagent card, such as from a storage drawer, or a cassette, and advances the multiple-profile reagent card through the reagent analyzer over a travelling surface via a card moving mechanism, typically one step at a time so that one test strip (or one row of reagent pads) are positioned at a sample-dispensing position and / or at one or more read position. Exemplary card moving mechanisms include a conveyor belt, a ratchet mechanism, a sliding ramp, or a card-gripping or pulling mechanism. As the multiple-profile reagent card is moved or travels along the travelling surface and is positioned at the sample-dispensing position, one or more pipettes (e.g., manual or automatic) deposits a volume of one or more samples on one or more of the reagent pads on the reagent card. Next, the reagent pads are positioned at one or more read positions and analyzed (e.g., manually or automatically) to gauge the test result. The reagent card is placed in the field of view of an imaging system, such as an optical imaging system, a microscope, or a photo spectrometer, for example, and one or more images of the reagent pads on the card (e.g., optical signals indicative of the color of the reagent pads) is captured and analyzed. Typically, the field of view of the imaging system is relatively large to allow for the capture of multiple images of the same reagent pad as the reagent card is moved or stepped across multiple read positions in the field of view of the imaging system. The field of view encompasses multiple read positions or locations, and each reagent pad is moved in a stepwise fashion through the read positions as the reagent card travels across the field of view of the imaging system. Because the reagent analyzer moves the card between various read positions in known intervals of time, the multiple images taken in the field of view of the imaging system allow the reagent analyzer to determine changes in the color of the reagent pad as a result of the reagent pad reacting with the sample at each read position as a function of the time it takes the pad to be moved to the respective read position, forexample. Finally, the used card is removed from the reagent analyzer, and is disposedof appropriately.

[0012] The performance of reflectance-measuring urinalysis reagent analyzersis based on how precisely the reagent analyzer can provide a measure of some analyte based on these reflectance measurements. Calibration procedures exist in which the reagent analyzer performs an assay test with a sample having a known amount of an analyte, and then comparing the results with known results. An assay test, however, involves applying a sample to a reagent pad and then running the test on the reagent analyzer. In a manufacturing setting, it is impractical to have operatorsperform assay testing to assess this performance. This type of performance testing may require special training and a specific lab environment that would add complexity to the process to ensure that the reagent analyzer is properly calibrated prior to introduction of the reagent analyzer into a lab environment. In addition, there may be many different types of assays that would need to be performed to fully assess the performance of the system.

[0013] A reagent analyzer that is used to measure reflectance measurementsfrom an assay test includes an illumination source to illuminate the reagent pad containing the sample and a camera to capture an image of the reagent pad containing the sample. Conventionally, non-reagent control test devices have been utilized to assess the precision and accuracy of the digital measurements of the camera itself in response to system error sources, having no direct correlation with the reagent measurements. The method used assumes that all system error sources affect each of the reagent results the same way, and that the effect is directly proportional to the effect measured on the control test device. For instance, 21145 Error Budget for Atlas Next Revision 1 states: “There are several error sources that are inherent to the camera and illumination. Since those errors contribute equally to the total error of each reagent, they do not require separate analysis per reagent.”

[0014] The referenced study aims to estimate the combined effect of all errorsources, both reagent and system-specific, to determine if the budgeted allowance for the variance in each error source is acceptable. Rather than study the system-specific error sources on both reagent and control test device tests, they only performed testing to assess this error using the control test devices. For predecessor devices, such as the Atlas Next, it was assumed that the variation in the digital signal measured during the control test device testing is directly proportional and translatable to the variation that would occur in measuring reagent tests under the same experimental conditions. For this reason, the variability estimates calculated from the control test device experiments were used directly in the calculation of the total allowable variability in each of the reagent results, even though there was no correlation shown between control test device measurements and reagent measurements under the same experimental conditions. In addition, this method assumed that a system error source has exactly the same effect on all reagent measurements at all concentration levels.

[0015] These previously mentioned assumptions fail to consider that there maybe interaction effects between these factors of system variation and reagent-related factors associated with using a wet sample undergoing complex chemistry.

[0016] Accordingly, a need exists in the art for a method of ensuring calibrationof the reagent analyzer to confirm that the reagent analyzer will properly measure an analyte based on reflectance measurements without having an operator perform an assay test to assess this performance. SUMMARY

[0017] To overcome the problems in the conventional methods noted above, inone embodiment, the presently disclosed method considers that within a reagent analyzer, the same mechanisms are utilized for measuring both reagent tests and control test device tests (i.e., using an image produced by a camera to measure a digital signal of a test under certain lighting), so any error source that significantly affects the analyzer’s measurement system will also affect both reagent and control test device test results. In some embodiments, this method models potential error sources for the reagent analyzer based on testing done using both reagent test devices and control test devices, allowing for the prediction of probability distributions for reagent tests under the combined effect of each error source as well as the equivalent probability distributions for the control test device tests under these same error sources. This method is used to set the acceptable ranges of the control test devices by predicting the reagent result distributions under the combined effect of all error sources, determining if these distributions allow the reagent analyzer to meet performance requirements, and setting the acceptable control test device ranges based on the equivalent resulting control test device result distributions under the same error sources. By considering the effect on reagent result distributions separately from the control test device results, the presently disclosed method considers any interaction effects with reagent factors, effects that may be different from reagent-to-reagent, and effects that may be different at different concentration levels. The presently disclosed reagent analyzer and method allows operators inside or outside of a lab environment to perform control tests on the reagent analyzer with control test devices, such as Munsell test devices and Ronchi test devices. The control tests measure the reflectance of the control test devices (without a volume of sample applied thereto) to assess the reagent analyzer’s clinical performance against the control test device test acceptance criteria. Because the control test device testacceptance criteria is set based on the correlation between the control test device models and models developed separately based on reagent testing, this control test device test acceptance criteria takes into consideration possible differences between these control test device measurements and clinical reagent test measurements to determine whether the reagent analyzer is properly calibrated.

[0018] In some embodiments, the present disclosure describes a method inwhich a series of patient tests with reagent test devices are executed on one or more first reagent analyzer to obtain a series of reagent test device measurements, the one or more first reagent analyzer having a plurality of first error sources and a plurality of first performance requirements. A series of control tests for control test devices are also executed on the one or more first reagent analyzer to obtain control test device measurements. Control test device test acceptance criteria are generated for control test device measurements of control tests to be run by the one or more first reagent analyzer on control test devices, the control test device test acceptance criteria being calculated with the reagent test device measurements and the control test device measurements and allowing for analyte measurement precision that meets the plurality of first performance requirements. The control test device test acceptance criteria is stored in a non-transitory computer readable medium coupled to a processor of a second reagent analyzer.

[0019] In some embodiments, the second reagent analyzer is mutuallyexclusive to the one or more first reagent analyzer. For example, the second reagent analyzer can be in production and located at a manufacturing site, and the one or more first reagent analyzer can be located in a lab environment.

[0020] In some embodiments, the control tests are first control tests, the controltest devices are first control test devices, and the control test device measurements are first control test device measurements. In these embodiments, a series of second control tests for second control test devices are executed on the second reagent analyzer to obtain second control test device measurements. The second control test device measurements are compared to the control test device test acceptance criteria. Then, first data responsive to the second control test device measurements being within the control element test acceptance criteria, and second data responsive to the second control test device measurements being outside of the control element test acceptance criteria is stored in the non-transitory computer readable medium.

[0021] In some embodiments, a set of patient models with the series of reagenttest measurements is created. The series of patient models include at least one model for each of the plurality of first error sources. The series of patient models are used to calculate a between reagent analyzer variability patient value.

[0022] In some embodiments, a set of control models is created with the seriesof control test device measurements, the series of control models include at least one model for each of the plurality of first error sources. The series of control models is used to calculate a between reagent analyzer variability control value.

[0023] In some embodiments, the second reagent analyzer has second errorsources and second operation performance metrics identical to the first error sources and first operation performance metrics of the one or more first reagent analyzer.

[0024] The foregoing Summary provides an overview of certain selectedimplementations or embodiments disclosed herein, and is not intended to describe every aspect, embodiment, implementation, feature, or advantage of the disclosure exhaustively or comprehensively. Therefore, this Summary should not be construed in such a way to limit the scope of this disclosure or to limit the scope of the claims. The details of one or more implementation or embodiment disclosed herein are set forth in the accompanying drawings and descriptions below. Other aspects, features, implementations, embodiments, and advantages will become readily apparent in view of the description, the drawings, and the claims set forth herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] To assist those of ordinary skill in the relevant art in making and usingthe inventive concepts disclosed herein, reference is made to the appended drawings and schematics, which are not intended to be drawn to scale, and in which like reference numerals are intended to refer to the same or similar elements for consistency. For purposes of clarity, not every component may be labeled in every drawing. Certain features and certain views of the figures may be shown exaggerated and not to scale or in schematic in the interest of clarity and conciseness. In the drawings:

[0026] FIG.1 is a front elevation view of an exemplary embodiment of a reagentanalyzer according to the inventive concepts disclosed herein, showing a transparent shield positioned in a field of view of an imaging system thereof.

[0027] FIG. 2 is a side elevation view of the reagent analyzer of FIG. 1.

[0028] FIG. 3 is an end elevation view of the reagent analyzer having thetransparent shield positioned within a slot formed in a housing according to the inventive concepts disclosed herein.

[0029] FIG. 4A is a bottom plan view of a circuit board having an aperturesurrounded by onboard light sources according to the inventive concepts disclosed herein to facilitate controlled illumination of the reagent card and reduce light scattering detected by the imaging system.

[0030] FIG. 4B is a block diagram of a reagent analyzer controller of a reagentanalyzer constructed in accordance with the present disclosure, and a side elevational view of a transparent shield positioned below a circuit board, the circuit board having an aperture surrounded by one or more illumination source according to the inventive concepts disclosed herein.

[0031] FIG. 5 is top plan view image of a sample holder and test device asviewed through the transparent shield in accordance with the inventive concepts disclosed herein.

[0032] FIG. 6 is a block diagram showing various error sources in a reflectancetype reagent analyzer such as the reagent analyzer depicted in FIG.1.

[0033] FIGS. 7A and 7B are graphs showing predicted versus actual lightvariation decodes.

[0034] FIGS. 8A and 8B are graphs showing Micro 9 predicted versus actuallight variation decodes.

[0035] FIGS. 9A and 9B are graphs showing predicted versus actual 10SGcalibration bar color decode results.

[0036] FIGS. 10A and 10B are graphs showing predicted versus actual Micro 9calibration bar color decode results.

[0037] FIG. 11 is a graph depicting an example of a quadratic regression fit toraw height variation data.

[0038] FIGS. 12A and 12B are graphs showing actual versus predicted 10SGdecode results for sample height variations.

[0039] FIGS. 13A and 13B are graphs showing actual versus predicted Micro 9decodes for sample height variation.

[0040] FIG. 14 is a graph showing an actual versus predicted indicator resultsfor 2mlU focus dataset.

[0041] FIG. 15 is a graph showing a quadratic regression for a MTF focusdataset.

[0042] FIG. 16 is a graph showing a fitted line plot for a MTF light error dataset.

[0043] FIG.17 is a graph showing actual versus predicted hCG indicator resultsfor a sample height error dataset.

[0044] FIG. 18 is a graph showing an actual versus predicted indicator resultsfor a sample height error dataset.

[0045] FIG. 19 is a graph showing a predicted versus actual hCG test indicatorresults for calibration bar color error dataset.

[0046] FIG. 20 is a graph showing a fitted line plot for a MTF versus mean rawcalibration bar measurements. DETAILED DESCRIPTION

[0047] Before explaining at least one embodiment of the inventive conceptsdisclosed herein in detail, it is to be understood that the inventive concepts are not limited in their application to the details of construction and the arrangement of the components or steps or methodologies set forth in the following description or illustrated in the drawings. The inventive concepts disclosed herein are capable of other embodiments or of being practiced or carried out in various ways. Also, it is to be understood that the phraseology and terminology employed herein is for the purpose of description and should not be regarded as limiting the inventive concepts disclosed and claimed herein in any way.

[0048] In the following detailed description of embodiments of the inventiveconcepts, numerous specific details are set forth in order to provide a more thorough understanding of the inventive concepts. However, it will be apparent to one of ordinary skill in the art that the inventive concepts disclosed herein may be practiced without these specific details. In other instances, well-known features have not been described in detail to avoid unnecessarily complicating the instant disclosure.

[0049] As used herein, the terms "comprises," "comprising," "includes,""including," "has," "having" or any other variation thereof, are intended to cover a non- exclusive inclusion. For example, a process, method, article, or apparatus that comprises a list of elements is not necessarily limited to only those elements but may include other elements not expressly listed or inherently present therein.

[0050] Unless expressly stated to the contrary, "or" refers to an inclusive or andnot to an exclusive or. For example, a condition A or B is satisfied by anyone of the following: A is true (or present) and B is false (or not present), A is false (or not present) and B is true (or present), and both A and B are true (or present).

[0051] In addition, use of the "a" or "an" are employed to describe elements andcomponents of the embodiments herein. This is done merely for convenience and to give a general sense of the inventive concepts. This description should be read to include one or at least one and the singular also includes the plural unless it is obvious that it is meant otherwise.

[0052] Further, as used herein any reference to "one embodiment" or "anembodiment" means that a particular element, feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment.

[0053] Analyte, as used herein, means a substance whose chemicalconstituents are being identified and measured. Exemplary analytes include leukocyte, nitrite, urobilinogen, protein, pH, blood, specific gravity, ketone, bilirubin, glucose, microalbumin, and creatinine.

[0054] Assay-related variability estimate, as used herein, means an estimate oferror from error sources that are not reagent analyzer specific. Examples of error sources that are not reagent analyzer specific include operator variability and consumable lot variability.

[0055] Between Instrument Variability, as used herein, means an estimate oftotal variability from reagent analyzer to reagent analyzer as estimated by the combined effect of error budgets from multiple reagent analyzer specific error sources.

[0056] Calibration bar, as used herein refers to a flat white plastic strip within ahousing of the reagent analyzer that spans the width of a reagent strip and is used as a reference measurement to compute percent reflectance for test devices (e.g., strips).

[0057] Composite variability, as used herein, means a combined effect ofbetween instrument variability and other assay-related variability estimates.

[0058] Control test device, as used herein, refers to a carrier havingpredetermined area(s) to be analyzed by a reagent analyzer that does not have a volume of sample deposited thereon. Exemplary control test devices include, but are not limited to Munsell strips and Ronchi strips.

[0059] CSN, as used herein, means Chek-Stik Negative: Negative 10SG stripsample.

[0060] CSP, as used herein, means Chek-Stik Positive: Positive 10SG stripsample.

[0061] Error budget, as used herein, means an allocation of error for each errorsource.

[0062] Error model, as used herein, refers to a mathematic model of an errorsource.

[0063] Error source, as used herein, means a system or process used in themeasurement of a patient test or a manufacturing test. Exemplary error sources include, but are not limited to, calibration bar measurements, sample image measurements, and dark image measurement. Calibration bar measurements include illumination (debris in optical pathway, LED variability), material (lot-to-lot process variation, within-lot process variation), camera (camera gain settings, and debris on camera chip). Exemplary sample image measurements include illumination (LED variability, debris in optical pathway, and sample height), camera (camera focus, and gain settings). In some embodiments, the error source does not include user error, damaged pads, reagent noise or measurements of the IR band.

[0064] K2, as used herein, means Kova 2: High-concentration solution usedwith MICRO9 strip tests.

[0065] K3, as used herein, means Kova3: Low-concentration solution used withMICRO9 strip tests.

[0066] Instrument-related variability estimates, as used herein, refers to all errorsources that cause changes to the way the reagent analyzer takes measurements.

[0067] Microalbumin 9 Urine Strip, as used herein, refers to a particular type oftest device, i.e., urine strip having multiple reagents configured to detect the following analytes in the following sequence: leukocyte, nitrite, blood, pH, creatinine, ketone, microalbumin, protein, glucose, and Black (i.e., a black color band on the Urine Strip that is used to identify the strip type). The microalbumin 9 Urine Strip also includes an infrared band following the glucose reagent for ensuring compliance.

[0068] MTF, as used herein, refers to Modulation Transfer Function: a functionused to quantify the quality of image focus by measuring the distinction between black and white line pairs in the image.

[0069] Munsell strip, as used herein, a particular type of test device used forevaluation of the reagent analyzer’s ability to measure reflectance. The Munsell Strips replicate a urine strip with grayscale pads, meant to match different grayscale Munsell standards, rather than reagent pads. Munsell Strips are available in different types including a N6 Munsell Strip, a N8 Munsell Strip and a N95 Munsell Strip. A N6 Munsell Strip is a Munsell strip with pads that are colored to match the Munsell N6 (dark gray) color standard. A N8 Munsell Strip is a Munsell strip with pads that are colored to match the Munsell N8 (light gray) color standard. A N95 Munsell strip is a Munsell strip with pads that are colored to match the Munsell N95 (white) color standard.

[0070] Reagent, as used herein, means a substance or mixture for use inchemical analysis of an analyte.

[0071] Reagent test device, as used herein, means a carrier having at least onereagent that is intended for use within the reagent analyzer. Exemplary reagent test devices include hCG cassettes, and reagent strips. The reagent(s) of the reagent test device may be used to identify and measure various analytes, including, but not limited to leukocyte, nitrite, urobilinogen, protein, pH, blood, specific gravity, ketone, bilirubin, glucose, microalbumin, and creatinine.

[0072] Ronchi strip, as used herein is a particular type of test device in stripform which contains alternating black and white line pairs which span the length of a 12-pad urine strip. The Ronchi strip is used in manufacturing to evaluate the system’s lens focus. In some embodiments, the Ronchi strip may contain 1.66 line pairs per millimeter.

[0073] Unique analyte result, as used herein, refers to an analyte result that isuniquely defined by a reagent type, an image number, a tested solution type, and a color channel.

[0074] 10SG Urine Strip, as used herein, refers to a urine strip having multiplereagents configured to detect the following analytes in the following sequence: leukocyte, nitrite, urobilinogen, protein, pH, blood, specific gravity, ketone, bilirubin, glucose. The 10SG Urine Strip also includes an infrared band following the glucose reagent for ensuring compliance.

[0075] As used herein “wet reagent test device” refers to a reagent test devicethat has a volume of sample deposited thereon such that the reagent in the reagent device may react with its target constituent if such constituent is present in the sample.A wet reagent test device may also have a volume of a negative control deposited thereon.

[0076] Finally, as used herein qualifiers such as “about,” “approximately,” and“substantially” are intended to signify that the item being qualified is not limited to the exact value specified, but includes some slight variations or deviations therefrom, caused by measuring error, manufacturing tolerances, stress exerted on various parts, wear and tear, and combinations thereof, for example.

[0077] The inventive concepts disclosed herein are generally directed to areagent analyzer for reagent test devices and methods for determining performance reading reagent test devices, and more particularly, but not by way of limitation, a reagent analyzer having a sample holder such that the imaging system is configured to capture images of the sample holder. In some embodiments, the reagent analyzer includes a processor. The processor is configured to receive an image and analyze pixels of the image. While the inventive concepts disclosed herein will be described primarily in connection with automatic reagent analyzers using multiple-profile reagent cards as the reagent test device, the inventive concepts disclosed herein are not limited to automatic reagent analyzers or to multiple-profile reagent cards. For example, a method according to the inventive concepts disclosed herein may be implemented with a manual reagent analyzer, or may be implemented with an automatic reagent analyzer using a reagent test device other than a multiple-profile reagent card, such as a lateral flow immunoassay, dip-and-read reagent test device, or a reel of reagent test devices on a substrate, and combinations thereof, as will be appreciated by a person of ordinary skill in the art having the benefit of the instant disclosure. Further, the inventive concepts disclosed herein may be implemented with any reagent device imaging system which has a field of view with at least one read position in the field of view.

[0078] In particular, a signal value indicative of a color of a reagent test device,such as a reagent pad, control line or test line, changes when the reagent test device is exposed to a sample. For a negative solution, the change in signal value is known (or can be measured) and therefore may become an optional offset signal value. Any change outside of the offset signal value is likely caused by a reaction with a clinical component that is being measured.

[0079] Referring now to FIGS. 1-3, shown therein is an exemplary embodimentof a reagent analyzer 10 according to the inventive concepts disclosed herein. Thereagent analyzer 10 may be an automatic reagent card analyzer, for example. Exemplary embodiments of automatic reagent card analyzers are described in detail in U.S. patent application Serial No.13 / 712,144, filed on December 12, 2012, and in PCT application No. PCT / US2012 / 069621, filed on December 14, 2012, the entire disclosures of which are hereby expressly incorporated herein by reference.

[0080] Generally, the exemplary reagent analyzer 10 may include a housing 14,having a slot 15, the housing 14 surrounding a cavity 18. The reagent analyzer 10 may also include, at least one rail 19, an imaging system 22 comprising at least a camera 26, a sample tray 30 having a sample holder 32 positioned within the cavity 18, a transparent shield 31, and a circuit board 34 having an aperture 38 and one or more illumination source 42a-n positioned within the cavity 18.

[0081] The housing 14 may be formed from one or more componentsconfigured to form the cavity 18 and support the at least one rail 19, imaging system 22, the sample tray 30, the transparent shield 31, and the circuit board 34. In one embodiment, the housing 14 is opaque to visible light. In another embodiment, the housing 14 is opaque to one or more wavelength of light generated by the one or more illumination source 42a-n. In one embodiment, the housing 14 may normalize ambient light. In other non-limiting embodiments, the housing 14 has the slot 15 within which the transparent shield 31 may be positioned into the housing 14, and from which the transparent shield 31 may be removed from within the housing 14.

[0082] The transparent shield 31 has at least one sidewall 33, at least one end35, a first surface 36 extending between the at least one sidewall 33 and the at least one end 35, a second surface 37 positioned opposite the first surface 36 extending between the at least one sidewall 33 and the at least one end 35, and an intermediate region 41 extending between the first surface 36 and the second surface 37. In one non-limiting embodiment, the transparent shield 31 has a first sidewall 33a, a second sidewall 33b positioned opposite the first sidewall 33a, a first end 35a, a second end 35b positioned opposite the first end 35a, the first surface 36 extending from the first end 35a to the second end 35b, a second surface 37 positioned opposite the first surface 36 extending from the first end 35a to the second end 35b, and the intermediate region 41 extending between the first surface 36 and the second surface 37. In one embodiment, the transparent shield 31 is transparent to visible light such that light may travel through the transparent shield 31 without appreciable scattering allowing objects positioned beyond the transparent shield 31 to be seen and imagedclearly. In some embodiments, the transparent shield 31 may have a degree of translucence such that light may travel through the transparent shield 31 with scattering allowing objects positioned beyond the transparent shield 31 to be seen with varying degrees of clarity. The first surface 36 and the second surface 37 may both be planar and substantially parallel so as to avoid magnifying visible light passing through the transparent shield 31. As will be discussed in more detail below, the transparent shield 31 is configured to protect the imaging system 22 from splatter or other debris resulting from movement of the sample tray 30 into and out of the housing 14. In some non-limiting embodiments, the transparent shield 31 may be movable within and out of the housing 14 through the slot 15. For example, the transparent shield 31 may have a grip (not shown) on at least one end 35 of the transparent shield 31. The grip may be a textured surface, e.g., a frost or an etching on the first surface 36 and / or the second surface 37, or may include a handle extending from and / or connected to the first surface 36 and / or the second surface 37, or the like. In some non-limiting embodiments, the transparent shield 31 has an aperture 43 positioned on an edge of the transparent shield 31 extending from the first surface 36 through the intermediate region 41 to the second surface 37.

[0083] In some non-limiting embodiments, the at least one rail 19 may bepositioned within the cavity 18, adjacent to the slot 15 such that upon the positioning of the transparent shield 31 within the slot 15, the surface 37 of the transparent shield 31 may be positioned on the at least one rail 19. In other non-limiting embodiments, the at least one rail 19 may be positioned within the cavity 18, adjacent to the slot 15 such that upon positioning the transparent shield 31 within the slot 15, the surface 36 of the transparent shield 31 may be positioned on the at least one rail 19. The transparent shield 31 may function as a barrier to prevent debris from the sample holder 32 from contacting the circuit board 34 and / or the imaging system 22.

[0084] The imaging system 22 includes the at least one camera 26 and issupported by the housing 14. In one embodiment, the imaging system 22 may be fixed to the housing 14 or fixed at a relative distance from the sample tray 30 or the transparent shield 31, for example. The imaging system 22 and / or the camera 26 may include one or more lens with a focal length selected to provide a field of view 40 to include at least the aperture 38 of the circuit board 34.

[0085] The imaging system 22 may be implemented and function as anydesired reader such that the field of view 40 of the imaging system 22 includessubstantially the entire aperture 38 of the circuit board 34, for example. The imaging system 22 may be supported at a location above, below, or beside the sample tray 30. In some embodiments, the field of view 40 may extend in a linear direction from the imaging system 22 to the aperture 38. In other embodiments, the field of view 40 may extend in a non-linear direction from the imaging system 22 to the aperture 38 due to the presence of one or more optical steering component in the field of view 40. Exemplary optical steering components include mirror(s), lens(es), beam splitter(s), or combinations thereof. The imaging system 22 may be configured to detect or capture an image or an optical signal indicative of a reflectance value or a color value of a reagent pad, a lateral flow assay, or the like, (shown in FIGS.5-7 and discussed in more detail below) positioned in the field of view 40 of the imaging system 22, for example. In other non-limiting embodiments, the imaging system 22 may be configured to detect or capture an image or an optical signal indicative of a reflectance value or a color value of the sample holder 32 positioned in the field of view 40 of the imaging system 22, through the transparent shield 31. It is to be understood, however that in some exemplary embodiments, the field of view 40 of the imaging system 22 may include only a portion of the aperture 38 of the circuit board 34. It is also to be understood, that in some exemplary embodiments, field of view 40 of the imaging system may include only a portion of the transparent shield 31. The camera 26 of the imaging system 22 may include any desired digital or analog imager, such as a digital camera, an analog camera, a CMOS imager, a diode, and combinations thereof. The imaging system 22 may also include a lens system, optical filters, collimators, diffusers, or any other optical-signal processing devices, for example. Further, the imaging system 22 is not limited to an optical imager in the visible spectrum, and may include an infrared imaging system, an ultra-violet imaging system, a microwave imaging system, an X-ray imaging system, and / or other desired imaging systems, for example. Non-exclusive examples of the imaging system 22 include optical imaging systems, spectrophotometers, gas chromatographs, microscopes, infrared sensors, and combinations thereof, for example.

[0086] In one embodiment, the imaging system 22 includes at least one camera26 and lens wherein the at least one camera 26 is an AR0239: CMOS Image Sensor, 2.3 MP, 1 / 2.7” and the lens is a DSL949 Sunex lens (Sunex Inc., Carlsbad, CA), both configured to maintain a large field of view 40 while keeping geometric image distortion low, thereby providing a resolution of 1080 pixels by 1920 pixels wherein each pixeldepicts approximately a 0.065mm square area of the sample tray 30 and / or the sample holder 32.

[0087] The sample tray 30 may be configured to adjust the location of thesample holder 32 within the field of view 40. The sample holder 32 may be configured to receive at least one of reagent test device 44, which may be a reagent card and a reagent card cassette, each having a sample 46. The sample 46 may be any bodily fluid, tissue, or any other chemical or biological sample, and combinations thereof, such as urine, saliva, or blood, for example. The sample 46 may be in liquid form and may contain one or more target constituents such as bilirubin, ketones, glucose, or any other desired target constituent, for example.

[0088] The circuit board 34 having the aperture 38 may be positioned within thecavity 18 and interposed between the imaging system 22 and the sample tray 30 such that the field of view 40 of the imaging system 22 is substantially unobstructed from the sample holder 32, the reagent test device 44, and / or the sample 46. The circuit board 34 is described in FIGS. 1-2, and FIG. 4B and in more detail below. In one embodiment, the circuit board 34 is positioned at a fixed location between the imaging system 22 and the sample tray 30; however, in another embodiment, the circuit board 34 may be adjusted to varying locations between the imaging system 22 and the sample tray 30. If the circuit board 34 is adjustable, a calibration routine (described below) would have to be performed after any adjustment. In other non-limiting embodiments, the circuit board 34 may be positioned between the imaging system 22 and the transparent shield 31, such that the field of view 40 of the imaging system 22 is substantially unobstructed from the sample holder 32, the reagent test device 44, and / or the sample 46.

[0089] Referring again to FIG. 1 and FIG. 2, the housing 14 may include aplurality of connected sidewalls 80, 82, 84 and 86 cooperating to surround the cavity 18. The sidewall 80 is spaced from the sidewall 82, and the sidewall 84 is spaced from the sidewall 86. The transparent shield 31 may be sized and dimensioned to traverse the cavity 18 between the sidewalls 80 and 82, and the sidewalls 84 and 86 thereby dividing the cavity 18 into a first portion 88 and a second portion 90. In some embodiments, the transparent shield 31 may have a length L between approximately 5 cm and approximately 21 cm. In some non-limiting embodiments, the transparent shield 31 may be a protecting device for the lens, wherein the transparent shield 31 may have a length of approximately 5 cm. In other non-limiting embodiments, thetransparent shield 31 may be a protecting device for the lens and the optical components, wherein the transparent shield 31 may have a length of approximately 14 cm. In some non-limiting embodiments, wherein the transparent shield 31 is a protecting device for the lens and the optical components, the transparent shield 31 may not be removable from the reagent analyzer 10.

[0090] Referring to FIG. 3, shown therein is an end elevational view of thetransparent shield 31 positioned within the slot 15 according to the inventive concepts disclosed herein. In some embodiments, the housing 14 has the slot 15, and the transparent shield 31 is positioned within the cavity 18 adjacent to the slot 15, the slot having a width W1 and a height d1, the transparent shield 31 having a width W2 less than the width W1 of the slot 15 and a thickness d2 less than the height d1 of the slot 15. In an exemplary, non-limiting embodiment the slot may have a width W1 of approximately 4.5 cm, a height d1 of approximately 0.2 cm, and the transparent shield 31 may have a width W2 of approximately 4.3 cm and a thickness d2 of approximately 0.1 cm. In some non-limiting embodiments, the height d1 of the slot 15 may be greater than 0.1 cm. In some non-limiting embodiments, the thickness d2 of the transparent shield 31 may be between approximately 0.1 cm and approximately .3 cm. In some embodiments, the width W1 of the slot 15 may be between approximately 1.7 cm and approximately 4.5 cm. In some non-limiting embodiments, the width W2 of the transparent shield 31 may be between approximately 1.5 and approximately 4.3 cm. In some embodiments, the transparent shield 31 is movably supported within the housing 14 and aligned with the slot 15 such that the transparent shield 31 is movable through the slot 15. In some embodiments, the surface 36 and the surface 37 of the transparent shield 31 are planar and parallel within the intermediate region 41 to avoid distorting or scattering light passing through the transparent shield 31. The transparent shield 31 may be separate from the imaging system 22 and configured to block debris originating from the sample tray 30 from coming into contact with the imaging system 22. The transparent shield 31 may be constructed of glass, ceramic, plastic, such as acrylic, polycarbonate, and the like.

[0091] The illumination source 42a-n may be implemented as one or more of alight emitting diode, a light bulb, a laser, an incandescent bulb or tube, a fluorescent light bulb or tube, a halogen light bulb or tube, or any other desired light source or object configured to emit an optical signal having any desired intensity, wavelength, frequency, or direction of propagation, for example. The illumination source 42a-n maybe attached to the circuit board 34 and may be oriented such that substantially the entire field of view 40 of the imaging system 22 is illuminated by the illumination source 42a-n. In some exemplary embodiments, the illumination source 42a-n may be operably coupled with a controller 144 (see FIG.4B - described in detail below) so that control and / or power signals may be supplied to the illumination source 42a-n by the controller 144. Desirably, the intensity of the optical signal emitted by the illumination source 42a-n is maintained substantially constant through the operation of the reagent analyzer 10, such as by control and power signals supplied by the controller 144. In one embodiment, the optical signals emitted by the illumination source 42a-n may be conditioned or processed by one or more optical or other systems (not shown), such as filters, diffusers, polarizers, lenses, lens systems, collimators, and combinations thereof, for example.

[0092] In some exemplary embodiments the one or more illumination source42a-n may be implemented, such as a first illumination source 42a and a second illumination source 42b, and the first illumination source 42a and the second illumination source 42b may have different locations and / or orientations thereby causing the first illumination source 42a and the second illumination source 42b to cooperate to illuminate substantially the entire field of view 40 of the imaging system 22. (e.g., substantially the entire sample holder 32 and / or sample 46). The first illumination source 42a and the second illumination source 42b may emit optical signals having different illumination intensities, for example.

[0093] In one embodiment, the sample holder 32 may be adapted to accept thereagent test device 44 in the form of a reagent card cassette having one or more multiple-profile reagent cards 124 therein, for example. An exemplary reagent test device 44 is shown in FIG.5 and described in more detail below. Each reagent test device 44 (detailed below) may include a substrate and one or more reagent pads positioned thereon, or otherwise associated therewith. In an exemplary embodiment, the reagent pads may include fluidic or microfluidic compartments (not shown).

[0094] Each reagent pad of the one or more reagent test device 44 may includea reagent configured to undergo a color change in response to the presence of a target constituent such as a molecule, cell, or substance in the sample 46 of a specimen deposited on the reagent pad. The reagent pads may be provided with different reagents for detecting the presence of different target constituents. Different reagents may cause one or more color change in response to the presence of a certainconstituent in the sample 46, such as a certain type of analyte. The color developed by a reaction of a particular constituent with a particular reagent may define a characteristic discrete spectrum for absorption and / or reflectance of light for that particular constituent. The extent of color change of the reagent and the sample 46 may depend on the amount of the target analyte present in the sample 46, for example.

[0095] The presence and concentrations of these target analytes in the sample46 may be determinable by an analysis of the color changes undergone by the one or more reagent pads at predetermined times after application of the sample 46 to the reagent pads and / or at predetermined read positions in the field of view of the imaging system 22, for example. This analysis may involve a color comparison of each reagent pad to itself at different time periods after application of the sample 46 and / or at different read positions in the field of view 40 of the imaging system 22.

[0096] Based upon an analysis of a magnitude of the optical signal detected bythe imaging system 22 the sample 46 may be assigned to one of a number of categories, e.g., a first category corresponding to no target constituent present in the sample 46, a second category corresponding to a small concentration of target constituent present in the sample 46, a third category corresponding to a medium concentration of target constituent present in the sample 46, and a fourth category corresponding to a large concentration of target constituent present in the sample 46, for example.

[0097] Further, the imaging system 22 may detect an optical signal indicative ofa color or a reflectance value of a reagent pad and / or a test strip at any time interval after a volume of sample 46 has been dispensed on the reagent test device 44, e.g., the reagent pad and / or test strip, and regardless of location of the particular reagent pad and / or test strip, for example. In one exemplary embodiment, a video, or a sequence of images may be captured of the reagent pad and / or test strip at a variety of time intervals after a volume of sample 46 is deposited on the reagent pad and / or test strip.

[0098] The imaging system 22 may be operated intermittently, continuously, orperiodically, to detect one or more reflectance signals indicative of the color or the reflectance value of the one or more test devices 44, e.g., reagent pads, at any time and at any position in the field of view of the camera 26, for example. In some exemplary embodiments, the imaging system 22 may capture an image indicative of the color or the reflectance value of the reagent test device 44, e.g., the reagent pad,prior to any sample 46 being deposited onto the reagent pad, or at any known time after a volume of sample 46 has been deposited onto the reagent pad, for example.

[0099] Referring now to FIG.4A, shown therein is a bottom plan view of a circuitboard having an aperture surrounded by onboard light sources according to the inventive concepts disclosed herein to facilitate controlled illumination of the reagent card and reduce light scattering detected by the imaging system 22.

[0100] Controlled illumination will be described herein by way of example asuniform illumination across an extent, i.e., length and width, of the sample holder 32 and / or sample 46 within acceptable limits. It should be understood, however, that the present disclosure is not limited to uniform illumination. The circuit board 34 is comprised of a substrate 60 having a bottom surface 61a and a top surface 61b, a plurality of conductive leads extending on or in the substrate 60, and the aperture 38 extending between the bottom surface 61a and the top surface 61b.

[0101] In one embodiment, shown in FIG. 4A, the one or more illuminationsource 42a-n is a plurality of LEDs 64a-n and one or more IR LED 68. The LEDs 64a- n shown in FIG.4A include twenty (20) visible light LEDs arranged as shown in FIG. 4A, and one or more IR LED 68. The LEDs 64a-n includes any LED that is needed to produce a substantially uniform light intensity across the sample holder 32 and / or reagent card 124 or reagent cassette. The IR LED 68 may be used to apply heat to the sample 46, or identify an ID pad on the reagent test device 44, for example. In one embodiment, the ID pad is utilized to correlate the sample 46 on the reagent test device 44 supported by the sample holder 32 with a data obtained by the reagent analyzer 10.

[0102] In one embodiment, the plurality of LEDs 64a-n are selected to providea fixed color, visible light, ultra-violet light, infrared light, or white light, or some combination thereof. In another embodiment, each LED 64a-n is positioned at an angle relative to the reagent card 124. In yet another embodiment, each LED 64a-n is positioned at one or more distance from the reagent test device 44 supported by the sample holder 32 such that a first LED 64 and a second LED 64 are different distances from the reagent test device 44 and / or the sample holder 32.

[0103] In some non-limiting embodiments, by adjusting the power level of eachLED 64a-n, a substantially uniform light intensity may be achieved. The substantially uniform light intensity may be between 85% - 100% uniform. The construction, useand calibration of the circuit board 34 are described in U.S. Serial Nos. 63 / 064,609; and 63 / 225,124.

[0104] The circuit board 34 as shown in FIG.4A depicts the bottom surface 61aof the circuit board 34 having one or more illumination source 42a. As placed within the reagent analyzer 10, the bottom surface 61a is oriented to face the sample tray 30 such that light produced by the one or more illumination source 42a-n may be directly shown onto the sample holder 32 and / or sample 46. The illumination sources 42a-n are connected to the plurality of conductive leads of the circuit board 34 such that the conductive leads provide electricity to each illumination source 42a-n. In one embodiment, the circuit board 34 further includes illumination source circuitry (not shown) connected to the plurality of conductive leads and configured to apply electricity independently to each illumination source 42a-n. For example, the illumination source circuitry may be configured to supply a first power to a first illumination source 42a and a second power to a second illumination source 42b wherein the first power and the second power are different, thereby causing a difference in illumination intensity across the sample. The illumination sources 42a-n are arranged such that the illumination intensity across the field of view 40 of the camera 26 is substantially uniform, thereby increasing accuracy of readings of color changes of the reagent pads as the reagent pads are illuminated with a substantially uniform intensity (depicted in more detail below and in FIGS.5-7). The substrate 60 of circuit board 34, as shown in FIG.4B, is substantially planar thereby causing each of the one or more illumination source 42a-n to be a similar distance from the sample tray 30. Depending upon the location of the illumination source 42a-n relative to the sample 46, the distance between the illumination source 42a-n and the sample 46 may be different for certain of the illumination sources 42a-n. However, in other embodiments, the circuit board 34 may be non-planar thereby causing one of more of the illumination source 42a-n to be located at different distances from the sample tray 30. In one embodiment, one or more illumination source 42a-n may be affixed to a standoff (not shown) where each standoff is affixed to the circuit board 34 and provides one or more conductive paths to a particular one of the illumination source 42a-n. When a standoff is used, this causes a portion of the one or more illumination source 42a-n to be closer to the sample 46 and / or the sample tray 30.

[0105] In one embodiment, the substrate 60 has a first region 62a, a secondregion 62b opposite the first region 62a and an intermediate region 62c between thefirst region 62a and the second region 62b. The one or more illumination source 42a- n may be affixed to the substrate 60 in each of the first region 62a, second region 62b, and intermediate region 62c, or some combination thereof. In one embodiment, a first power may be applied to the one or more illumination source 42a-n within the first region 62a and within the second region 62b thereby causing the one or more illumination source 42a-n within the first region 62a and within the second region to provide a first illumination intensity and a second power may be applied to the one or more illumination source 42a-n within the intermediate region 62c thereby causing the one or more illumination source 42a-n within the intermediate region 62c to provide a second illumination intensity, the first power and the second power being different and the first illumination intensity and the second illumination intensity being different.

[0106] The aperture 38 of the circuit board 34 extends from the top surface 61bto the bottom surface 61a to provide an opening for the field of view 40 of the imaging system 22 to pass through from the imaging system 22 through the transparent shield 31 to the sample holder 32 and provide the camera 26 with a controlled view of the reagent test device 44 associated with the sample holder 32. The aperture 38 may be further configured such that the bottom surface 61a of the circuit board 34 may include one or more illumination source 42a-n on each side of the aperture 38. In one embodiment, the aperture 38 is located substantially within the intermediate region 62c. In one embodiment, the aperture 38 has a first major axis and a first minor axis and the sample holder 32 has a second major axis and a second minor axis wherein the first major axis is aligned with the second major axis. While the aperture 38 is depicted as a rectangle in FIG. 4A for providing a controlled view of a rectangular reagent test device, it is understood that the aperture 38 may be configured of any shape such that the field of view 40 is a controlled view of the sample tray 30 and the illumination source 42 can be calibrated to provide a substantially uniform illumination of the sample 46. In the example of FIG.4A, the aperture 38 does not extend to an edge of the circuit board 34.

[0107] In one embodiment, the aperture 38 extends to an edge of the circuitboard 34 without bisecting the circuit board 34 whereas in another embodiment, the aperture 38 extends through the entire circuit board 34, bisecting the circuit board into a first half and a second half, wherein the first half and the second half are mounted at separate locations and supported by the housing 14 such that the field of view 40 is a controlled view of the sample tray and the illumination source 42.

[0108] In some non-limiting embodiments shown in FIG. 5, the reagent testdevice 44 may include a substrate 128 and one or more, or a plurality of reagent pads 132a-n positioned thereon, or otherwise associated therewith. The substrate 128 may be constructed of any suitable material, such as paper, photographic paper, polymers, fibrous materials, and combinations thereof, for example. The reagent pads 132a-n may be arranged in a grid-like configuration on the substrate 128 so as to define one or more test strip, for example. In an exemplary embodiment, the reagent pads 132a- n may include fluidic or microfluidic compartments (not shown). The reagent pads 132a-n may be spaced apart a distance from one another so that the test strips are spaced apart such that adjacent test strips and / or reagent pads 132a-n may be simultaneously positioned at separate positions within the field of view 40 of the imaging system 22, for example. The reagent test device 44 may be a multiple-profile reagent card having multiple reagent pads having different reagents and / or multiple different test strips. Further, in some exemplary embodiments, the reagent test device 44 may include one or more calibration chips or reference pads, which may have no reagent and may serve as color references, for example. In another embodiment, the reagent card 44 includes an ID pad having an identifier visible under IR light.

[0109] Each reagent pad 132a-n may include a reagent configured to undergoa color change in response to the presence of a target constituent such as a molecule, cell, or substance in the sample 46 of a specimen deposited on the reagent pad 132a- n. The reagent pads 132a-n may be provided with different reagents for detecting the presence of different target analytes. Different reagents may cause one or more color change in response to the presence of a certain analyte in the sample 46, such as a certain type of analyte. The color developed by a reaction of a particular constituent with a particular reagent may define a characteristic discrete spectrum for absorption and / or reflectance of light for that particular constituent. The extent of color change of the reagent and the sample may depend on the amount of the target analyte present in the sample 46, for example.

[0110] The color change may be read by the imaging system 22. Signalsindicative of the color of the reagent pads 132a-n may be received and / or captured in an image by the imaging system 22, which may analyze the signals and determine a change in the color of the reagent pad 132a-n as a result of the reagent pad 132a-n reacting with the volume of sample 46 deposited thereon. Such color change may be analyzed as a function of the read position of the reagent pad 132a-n when the opticalsignal or image indicative of the color of the reagent pad 132a-n was detected and / or as a function of the known duration of time the volume of sample 46 has been deposited onto the reagent pad 132a-n, and combinations thereof, for example. The color change may be interpreted as a quantitative, qualitative, and / or semi-qualitative indication of the presence and / or concentration or amount of a target constituent in the volume of sample 46 deposited on the reagent pad 132a-n as described above.

[0111] Referring now to FIG. 4B, shown therein is an analyzer diagram 140depicting the reagent analyzer 10 including the transparent shield 31 and an analyzer controller 144. The analyzer controller 144 has at least a processor 148 and a non- transitory computer readable memory 152. The memory 152 may store computer executable instructions that, when executed by the processor 148, causes the processor 148 to communicate with and / or be operably coupled to other elements of the reagent analyzer 10. While the analyzer controller 144 is depicted separately from the reagent analyzer 10, it is understood that in some embodiments, the analyzer controller 144 may be integrated into the reagent analyzer 10, such as, by way of example only, the analyzer controller 144 may be an additional component of the reagent analyzer 10 or may be integrated with another component of the reagent analyzer 10, for example, the circuit board 34.

[0112] In one embodiment, the imaging system 22 may be operably coupledwith the analyzer controller 144 and / or the processor 148 so that one or more power and / or control signals may be transmitted to the camera 126 and / or to the one or more illumination source 42a-n by the controller 144, and so that one or more signals may be transmitted from the camera 126 to the processor 148, for example. The analyzer controller 144 may be configured to gauge test results as the reagent test device 44 is sampled within the reagent analyzer 10, for example, by receiving one or more signals from the camera 126. The camera 126 may be configured to detect or capture one or more optical or other signals through the transparent shield 31 that are indicative of a reflectance value of the reagent test device 44, such as a reagent pad, and to transmit a signal indicative of the reflectance value of the reagent test device 44, e.g., the reagent pad, to the processor 118, for example. One or more optical signals having wavelengths indicative of a reflectance value of the reagent pads and / or the reagent test device 44 may be detected through the transparent shield 31 by the camera 126 at each read position, for example. The camera 126 may detect an optical signal through the transparent shield 31 indicative of a reflectance value of a reagentpad 132 and / or test strip at any desired read position, location, or area within the field of view 40, or any other desired location or area or multiple locations or areas, for example. The signal transmitted to the processor 148 by the camera 126 may be an electrical signal, an optical signal, and combinations thereof, for example. In one embodiment, the signal is in the form of an image file having a matrix of pixels, with each pixel having a color code indicative of a reflectance value. Each pixel includes at least one color space pixel value for at least one channel conforming to a color space. In some embodiments, each pixel includes a plurality of channels conforming to a color space, such as RGB. The pixel values from the plurality of channels can be analyzed separately as described herein, or averaged to form a single value. When the pixel values are averaged, the single value would represent a gray scale pixel value conforming to a single channel.

[0113] In an exemplary embodiment, the image file may have two or morepredetermined regions of pixels, each predetermined region of pixels corresponding to a read position of one of the reagent pads 132, a calibration bar and / or the control test device in the field of view 40 of the camera 126. In one embodiment, the processor 148 may store the signal transmitted and or the image file in one or more database 156 and / or in the memory 152.

[0114] The processor 148 may determine the reflectance value or the colorchange of reagent pad and / or a test strip along with a sample (e.g., urine) disposed on the reagent pad and / or test strips based on the signals detected by the camera 126, for example. Each optical or other signal indicative of one or more reflectance value readings detected by the camera 126 may have a magnitude relating to a different wavelength of light (i.e., color). The color of the sample(s) and / or the reaction of the one or more reagents with a target analyte in the reagent pad 132 may be determined based upon the relative magnitudes of the reflectance signals of various color components, for example, red, green, and blue reflectance component signals. For example, the color of each reagent pad may be translated into a standard color model, which typically includes three or four values or color components (e.g., RGB color model, including hue, saturation, and lightness (HLS) and hue, saturation, and value (HSV) representation of points and / or CMYK color model, or any other suitable color model) whose combination represents a particular color. In some embodiments the camera 126 may detect multiple optical signals at each read position, with each detected signal having one or more color components, such as a red componentsignal, a green component signal, and a blue component signal, for example, and each of the component signals may be transmitted to the processor 148. In some exemplary embodiments, the camera 126 may detect a single optical signal at each read position, and the processor 148 may translate a signal received from the camera 126 into separate color component signals such as a red component signal, a green component signal, and a blue component signal, for example.

[0115] In one embodiment, the methods described herein for calibrating thereagent analyzer 10 or determining whether the reagent analyzer 10 is performing within proper operating specifications may be implemented as a set of processor executable instructions or logic stored in the non-transitory computer readable medium, which instructions or logic when executed by the processor 148, cause the processor 148 to determine whether or not the reagent analyzer 10 is performing within proper operating specifications. The method for determining whether the reagent analyzer 10 is performing within proper operating specifications may be carried out periodically such as at a preset internal time as desired according to specific quality control procedures applicable to the reagent analyzer 10, and combinations thereof, for example.

[0116] In some embodiments, the processor 148 is further configured to have aset of processor executable instructions or logic stored in the non-transitory computer readable medium. wherein when the instructions or logic are executed by the processor 148, cause the processor 148 to cause an action on at least one periodic interval selected from a group consisting of: initiate an alert in a form perceivable by a human. In some embodiments, the periodic interval is based on a period of time or a number of tests performed by the reagent analyzer 10.

[0117] Referring now to FIG. 5, shown therein is exemplary image 170aillustrating a top plan view of the sample holder 32 and reagent test device 44, in the form of the reagent test device 44, as viewed through the exemplary transparent shield 31 positioned within the housing 14 of the reagent analyzer 10 in accordance with the present disclosure.

[0118] In some non-limiting embodiments, the processor 148 is furtherconfigured have a set of processor executable instructions or logic stored in the non- transitory computer readable medium, wherein when the instructions or logic are executed by the processor 148, cause the processor 148 to cause an action on at least one periodic interval selected from a group consisting of: initiate an alert in a formperceivable by a human. In some embodiments, the periodic interval is based on a period of time or a number of tests performed by the reagent analyzer 10.

[0119] To better understand the performance of the reagent analyzer 10, it isdesired to break the reagent analyzer 10 into a set of processes that are utilized in the measurement of patient (reagent test device) and control test device tests and understand which error sources could contribute error to these error sources. In one embodiment, the present disclosure describes a method for creating control test device test acceptance criteria. The control test device test acceptance criteria is saved in the one or more database 156 and / or in the memory 152 and is used by the processor 148 to compare the results of control tests measuring the reflectance of the control test devices (without a volume of sample applied thereto) to the control test device test acceptance criteria. This comparison assesses the reagent analyzer’s clinical performance against the control test device test acceptance criteria thereby providing an indication of the reagent analyzer is meeting performance expectations. Then, first data responsive to the control test device measurements being within the control element test acceptance criteria, and second data responsive to the second control test device measurements being outside of the control element test acceptance criteria is stored in the memory 152 or the database 156. The first data and the second data can be used by the processor 148 to perform an action. For example, the processor 148 can check to see if the first data is stored within the memory 152 or the database 156. If so, the processor 148 can generate and transmit a signal to circuitry to provide an alert to notify a user that the reagent analyzer 10 is properly calibrated. The processor 148 can also check to see if the second data is stored within the memory 152 or the database 156. If so, the processor 148 can generate and transmit a signal to circuitry to provide an alert to notify a user that the reagent analyzer 10 is not properly calibrated. The circuitry for generating an alert will vary depending upon the type of alert to be provided. For example, the alert may be provided in visual, audio, or haptic form. The circuitry for generating a visual alert may include a light source, such as an LED, amplifier or the like. The circuitry for generating an audio alert may include a speaker, amplifier or the like. The circuitry for generating a haptic form of alert may include an actuator, such as a solenoid capable of moving or vibrating a device.

[0120] The control test device test acceptance criteria can be determined usinga 12 step process described below.

[0121] First, experiments are setup to test patient and control test device resultsin response to reagent analyzer-related error sources. Patient test experiments using wet reagent test devices are tested by the reagent analyzer 10 for all supported analytes using urine samples with known analyte concentrations at a high and low concentration level. Control test device experiments may be tested for all control test devices supported by the reagent analyzer 10.

[0122] Second, results from patient test experiments are used to create a set ofmodels, = { 1, 2,…, }, to predict patient test results in response to thesesources of error. In some embodiments, there is one model for each reagent analyzer 10 related error source. For each analyte, there may be two models, one model for high concentration and another model for low concentration. Once the tests from the patient test experiments is obtained, the method can either directly correlate final patient results (decodes) with a particular error source using numerical methods, such as regression analysis; or, the method correlates intermediate results with a particular error source using numerical methods, such as regression analysis. The final patient results can then be predicted from the intermediate results. The intermediate result is a result that is used in a determination of the final patient results (such as how the reflectance values at different image times are used to calculate the final patient result), but not the final patient result itself.

[0123] Third, results from control test device test experiments are used to createa set of models, = { 1, 2,…, }, to predict control test device results inresponse to these sources of error. A model for each reagent analyzer error source (illumination, ambient lighting, etc.) is created for each type of control test device.

[0124] Fourth, a set of error sources, = { 1, 2,…, }, is initialized where eachvalue in the set is a tuple representing operating limits for that particular error source. For instance, the ambient light specification says the reagent analyzer can operatewithin 300 and 1000 lux of ambient light. Therefore, 1=(300,1000).

[0125] Fifth, use the patient test models to predict patient test results at the lowand high operating limit for each error source. More formally, calculate ( ) for allintegers from 1 to n error sources.

[0126] Sixth, calculate the range of patient test results, , for each error sourceas the difference between the predicted patient test results at the low and high limit.

[0127] Seventh, estimate patient test result variability, ={ 1, 2,… }, as astandard deviation for each error source. This can be calculated as the predicted range of patient test result variation divided by 4 for each error source (range rule of SD*). V = R / 4.

[0128] Eighth, Calculate the between instrument variability, , as theroot mean square (RMS), for example, of the individual standard deviation estimates for each error source.

[0129] More formally, in some embodiments, =

[0130] In some embodiments, the BIpatient is calculated for all analytes at allconcentration levels. A single BI value represents the between instrument variability for a particular analyte at a particular concentration level.

[0131] Ninth, steps 5-8 are repeated using the control test device modelsinstead to predict total control test device test variability, . Note thatmay be calculated for all of the control test devices supported by the reagent analyzer 10.

[0132] Tenth, given that represents only instrument-related errorsources, combine the BIpatient value with assay-related error sources as a prediction for the maximum patient result variability. Use this maximum patient test result variability to determine if the reagent analyzer 10 meets performance requirements.

[0133] Eleventh, if the estimated maximum patient test result variability allowsthe reagent analyzer to meet performance requirements, then representsthe maximum amount of control test device result variability expected to be seen from instrument to instrument while still being able to meet performance requirements. Assuming the population of control test device results is normally distributed with some mean, , we determine the maximum limit we would want to see the results deviatefrom this mean as ±4* . The maximum limit can be adjusted based on howstrictly it is desired to screen for outliers. In this example, anything outside of + / - 4SD is likely an outlier caused by an excess of system error sources. But, this can be made wider or narrower depending on a desired sensitivity.

[0134] Twelfth, set the control test device acceptance criteria used inmanufacturing as a range within [ 4 , +4 ], for example. Theseranges are set narrower than the maximum limits, as the manufacturing tests may beperformed on reagent analyzers that are unused – this accounts for additional deviation over time over the life of the reagent analyzer. These ranges are referred to herein as control test device test acceptance criteria.

[0135] For both the patient test using the wet reagent test device, and thecontrol test device test, the presently disclosed inventive concepts are measuring the reflected light, based on an image of the wet reagent test device or the control test device, to produce a result. So, in theory, if there is something wrong with the reagent analyzer 10 that affects the reagent analyzer’s measurement system (e.g., too much ambient light, weird lighting, etc.) the test results will show a deviation in both the control test device results AND the reagent test device results. Although the patient test using the wet reagent test devices involve complex chemistry and is ultimately different than the control test device test, the mechanisms and error sources within the reagent analyzer 10 for measuring both of these types of test devices are the same, so any error in the way the reagent analyzer 10 is taking measurements manifests in the results of both types of tests.

[0136] If there is enough of a deviation in the reagent analyzer 10’smeasurement system from these error sources, it is expected that patient results would not meet performance requirements for the reagent analyzer 10. The deviation may be set such that < 99% of results fall within + / - the control test device test acceptance criteria. Given that a deviation to the measurement system can be detected in both the patient test taken with the reagent test device and control test device test taken with the control test device based on the logic above, then there must be some level of deviation in the control test device test results that represents the point at which there is too much deviation in the measurement system to allow for patient test results to meet desired performance requirements. The presently disclosed method determines these levels of deviation, which are referred to herein as control test device test acceptance criteria and can then be set as the pass / fail criteria for the control test device test.

[0137] Because the reagent analyzer 10 can detect deviations to themeasurement system using the control test devices and patient test results using the reagent test device, the Mpatient and Mcontrol models are used to predict what these results would look like given the maximum allowable deviation to the measurement system. The Mpatient and Mcontrol models are used to estimate the results across the operating range for each error source. Statistics can then be used to determinewhat the overall variation in results would look like when the contributions of different error sources are combined. This provides a range of patient test result values that represent the range of values to expect to see if the reagent analyzer is operated within its expected ranges for each error source. Likewise, the Mcontrol models provide a range of control test device test values expected to be seen if the reagent analyzer is operated under its expected ranges for each error source.

[0138] Assuming this range of patient test results is acceptable (meetsperformance requirements), then the range of control test device results is also acceptable. Again, the deviations in the measurement system manifest in the results of both control test device and patient test results. If a control test device test result is outside of this acceptable range, this would indicate that the combined effect of all error sources caused a deviation to the control test device test result that is outside the range, i.e., control test device test acceptance criteria, we would expect if we were operating under normal conditions. Therefore, it would be expected that the patient test results would also be at risk, since these limits directly relate to the total variation in patient test results expected to be seen under normal operating conditions. Therefore, the control test device (pass / fail) test acceptance criteria are based on the range of acceptable control test device results as previously mentioned.

[0139] Set forth below is a description of exemplary error sources, error modelsfor the error sources, measurements, and exemplary calculations of a maximum tolerable error from each error source such that the combined effect of each error source allows for acceptable performance. This allocation of error for each factor is called the error budget.

[0140] The combined effect of each of the different error sources of the reagentanalyzer 10 is referred to as the between-instrument (BI) variability. This estimate is combined with other assay-related variability estimates to determine if the reagent analyzer 10 is expected to meet performance requirements with all variability sources considered. Exemplary manners to calculate this BI estimate and adjust the error budget for the reagent analyzer 10 if the calculated values are too high are disclosed. These BI estimates are calculated for both patient (i.e., reagent test device) and control test device tests.

[0141] Given that different error sources should be limited to meet the proposederror budget and BI estimates, the following describes how to establish differentthresholds that flag results that may contain excess error from one or multiple error sources.

[0142] The following description by way of example describes a method fordetermining expected performance of the reagent analyzer 10 with regards to the measurement of reagent test devices and control test devices.

[0143] To determine which error sources contribute to the reagent analyzererror budget, the reagent analyzer 10 is broken down into the raw measurements that are taken by the reagent analyzer 10 and further broken down into a list of error sources that could potentially affect these measurements. An exemplary breakdown of error sources for a reflectance based reagent analyzer is illustrated in Figure 6.

[0144] To collect test results to create error models for each error source, bothcontrol test device and reagent test device test were performed in a collection of controlled experiments. For error sources included in the urine strip error budget, for example, these error sources may be tested in experiments using both control test devices and reagent test devices. Error sources included in the hCG error budget, for example, were tested on both hCG cassettes and Ronchi strips. Note that some error sources may not be directly tested and instead may be either simulated or extrapolated from the other experimental data.

[0145] For most of the tests described below, the number of samples (i.e. thenumber of levels tested for each error source) is also described. The number of samples can be varied based upon the ability of the reagent analyzer 10 to perform the test given software constraints or by limited access to samples. In some embodiments, lower sample sizes may be justified by high correlations (R-squared) present in the data, visual inspection of the test results using scatter plots, and strong hypotheses for the behavior of the data in response to these error sources given in- depth knowledge of the reagent analyzer 10. Once control element test acceptance criteria is created, as discussed herein, the control element test acceptance criteria is stored on other reagent analyzers 10 so that performance of the other reagent analyzers 10 can be tested with control test devices to determine whether the other reagent analyzers10 are calibrated.

[0146] To study the effect of varying calibration bar color on instrumentmeasurement system performance, eight different calibration bars were installed in a single reagent analyzer to collect reagent test device and control test device results for each configuration. Four of the calibration bars were sampled from a first lot ofcalibration bar material, and four were sampled from a second lot of calibration bar material. Both sample sets contained within-lot variability due to differences in molding and coloring processes.

[0147] The following tests were run for each of the calibration bars tested in thisexperiment.

[0148] 3 x 10SG, CSN

[0149] 3 x 10SG, CSP

[0150] 3 x MICRO9, K2

[0151] 3 x MICRO9, K3

[0152] 1 x N6

[0153] 1x N8

[0154] 1 x N95

[0155] Each calibration bar, in this example, was tested on the same reagentanalyzer 10 under the same settings.

[0156] To study the effect of light intensity variation on measurementperformance for the reagent analyzer 10, six sets of LED settings were used to perform tests on Munsell type control test devices and reagent test devices. These settings were chosen to first test a set of datapoints ranging from the minimum and maximum light intensities, and second to test if there are differences between the effect of pad- to-pad light variation and the effect of differences in light on the same pad between tests.

[0157] The following tests were run for each of the LED settings tested in thisexperiment.

[0158] 3 x 10SG, CSN

[0159] 3 x 10SG, CSP

[0160] 3 x MICRO9, K2

[0161] 3 x MICRO9, K3

[0162] 1 x N6

[0163] 1x N8

[0164] 1 x N95

[0165] To study the effect of variability in the sample height on measurementperformance of the reagent analyzer 10, samples were tested at four different heights: the normal strip height, a strip with an additional backing, a strip with two additional backings, and a strip with three additional backings. Each backing may beapproximately 0.25 mm in thickness. It was found to be difficult to stack more than three additional backings in the sample tray, so the height was limited to three backings for this example.

[0166] The following tests were run for each of the LED settings tested in thisexperiment.

[0167] 3 x 10SG, CSN

[0168] 3 x 10SG, CSP

[0169] 3 x MICRO9, K2

[0170] 3 x MICRO9, K3

[0171] 1 x N6

[0172] 1x N8

[0173] 1 x N95

[0174] The effect of focus by the camera sensor on the measurement of hCGindicator values was tested by running hCG tests and Ronchi tests at eight different lens positions, where the increment represented approximately one sixteenth the total camera circumference.

[0175] The range of camera positions was limited by the camera's ability todetect the hCG cassette - at lower levels of focus, the camera was unable to detect the hCG cassette, so the full test sequence could not be executed. The granularity of the adjustments was limited by the size of the camera, allowing only eight positions to be tested.

[0176] At each lens position, the following samples were tested:

[0177] 1 x hCG cassette, 2 mIU (negative)

[0178] 1 x hCG cassette, 25 mIU (borderline)

[0179] 1 x hCG cassette, 100 mIU (positive)

[0180] 1 x Ronchi

[0181] To study the effect of varying LED intensity on hCG and Ronchi testresults, tests were performed using LED settings ranging from 60% to 70% of the maximum intensity with increments of 1%. The range of LED settings is limited by multiple factors. Firstly, at settings below 60%, it was found that the reagent analyzer 10 could not detect the hCG cassette. In addition, values above 70% led to specular reflectance of the calibration bar, causing non-linear effects. In addition, the granularity of the LED intensity settings is limited to 1%, allowing for 10 levels of LED settings to be tested.

[0182] For each set of LED settings, the following samples were tested:

[0183] 1 x hCG cassette, 2 mIU (negative)

[0184] 1 x hCG cassette, 25 mIU (borderline)

[0185] 1 x hCG cassette, 100 mIU (positive)

[0186] 1 x Ronchi

[0187] Given models to predict the hCG and strip error as a function ofcalibration bar reflectance (as presented in the Results section), predictions of how much the calibration bar color is expected to vary within a single lot of calibration bars were made. The calibration bar color tolerance is presented in terms of delta L*a*b* parameters. Thus, a study was performed to correlate white color chips with known L*a*b* parameters to reflectance measurements taken by the reagent analyzer 10.

[0188] Seven white color chips with different color parameters were tested aspart of this study. For each test, the calibration bar was replaced with a different color chip and a Munsell test (e.g., control test device) was performed to allow for the collection of raw calibration bar results.

[0189] There are multiple methods that are utilized to determine the influenceof the error sources on decode results: in some cases, experimental results show a direct relationship between the error source and decode results, and the models are derived from these results. In some cases, the influence of the error source can be fully determined by the error sources' mathematical relationship described by the decode equation, so the models are derived based on simulated values. Lastly, some error sources are like others in the way that they affect certain raw signal measurements, so it is possible to reuse experimental results in combination with simulated results to recalculate decode values and develop models based on these results.

[0190] In discussing the model results, the experimental results, error modelmethods, and model results are presented for each error source. The amount of error that is allocated for each error source is presented. The between-instrument error estimate is calculated based on the total of these different error estimates and is presented below.

[0191] The data collected from the urine strip light variation experiment wasused to generate linear regression models for each analyte dataset, where the decode was set as the response variable and the LED setting was set as a predictor. Asummary of these results for both the 10SG and Micro 9 strip experiments are presented in the following table 1 and table 2.Table 1: Linear Regression Model for 10SG strip analytesTable 2: Linear Regression Model Results for Micro 9 Strip Analytes

[0192] Although some datasets are found to have a high R-squared value, mostdata show little correlation with the LED light output. This result indicates that much of the variability in the decode result s cannot be explained solely by the LED light output. In addition, it is found that that many of the p-values exceed a significance level of0.05, indicating the light has little significance on the decode results compared to the overall variability in the datasets.

[0193] Given the lack of correlation displayed in these models, a differentmethod was utilized to determine the effect of varying LED settings on the decode results. The decode results are calculated from a series of raw signal measurements in different color channels at different time points. If these raw measurements are highly correlated with light output, then models for these raw measurements can be used to predict raw signals independent of other variability error sources, and these predicted raw signals can be used to directly calculate the decode variability as a response to light variability.

[0194] Thus, to model the light error, it is desired to determine the effect of lightvariability on both the dark-subtracted calibration bar measurements, and dark- subtracted raw sample measurements (at each time point and in each color channel). However, the raw calibration bar signal can be used to describe the strength of the lighting, so regression models were generated for each unique raw analyte dataset with the calibration bar measurements as predictors. The r-squared values for each of the regression models for both the 10SG and Micro 9 results are presented in table 3 and table 4.Table 3: Linear regression model results for unique raw 10SG sample measurementsTable 4: Linear regression model results for unique raw Micro 9 sample measurements.

[0195] As seen in the tables, the raw dark-subtracted signals are highlycorrelated, meaning most of the variability seen in the sample measurements can be attributed to lighting variability. Given this high correlation, it is possible to predict sets of unique raw sample measurements for a given light output, where the light output is expressed as a calibration bar measurement. Because the raw sample models predict results solely as a response of light variability, the calculated decode results from these predictions contain only variability from lighting as well.

[0196] The accuracy of the decode predictions using this method can beassessed by comparing predicted and experimental decode results. This comparison is made by predicting decode results using the dark-subtracted raw calibration bar values from the experimental results as inputs to each of the error models. The resulting predicted raw reagent results are normalized by the experimental calibrationbar values, and the normalized results are used to calculate the predicted decode results. These predicted decode results are compared to the collected decode results.

[0197] From figures 7A, 7B, 8A and 8B, it is seen that the predicted decoderesults fall within the range of the actual results. The decode results is a numerical value corresponding with the level of analyte concentration in the sample as measured by the reagent analyzer 10. The range of decode values for a particular analyzer is categorized, such that each category corresponds to a semi-quantitative measurement of the analyte concentration. It is noted that in some cases the spread of the predicted results is less than that of the actual results – this outcome illustrates how the variability due to sources other than lighting was eliminated using these error models. For results where the actual and predicted decodes have similar spreads, such as with K3 blood and CSN specific gravity, it is found that the r-squared values in the experimental models are higher, indicating that much of the variability can be explained by the lighting. Therefore, these raw analyte error models can ultimately be used to predict decode results as a response to lighting variability.

[0198] A similar method for calculating the lighting effect was performed for theMunsell results, where the raw calibration bar values were used as inputs to linear regression models for the raw Munsell results. The model r-squared results are described in the following table 5.Table 5: Linear regression model results for normalized Munsell lighting variation

[0199] Again, the model can be used to accurately predict raw Munsell resultsfrom raw calibration bar results, and thus the models are sufficient for predicting normalized Munsell lighting error.

[0200] As was done with the light variability data, the decode results from thecalibration bar color experiments are utilized to generate linear regression models, where the categorical predictor for the model is the calibration bar type.Table 6: Linear regression model results for 10SG calibration bar color variationTable 7: Linear regression model results for Micro 9 calibration bar color variation

[0201] In most cases, the correlation between the decode results and thecalibration bar seems to be somewhat high. However, most of the error models do not have high enough R-squared values to allow for confident predictions of the decode results in response to calibration bar color variability. In addition, these models utilizecategorical predictors, meaning predictions can only be made for the calibration bars used in this study. To generalize the predictions, it is desired to create regression models that are based on the raw calibration bar values.

[0202] The approach to determining the effect of calibration bar color variabilityon reagent decode results is like that used in the lighting variability models. Because there is a known mathematical relationship between the raw calibration bar signal and the decode results, predictions of the decode results can be made for calibration bar color variation if the total raw calibration bar measurement variation is known.

[0203] To make these predictions, the experimental results are used todetermine the average raw analyte value for each unique result. Then, the experimental raw calibration bar measurements are used to normalize these averages, and general linear models are created to describe the relationship between the raw calibration bar signal and the calculated normalized results. Because the unique raw analyte measurements in a unique dataset are held constant, the resulting normalized datasets have a perfect linear relationship with the raw calibration bar variation – this variation in the normalized results is solely due to the calibration bar signal, so any reagent noise is eliminated.

[0204] These calibration bar color variation models can then be used to predictsets of normalized reagent results for a given set of raw calibration bar inputs. The resulting predictions can then be used to predict decode results which have variation that is directly attributed to raw calibration bar variation.

[0205] As was previously done, the experimental decodes are plotted againstthe predicted decodes, where the predicted datasets were generated using the measured raw calibration bar values from the experiment.

[0206] Again, as shown in FIGS. 9A, 9B, 10A, and 10B, it is found that themodels accurately predict the decode values, with spreads that are similar or smaller than the experimental results. Because a significant amount of the experimental variation could be explained by the calibration bar signal variation, it is seen that many of the predicted and experimental decode spreads are similar.

[0207] A similar method was used to predict the effect of calibration bar colorvariation on normalized Munsell results – all raw Munsell values in each dataset were set to the average of the set and then renormalized by the experimental raw calibration bar values. The resulting normalized Munsell datasets only contain variation for the calibration bar color variation.

[0208] The decode results from the sample height variation experiments wereused to create regression models, where the change in height of the sample was set as a predictor for the models. A summary of the model results is displayed in the following Table 8 and Table 9. Reagent Solution Effect pValue RsquaredTable 8: 10SG sample height variation linear regression model results Reagent Solution Effect pValue RsquaredTable 9: Micro 9 sample height variation linear regression model results

[0209] From these results, it is found that many of the datasets are highlycorrelated with the height of the sample, and many p-values indicate the height to be a significant source of variability in the results. However, it is desired to eliminate as much noise from the data as possible to capture the variability that is purely due to sample height variation. It is found that many normalized reagent datasets display a quadratic relationship with the height of the sample – using this model allows for higher R-squared values than a linear model. Also, each set of repetitions at a single height within the normalized datasets are averaged, eliminating further repetition-to-repetition noise. As shown in Figure 11, the resulting datasets are highly correlated with the height of the sample, as illustrated in the following example.

[0210] Summaries of the model R-squared results are presented in Table 10and Table 11.Table 10: Summary of regression model results for 10SG height variation raw dataTable 11: Summary of regression model results for Micro 9 height variation raw data

[0211] From Table 11, it is found that most of the error models have high R-squared values, indicating that much of the variability in the normalized datasets can be explained by the quadratic relationship with sample height. These models can be used to predict sets of normalized reagent values from a single sample height value. However, some of the models still have a very low R-squared, indicating that the effect of sample height may be insignificant compared to noise in the reagent sample.

[0212] As previously done, the model accuracy in predicting decode results isassessed by comparing the experimental decode results to the predicted results using the experimental heights as inputs to the model. These results are illustrated in Figures 12A, 12B, 13A and 13B.

[0213] Again, as shown in FIGS. 12A, 12B, 13A and 13B, the models seem toaccurately predict the decode values for each reagent at each solution level. Because the R-squared values for the experimental decode versus height models were mostly high, it is known that many of the results should have variability mostly due to the sample height variation – this result is confirmed in figures 12A, 12B, 13A, and 13B,as the predicted results due to sample height variability have ranges close to the ranges of the experimental decode values.

[0214] The same approach is utilized to determine the effect of height variationon normalized Munsell results. The R-squared results for these models are summarized in the following Table 12.Table 12: Normalized Munsell versus height regression model results

[0215] It is found that the normalized Munsell results are highly correlated withthe height of the sample.

[0216] Although an experiment to test the effects of white balance error ondecode and Munsell results was not performed, the light variability models can be utilized to approximate these effects. The effect from variability in camera gain settings on results is like the effect from lighting in that both error sources affect the raw calibration bar and raw sample signal in a proportional way. The main difference between these error mechanisms is that white balance error affects each color channel independently, whereas lighting error influences all three of the color channels.

[0217] To incorporate this difference into the models, a set of raw calibrationbar results were generated, where the differences between the color channels have randomly generated values between -10 and 10. The raw calibration bar values are input into the light variability models to predict raw reagent results, and the predicted raw reagent are normalized by the generated raw calibration bar values. The resulting normalized values are used to predict decode results. Two linear models are created using these predicted decode results – one with the delta between the raw red and green calibration bar values as the predictor, and the other with the delta between the raw red and blue calibration bar values as the predictor. Summaries of the model R- squared values are presented in the following tables 13 and table 14.Table 13: 10SG white balance regression model resultsTable 14: Micro 9 white balance regression model results

[0218] For the reagents that use both the red and green color channels, it isfound that the delta RG models have high correlations. For the models that use both the red and blue color channels, it is found that the delta RB models have high correlations.

[0219] The same method was used to predict the variation in the normalizedMunsell results because of white balance variation. However, additional models were not needed to describe decode variation, so the only models used to make predictions were the light variation models.

[0220] The error models used to determine the effect of ambient light ingresson the reagent test device and Munsell test results were generated from Munsell data collected in environments with lighting ranging from illuminance values of 300 to 80000 lux. Using this data, regression models were created to predict dark imagemeasurements at each reagent pad location for different illuminance levels. The R- squared values for each of these models are summarized in table 15.Table 15: Light ingress dark image regression model results

[0221] Given the strong correlation observed between the illuminance levelsand the dark measurements, it is possible to determine the effect of ambient light variation on analyte decode results. This effect is determined by predicting dark measurements for a given ambient illuminance level, subtracting out these predicted values from the experimental raw analyte and raw calibration bar measurements, renormalizing these predicted raw analyte values by the predicted raw calibration bar values, and using these predicted normalized values to calculate decode values.

[0222] The effect of ambient lighting variation on normalized Munsell values isdetermined by directly correlating the normalized values with the measured illuminance values using linear regression models. The R-squared values observed from these linear regression models are presented in the following table 16.Table 16: Light ingress Munsell regression model results

[0223] The results from the hCG focus experiment were used to create linearregression models correlating the hCG indicator results with the mean MTF value in the hCG measurement window. The following table 17 summarizes the results of these models.Table 17: hCG test indicator focus regression model results

[0224] It is found that the camera focus has a significant effect (P-Value < 0.05)on both the 25 mIU and 100 mIU results. Thus, these error models are used to predict the effect of camera focus for these indicator results. On the other hand, the 2 mIU indicator results had little correlation with camera focus, so an alternative approach was established to determine the effect of camera focus.

[0225] It was found that the camera focus has a significant effect on both theminima and peak value results from the 2 mIU dataset, as displayed in the linear regression model results in the following table 18.Table 18: Peak value and minima focus regression model results for 2 mIU dataset

[0226] The peak value and minima represent the peak signal and minimumsignal measured in the region of the test indicator, where the reported indicator value is the difference between these two values. Given the significant effect of the camera focus on both values, it is possible to use these models to predict the effect on the indicator value. Using these models, both the minima and the peak value are calculated for each experimental mean MTF datapoint. The predicted minima are subtracted from the peak values, resulting in the predicted indicator values. Thee predicted values are plotted with the measured indicator values in Figure 14.

[0227] From Figure 14, it is seen that the noise in the measured indicatordataset is much greater than the predicted effect from camera focus variation.

[0228] In addition to the test line indicator models, linear regression modelswere created to determine the effect of camera focus on the reference and control line indicator values. A summary of these models is displayed below in Table 19.Table 19: Reference and control indicator regression models for focus variation

[0229] For both the reference and control line datasets, it is found that the effectof camera focus on measured indicator values is negligible compared to the random repetition-to-repetition noise.

[0230] Lastly the effect of camera focus on Ronchi MTF values was assessedby creating a quadratic regression model of the mean MTF results (measured in the hCG measurement region) as a function of camera height. The regression results are depicted graphically in Figure 15.

[0231] It is found that adjustments to the camera position have only a slighteffect on the MTF values at higher focus levels, but this effect greatly increases as the camera deviates from the ideal position.

[0232] To assess the impact of lighting variability on hCG test indicator values,linear regression models were created for the measured indicator values as a functionof the average raw calibration bar values in the hCG measurement region. The results of these models for each hCG concentration level are summarized in Table 20.Table 20: hCG test indicator regression model results for lighting error

[0233] It is found that the effect from light variation on test indicator results isnegligible compared to the random noise observed in the dataset. Thus, light variability is disregarded as a cause for variation in test indicator results.

[0234] Similar regression error models were created for both the reference andcontrol indicator values as a function of the measured raw calibration bar signal. These results are displayed in Table 21.Table 21: hCG reference and control indicator regression model results for lighting error

[0235] Again, the effect of lighting on the measured indicator value is notsignificant for either result type.

[0236] Lastly, the results from the Ronchi light variability testing were used togenerate a linear regression model of the average MTF value in the hCG region as a function of the average raw calibration bar value in the hCG region. The results of this model are as follows in Table 22. R-sq P-Value0.4467 0.025Table 22: MTF regression model results for lighting error

[0237] From this model, it is found that the lighting has a significant effect onthe MTF result. The linear correlation between the lighting and the MTF values is somewhat low, however a graphical depiction of the data in Figure 16 displays a clear linear relationship between the two with a few outlier datapoints.

[0238] This model is used to determine the amount of variability in MTF resultsthat must be budgeted for lighting variability.

[0239] The data from the hCG and Ronchi focus experiment was utilized tosimulate the effect of varying sample height on indicator and MTF results. It is known that varying the sample height causes a variation in the sample measurement while the calibration bar measurements remain roughly the same. Thus, sample heightvariation was simulated by replacing each set of calibration bar images from the focus experiment with a single set of calibration images, and the image sets were reanalyzed using the system software. The results from the simulated hCG datasets were used to create linear regression error models, where the indicator values were regressed against the mean MTF value in the hCG region. The following is a summary of the regression model results.Table 23: hCG test indicator regression model results for sample height error

[0240] For both the 25 mIU and 100 mIU datasets, it was found that the sampleheight had a significant effect on indicator results. However, to better understand this effect, linear regression error models were created to correlate the minima and peak values with the mean MTF values, as was done with the focus study. The results of these regression models are shown in Table 24.Table24: hCG minima and peak value regression model results for sample height error

[0241] Given the much higher correlation seen in these error models comparedto the indicator value error models, these error models were used to predict the effect of sample height variation on indicator results. Again, the indicator results were predicted by utilizing the error models to predict minima and peak values for the measured mean MTF values and subtracting the predicted minima from the predicted peak values to get the predicted indicator values. Linear regression error models, for example, were created for these predicted indicator values to determine the effect of sample height variation on indicator results. An example of the predicted indicator results compared to the actual indicator results is illustrated in Figure 17.

[0242] The same method for predicting the effect of sample height variability ontest indicator results was utilized for the reference and control line indicator results, where linear regression models, for example, were created for both the minima andpeak values as a function of mean MTF. The results of these models are summarized in Table 25.Table 25: Reference and control minima and peak value regression model results for sample height error

[0243] From the model results, it is found that the sample height variation hasa significant effect on the reference and control minima and peak values. Again, these models are used to predict minima and peak values for the measured mean MTF values, and the predicted indicator results are calculated from the difference between the peak values and minima. The predicted indicator results are used to create regression error models, for example, to estimate the effect of sample height variation on reference and control indicator results. An example of the predicted indicator values compared to the actual indicator values is displayed in Figure 18.

[0244] As was the case with the test indicator results, much of the effect fromsample height variation is hidden by random noise.

[0245] Lastly, the effect of sample height variation on MTF results wasestimated by creating a linear regression model of the mean MTF value in the hCG region as a function of sample height. Given that the height of the camera changes by about 0.5mm per revolution, it was determined that the camera height changed by approximately 0.03mm per increment in the test. The results of the regression correlating the average MTF values to this change in height are summarized in Table 26.Table 26: MTF versus sample height regression model results.

[0246] This model is utilized to determine the amount of error in the MTF that isexpected given the tolerancing in the sample height.

[0247] Like the sample height variability analysis, the effect of calibration barcolor variation on hCG indicator results and Ronchi results was evaluated using image sets with simulated calibration bar color variation. This variation was simulated usingthe image sets from the light variability study, where the sample images from each test were replaced with a single set of sample images. The resulting image sets effectively represent calibration bar color variation, as the calibration bar images change from the light variation, and the sample images remain constant.

[0248] These new image sets were reprocessed, and the results were used tocreate linear regression error models of the test indicator values as a function of the average green calibration bar values measured in the hCG window region. The model results are summarized in the following Table 27.Table 27: hCG test indicator regression model results for calibration bar color error

[0249] As was seen in some of the previous analyses, a stronger correlationcould be made between the peak values and minima values with the calibration bar measurements. The results of these linear regressions are displayed as follows in Table 28.Table 28: hCG test indicator peak value and minima regression model results for calibration bar color error

[0250] Given the high correlation seen in these results, these error models wereutilized to predict peak values and minima values from the mean calibration bar measurements. The predicted minima values were then subtracted from the predicted peak values to get the predicted indicator values. Again, these predicted values were used to create linear regression error models, for example, to accurately describe the effect of calibration bar color variation on test indicator results. An example comparison of the predicted and actual indicator results is displayed in Figure 19.

[0251] The same analysis method was performed for both the reference andcontrol indicator values. The results of these regression models are presented below in Table 29.Table 29: Reference and control peak value and minima regression model results for calibration bar color variation

[00252] Again, the models were used to predict minima and peak values, andthe predicted values were subtracted to get the predicted indicator values.

[00253] Lastly, the data from the simulated Ronchi image sets was utilized tocreate a linear regression error model of the mean MTF value in the hCG measurement window region as a function of the mean calibration bar color (green channel) in that region. The results of this error model are summarized in Figure 20.

[00254] As seen from these results, the MTF results do not seem to besignificantly correlated with variation in calibration bar color.

[00255] The results from the raw calibration bar color precisions study were usedto create linear regression models, where the raw calibration bar measurements in each color channel were correlated with each color parameter. It was found that the red and green channel results are highly correlated with the a* parameter, and the blue channel results are highly correlated with the L* parameter. The results of these error models are summarized in the following table 30.Table 30: Raw calibration bar color versus delta L*a*b* parameter regression model results

[0256] Given the high correlation and significant effects seen in these results,the models can be used to predict the amount of raw calibration bar measurement variation from a given delta L*a*b* specification on the calibration bar color.

[0257] The error budget estimates for each error source were calculated usingboth instrument specifications and the error models outlined in the previous sections. The error budget is split into the hCG error budget and the reagent test device error budget, which accompany Ronchi and Munsell manufacturing test error budgets. An estimation for the sum of the total error contributed from each error source is calculated as an estimate for the between-instrument variability. The associated total error estimates calculated from the control test device tests are utilized to determinemaximum thresholds for manufacturing test acceptance criteria, where values outside these bounds indicate a problem with system performance.

[00258] The following table 31 displays the specifications used in this study thatdictate the total amount of error to be budgeted for each error source.Table 31: Specifications for controls on each error source

[00259] The “Control” column in this table specifies the implemented control ineither the system software, part specifications, or manufacturing tests that limit each error source.

[00260] These lower and upper specifications are input into the previouslymentioned error models to determine the range of results caused by each error source. Note that the specifications can be changed and result in the same predicted decode ranges if the range of the lower and upper specifications stays the same. The following tables 32A and 32B show the predicted decode ranges for each of the reagent test device analytes for different solution levels.Table 32A Predicted Range of 10SG and Micro 9 Decode Error for Each Error SourceTable 32B Predicted Range of 10SG and Micro 9 Decode Error for Each Error

[0261] Given the decode range estimates from each error source, it is possibleto conservatively estimate the standard deviations caused by each error source using the range rule for standard deviations (i.e., dividing the range by four). The BI variability for each analyte at each solution level is estimated by calculating the root mean square of these standard deviations. The estimated BI results are presented in the following table 33.Table 33: BI estimates for 10SG and Micro 9 decodes

[0262] To better compare to the performance of the predecessor system, it wasrequested to also calculate the BI variability without the dirty calibration bar error, as this error source was not included for the previous systems.

[0263] In addition to the urine strip analyte results, error source estimationswere calculated for the normalized Munsell channels for each Munsell type. The following tables 34A and 34B displays the estimated ranges of normalized Munsell result variation caused by each error source.Table 34A – Predicted range of normalized Munsell error for each error sourceTable 34B – Predicted range of normalized Munsell error for each error source

[0264] Again, these range estimates were divided by 4 to estimate the standarddeviations, and the RMS of the standard deviations was computed to get the BI variability in terms of normalized Munsell test results. In addition, the maximum tolerable range for the Munsell acceptance criteria, representing the maximum amount of Munsell variation that can be tolerated before failing performance requirements, was estimated by multiplying the BI estimates by 4. These values are conservative estimates for the range of normalized Munsell results given the combined effects of the different error sources. Both the BI estimates and maximum range estimates are shown in the following table 35.Table 35: BI estimates for normalized Munsell results

[0265] In addition to the reagent and Munsell strip, the hCG and Ronchi errorbudget estimates were calculated using the corresponding error models. The following table 36 displays the estimated ranges of indicator variance caused by each error source.Table 36: Estimated ranges of hCG indicator result error for each error source

[0266] Note that the light variation was not included in these results, as themodels showed that the light had a negligible effect on indicator variation.

[0267] Again, the estimated ranges were divided by 4 to estimate the standarddeviations, and the RMS of the standard deviations was calculated to estimate the BI variability. These estimates are displayed in the following table 37. Concentration Est. BI2 mIU 5.121225 mIU 20.1044100 mIU 47.2834Reference Peak 23.7125Control Peak 30.3735Table 37: Estimated BI for hCG indicator results

[0268] Lastly, the same process was performed to estimate the standarddeviations in MTF values caused by each error source, with the BI variability calculated from the estimations. A summary of these results is shown in Table 38.Table 38: MTF error estimates, in terms of standard deviation, for each error source

[0269] It is noted from these results that the calibration bar error sources arenot included – the Ronchi calibration bar color variation models showed that the calibration bar color variation had no significant effect on the Ronchi results.

[0270] The BI estimates for both hCG and urine strip analytes at differentconcentration levels can be combined with other assay-related error source estimates to calculate the guard bands for the desired performance requirements. Given that the desired performance requirements leads to acceptable results with these guard bands, then the total budgeted error for each error source is also acceptable. With this assumption, it is recommended that the range of the limits, i.e., control test device test acceptance criteria for each of the error source controls, including manufacturing controls, software limits, and specifications, in some embodiments are set to or within the range of the limits presented in Table 31. In this example, any limits set outside ofthose presented in this table could put the reagent analyzer 10 at risk of unacceptable performance.

[0271] In addition, total error estimates for manufacturing test results werecalculated. Because these error estimates were calculated given the same inputs as the assay result error estimates, these values effectively represent the maximum range that the manufacturing test results can vary while still allowing for acceptable analyte test performance. Thus, the total range values presented in Table 35 and Table 38, for example, represent the maximum range that the manufacturing test acceptance criteria should have in some embodiments. The total ranges presented in these tables are referred to as performance limits, and the total ranges allowed by the manufacturing testers are referred to as the control test device test acceptance criteria. It is recommended that the control test device test acceptance criteria are set tighter than the performance limits to allow for additional instrument variation caused by usage.

[0272] Variations in levels of camera focus mainly affect the ability of theinstrument to distinguish indicators on the hCG cassette. However, it is also possible to evaluate the effect of focus variation on reagent test device results and budget for this effect.

[0273] To evaluate the effect of camera focus variation on reagent test devices,reagent test device image datasets from development trial studies were altered to contain image noise to simulate the blurring effect at lower camera focus levels. This noise was created using 2-D Gaussian filtering with a standard deviation of 2.5. This filtering effectively decreased the average focus for each dataset by approximately 15%, as proven by comparing Ronchi results before and after this filtering. This method for simulating lower focus levels was chosen over physical testing to separate out the effects from camera focus and camera height, as the height of the camera changes as lens is adjusted during focusing.

[0274] Each of the lower-focus datasets were reanalyzed using the automatedpipeline to determine the effect on decode results. The resulting decodes were compared to the decode results from the unaltered dataset. A summary of these results are shown in Table 39.Table39: Summary of decode comparison results for strip region focus study

[0275] In table 39, the mean delta is a calculation of the mean differencebetween the original and simulated decode values. In other words, this value represents the average amount that the decode changed for each analyte due to the differences in focus between the datasets. Likewise, the SD delta represents the standard deviations of these difference, showing how variable the calculated deltas were for each analyte dataset.

[0276] To evaluate the significance of decreased focus on analyte results, thesemean deltas can be used to estimate the impact to the BI calculations presented in Table 33. As previously done, the range rule for standard deviation is applied to the mean values to conservatively estimate the variability due to decreased focus. These standard deviations can then be combined with the previously calculated BI estimates by taking the root mean square of the original BI estimate and the estimated SD for each analyte. A summary of these calculations is presented in the following table 40.Table 40: Focus variability impact to BI estimates.

[0277] In table 40, Est. BI represents the original BI estimates, and New BIrepresents the new estimates containing the focus variability. The Min Bin Widthcolumn shows the minimum decode bin width for the associated analyte, and the Percent Bin Width shows how much the BI estimate changed as a percentage of the minimum decode bin width. In table 40, it is found that most BI estimates do not change at all when the estimates are rounded to one decimal place as was done in the calculations for FFU requirements. For those BI estimates that do change with the focus estimates included, the change is insignificant – these changes in BI values are less than 0.2% of the minimum width of the associated decode bins.

[0278] The following is a number list of non-limiting illustrative embodiments ofthe inventive concept disclosed herein:

[0279] Illustrative Embodiment 1. A reagent analyzer, comprising:a housing surrounding a cavity, the housing being opaque to visible light; an Illumination source in the housing; an imaging system having a camera sensor with a field of view within the cavity, the camera sensor configured to generate pixel data conforming to a color space having at least one color channel; a sample tray positioned within the cavity, the sample tray having a sample holder within the field of view of the camera sensor, the sample tray being positioned a distance away from the camera sensor, the sample tray holding a control test device and a calibration bar within the field of view of the camera; a processor coupled to a non-transitory computer readable medium storing computer executable instructions that when executed by the processor cause the processor to: actuate the imaging system to cause the camera sensor to capture an image of the control test device and the calibration bar, the image being a reflectance measurement of the control test device; analyze a first region of interest of the image to determine first color space pixel values of the control test device within the image for the at least one color channel; analyzing a second region of interest of the image to determine second color space pixel values of the calibration bar within the image for the at least one color channel; calculate normalized values of the at least one color channel by determining a percent reflectance as a ratio of the first color space pixel values to the second color space pixel values;compare the normalized values to control test device test acceptance criteria calculated with control test device measurements and reagent test device measurements; and store in the non-transitory computer readable medium first data responsive to the normalized values being within the control test device test acceptance criteria, and second data responsive to the normalized values being outside of the control test device test acceptance criteria.

[0280] Illustrative Embodiment 2. The reagent analyzer of illustrativeembodiment 1 wherein the at least one color channel includes at least three color channels.

[0281] Illustrative Embodiment 3. The reagent analyzer of illustrativeembodiment 1 or 2, wherein the three color channels are red, green and blue.

[0282] Illustrative Embodiment 4. The reagent analyzer of illustrativeembodiment 1 through 3, wherein the control test device test acceptance criteria are calculated with a patient model derived from the reagent test measurements, and a control model derived from the control test device measurements.

[0283] Illustrative Embodiment 5. The reagent analyzer of illustrativeembodiment 1 through 4, wherein the computer executable instructions that when executed by the processor cause the processor to perform an action responsive to the first data being stored in the non-transitory computer readable medium.

[0284] Illustrative Embodiment 6. The reagent analyzer of illustrativeembodiment 5, wherein the action is generating an alert.

[0285] Illustrative Embodiment 7. The reagent analyzer of illustrativeembodiment 1 through 6, wherein the computer executable instructions that when executed by the processor cause the processor to perform an action responsive to the second data being stored in the non-transitory computer readable medium.

[0286] Illustrative Embodiment 8. The reagent analyzer of illustrativeembodiment 7, wherein the action is generating an alert.

[0287] Illustrative Embodiment 9. A method, comprising:executing a series of patient tests with reagent test devices on one or more first reagent analyzer to obtain a series of reagent test device measurements, the one or more first reagent analyzer having a plurality of first error sources and a plurality of first performance requirements;executing a series of control tests for control test devices on the one or more first reagent analyzer to obtain control test device measurements; generating control test device test acceptance criteria for control test device measurements of control tests to be run by the one or more first reagent analyzer on control test devices, the control test device test acceptance criteria being calculated with the reagent test device measurements and the control test device measurements and allowing for analyte measurement precision that meets the plurality of first performance requirements; and storing the control test device test acceptance criteria in a non-transitory computer readable medium coupled to a processor of a second reagent analyzer.

[0288] Illustrative Embodiment 10. The method of illustrative embodiment 9,wherein the second reagent analyzer is mutually exclusive to the one or more first reagent analyzer.

[0289] Illustrative Embodiment 11. The method of illustrative embodiment 9through 10, wherein the control tests are first control tests, wherein the control test devices are first control test devices, and wherein the control test device measurements are first control test device measurements, and further comprising the steps of: executing a series of second control tests for second control test devices on the second reagent analyzer to obtain second control test device measurements; compare the second control test device measurements to the control test device test acceptance criteria; and store in the non-transitory computer readable medium first data responsive to the second control test device measurements being within the control test device test acceptance criteria, and second data responsive to the second control test device measurements being outside of the control test device test acceptance criteria.

[0290] Illustrative Embodiment 12. The method of illustrative embodiment 9through 11, further comprising creating a set of patient models with the series of reagent test measurements, the series of patient models including at least one model for each of the plurality of first error sources, and using the series of patient models to calculate a between reagent analyzer variability patient value.

[0291] Illustrative Embodiment 13. The method of illustrative embodiment 9through 12, further comprising creating a set of control models with the series of controltest device measurements, the series of control models including at least one model for each of the plurality of first error sources, and using the series of control models to calculate a between reagent analyzer variability control value.

[0292] Illustrative Embodiment 14. The method of illustrative embodiment 9through 13, wherein the second reagent analyzer has second error sources and second operation performance metrics identical to the first error sources and first operation performance metrics of the one or more first reagent analyzer.

[0293] Illustrative Embodiment 15. A reagent analyzer, comprising:a housing surrounding a cavity, the housing being opaque to visible light; an Illumination source in the housing; an imaging system having a camera sensor with a field of view within the cavity, the camera sensor configured to generate pixel data conforming to a color space having at least one color channel; a sample tray positioned within the cavity, the sample tray having a sample holder within the field of view of the camera sensor, the sample tray being positioned a distance away from the camera sensor, the sample tray holding a control test device and a calibration bar within the field of view of the camera; a processor coupled to a non-transitory computer readable medium storing computer executable instructions that when executed by the processor cause the processor to: actuate the imaging system to cause the camera sensor to capture an image of the control test device and the calibration bar, the image being a reflectance measurement of the control test device; analyze a first region of interest of the image to determine first color space pixel values of the control test device within the image for the at least one color channel; compare at least one value based on the first color space pixel values to control element test acceptance criteria calculated with control test device measurements and reagent test device measurements; and store in the non-transitory computer readable medium first data responsive to the normalized values being within the control test device test acceptance criteria, and second data responsive to the normalized values being outside of the control test device test acceptance criteria.

[0294] Illustrative Embodiment 16. The reagent analyzer of illustrativeembodiment 15 wherein the at least one color channel includes at least three color channels.

[0295] Illustrative Embodiment 17. The reagent analyzer of illustrativeembodiment 15 through 16, wherein the three color channels are red, green and blue.

[0296] Illustrative Embodiment 18. The reagent analyzer of illustrativeembodiment 15 through 17, wherein the control test device test acceptance criteria are calculated with a patient model derived from the reagent test measurements, and a control model derived from the control test device measurements.

[0297] Illustrative Embodiment 19. The reagent analyzer of illustrativeembodiment 15 through 18, wherein the computer executable instructions that when executed by the processor cause the processor to perform an action responsive to the first data being stored in the non-transitory computer readable medium.

[0298] Illustrative Embodiment 20. The reagent analyzer of illustrativeembodiment 19, wherein the action is generating an alert.

[0299] Illustrative Embodiment 21. The reagent analyzer of illustrativeembodiment 15 through 20, wherein the computer executable instructions that when executed by the processor cause the processor to perform an action responsive to the second data being stored in the non-transitory computer readable medium.

[0300] Illustrative Embodiment 22. The reagent analyzer of illustrativeembodiment 21, wherein the action is generating an alert.

[0301] From the above description, it is clear that the inventive conceptsdisclosed herein are well adapted to carry out the objects and to attain the advantages mentioned herein as well as those inherent in the inventive concepts disclosed herein. While exemplary embodiments of the inventive concepts disclosed herein have been described for purposes of this disclosure, it will be understood that numerous changes may be made which will readily suggest themselves to those skilled in the art and which are accomplished within the scope of the inventive concepts disclosed and as defined in the appended claims.

Claims

What is claimed is:

1. A reagent analyzer, comprising:a housing surrounding a cavity, the housing being opaque to visible light; an Illumination source in the housing; an imaging system having a camera sensor with a field of view within the cavity, the camera sensor configured to generate pixel data conforming to a color space having at least one color channel; a sample tray positioned within the cavity, the sample tray having a sample holder within the field of view of the camera sensor, the sample tray being positioned a distance away from the camera sensor, the sample tray holding a control test device and a calibration bar within the field of view of the camera; a processor coupled to a non-transitory computer readable medium storing computer executable instructions that when executed by the processor cause the processor to: actuate the imaging system to cause the camera sensor to capture an image of the control test device and the calibration bar, the image being a reflectance measurement of the control test device; analyze a first region of interest of the image to determine first color space pixel values of the control test device within the image for the at least one color channel; analyzing a second region of interest of the image to determine second color space pixel values of the calibration bar within the image for the at least one color channel; calculate normalized values of the at least one color channel by determining a percent reflectance as a ratio of the first color space pixel values to the second color space pixel values; compare the normalized values to control test device test acceptance criteria calculated with control test device measurements and reagent test device measurements; and store in the non-transitory computer readable medium first data responsive to the normalized values being within the control test device test acceptance criteria, and second data responsive to the normalized values being outside of the control test device test acceptance criteria.

2. The reagent analyzer of claim 1 wherein the at least one color channel includesat least three color channels.

3. The reagent analyzer of claim 2, wherein the three color channels are red,green and blue.

4. The reagent analyzer of claim 1, wherein the control test device test acceptancecriteria are calculated with a patient model derived from the reagent test measurements, and a control model derived from the control test device measurements.

5. The reagent analyzer of claim 1, wherein the computer executable instructionsthat when executed by the processor cause the processor to perform an action responsive to the first data being stored in the non-transitory computer readable medium.

6. The reagent analyzer of claim 5, wherein the action is generating an alert.

7. The reagent analyzer of claim 1, wherein the computer executable instructionsthat when executed by the processor cause the processor to perform an action responsive to the second data being stored in the non-transitory computer readable medium.

8. The reagent analyzer of claim 7, wherein the action is generating an alert.

9. A method, comprising:executing a series of patient tests with reagent test devices on one or more first reagent analyzer to obtain a series of reagent test device measurements, the one or more first reagent analyzer having a plurality of first error sources and a plurality of first performance requirements; executing a series of control tests for control test devices on the one or more first reagent analyzer to obtain control test device measurements; generating control test device test acceptance criteria for control test device measurements of control tests to be run by the one or more first reagent analyzer on control test devices, the control test device test acceptance criteria being calculated with the reagent test device measurements and the control test device measurements and allowing for analyte measurement precision that meets the plurality of first performance requirements; andstoring the control test device test acceptance criteria in a non-transitory computer readable medium coupled to a processor of a second reagent analyzer.

10. The method of claim 9, wherein the second reagent analyzer is mutually exclusive to the one or more first reagent analyzer.

11. The method of claim 9, wherein the control tests are first control tests, wherein the control test devices are first control test devices, and wherein the control test device measurements are first control test device measurements, and further comprising the steps of: executing a series of second control tests for second control test devices on the second reagent analyzer to obtain second control test device measurements; compare the second control test device measurements to the control test device test acceptance criteria; and store in the non-transitory computer readable medium first data responsive to the second control test device measurements being within the control test device test acceptance criteria, and second data responsive to the second control test device measurements being outside of the control test device test acceptance criteria.

12. The method of claim 10, further comprising creating a set of patient models with the series of reagent test measurements, the series of patient models including at least one model for each of the plurality of first error sources, and using the series of patient models to calculate a between reagent analyzer variability patient value.

13. The method of claim 12, further comprising creating a set of control models with the series of control test device measurements, the series of control models including at least one model for each of the plurality of first error sources, and using the series of control models to calculate a between reagent analyzer variability control value.

14. The method of claim 9, wherein the second reagent analyzer has second error sources and second operation performance metrics identical to the first error sources and first operation performance metrics of the one or more first reagent analyzer.

15. A reagent analyzer, comprising: a housing surrounding a cavity, the housing being opaque to visible light; an Illumination source in the housing;an imaging system having a camera sensor with a field of view within the cavity, the camera sensor configured to generate pixel data conforming to a color space having at least one color channel; a sample tray positioned within the cavity, the sample tray having a sample holder within the field of view of the camera sensor, the sample tray being positioned a distance away from the camera sensor, the sample tray holding a control test device and a calibration bar within the field of view of the camera; a processor coupled to a non-transitory computer readable medium storing computer executable instructions that when executed by the processor cause the processor to: actuate the imaging system to cause the camera sensor to capture an image of the control test device and the calibration bar, the image being a reflectance measurement of the control test device; analyze a first region of interest of the image to determine first color space pixel values of the control test device within the image for the at least one color channel; compare at least one value based on the first color space pixel values to control element test acceptance criteria calculated with control test device measurements and reagent test device measurements; and store in the non-transitory computer readable medium first data responsive to the normalized values being within the control test device test acceptance criteria, and second data responsive to the normalized values being outside of the control test device test acceptance criteria.

16. The reagent analyzer of claim 15 wherein the at least one color channel includes at least three color channels.

17. The reagent analyzer of claim 16, wherein the three color channels are red, green and blue.

18. The reagent analyzer of claim 15, wherein the control test device test acceptance criteria are calculated with a patient model derived from the reagent test measurements, and a control model derived from the control test device measurements.

19. The reagent analyzer of claim 15, wherein the computer executable instructions that when executed by the processor cause the processor to perform an actionresponsive to the first data being stored in the non-transitory computer readable medium.

20. The reagent analyzer of claim 19, wherein the action is generating an alert.

21. The reagent analyzer of claim 15, wherein the computer executable instructions that when executed by the processor cause the processor to perform an action responsive to the second data being stored in the non-transitory computer readable medium.

22. The reagent analyzer of claim 21, wherein the action is generating an alert.

Images