HK1218668B - System and method for determining the concentration of an analyte in a sample fluid - Google Patents

Description

Systems and methods for determining the concentration of an analyte in a sample fluid

This application is a divisional application. The name of the invention of the parent application is “System and method for determining the concentration of an analyte in a sample fluid”, the application date is December 23, 2010, and the application number is 201080060168.1.

Technical Field

The present invention relates to a measurement method and apparatus for use in measuring the concentration of an analyte in a fluid. The present invention more particularly, but not exclusively, relates to a method and apparatus that can be used to measure the concentration of glucose in blood.

Background Art

Measuring the concentration of a substance, especially in the presence of other contaminants, is important in many fields, and especially in medical diagnostics. For example, the measurement of glucose in body fluids such as blood is crucial for the effective treatment of diabetes.

Diabetes treatment typically involves two types of insulin therapy: basal and bolus. Basal insulin involves continuous, for example, timed-release insulin. Bolus insulin therapy provides additional doses of faster-acting insulin to regulate blood sugar fluctuations caused by various factors, including the metabolism of sugars and carbohydrates at meal times. Proper regulation of blood sugar fluctuations requires accurate measurement of blood glucose concentrations. Failure to do so can result in extreme complications, including blindness or impaired circulation in the limbs, which can ultimately cause a diabetic to lose the use of his or her fingers, hands, feet, etc.

A variety of methods are known for measuring the concentration of an analyte (e.g., glucose) in a blood sample. Such methods generally fall into one of two categories: optical and electrochemical. Optical methods generally involve reflectance or absorbance spectroscopy to observe spectral shifts in a reagent. These shifts are caused by a chemical reaction that produces a color change that indicates the analyte concentration. Alternatively, electrochemical methods generally involve an amperometric or coulometric response that indicates the analyte concentration. See, e.g., U.S. Patent No. 4,233,029 to Columbus, U.S. Patent No. 4,225,410 to Pace, U.S. Patent No. 4,323,536 to Columbus, U.S. Patent No. 4,008,448 to Muggli, U.S. Patent No. 4,654,197 to Lilja et al., U.S. Patent No. 5,108,564 to Szuminsky et al., U.S. Patent No. 5,120,420 to Nankai et al., U.S. Patent No. 5,128,015 to Szuminsky et al., U.S. Patent No. 5,243,516 to White, U.S. Patent No. 5,430,016 to Diebold et al. No. 7,999 to Pollmann et al., U.S. Patent No. 5,288,636 to Pollmann et al., U.S. Patent No. 5,628,890 to Carter et al., U.S. Patent No. 5,682,884 to Hill et al., U.S. Patent No. 5,727,548 to Hill et al., U.S. Patent No. 5,997,817 to Crismore et al., U.S. Patent No. 6,004,441 to Fujiwara et al., U.S. Patent No. 4,919,770 to Priedel et al., U.S. Patent No. 6,054,039 to Shieh, and U.S. Patent No. 6,645,368 to Beaty et al., all of which are incorporated herein by reference in their entirety.

For the sake of user convenience, reducing the time required to display an indication of the glucose level in a blood sample has been a goal of system designers for many years. Test times have been reduced from early colorimetric products, which took approximately two minutes to display a reading, to approximately 20 to 40 seconds. Recently, test times shorter than ten seconds have been described (see, for example, U.S. Patent Nos. 7,276,146 and 7,276,147), and several products currently on the market advertise test times of approximately five seconds. Minimum test times of less than two seconds have been discussed in various patent applications (see, for example, U.S. Patent Application Publication Nos. 2003/0116447A1 and 2004/0031682A1). However, the practical utility of short test times has not been fully achieved with these technologies, as results are substantially affected by confounding interferences.

A significant limitation of electrochemical methods for measuring the concentration of chemicals in blood is the effect of confounding variables on the diffusion of the analyte and the actual composition of various active components. Examples of limitations on the accuracy of blood glucose measurements include changes in blood composition or state (rather than the aspect being measured). For example, changes in hematocrit (red blood cell concentration) or the concentration of other chemicals in the blood can affect the signal generated by a blood sample. Temperature changes in a blood sample are another example of a confounding variable in measuring blood chemistry. In applications where the results are not compensated for other variables or interferents, such as hematocrit and temperature, the practicality of reported blood glucose responses after short test times is questionable.

With respect to the hematocrit in a blood sample, prior art methods have relied on separating red blood cells from the plasma in the sample, for example, using glass fiber filters or reagent membranes containing pore-forming agents that only allow plasma to enter the membrane. Separation of red blood cells using glass fiber filters increases the size of the blood sample required for measurement, contrary to the expectations of test meter customers. Porous membranes are only partially effective at reducing the hematocrit effect and must be used in combination with increased delay times and/or AC measurements (see below) to achieve the desired accuracy.

Prior art methods have also attempted to reduce or eliminate hematocrit interference by using DC measurements that include longer incubation times of the sample on the test strip reagents, thereby reducing the magnitude of the effect of the sample hematocrit on the measured glucose value. Such methods also suffer from greatly increased test times.

Other attempts to reduce or eliminate hematocrit and temperature interferences are taught in U.S. Patent No. 7,407,811 and the disclosures of the parent case of the present application. A low-amplitude AC potential is applied to a sample to determine certain sample characteristics based on the phase angle (also referred to herein as "phase") and admittance information from the current response to the AC excitation signal. As taught, multiple frequencies of the AC excitation signal are applied in consecutive blocks, followed by a conventional DC excitation signal. However, those disclosures indicate the inventors' belief that there is a minimum time limit for applying each frequency in order to obtain useful, known, and reasonably reproducible information from both the AC and DC excitation signals. Despite that, the shortest total test time achievable from a complete AC method is practically 3 seconds. Alternatively, to achieve practical analysis in less than 3 seconds, a limit is imposed on the number of frequency blocks used during AC excitation—that is, 2 blocks instead of 4. However, reducing the number of frequency blocks used can negatively impact the accuracy level achievable when correcting for multiple interferents (e.g., hematocrit and temperature). As taught in these previous disclosures of AC excitation, correction of indicated glucose for multiple interferents can be achieved by obtaining multiple correction factors, such as phase and/or admittance response data at multiple frequencies derived from AC signal excitation. This is particularly beneficial when the multiple correction factors measure separate or different aspects of the interferents, or when it is affected by one interferent but not another.

Furthermore, the correction factors used to determine the expected analyte concentration, or even the measurement results, can also be used to calculate and optionally report additional parameters, such as the hematocrit level or hematocrit range of the blood. By reducing the number of potential correction factors, for example, by measuring the phase and/or admittance of only two frequencies from the AC excitation rather than three, four, or more, potentially useful information may be discarded. For example, information such as the hematocrit level or hematocrit range may be useful information to users, especially healthcare providers in clinical settings, where patients who are more sensitive to medically significant abnormal hematocrits due to disease or treatment may be identified during routine blood glucose testing. For example, providing the hematocrit level in addition to the glucose concentration would be a valuable piece of information in certain settings that may be lost as a result of solutions proposed by the prior art.

Therefore, there is a need for a system and method that more accurately measures blood glucose, even in the presence of confounding variables, including variations in hematocrit, temperature, and concentrations of other chemicals in the blood. Furthermore, there is a need for such a system and method that has a test time of less than 2 seconds. Similarly, there is a need for a system and method that accurately measures any medically significant component of any biological fluid with a test time of less than 2 seconds. It is an object of the present invention to provide such a system and method.

Summary of the Invention

In one embodiment, a method for determining the concentration of a medically significant component of a biological fluid is disclosed, comprising the steps of applying a first signal having an AC component to the biological fluid; measuring a first current response to the first signal; applying a second signal comprising a DC signal to the biological fluid; measuring a second current response to the second signal; combining the first and second responses; and determining an indication of the concentration of the medically significant component. In other embodiments, the time required to complete each step does not exceed approximately 2 seconds. In other embodiments, the total systematic error from the method does not exceed approximately 10%. In other embodiments, the first signal comprises an AC signal comprising a multi-frequency excitation waveform, wherein the different AC frequencies are generally applied simultaneously rather than sequentially to minimize the time required to complete application of the first and second signals.

The present invention can be used with a variety of medically significant components (or analytes, as they are also called, such as glucose, lactate, cholesterol, triglycerides, etc., with glucose being the most prominent analyte) and biological fluids (or sample fluids), such as blood, serum, plasma, urine, etc., with blood being the most typical example.

Other embodiments of the systems and methods will be apparent from the description herein and as set forth in the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be further described, by way of example only, with reference to the accompanying drawings, in which:

1 is a graph of current versus time for measurements using a biosensor having a reagent layer thickness of approximately 3.6 μm, parameterized for the time between sample application and the time of DC excitation.

2 is a table showing glucose, hematocrit, and temperature levels used in the first covariation study described herein.

FIG3 illustrates in tabular form the excitation signal distribution and timing used for the first study described herein.

FIG4 graphically illustrates the excitation signal distribution and timing used for the first study described herein.

5 is a graph of normalized error versus reference glucose for uncorrected measurement data from the first study described herein.

6 is a graph of normalized error versus reference glucose for the data of FIG. 5 corrected using the methods described herein.

7 is a graph of current versus time for measurements using a biosensor having a reagent layer thickness of approximately 1.6 μm, parameterized for the time between sample application and application of a DC excitation.

8 is a graph of admittance versus time showing the stabilization of the AC response in the first study described herein.

9 is a graph showing the cross-sectional thickness of a biosensor reagent strip in the first study described herein.

10 is a table showing glucose, hematocrit, and temperature levels of whole blood samples used in the first study described herein.

FIG. 11 illustrates the excitation signal distribution and timing used for the second study described herein.

12 is a table showing the measured performance of three reagent thicknesses used in the second study described herein.

13 is a graph of normalized error versus baseline glucose levels for the second study described herein.

14 is a Clark error grid showing predicted glucose versus reference glucose for uncorrected DC data obtained in the second study described herein.

15 is a Clark error grid showing predicted glucose versus reference glucose for the DC data of FIG. 14 corrected using AC measured data.

16 is a graph of one embodiment multi-sine excitation waveform used in the third study described herein.

17A is a table of 200 ms admittance and phase response data for a third covariate glycemic measurement study obtained using the methods disclosed herein.

17B is a graph of admittance magnitude versus hematocrit from the data table of FIG. 17A .

FIG17C is a graph of phase contrast hematocrit from the data table of FIG17A.

18 is a table showing both uncorrected blood glucose measurement estimates at several testing times in the third study described herein, as well as measurement estimates for the same data corrected using the methods disclosed herein.

19 is a graph of normalized error versus reference glucose for uncorrected measurement data from the third study described herein.

20 is a graph of normalized error versus reference glucose for the data of FIG. 19 corrected using the methods disclosed herein.

21 is a Clark error grid showing predicted glucose versus reference glucose for both the uncorrected data of FIG. 19 and the corrected data of FIG. 20 .

22 is a table showing both uncorrected blood glucose measurement estimates at several testing times in the fourth study described herein, as well as measurement estimates for the same data corrected using the methods disclosed herein.

23 is a graph of normalized error versus reference glucose for uncorrected measurement data from the fourth study described herein.

24 is a graph of normalized error versus reference glucose for the data of FIG. 23 corrected using the methods disclosed herein.

25 is a Clark error grid showing predicted glucose versus reference glucose for both the uncorrected data of FIG. 23 and the corrected data of FIG. 24 .

26 is a table showing target versus actual values for the results of the fourth covariation study described herein.

27 is a graph of admittance magnitude versus hematocrit from the fourth covariation study described herein.

28 is a graph of phase contrast hematocrit from the fourth covariation study described herein.

29 is an exemplary current response resulting from a measurement sequence comprising a multi-frequency AC excitation waveform followed by a DC excitation, performed on a whole blood sample having a target glucose concentration of 93 mg/dL and a hematocrit of 70%.

30 is a table of estimated and measured dry coating film thickness as a function of coating weight for known wetting agent application processes.

31 is a graph of admittance magnitude versus hematocrit from the fifth covariation study described herein.

32 is a graph of DC current responses measured at different test times and covaried by hematocrit.

33 is a graph of normalized error versus reference glucose for uncorrected DC measurement data measured at 900 ms from the fifth study described herein.

34 is a graph of normalized error versus reference glucose for DC measurement data measured at 900 ms and corrected according to the fifth study described herein.

35 is a graph of normalized error versus reference glucose for DC measurement data measured at 1100 ms and corrected according to the fifth study described herein.

36 is a graph of normalized error versus reference glucose for DC measurement data measured at 1500 ms and corrected according to the fifth study described herein.

37 is a graph of normalized error versus reference glucose for DC measurement data measured at 2000 ms and corrected according to the fifth study described herein.

38 is a graph of normalized error versus reference glucose for DC measurement data measured at 2500 ms and corrected according to the fifth study described herein.

39 is a graph of normalized error versus reference glucose for DC measurement data measured at 3000 ms and corrected according to the fifth study described herein.

40 is a table showing TSE for corrected responses at different DC test times according to the fifth study described herein.

41 is a Clark error grid showing predicted glucose versus reference glucose for uncorrected DC response data at 900 ms according to the fifth study described herein.

42 is a Clark error grid showing predicted glucose versus reference glucose for DC response data at 900 ms corrected by response data from one AC frequency according to the fifth study described herein.

43 is a Clark error grid showing predicted glucose versus reference glucose for DC response data at 900 ms corrected by response data from another AC frequency according to the fifth study described herein.

44 is a Clark error grid showing predicted glucose versus reference glucose for DC response data at 900 ms corrected by response data from two AC frequencies according to the fifth study described herein.

DETAILED DESCRIPTION

For the purpose of promoting understanding of the principles of the invention, reference will now be made to the embodiments illustrated in the drawings, and specific language will be used to describe those embodiments. However, it should be understood that it is not intended to limit the scope of the invention. As normally thought by those skilled in the art, it is contemplated that changes and modifications in the equipment shown and further applications of the principles of the invention as illustrated herein will be protected. In particular, although the present invention discussed in terms of blood glucose testing equipment and measuring methods, it is contemplated that the present invention can be used together with the equipment for measuring other analytes and other sample types. Such alternative embodiments require some adaptation to the embodiments discussed herein, which will be apparent to those skilled in the art.

The system and method according to the present invention allow for accurate measurement of analytes in fluids in ultrafast test times (i.e., no more than about 2 seconds). In particular, the measurement of the analyte is accurate despite the presence of interfering substances that would otherwise cause errors. For example, a blood glucose meter according to the present invention measures the concentration of blood glucose in a whole blood sample without the errors typically caused by variations in the hematocrit level and temperature of the sample. Accurate measurement of blood glucose is invaluable for predicting other complications of blindness, loss of circulation, and improper regulation of blood glucose in diabetic patients. An additional advantage of the system and method according to the present invention is that measurements can be performed more quickly and with much smaller sample volumes, making it more convenient for diabetics to measure their blood glucose. Similarly, accurate and rapid measurement of other analytes in blood, urine, or other biological fluids provides improved diagnosis and treatment of a wide range of medical conditions.

It should be appreciated that electrochemical glucose meters typically (but not always) measure the electrochemical response of a blood sample in the presence of a reagent. The reagent reacts with glucose to produce charge carriers not otherwise present in the blood. Therefore, the electrochemical response of blood in the presence of a given signal is intended to depend primarily on the concentration of glucose. However, the electrochemical response of blood to a given signal can also depend on other factors, including hematocrit and temperature. See, for example, U.S. Patents Nos. 5,243,516; 5,288,636; 5,352,351; 5,385,846; and 5,508,171, which discuss the confounding effects of hematocrit on glucose measurements and are incorporated herein by reference in their entireties. Furthermore, certain other chemicals, including, for example, uric acid, bilirubin, and oxygen, can affect the transfer of charge carriers through a blood sample, thereby causing errors in glucose measurements.

Various embodiments disclosed herein relate to systems and methods that allow for shorter test times while still delivering an analyte measurement (e.g., glucose or another fluid sample analyte) that is corrected for confounding interferents (e.g., hematocrit and temperature, or other interferents). The systems and methods disclosed herein enable test times of less than two seconds, including less than one second. As used herein, "total test time" is defined as the length of time from sample detection (or sample dose adequacy, if both are detected) at which the first electrical signal is applied to the sample to the acquisition of the last measurement used in the concentration determination calculation.

In addition to shorter total test times, embodiments disclosed herein result in analyte measurements with lower total systematic error, or "TSE." TSE generally comprises a combined measure of the accuracy and precision of a system and method. It is typically calculated as (absolute bias) + 2*(precision), where bias = the mean of the standardized errors; precision = StdDev (standardized error). Standardized error is typically calculated relative to a standard reference value. For example, in the context of blood glucose measurement, for a reference glucose sample less than or equal to 75 mg/dl, standardized error = (predicted glucose - reference glucose); but for a reference glucose sample greater than 75 mg/dl, standardized error = (predicted glucose - reference glucose) * 100 / (reference glucose).

As used herein, the phrase "signal having an AC component" refers to a signal that has some alternating potential (voltage) component. For example, the signal can be an "AC signal" with 100% AC potential (voltage) and no DC component; the signal can have AC and DC components separated in time; or the signal can be AC with a DC offset (AC and DC signals are superimposed). In the latter case, the signal can still be described as having an AC component, even if the polarity of the variable potential does not alternate.

Examples 1 and 2 describe details of experiments in which the total test time is reduced. In each example, AC blocks are used to generate correction data that will be algorithmically combined with the DC measurement results, similar to the known measurement sequences used in the ACCU-CHEK® Aviva instrument. That is, multiple AC potential frequencies are applied in a continuous manner, and the current response and other measurement data are determined for each frequency. However, in Examples 1 and 2, the total test time is reduced by reducing the time used for each continuous AC frequency block and the time used for the DC block. Example 1 details an experiment using these reduced time blocks in a covariation study using a biosensor with a general reagent layer thickness. Example 2 details an experiment with reduced time in a covariation study using a biosensor with a variable reagent layer thickness.

The measurement sequences used for Examples 1 and 2 were performed using an in-house data acquisition test station (DATS potentiostat), which included a group of blood glucose meters configured as a multi-meter test station, using modified code keys to program the desired measurement parameters. While the meters could be programmed and configured using a variety of methods and durations for the test sequence, several limitations existed, such as the selection of available frequencies pre-programmed in the meter hardware. This in-house test station will be referred to as "DATS" hereinafter.

Certain embodiments of the present invention disclosed herein generally utilize the collection of AC test data at multiple frequencies within a short period of time by using a multi-frequency excitation waveform technique. Examples 3 and 4 describe details of experiments in which multi-frequency excitation waveforms were used. These multi-frequency excitation waveforms are formed by adding together multiple individual waveforms of varying frequencies, such that the fluid sample is excited by multiple frequencies simultaneously. Multi-frequency excitation waveforms allow not only short measurement times but also adaptive measurement sequences because AC signal data collection does not permanently alter the sensed chemical effects in the way that DC measurements do due to the alternating polarity of the applied excitation. Furthermore, in accordance with the methods disclosed in co-pending published U.S. patent applications US-2004-0157339-A1, US-2004-0157337-A1, 2004/0157338-A1, US-2004-0260511-A1, US-2004-0256248-A1, and US-2004-0259180-A1, additional frequencies of the AC signal are applied at a low excitation AC potential to generate a non-Faradaic current response whereby the phase angle provides an indication of certain interfering factors, from which one or more interferent corrections can be determined and used to more accurately determine the analyte concentration in the fluid sample.

The resulting sample responses can then be measured and each excitation frequency component derived using Fourier transform techniques such as the Discrete Fourier Transform (DFT). While the various examples disclosed herein utilize multi-sinusoidal excitation waveforms, those skilled in the art will recognize that multi-frequency waveforms can be constructed using individual waveforms having any desired shape, such as triangular, square, sawtooth, delta, etc., to name a few non-limiting examples. The component AC waveforms used to generate the multi-frequency waveform can each have any desired frequency and any desired amplitude. The use of multi-frequency techniques not only shortens the time required to collect the desired data (because AC measurements are performed simultaneously rather than continuously), but also allows for better correlation for corrections, as the sample varies less during the data collection period corresponding to each temporal frequency. Furthermore, AC measurements can be performed closer in time to DC measurements. The improved correlation between the sample states during each AC and DC measurement allows for better interference compensation, even when the sample is not in a steady state.

The measurements for Examples 3 and 4 were performed using an electrochemical test bench based on a VXI component configuration from Agilent, and are programmable to apply AC and DC potentials to the sensor in a requested combination and sequence and to measure the current response of the resulting sensor. Data were transferred from the electrochemical analyzer to a desktop computer for analysis using Microsoft® Excel®. Measurements can be performed using any commercially available programmable potentiostat with a suitable frequency response analyzer and a digital signal acquisition system. For commercial use, this method can be performed in a dedicated low-cost handheld measurement device, such as the ACCU-CHEK® AVIVA ^™ blood glucose meter, in which firmware is configured to enable application of an AC signal with a multi-frequency waveform. In such cases, measurement parameters can be included in or provided to the firmware of the meter, and measurement sequences and data evaluation can be automatically performed without user interaction. For example, using a programmable potentiostat as described above, measurements and analysis results can be performed in a manner such that a total test time of less than 2 seconds after a sample containing an analyte is applied to the biosensor and detected with the device is possible. Similarly, the firmware of the ACCU-CHEK® AVIVA ^™ blood glucose meter can be provided with measurement parameters configured and arranged to cause the measurement sequence to occur within the same time period, i.e., a total test time of less than 2 seconds after the sample containing the analyte is applied to the biosensor and detected by the meter. When the evaluation of the measurement data is complete, typically 25 to 50 milliseconds after the last measurement result is obtained, the measurement result can be displayed on the meter's digital display.

Example 1—Continuous Multiple AC Frequency Tests with Fast Total Test Time

U.S. Patent No. 7,407,811 teaches the use of a multi-frequency AC block followed by a continuous application of a DC block. For example, Example 5 described in U.S. Patent No. 7,407,811 utilizes a continuous application of an AC excitation followed by a DC excitation. The excitation i signal includes a 10 kHz AC signal applied for approximately 1.8 seconds, a 20 kHz AC signal applied for approximately 0.2 seconds, a 2 kHz AC signal applied for approximately 0.2 seconds, a 1 kHz AC signal applied for approximately 0.2 seconds, and a DC signal applied for approximately 0.5 seconds. The total test time is 3.0 seconds.

In Example 6 of the patent, it is desired to achieve a total test time as low as 1.1 seconds using the same test strip design used in Example 5 of the patent. To achieve this, the inventors do not believe they can simply apply the continuous stimulation of Example 5 for a shorter period of time. As described in the patent:

"Using the same test strip 1700 and reagents as described above for Example 5, the excitation profile shown in FIG. 24 was utilized in order to reduce the overall test time. As described with respect to Example 5, it was determined that the phase angles at 20 kHz and 10 kHz correlated most closely with the hematocrit estimate. Therefore, in Example 6, it was decided to limit the AC portion of the excitation to two frequencies in order to reduce the overall test time. To achieve a further reduction in overall test time, a 10 kHz AC excitation (i.e., an AC signal with a DC offset) was applied simultaneously with the DC signal, the theory being that this combined mode allowed simultaneous results to be collected for DC current, AC phase, and AC admittance, providing the fastest possible results. Therefore, the 20 kHz signal was applied for 0.9 seconds. Thereafter, after a 0.1 second interval, the 10 kHz and DC signals were applied simultaneously for 1.0 second."

(U.S. Patent No. 7,407,811, column 23, lines 23-40.) The inventors of U.S. Patent No. 7,407,811 therefore believed that in order to reduce the total test time to below 3.0 seconds, it would be necessary to remove two of the AC excitation blocks (those at 2 kHz and 1 kHz) and apply one of the remaining two AC excitation blocks simultaneously with the DC excitation.

One reason for this belief in the prior art is illustrated in Figures 1 and 7, where the DC response of the applied sample to the measurement of the reagent chemistry is shown for various tests in which the application timing of the AC excitation signal after sample application is varied. It can be seen that when the DC excitation is applied very quickly after sample application, the response does not show the expected Cottrellian decay, making accurate determination of the sample glucose concentration impossible for rapid test times. This is because the enzyme and mediator availability, hydration, and diffusion within the reagent layer limit how long DC measurements can be made in a reproducible manner. Reagent hydration and coating uniformity between sensors are significant factors in how quickly the DC response can be measured.

We have found that shorter AC times are possible because even at these early times, information is present in the AC response data that can be correlated with interferences such as hematocrit. While there is some stabilization of the AC within the first 100 ms, even the short AC signal correlates well with the hematocrit interference. Using information collected from the expected glucose DC response at the same interval for all expected frequencies enables good correlation or correction of the DC glucose response with the AC-measured hematocrit interference.

This Example 1 was constructed to demonstrate the feasibility of running prior art continuous multi-AC frequency testing techniques at a faster rate using a sensor slot die coated with a uniform reagent formulation containing glucose oxidase. (Slot die coating of the uniform reagent is described in U.S. Patent Application Publication No. 2005/0008537, which is incorporated herein by reference in its entirety.) The sensor electrodes were fabricated by sputtering gold (~50 nm) onto Melinex 329, followed by laser ablation of a chromium mask on quartz to pattern the conductive layer to define the working and dose adequacy electrodes. The structure used was similar to that shown in FIG33 of U.S. Patent No. 7,407,811. The electrodes included a pair of dose adequacy electrodes independent of a pair of measurement electrodes. Measurements were performed using a DATS. The benefits of the DATS configuration include ease of setup and rapid, multi-channel data collection in an environmental chamber at varying temperatures. Using a DATS that includes existing meters configured for use with the AC excitation measurement method does not impose certain limitations on programmability, including specific transition times between blocks and available AC frequencies, and essentially all basic, non-programmable aspects of the meter for use with the DATS. However, using such existing meters is beneficial in that the continuous multiple AC frequency method examined in Example 1 (and in Example 2 below) can be backward compatible for use with existing meters.

Covariation studies were performed using whole blood samples with seven different target glucose concentrations (50, 100, 150, 250, 300, 450, and 550 mg/dL), three different target hematocrit concentrations (25%, 45%, and 70%), and five different temperatures (8, 14, 24, 35, and 42°C). The table in Figure 2 details the composition of the whole blood samples used in this Example 1.

AC data was collected using a continuously applied AC excitation signal at 10 kHz, 20 kHz, 2 kHz, and 1 kHz at 9 mV RMS. Then, after a 100 ms open circuit, a 450 mV DC potential was applied starting at 1300 ms. DC measurement data was collected every 100 ms starting at 1400 ms, and 1525 ms of DC data points were analyzed in this Example 1 (i.e., the test utilized a total test time of 1.525 seconds). (DC data points were taken at a time later than the total test time to confirm durability for shorter total test times. Since durability was confirmed for total test times of, for example, 1.525 seconds or less, DC data points at longer times were not used to calculate the final results.) The table of FIG3 lists the excitation signal composition and timing, while FIG4 presents this data graphically in a general manner.

FIG5 illustrates the normalized error for approximately 1600 data points for all 105 covariate samples ([G], % HCT, °C) versus reference glucose using only the uncorrected DC measurement acquired at 1525 ms. Determining predicted glucose from the DC measurement was performed using well-known prior art techniques. As will be readily apparent to one skilled in the art, the measurement system performs extremely poorly with such a short DC-only test time, with a total systematic error of 51%.

As shown in Figures 3 and 4, the AC excitation potentials (1 to 5) used in this Example 1 were applied continuously. The sequence began after the first application of a 10 kHz signal (not shown) to detect sample application (dose detection) and determine the fill of the capillary test chamber (sample adequacy). The use of AC measurements for droop detection and dose adequacy is described in U.S. Patent No. 7,597,793, which is incorporated herein by reference.

After the sample adequacy is determined, a 300 ms 10 kHz block is applied to obtain AC stability, followed by four additional AC data blocks at 10 kHz, 20 kHz, 2 kHz and 1 kHz signals, each with a 100 ms duration. All times herein are relative to the start of the detection of sample dose adequacy. In addition, block 1 is retained in the sequence of FIG3 primarily for backward compatibility with the ACCU-CHEK® AVIVA ^™ related meter failsafe, which is unrelated to the present invention. In addition, block 1 is used to stabilize AC before the next block at the same frequency. One of the goals of some of this work is to illustrate a short test time for backward compatibility of the product platform for the ACCU-CHEK® AVIVA ^™ meter. Additional experiments were performed using only two frequencies to verify the limits of the effectiveness of the correction in time versus continuous AC/DC. See, for example, FIG8.

After the AC measurement, the measurement electrodes were then held open circuit for 100 ms, followed by the application of a 450 mV DC signal. Between each stimulus block, there was a 75 ms period consisting of a 50 ms pre-stabilization period and a 25 ms tail data transfer period. The test time was evaluated at 1525 ms (test time = 1.525 seconds using a DC + 0.025 s transfer time at 1.5 seconds).

AC admittance values were captured for each of the four 100 ms AC excitation blocks in order to correct the DC glucose measurements for the interfering effects of hematocrit and temperature using the following equations:

Predicted glucose = INT + Yi2*Y2 + Pi2*P2 + Yi3*Y3 + Pi3*P3 + Yi4*Y4 + Pi4*P4 + Yi5*Y5 + Pi5*P5 + exp(SLOPE + Ys2*Y2 + Ps2*P2 + Ys3*Y3 + Ps3*P3 + Ys4*Y4+ Ps4*P4 + Ys5*Y5 + Ps5*P5)*DC**POWER (Equation 1)

Among them: Yi2, Yi3, Yi4, Yi5, Ys2, Ys3, Ys4 and Ys5 are constants

Pi2, Pi3, Pi4, Pi5, Ps2, Ps3, Ps4 and Ps5 are constants

Y2 is the admittance amplitude at 10 kHz (second block)

Y3 is the admittance amplitude at 20 kHz

Y4 is the admittance amplitude at 2 kHz

Y5 is the admittance amplitude at 1 kHz

P2 is the phase angle at 10 kHz (second block)

P3 is the phase angle at 20 kHz

P4 is at a phase angle of 2 kHz

P5 is the phase angle at 1 kHz

INT is the intercept

SLOPE is the slope

DC is the uncorrected glucose response predicted from DC measurements

POWER = Const + Yp2*Y2 + Pp2*P2 + Yp3*Y3 + Pp3*P3 + Yp4*Y4 + Pp4*P4 +Yp5*Y5 + Pp5*P5.

Equation 1 demonstrates that the system's dose-response can be approximated using a power model. The slope and power of this power model are affected by covariations such as temperature and hematocrit. Since the AC measurements (admittance and phase) are sensitive to these covariations, they are used to compensate for the covariation effects in terms of slope and power. Parameter estimates are established by performing parameter estimation using data collected while glucose, temperature, and hematocrit covaried. The DC value in this example was selected from a single measured DC point, and Equation 1 is specific to a single DC value. For more DC values, i.e., more than one current response measurement taken during the DC block, a more general expression is:

Predict glucose = ba0 + a1*Ieff + a2*Peff + a3*Yeff + ( b4 + exp( b0 + b2*Peff + b3*Yeff))*Ieff**( c0 + c2*Peff + c3*Yeff)

Where: Ieff= bV0 + bV1*DC1 + bV2*DC2 + bV3*DC3 + bV4*D4 + bV5*DC5 +bV6*DC6

Peff= bP0 + bP1*P1 + bP2*P2 + bP3*P3 + bP4*P4 + bP5*P5 + bP6*P6

Yeff= bY0 + bY1*Y1 + bY2*Y2 + bY3*Y3 + bY4*Y4 + bY5*Y5 + bY6*Y6.

The use of AC admittance magnitude and phase data to correct DC glucose response data for the effects of hematocrit and temperature is discussed in detail in US Patent No. 7,407,811.

Similar to FIG5 , FIG6 also illustrates the normalized error versus reference glucose for all 105 samples, except that the DC measurement taken at 1525 ms has been corrected using the AC measurement and method discussed above. Such correction allows the measurement system to compensate for the interfering effects of hematocrit and temperature. As can be seen, all measurements now fall within a normalized error of +/- 15%, with a total system error of 9.4%, all with a test time of only 1.525 seconds.

This Example 1 thus demonstrates that using multiple AC excitation frequencies in series, it is possible to achieve an extremely short test time of 1.525 seconds in order to probe a sample and measure interferents that prevent accurate assessment of glucose values, and to correct the measured glucose values to remove the effects of these interferents on the measurement. This surprising result contradicts the teachings of the prior art noted above.

Control of reagent thickness

Certain embodiments disclosed herein, including the embodiments illustrated by the description of Example 2 below, also utilize accurate control of biosensor reagent thickness by using a uniform method of applying the reagent to the biosensor surface, such as by slot die coating. The reagent coatings disclosed herein are generally about 1.6 to 5 μm in thickness. The uniformity of the reagent coating and the resulting uniform dissolution/hydration of the reagent film with the fluid sample enable reproducibility that correlates well with the AC measurement results to provide accurately compensated glucose. Uneven reagent thickness is detrimental to achieving faster methods and improved performance, especially over short time periods, due to greater variability in the measurement results. In order to obtain robust performance, we seek very uniform films. For descriptions and disclosures of methods related to uniformly coated films, see U.S. Patent Application Publication No. 2005/0008537 referenced above.

Figures 1 and 7 illustrate the measured DC response of samples applied to reagent chemistries for various tests, where the timing of application of the DC excitation signal after sample application varied from approximately 75 ms to 1400 ms. The hydration of Figure 1 is not as uniform across the bars as the thinner reagent in Figure 7. Applying the DC excitation too quickly after sample application can result in a non-Cottrellian response and, therefore, inaccurate glucose concentration measurements.

Using the coating method generally described in U.S. Patent Application Publication No. 2005/0008537, the reagent coating for Fig. 1 (50 g/m ² coating weight) is about 3.6 μm and for Fig. 7 (20 g/m ² coating weight) is about 1.6 μm. It can be seen that when applying DC excitation very quickly after applying the sample to the thinner reagent layer, the response begins to show the expected Cottrellian-shaped decay characteristic more quickly, thereby making the accurate determination of the sample glucose concentration in the fast test time possible. This is because the enzyme availability, hydration and diffusion in the reagent layer are achieved with thinner and or more uniform reagent layer thickness, which limits how long it can be measured by DC. Figure 30 illustrates the coating weight setting and the actual measurement table of the dry coating film thickness using the wetting reagent coating method described in the US 2005/0008537 disclosure. From this disclosure and about those skilled in the art will recognize the required equipment operating parameters for each coating weight.

Techniques and methods that can be used to form thin reagent strips on biosensors are disclosed in US Patent Application Publication Nos. 2005/0016844 and 2005/0008537, the disclosures of which are hereby incorporated by reference herein in their entireties.

Fig. 8 summarizes the test carried out on the biosensor with the reagent coating formed thereon with the thickness of 4 μ m, 2.5 μ m and 2.0 μ m.Table 1 illustrates the general composition of the wetting reagent that is coated on the biosensor used in Fig. 8.This reagent is similar to the reagent of ACCU-CHEK® AVIVA ^™ biosensor, but is prepared with milled silica.Because unmilled silica may have the problem of particle size that will be unfavorable for thinner coating, milled silica is used to reduce the average particle size of silica.They are coated with different coating weights, cause different measured thicknesses.The target is to start from having some aforementioned optimized reagent quality that is used for glucose biosensor at least for upper thickness level.Then, by using identical reagent quality to adjust coating weight, use slot die coating method to prepare the reagent thickness from about 4 μ m to 2 μ m.By making reagent in this way, also reduced the concentration of active ingredient that was optimized at first for thicker coating weight.

Table 1

Blood sample is applied to each biosensor and along with reagent and sample are hydrated, the AC excitation frequency of 2 kHz or 20 kHz is applied to the biosensor. Admittance data is measured every 100ms in one second and is plotted in Figure 8. As can be seen, AC admittance is less than stabilizing in 400ms after sample is applied, and the AC data showing 100ms are suitable for revising the DC glucose test obtained using the program disclosed below. From the data represented in Figures 1 and 7, it is apparent that reagent stabilizes quite quickly after sample is applied. About the film being tested, thinner reagent provides AC response that stabilizes faster and in a more reproducible manner.

The ability of realizing fast test time as disclosed herein is greatly affected by the hydration rate of the enzyme and mediator in the reagent film and the diffusion rate of the reaction product to the electrode surface below the reagent film. The use of the slot die coating method disclosed in U.S. Patent Application Publication No. 2005/0016844 and 2005/0008537 for depositing the reagent layer allows for faster and more reproducible dissolution, filling time and hydration distribution deposition of uniform thin film reagents. Figure 9 illustrates the surface profilometry measurement of thin film reagents deposited to a target thickness of 2.5 μm (nominal coating weight = 30 g/m ² ) on a biosensor using these methods. As can be seen, the average thickness in the central B zone of the reagent strip is 2.46 μm. With a thin reagent film (in one embodiment, about 1.6 to 10 μm in thickness, and in other embodiments, about 1.6 to 5 μm in thickness), enzyme availability, hydration and diffusion are more quickly and more uniformly demonstrated.

Thin films benefit measurements in terms of faster hydration, allowing for faster measurements after sample application. AC stability appears to be less affected by film thickness than DC response. When the film is thinner, more Cottrellian-like behavior is observed at earlier times in response to DC excitation. This can be seen by comparing Figures 1 and 7. Figure 1 shows the current response of a thicker film, i.e., 50 g/ ^m2 , which provides a variable early I vs. T trace when DC excitation begins approximately 100 to 700 ms after sample adequacy is detected. In contrast, as shown in Figure 7, for the 20 g/ ^m2 film over the same time range, the current response follows a good trend. In addition, approximately 300 ms after the DC potential is applied, the I vs. T response becomes more Cottrellian-like. However, there are certain limitations regarding thin films that need to be considered. A minimum amount of enzyme is required on both sensors to obtain a linear response and maintain the required long-term stability of the sensor. The film in this example has proportionally less enzyme because it is made from the same reagent mass, as it is made thinner. The lower limit of film thickness generally depends on the concentration of the enzyme in order to provide the reagent quality of appropriate response and stability. It is also understood that there will be a certain lower limit of thickness, wherein the variability of the coating method and substrate thickness will not provide a uniform coating thickness. For other problems and disclosures about uniform and homogeneous film thickness control, see, for example, U.S. Patent Application Publication No. 2005/0008537 referenced above.

Example 2—Continuous Multiple AC Frequency Testing with Fast Total Test Time and Varying Reagent Thickness

Covariation studies were performed testing multiple whole blood samples for glucose concentration using electrodes and test structures similar to those in the ACCU-CHEK® AVIVA ^™ , available from Roche Diagnostics, Inc., Indianapolis, Indiana, USA. A pyrroloquinoline quinone-dependent glucose dehydrogenase (PQQ-GDH)-based reagent with a composition identical or substantially similar to that from Table 1 (above) was applied to the biosensor at one of three thicknesses: 2 μm, 2.5 μm, and 4 μm. As detailed in FIG10 , covariation studies were performed using whole blood samples similar to those in Example 1, but with six glucose concentrations, five hematocrit levels, and five concentrations.

As detailed in FIG11 , the AC excitation potential for this Example 2 was applied continuously. A 10 kHz dose detection and sample adequacy method (not shown) was followed by a 300 ms 20 kHz signal, followed by 100 ms application of 20 kHz, 10 kHz, 2 kHz, and 1 kHz signals. The measuring electrodes were then held open circuit for 100 ms, followed by application of a 550 mV DC signal. Due to preset timing parameters in the existing DATS instrument, there was a 50 ms settling delay and a 25 ms tail data transmission period between each excitation block. Measurements of the response to the DC signal were taken over the total test time, starting at approximately 1500 ms, and measured at 100 ms intervals. AC admittance values were captured for each AC excitation block in order to correct the DC glucose measurements for the interfering effects of hematocrit and temperature using the following equation:

Predicted glucose = INT + Yi2*Y2 + Pi2*P2 + Yi3*Y3 + Pi3*P3 + Yi4*Y4 + Pi4*P4 + Yi5*Y5 + Pi5*P5 + exp(SLOPE + Ys2*Y2 + Ps2*P2 + Ys3*Y3 + Ps3*P3 + Ys4*Y4+ Ps4*P4 + Ys5*Y5 + Ps5*P5)*DC**POWER (Equation 2)

Where: Yi2, Yi3, Yi4, Yi5, Ys2, Ys3, Ys4 and Ys5 are constants

Pi2, Pi3, Pi4, Pi5, Ps2, Ps3, Ps4 and Ps5 are constants

Y2 is the admittance amplitude at 20 kHz (second block)

Y3 is the admittance amplitude at 10 kHz

Y4 is the admittance amplitude at 2 kHz

Y5 is the admittance amplitude at 1 kHz

P2 is the phase angle at 20 kHz (second block)

P3 is the phase angle at 10 kHz

P4 is at a phase angle of 2 kHz

P5 is the phase angle at 1 kHz

INT is the intercept

SLOPE is the slope

DC is the uncorrected glucose response predicted from DC measurements

POWER = Const + Yp2*Y2 + Pp2*P2 + Yp3*Y3 + Pp3*P3 + Yp4*Y4 + Pp4*P4 +Yp5*Y5 + Pp5*P5.

It will be appreciated that Equation 2 is substantially identical to Equation 1 from Example 1. The primary difference is the order in which the different frequencies are applied, with Example 1 applying a 10-20-2-1 kHz sequence and Example 2 applying a 20-10-2-1 kHz sequence.

Well-known prior art techniques were used to determine the uncorrected glucose response from the DC measurement (i.e., not corrected for the interfering effects of hematocrit and temperature). This DC glucose response was then corrected for the interfering effects of hematocrit and temperature using the AC admittance magnitude and phase measurement data detailed above in Equation 2. The total systematic error, bias, accuracy, and NVar were calculated for each and tabulated in FIG12 . As can be seen, the total systematic error for all three reagent thicknesses was very good at a total test time as low as 1.525 seconds.

As mentioned above, the total systematic error or TSE is a combined measure of the accuracy and precision of a system. It is usually defined as: (absolute bias) + 2*(precision). The details are as follows:

Bias = mean of the standardized errors;

Accuracy = StdDev (standardized error);

in

When baseline glucose <= 75 mg/dl, standardized error = (predicted glucose - baseline glucose); and

When baseline glucose is > 75 mg/dL, the standardized error = (predicted glucose - baseline glucose) * 100 / (baseline glucose).

FIG13 illustrates the normalized error for DC measurement data corrected using AC measurements, as detailed above, versus baseline glucose values. Only DC measurements taken at 1500 ms (+25 ms for transmission) were used; therefore, this data represents an actual total test time of 1.525 seconds. The interfering effects of hematocrit and temperature were significantly reduced, resulting in a total systematic error of 10.0% for the entire covariation study.

FIG14 is a Clark error grid showing predicted glucose values versus baseline glucose values for all uncorrected DC glucose measurements taken at 1525 ms. Clarke Error Grid Analysis (EGA) was developed in 1987 to quantify the clinical accuracy of a patient's estimate of their current blood glucose compared to the blood glucose value obtained in their meter. See Clarke WL, Cox D, Gonder-Frederick LA, Carter W, Pohl SL: Evaluating clinical accuracy of systems for self-monitoring of blood glucose . Diabetes Care 10:622–628, 1987. The Clark error grid has since been used to quantify the clinical accuracy of blood glucose estimates generated by test meters compared to baseline values. EGA is generally accepted as the standard method for determining the accuracy of blood glucose meters.

The Clark error grid breaks down the scatter plot of test results from the benchmark and evaluation glucose meters into five regions. Region A is those values within the 20% of the benchmark sensor, Region B contains points outside the 20% that will not result in inappropriate treatment, Region C is those points that will result in unnecessary treatment, Region D is those points that indicate a potentially dangerous failure to detect hypoglycemia, and Region E is those points that will confuse treatment for hyperglycemia with high blood sugar, or vice versa. In Figure 14, the dashed line also indicates values within the 15% of the benchmark sensor.

As can be easily seen in Figure 14, the uncorrected glucose value falls well outside the +/- 15% error window, which is the expected error window set forth in the Clark error grid. In a glucose test meter according to general industry practice for blood glucose monitoring systems and according to FDA guidelines, this level of accuracy would be considered unacceptable.

FIG15 is a Clark error grid showing the same DC test data shown in FIG14 , except that the data has been corrected for the interfering effects of hematocrit and temperature using the methods described above. As can be easily seen in FIG15 , the performance of the measurement system corrected for hematocrit and temperature using AC measurement data is much superior to using only DC measurements to predict glucose values at extremely fast total test times.

As can be seen from Example 2 above, the use of a thin reagent film (e.g., approximately 1.6-5 μm thick) enables the ability to perform accurate glucose determinations corrected for the interfering effects of hematocrit and temperature with a total test time of less than 2 seconds. The uniformity of the reagent coating and the resulting uniform dissolution/hydration of the reagent film with the fluid sample are believed to enable reproducibility that correlates well with AC measurement results to provide accurately compensated glucose test results.

Based on Examples 1 and 2, it has become apparent that despite prior understanding in the art, shorter test times can be achieved with shortened continuous AC blocks and/or by using fewer continuous AC frequencies. However, using more frequencies can provide benefits in measurement correction, particularly when correcting for multiple variables or when it is desirable to provide an indication of the level or general range of one or more such variables in addition to the analyte measurement result. To achieve this, and still achieve the shortest possible test, the use of multi-frequency excitation waveforms, such as those described in Examples 3 and 4, was explored.

Multi-frequency excitation

As described herein, certain embodiments disclosed herein utilize the collection of AC test data at multiple frequencies within a shorter period of time by using multi-frequency excitation waveform techniques. These multi-frequency excitation waveforms are formed by summing together multiple individual waveforms of varying frequencies so that a fluid sample is excited by multiple frequencies substantially simultaneously rather than sequentially.

The resulting sample response can then be measured, and this measurement will contain the sample response for all excitation frequencies. The specific contribution from each excitation frequency component can then be derived using Fourier transform techniques, such as the Discrete Fourier Transform (DFT). While the various examples disclosed herein utilize multi-sinusoidal excitation waveforms, those skilled in the art will recognize that multi-frequency waveforms can be constructed using individual waveforms having any desired shape, such as triangular, square, sawtooth, delta, etc., to name a few non-limiting examples. The component AC waveforms used to generate the multi-frequency waveform can each have any desired frequency and amplitude. The use of multi-frequency techniques not only shortens the time required to collect the desired data (because AC measurements are performed simultaneously rather than continuously), but also allows for better correlation for corrections, as the sample varies less during the data collection period corresponding to each temporal frequency. This is particularly true for tests utilizing very fast total test times, where measurements are taken very shortly after sample application, while the sample is still undergoing diffusion and reaction with reagent chemistry. Furthermore, AC measurements can be performed closer in time to DC measurements. Better correlation between AC and DC allows for better compensation for interferents, even when the sample is not in steady state.

An exemplary prior art measurement sequence for a blood glucose testing system corrected for the interfering effects of hematocrit and temperature, such as those disclosed in U.S. Patent No. 7,407,811, is as follows:

Step 1: Apply blood to the biosensor in the meter.

Step 2: Acquire AC measurements of the sample for dip detection and/or dose adequacy.

Step 3: AC measurements are acquired over a period of time to allow calculation of correction factors for hematocrit and temperature. In many cases, multiple AC excitation frequencies are applied to the sample in succession.

Step 4: Obtain DC measurements to measure the raw (uncorrected) glucose response.

Figure 5: Correction factors derived from AC measurements are used to compensate the raw DC response for hematocrit and temperature effects.

Step 6: Display the measurement results to the user.

This procedure has certain disadvantages relative to obtaining measurements in less than 2 seconds. While accurate measurements can be obtained by correcting the raw DC glucose measurement with AC-derived data regarding hematocrit and temperature, the additional time required to collect the AC data extends the overall test time and also temporally separates the various AC and DC measurements used to arrive at the final measurement. This temporal separation of the various AC and DC measurements can be beneficial in some cases because the sample being tested continues to undergo chemical reactions with the reagent, and the reagent is hydrated during this time. That is, in a measurement sequence in which the AC signal is applied continuously with different waveform frequencies, the admittance and phase data for each frequency, while still useful for correcting subsequent raw DC response measurements, is not ideal because each data point is acquired at a different time during the progression of the sample-reagent hydration-reaction dynamics. By applying all frequencies simultaneously within the AC excitation waveform, the admittance and phase data for each frequency remain individually distinguishable and advantageously relate to the same state of the sample-reagent dynamics.

The current response measured starting with the application of an exemplary multi-frequency AC excitation waveform, followed by the application of a DC signal, is illustrated in Figure 29. For Examples 3 and 4, data acquisition was performed using an electrochemical test bench constructed based on VXI components from Agilent and which is programmable to apply AC and DC potentials to the sensor in requested combinations and sequences and measure the resulting current response of the sensor.

Example 3—Multi-Frequency AC Testing with Fast Total Test Time

The measurements performed for Example 3 were made with similar electrode structures as the ACCU-CHEK® AVIVA ^™ biosensors and reagents of the same or similar composition as set forth in Table 1 (above). These sensors were fabricated using a combination of processes including sputtering, laser ablation, reagent slot die coating, and lamination using generally the same techniques as the ACCU-CHEK® AVIVA ^™ biosensors.

The measurement sequence consists of three basic blocks. The first measurement block (not shown) utilizes a 10240 Hz sine wave excitation applied to the test strip to detect sample dose adequacy (sufficient filling of the capillary test chamber to perform the measurement). The use of AC measurements for dip detection and dose adequacy is described in the above-referenced U.S. Patent No. 7,597,793.

After sufficient samples have been detected, a second measurement block is initiated within a short time interval (as described in detail below) using a multi-sine (also called multi-tone) waveform to simultaneously collect AC admittance magnitude and phase data for each frequency of interest. The multi-sine waveform used in this Example 3 is constructed by summing sine waves at four frequencies (1024, 2048, 10240, and 20480 Hz). These frequencies were chosen because they are known to be useful for interfering agent correction, as described in the applicant's prior disclosures regarding the use of AC excitation referenced above. The higher frequency ranges of approximately 20 and 10 kHz are known to provide useful correction for hematocrit. Due to their known potential for useful discrete measurements, the lower frequency ranges of approximately 1 and 2 kHz are included. Generally, this frequency combination allows correction for multiple parameters, such as hematocrit and temperature. It should be understood that these values need not be specifically 20 kHz, for example, but rather within a range that allows for reasonable measurement of interfering agents independently of the glucose response being corrected. A higher frequency may be more correlated with one interferent, such as hematocrit, while another frequency may be more correlated with another interferent. Optimizing the frequency or frequency combination that provides the best overall corrected response would be useful and, in light of this disclosure, is well within the skill of those skilled in the art. However, when working with a multi-frequency AC waveform to reduce the time required to collect response data from multiple frequencies while still providing a short overall test time that provides good correction, it may be determined that using frequencies within these known ranges would be useful, so as to rely on past experience. Additionally, prior experience has shown that data from more than one frequency can better correct for multiple interferents than measurements at only one frequency. Four frequencies may be selected here so that previously programmed data analysis routines can be used. However, two, three, or even five or more frequencies, for example, may also provide adequate correction. Some discrete AC methods have been implemented using only two AC frequencies.

For Example 3, the multi-sine waveform consists of one cycle of a 1024 Hz signal, two cycles of a 2048 Hz signal, 10 cycles of a 10240 Hz signal, and 20 cycles of a 20480 Hz signal. The peak amplitude is set at 12.7 mV, but due to the multi-sine nature of the signal, the actual RMS value will be significantly lower. (RMS is the root mean square value of the waveform, SQRT[(1/N)*SUM(x ² )].) This waveform includes 16,000 data points that are input to the digital-to-analog converter and is shown in Figure 16.

One benefit of using a multi-sine excitation waveform is that the AC measurement time required to collect data for all four frequencies is reduced because the measurements are taken simultaneously. Another benefit of a multi-sine excitation waveform is that the AC measurement data for all frequencies is collected simultaneously and is therefore less affected by the fact that the sample changes as it reacts with the reagent.

The multi-sine waveform was applied to the specimen for 300 ms after an indication of dose adequacy and analyzed at 100 ms intervals. While this Example 3 utilized a 300 ms measurement period, longer, shorter, and even variable time periods can be employed with similar results. Generally, to achieve a total test time of two seconds or less, the multi-sine measurement period in one embodiment ranges from 100 ms to 1900 ms. With the ACCU-CHEK® AVIVA ^™ test configuration used, 200 to 500 ms is sufficient to provide a reproducible AC response from the specimen.

Although various excitation frequencies are applied simultaneously to the sample using a multi-sinusoidal signal, the response attributable to each frequency component can be extracted from the AC measurement data using an appropriate mathematical function, such as a fast Fourier transform (FFT) or discrete Fourier transform (DFT), or other mathematical techniques, as will be appreciated by those skilled in the art. In this Example 3, a DFT is used to extract admittance magnitude and phase data for each frequency. The extracted admittance data for each frequency is shown in FIG17A for a time point associated with 200 ms after dose adequacy for all nine samples tested. A plot of phase versus hematocrit at each frequency for Example 3 is illustrated in FIG17B , and a plot of admittance magnitude versus hematocrit at each frequency for Example 3 is illustrated in FIG17C .

The second measurement block consisted of a 550 mV DC signal applied to the sample to obtain a raw (uncorrected) predicted glucose reading, as known in the art. Four DC time points were extracted from the measurement data as 100 ms average data points with data points ending at 500, 600, 1000, and 1500 ms (i.e., for a total test time of 0.5, 0.6, 1.0, and 1.5 seconds).

Nine whole blood samples were prepared for the covariation study using target glucose concentrations of 90, 250, and 600 mg/dL and target hematocrit values of 20, 45, and 70%. For each sample tested, a nonlinear fit was used to analyze each DC time point and 300 ms AC admittance magnitude and phase data were used to calculate the predicted glucose response that was compensated for the effects of hematocrit and temperature using the following equation:

Predicted glucose = INT + Yi1*Y1 + Pi1*P1 + Yi2*Y2 + Pi2*P2 + Yi3*Y3 + Pi3*P3 + Yi4*Y4 + Pi4*P4 + exp(SLOPE + Ys1*Y1 + Ps1*P1 + Ys2*Y2 + Ps2*P2 + Ys3*Y3+ Ps3*P3 + Ys4*Y4 + Ps4*P4)*DC**POWER (Equation 3)

Where: Yi1, Yi2, Yi3, Yi4, Ys1, Ys2, Ys3 and Ys4 are constants

Pi1, Pi2, Pi3, Pi4, Ps1, Ps2, Ps3 and Ps4 are constants

Y1 is the admittance amplitude at 1024 Hz

Y2 is the admittance amplitude at 2048 Hz

Y3 is the admittance amplitude at 10240 Hz

Y4 is the admittance amplitude at 20480 Hz

P1 is the phase angle at 1024 Hz

P2 is the phase angle at 2048 Hz

P3 is the phase angle at 10240 Hz

P4 is at a phase angle of 20480 Hz

INT is the intercept

SLOPE is the slope

DC is the uncorrected glucose response predicted from DC measurements

POWER is = Const + Yp1*Y1 + Pp1*P1 + Yp2*Y2 + Pp2*P2 + Yp3*Y3 + Pp3*P3+ Yp4*Y4 + Pp4*P4.

Again, this equation is of the same form as Equations 1 and 2, but as can be seen, the variables range from Y1 to Y4 and P1 to P4 for simultaneous AC at 200 ms, rather than Y2-Y5 and P2-P5 used in Equations 1 and 2.

As discussed above, the use of AC admittance magnitude and phase data to correct DC glucose response data for the effects of hematocrit and temperature is discussed in detail in US Patent No. 7,407,811.

Well-known prior art techniques were used to determine the uncorrected glucose response from the DC measurement results (i.e., not corrected for the interfering effects of hematocrit and temperature). This DC glucose response was then corrected for the interfering effects of hematocrit and temperature using the AC admittance magnitude and phase measurement data, as detailed above in Equation 3. Total systematic error (TSE), bias, accuracy, and NVar were calculated for each total test time (for both corrected and uncorrected results) and tabulated in FIG18 . As can be readily seen, the performance of the measurement system corrected for hematocrit and temperature using AC measurement data is significantly superior to predicting glucose values using only DC measurements. Furthermore, simultaneously acquiring AC measurement data for multiple excitation frequencies allows for extremely fast total test times, with measurements showing very good TSE values for total test times of 1.5 seconds, 1.0 seconds, 0.6 seconds, and 0.5 seconds.

FIG19 illustrates the normalized error versus the baseline glucose value for the uncorrected glucose measurement, and a significant dependence on the hematocrit value can be seen in this data. The total systematic error is 53.8%. FIG20 illustrates the normalized error versus the baseline glucose value for the same measurement, only this time, as detailed above, the DC measurement data has been corrected using the AC measurement data. It is clear that the interfering effect of the hematocrit has been substantially reduced, with a total systematic error of 14.2%. This reduction was achieved using only the AC measurement data taken 500 ms after the dose adequacy indication.

FIG21 is a Clark error grid showing predicted glucose values versus baseline glucose values for all of the above 500 ms data points, both corrected and uncorrected. As can be readily seen, the uncorrected glucose values fall well outside the +/- 15% error window, while the corrected data are all within this limit. Thus, the use of multi-frequency excitation to achieve a half-second glucose total test time is demonstrated.

The above data for this Example 3 clearly shows that the use of the multi-frequency excitation technique disclosed herein allows for extremely short test times by allowing a sample to be excited by multiple frequencies simultaneously and measuring the sample's response to those frequencies simultaneously. Even at a half-second total test time, the data provides a significant reduction in DC measurement errors caused by interferers and allows essentially instantaneous measurement results to be reported to the user with measurement accuracy that is well within accepted industry standards.

Example 4—Multi-Frequency AC Testing with Fast Total Test Time

The measurements performed for Example 4 were performed using the same electrode configuration and reagents and the same measurement sequence as for Example 3. However, whereas the Example 3 measurements were performed on samples having three different target analyte concentrations and three different hematocrit levels for each concentration, the Example 4 measurements were performed on samples having seven different target analyte concentrations and three hematocrit levels for each concentration.

As discovered in Example 3, the benefit of using a multi-sine excitation waveform is that the AC measurement time required to collect data for all four frequencies is reduced from the measurements of Example 4 because the measurements are performed simultaneously. Another benefit of a multi-sine excitation waveform is that the AC measurement data for all frequencies is collected simultaneously and is therefore less affected by the fact that the sample changes as it reacts with the reagent.

Twenty-one whole blood samples were prepared for the covariation study using target glucose concentrations of 50, 100, 140, 250, 300, 450, and 550 mg/dL and target hematocrit values of 20, 45, and 70%. For each sample tested, each DC time point was analyzed using a nonlinear fit and the 300 ms AC admittance magnitude and phase data were used to calculate the predicted glucose response compensated for the effects of hematocrit and temperature using the following equation:

Predicted glucose = INT + Yi1*Y1 + Pi1*P1 + Yi2*Y2 + Pi2*P2 + Yi3*Y3 + Pi3*P3 + Yi4*Y4 + Pi4*P4 + exp(SLOPE + Ys1*Y1 + Ps1*P1 + Ys2*Y2 + Ps2*P2 + Ys3*Y3+ Ps3*P3 + Ys4*Y4 + Ps4*P4)*DC**POWER (Equation 4)

Where: Yi1, Yi2, Yi3, Yi4, Ys1, Ys2, Ys3 and Ys4 are constants

Pi1, Pi2, Pi3, Pi4, Ps1, Ps2, Ps3 and Ps4 are constants

Y1 is the admittance amplitude at 1024 Hz

Y2 is the admittance amplitude at 2048 Hz

Y3 is the admittance amplitude at 10240 Hz

Y4 is the admittance amplitude at 20480 Hz

P1 is the phase angle at 1024 Hz

P2 is the phase angle at 2048 Hz

P3 is the phase angle at 10240 Hz

P4 is at a phase angle of 20480 Hz

INT is the intercept

SLOPE is the slope

DC is the uncorrected glucose response predicted from DC measurements

POWER is = Const + Yp1*Y1 + Pp1*P1 + Yp2*Y2 + Pp2*P2 + Yp3*Y3 + Pp3*P3+ Yp4*Y4 + Pp4*P4.

Again, this equation is of the same form as Equations 1 and 2, but as with Equation 3 from Example 3, one can see that for simultaneous AC at 200 ms the variables range from Y1 to Y4 and P1 to P4, rather than Y2 to Y5 and P2 to P5 used in Equations 1 and 2.

As discussed in Example 3, the use of AC admittance magnitude and phase data to correct DC glucose response data for the effects of hematocrit and temperature is discussed in detail in U.S. Patent No. 7,407,811. A plot of admittance magnitude versus hematocrit at each frequency for Example 4 is illustrated in FIG27, and a plot of phase versus hematocrit at each frequency for Example 4 is illustrated in FIG28.

Well-known prior art techniques were used to determine the uncorrected glucose response from the DC measurement results (i.e., not corrected for the interfering effects of hematocrit and temperature). This DC glucose response was then corrected for the interfering effects of hematocrit and temperature using the AC admittance magnitude and phase measurement data, as detailed above in Equation 4. Total systematic error (TSE), bias, accuracy, and NVar were calculated for each total test time (for both corrected and uncorrected results) and tabulated in FIG22 . As can be readily seen, the performance of the measurement system corrected for hematocrit and temperature using AC measurement data is significantly superior to predicting glucose values using only DC measurements. Furthermore, simultaneously acquiring AC measurement data for multiple excitation frequencies allows for extremely fast total test times. As shown in FIG22 , the measurement results exhibit very good TSE values at total test times of 1.525 seconds, 1.025 seconds, 0.725 seconds, and 0.625 seconds.

FIG23 illustrates the normalized error for the uncorrected glucose measurement versus the baseline glucose value, and a significant dependence on the hematocrit value can be seen in this data. The total systematic error is 47.5%. FIG24 illustrates the normalized error for the same measurement versus the baseline glucose value, only this time, as detailed above, the DC measurement data has been corrected using the AC measurement results. It is clear that the interfering effect of the hematocrit has been substantially reduced, with a total systematic error of 10.2%. The measurement data for each of the 21 such measurement runs can be seen in FIG26.

FIG25 is a Clark error grid showing predicted glucose values versus baseline glucose values for all 725 ms data points, both corrected and uncorrected. As can be readily seen, the majority of the uncorrected glucose values fall well outside the +/- 15% error window, while the corrected data are all within this limit. Thus, the use of multi-frequency excitation to achieve a total glucose test time of less than three-quarters of a second is demonstrated.

The above data for this Example 4 clearly shows that the use of the multi-frequency excitation technique disclosed herein allows for extremely short test times by allowing a sample to be excited simultaneously with multiple frequencies and simultaneously measuring the sample's response to those frequencies. Even with a total test time of less than three-quarters of a second, the data provides a significant reduction in DC measurement errors caused by interferers and allows essentially instantaneous measurement results to be reported to the user with measurement accuracy that is well within accepted industry standards.

Example 5—Continuous Multiple AC Frequency Testing with Fast Total Test Time at Various DC Time Points

This Example 5 was performed similarly to Example 2 (above), using a test sensor with a reagent film thickness applied based on a nominal 30 g/ ^m² coating weight, as shown in Figure 30 , which corresponds approximately to a thickness of 2.45 μm. However, unlike Example 2, data acquisition was performed using an electrochemical test bench constructed based on VXI components from Agilent, which is programmable to apply AC and DC potentials to the sensor in requested combinations and sequences and measure the resulting sensor current response. This was done because, as discussed with respect to Examples 1 and 2, the existing instrumentation used with DATS for those measurements includes preset parameters, including mandatory time blocks required for waveform stabilization, tail transmission after each frequency block, and a preset "skip" period in the initial 100 ms of DC signal application, during which the current response cannot be measured. However, for this Example 5, it was desirable to apply multiple AC frequencies in succession without the constraints of the preset timing conditions imposed by the existing DATS instrumentation. The Agilent-based test bench used for Examples 3 and 4 provides the flexibility to program the desired measurement sequence in this manner.

The purpose of Example 5 was to explore different ways in which a set of four short 200 ms AC excitation blocks, applied sequentially and having frequencies of 20, 2, 10, and 1 kHz, could be used to modify the DC glucose response co-varying with hematocrit at a single, generally uniform reagent film thickness. The AC excitation block was applied starting at time zero, which is the time when a sufficient sample dose was detected. Thus, the AC excitation block began at time zero, had no open periods between them, and ended at approximately 800 ms, at which time the DC excitation was applied.

DC response data was collected starting at 800 ms and continuing until 3700 ms. This data set was used to analyze data with varying AC and DC parameters. The goal was to determine whether good performance could be achieved with this uniform film within a short test time, and to determine the effect of applying one or more AC responses for correction. FIG31 shows the AC response versus hematocrit for the four AC frequencies measured. All of this data demonstrates an inverse relationship between the admittance response and increasing hematocrit. As can be seen, the measurements at the higher frequencies of 20 kHz and 10 kHz show generally similar responses, and the measurements at the lower frequencies of 2 kHz and 1 kHz show generally similar responses. However, the higher frequencies have a greater hematocrit versus admittance relationship.

Figure 32 shows the uncorrected DC response data collected for this covariation study, and it is clear that at each DC test time, there is a variable current response associated with the changing hematocrit level. Figure 33 shows a typical response to baseline glucose values for the uncorrected normalized error for the covariation study using the DC response measured at 900 ms. This uncorrected response has a TSE of 41%. However, as shown in Figure 34, when the DC response at 900 ms is corrected using AC response data from only two frequencies, 20 kHz and 2 kHz, there is a significant improvement in TSE, which has been reduced to 7.1%. Figures 35-39 show normalized error plots for the corrected DC response measured at 1100 ms, 1500 ms, 2000 ms, 2500 ms, and 3000 ms, respectively, again corrected using only AC response data from the continuously applied frequencies of 20 kHz and 2 kHz of the AC signal.

Figure 40 shows a table of TSE as a function of variable DC test time for DC response data corrected in two ways, the first with 20 kHz and 2 kHz AC response data and the second with 10 kHz and 1 kHz AC response data. The AC response data at 20 kHz and 2 kHz provides better correction (in terms of TSE) for this test configuration than the 10 kHz and 1 kHz response data.

Figure 40 also indicates that at the shortest test time of 900 ms, the TSE is actually better than at longer times; that is, there is an increase in TSE as the DC measurement time increases, but this is then followed by a decrease in TSE at times much longer than 3000 ms. It is believed that the TSE is lower at shorter DC measurement times because, at these times, the time between the measurement of the AC response (used to obtain the correction factor) and the measurement of the DC response (used to obtain analyte-related data) is only approximately 100 to 900 ms. That is, AC response data is obtained at approximately 200 ms, 400 ms, 600 ms, and 800 ms, while the shorter DC response data is obtained at 900 and 1100 ms. Therefore, the correction factor response and analyte response correlate best when membrane hydration and reaction properties are nearly identical. At shorter DC response measurement times, the measurements are closer to the electrode surface, which has a short diffusion distance, where effects due to membrane hydration and swelling are less present.

As measurements are taken at moderately longer DC response times, TSE increases because the AC correction factor and DC response become further apart (less correlated) because the membrane is rapidly hydrating and swelling, and the DC response is being measured in this rapidly changing region. However, at even longer DC measurement times, such as 3000 ms, TSE drops back down as the actual hydration and swelling begin to stabilize, resulting in DC values with less variability and requiring less correction with the AC correction factor. Therefore, at these longer measurement times, TSE appears to improve to values close to those at earlier DC response measurement times. Typically, AC/DC response methods taught in prior art publications measure DC response data when the DC response is most stable, which is typically later, and therefore loses some correlation between the correction factor and analyte response. Here, we demonstrate that it is possible to measure DC response at earlier measurement times and still obtain an acceptable analyte response with the added benefit of reduced test time. In the case of this Example 5, the total test time was less than 1 second (i.e., 900 ms).

It is also believed that the AC correction factors disclosed herein not only correct for hematocrit effects, but also for variability in reagent film state or other sources of error. In the examples described herein, the electrodes used to detect AC signal responses are the same as those used for DC signal responses and are therefore coated with a reagent film. As a result, all AC response measurements are affected by the state (e.g., thickness, expansion) of the reagent film with the applied liquid sample.

Another way to view this data is from the corresponding Clark error grid. Figure 41 shows the error grid for the uncorrected DC response data at the 900 ms DC measurement time. Figures 42-44 show the same 900 ms DC measurement data corrected with AC response data for only 20 kHz (Figure 42), only 2 kHz (Figure 43), and both 20 kHz and 2 kHz (Figure 44).

The data from Example 5 supports the finding that analyte measurements with good TSE can be achieved at short total measurement times between 900 ms and 3000 ms.

Example 5 was not covaried with temperature as was done in Example 2 because the electrochemical test bench was less dedicated to performing "environmental" studies. Therefore, the AC signal responses and correction factors determined from those responses in this example did not contain information about sample temperature variations and corrections as shown in Example 2. However, the AC method using four AC frequencies was shown to correct for both hematocrit and temperature variations, and the measurement method of Example 5 would be sufficient to do so with a test time of less than 1000 ms.

For the purposes of the examples disclosed herein, the applied DC excitation is described and generally shown as being applied to a single block of potential for a duration. DC response data can be acquired throughout the duration or only at one or a few points within the duration. However, not shown or described is the application of a DC excitation comprising two or more shorter pulses of DC excitation with response data measured for each such pulse. While none of the examples herein illustrate the use of such DC excitation, it is believed that both the AC waveforms, continuous and multi-frequency (simultaneous) waveforms described herein are capable of modifying response data obtained from pulsed DC excitation.

The features disclosed in the above description, the claims and the drawings can be essential both individually and in any combination with one another for the realization of the invention in its various embodiments.

It should be noted that terms like "preferably," "generally," and "typically" are not intended herein to limit the scope of the claimed invention or to imply that certain features are critical, essential, or even important to the structure or function of the claimed invention. Rather, these terms are merely intended to highlight alternative or additional features that may or may not be utilized in a particular embodiment of the invention.

For the purposes of describing and defining the present invention, it should be noted that the term "substantially" is utilized herein to represent the inherent degree of uncertainty that may be attributed to any quantitative comparison, value, measurement, or other representation. The term "substantially" is also utilized herein to represent the degree by which a quantitative representation may vary from a stated reference without resulting in a change in the basic function of the subject matter at issue.

The present invention has been described in detail by reference to specific embodiments of the present disclosure, but it will be apparent that modifications and variations can be made without departing from the scope of the invention as defined in the appended claims. More specifically, although certain aspects of the present invention are identified herein as preferred or particularly advantageous, it is contemplated that the present invention is not necessarily limited to these preferred aspects of the invention.

Claims (18)

1. A method for determining the glucose concentration in a whole blood sample, comprising: a) Apply a signal to the sample; b) Measure the response to the signal; and c) The response is used at least to determine the analyte concentration of the sample, wherein the determination has a total test time of 2.0 seconds or less; The signal has an AC component comprising a multi-frequency excitation waveform, wherein the frequency of the AC component is applied with a low excitation AC potential to generate a non-Radidas current response. 2. The method of claim 1, wherein the determination has a total test time between 0.5 seconds and 1.525 seconds, and wherein the method has a total systematic error of less than 11%. 3. A method for determining the concentration of an analyte of interest in a biological sample, comprising: a) Apply a first signal having a first AC component to the sample; b) Measure the first response to the first signal; c) Apply a second signal having a second AC component to the sample; d) Measure the second response to the second signal; e) Apply a third signal with a third AC component to the sample; f) Measure the third response to the third signal; g) Apply a fourth DC signal to the sample; h) Measure the fourth response to the fourth signal; and e) At least the first, second, and third responses are used to determine the analyte concentration of the sample to correct for the effect of at least one interfering substance within the sample, wherein the determination has a total test time of 2.0 seconds or less. The frequencies of the first, second, and third AC components are applied with a low excitation AC potential to generate a non-Radida current response. 4. The method of claim 3, wherein the determination has a total test time between 0.5 seconds and 1.525 seconds. 5. The method of claim 3, wherein the first, second and third signals each comprise AC signals of different frequencies. 6. The method of claim 3, wherein the interfering agent comprises one or both of hematocrit and temperature. 7. A method for determining the concentration of an analyte of interest in a biological sample, comprising: a) Apply a first signal having a first AC component to the sample; b) Measure the first response to the first signal, wherein the first signal is applied no more than 100 ms before the first response is measured; c) Apply a second signal having a second AC component to the sample; d) Measure the second response to the second signal, wherein the second signal is applied for no more than 100 ms before the second response is measured; e) Apply a third signal with a third AC component to the sample; f) Measure the third response to the third signal, wherein the third signal is applied for no more than 100 ms before the third response is measured; g) Apply a fourth signal with a fourth AC component to the sample; h) Measure the fourth response to the fourth signal, wherein the fourth signal is applied for no more than 100 ms before the fourth response is measured; i) Apply a fifth DC signal to the sample; j) Measure the fifth response to the fifth signal; and k) At least the first, second, third, and fourth responses are used to determine the analyte concentration in the sample to correct for the fifth response in response to the influence of at least one interfering substance in the sample. The determination has a total test time between 0.5 seconds and 2.0 seconds, wherein the frequencies of the first, second, third and fourth AC components are applied with a low excitation AC potential to generate a non-Radidas current response. 8. The method of claim 7, wherein the interfering agent comprises one or both of hematocrit and temperature. 9. The method of claim 7, wherein the first signal is a 10 kHz AC signal, the second signal is a 20 kHz AC signal, the third signal is a 1 kHz AC signal, and the fourth signal is a 2 kHz AC signal. 10. The method of claim 7, wherein the fifth signal is applied for no more than 200 ms before the fifth response is measured. 11. The method of claim 7, wherein the method has a total systematic error between 7.5% and 11%. 12. A method for determining the concentration of an analyte of interest in a biological sample, comprising: a) Apply the sample to a reagent film with a thickness between 1.6 μm and 5 μm; b) Apply a signal to the sample; c) Measure the response to the signal; and d) Use at least the response to determine the analyte concentration of the sample, wherein the determination has a total test time of 2.0 seconds or less. The signal is a multi-frequency AC signal with an AC component, wherein the frequency of the AC component is applied with a low excitation AC potential to generate a non-Radidas current response. 13. The method of claim 12, wherein the determination has a total test time between 0.6 seconds and 1.525 seconds. 14. The method of claim 12, further comprising the following steps: e) Apply a second signal to the sample; and f) Measure the second response to the second signal; Step (d) includes using at least the first response to determine the analyte concentration of the sample in order to correct the second response for the effect of at least one interfering substance in the sample. 15. The method of claim 14, wherein the second signal is a DC signal. 16. A method for determining the concentration of an analyte of interest in a biological sample, comprising: a) Provide an electrochemical biosensor with a reagent coating having a thickness of less than 4 µm to an electrode system configured for use in electrochemical analysis; b) Provide a biological sample to a biosensor configured to allow the sample to come into contact with a reagent coating; c) Detect the time when the sample comes into contact with the reagent coating using an electrical method; d) Apply a first signal to the sample; e) Measure the first response to the first signal; f) Apply a second signal to the sample; g) Measure the second response to the second signal; and h) Use at least the first and second responses to determine the analyte concentration of the sample; The first signal includes an AC component having at least two consecutively applied frequencies, and the first response includes response data for each frequency, the response data including correction factor information. The second signal includes a DC signal, and the second response includes a current response indicating the concentration of the analyte. Specifically, a first signal is applied when the sample comes into contact with the coating for a first time period from 400 ms to 800 ms; Specifically, a second signal is applied immediately after the period used for the first signal, for a second time period ranging from 100 ms to 2200 ms; and The determination of a total test time of 3.0 seconds or less and a TSE of 9.0% or less is included. The frequency of the AC component is applied with a low excitation AC potential to generate a non-Radidas current response. 17. The method of claim 16, wherein the second time period is from 100 ms to 300 ms, the total test time is 1.1 seconds or less, and the TSE is 8.5% or less. 18. The method of claim 16, wherein the AC component comprises four consecutively applied frequencies, each frequency being applied for no more than 200 ms, such that the first time period is no more than 800 ms, and wherein the second time period is 100 ms, the total test time is 900 ms, and the TSE is 8.0% or less.