US20140051953A1 - Adaptive calibration system for spectrophotometric measurements - Google Patents

Abstract

This disclosure describes, among other features, systems and methods for customizing calibration curves, parameter algorithms, and the like to individual users. An initial calibration curve generated based on a population can be used as a starting point in an algorithm for measuring a physiological parameter such as glucose. The measurement algorithm can determine one or more initial measurement values for a user based on the initial calibration curve. In certain embodiments, one or more alternative measurements, such as invasive or minimally invasive measurements, can periodically or sporadically be input into the measurement algorithm. The algorithm can use the alternative measurements to adapt the calibration curve to the individual. As a result, measurements for the individual can more accurately reflect the individual's actual parameter values.

Description

RELATED APPLICATIONS
• This application is a divisional of U.S. application Ser. No. 12/891,428, filed Sep. 27, 2010, entitled “Adaptive Calibration System for Spectrophotometric Measurements,” which claims priority from U.S. Provisional Patent Application No. 61/246,288 filed Sep. 28, 2009, entitled “Adaptive Calibration System for Spectrophotometric Measurements,” and from U.S. Provisional Patent Application No. 61/257,722, filed Nov. 3, 2009, entitled “Adaptive Calibration System for Spectrophotometric Measurements.” Each of the foregoing applications are hereby incorporated by reference in their entirety.
BACKGROUND
• The standard of care in caregiver environments includes patient monitoring through spectroscopic analysis using, for example, a pulse oximeter. Devices capable of spectroscopic analysis generally include a light source(s) transmitting optical radiation into or reflecting off a measurement site, such as, body tissue carrying pulsing blood. After attenuation by tissue and fluids of the measurement site, one or more photodetection devices detect the attenuated light and output one or more detector signals responsive to the detected attenuated light. A processor can process the one or more detector signal and output a measurement reflective of a blood constituent of interest, such as glucose, oxygen, methemoglobin, total hemoglobin, among other physiological parameters.
• In noninvasive devices and methods, a sensor is often adapted to position an appendage such as a finger proximate a light source and a light detector. For example, noninvasive sensors often include a clothespin-shaped housing that includes a contoured bed conforming generally to the shape of a finger.
SUMMARY OF CERTAIN EMBODIMENTS
• This disclosure describes embodiments of noninvasive methods, devices, and systems for measuring a blood constituent or analyte, such as oxygen, carbon monoxide, methemoglobin, total hemoglobin, glucose, proteins, lipids, a concentration percentage thereof (e.g., saturation), or for measuring many other physiologically relevant patient characteristics. These characteristics can relate, for example, to pulse rate, hydration, trending information and analysis, patient wellness, and the like.
• In certain embodiments, a device capable of producing a signal responsive to light attenuated by tissue at a measurement site includes an optical sensor that can emit light on tissue of a living person, detect the light after attenuation by the tissue, and output a signal responsive to the attenuated light. The device can further include a processor that can receive the signal from the optical sensor, process the signal with a measurement algorithm to determine a first measurement of a physiological parameter, receive a second measurement of the physiological parameter from an alternative source, and adaptively adjust the measurement algorithm based at least partly on the second measurement.
• In certain embodiments, a method of determining whether to recommend an alternative measurement of a physiological parameter can include obtaining a noninvasive measurement of a physiological parameter using an optical sensor, receiving an alternative measurement of the physiological parameter, where the alternative measurement can be generated by an alternative sensor, analyzing the noninvasive and alternative measurements to determine whether a condition has been met, and in response to the condition being met, outputting an indication that a new measurement should be obtained from the alternative sensor.
• For purposes of summarizing the disclosure, certain aspects, advantages and novel features of the inventions have been described herein. It is to be understood that not necessarily all such advantages can be achieved in accordance with any particular embodiment of the inventions disclosed herein. Thus, the inventions disclosed herein can be embodied or carried out in a manner that achieves or optimizes one advantage or group of advantages as taught herein without necessarily achieving other advantages as can be taught or suggested herein.
BRIEF DESCRIPTION OF THE DRAWINGS
• Throughout the drawings, reference numbers can be re-used to indicate correspondence between referenced elements. The drawings are provided to illustrate embodiments of the inventions described herein and not to limit the scope thereof.
• FIG. 1 illustrates an embodiment of a calibration system;
• FIG. 2 illustrates an embodiment of a monitoring system that can implement the calibration systemFIG. 1;
• FIG. 3 illustrates another embodiment of a monitoring system that can implement the calibration systemFIG. 1;
• FIG. 4 illustrates an embodiment of a physiological monitor that may be used in the monitoring systems ofFIG. 2 or3;
• FIGS. 5A through 5C illustrate embodiments of glucose monitoring systems;
• FIGS. 6 and 7 illustrate additional embodiments of glucose monitoring systems;
• FIGS. 8 and 9 illustrate embodiments of parallel engines for glucose monitoring systems;
• FIGS. 10 and 11A illustrate embodiments of processes for determining whether to suggest obtaining an alternative measurement of a physiological parameter;
• FIG. 11B illustrates an embodiment of a process for adjusting an alternative glucose measurement based at least in part on a hemoglobin measurement; and
• FIG. 12 illustrates another embodiment of a data collection system.
DETAILED DESCRIPTION
• Noninvasive optical sensors can use spectrophotometry techniques to measure a variety of blood constituents, including for example, glucose, oxygen saturation, hemoglobin, methemoglobin, carboxyhemoglobin, other hemoglobin species, concentrations of the same, and the like. In addition, noninvasive optical sensors can also be used to measure a variety of other physiological parameters, including pulse rate, perfusion, and the like. An optical sensor can include one or more emitters that shine light through tissue of a living person, such as through a finger, toe, or foot. One or more detectors can receive the transmitted light after attenuation by the tissue and can generate one or more signals responsive to the attenuated light. A processor can process the one or more signals to derive measurements of one or more physiological parameters.
• Noninvasive optical sensors can be calibrated empirically by obtaining measurements from a population of users. By comparing the noninvasive measurements with measurements of known parameter values in the population, a calibration curve can be generated. Because this initial calibration curve may be generated based on a population of individuals, it may not accurately reflect parameter levels for certain individuals. The optimal calibration curve for one individual can differ from the initial curve because of skin variations resulting from pigmentation, UV damage, age, erythema, or the like. Different calibration curves can also result from fingernail variations and variations in individuals' hemoglobin species.
• This disclosure describes, among other features, systems and methods for customizing calibration curves, parameter algorithms, and the like to individual users. In certain embodiments, an initial calibration curve generated based on a population can be used as a starting point in an algorithm for measuring a physiological parameter such as glucose. The measurement algorithm can determine one or more initial measurement values for a user based on the initial calibration curve. In certain embodiments, one or more alternative measurements, such as invasive or minimally invasive measurements, can periodically or sporadically be input into the measurement algorithm. The algorithm can use the alternative measurements to adapt the calibration curve to the individual. As a result, measurements for the individual can more accurately reflect the individual's actual parameter values.
• In the example context of glucose, measurement information from a finger-prick glucose meter or from another glucose sensor can be supplied to a noninvasive glucose device. For instance, a user could manually input a measurement obtained from a finger prick glucose meter into the noninvasive glucose device, or the noninvasive device might have a built-in finger prick meter. An adaptive glucose algorithm in the noninvasive device can adaptively recalibrate itself based on measurements received from the finger prick meter. In other embodiments, noninvasive parameter measurements can be used to calibrate or adjust invasive or minimally-invasive measurements.
• For purposes of illustration, the remainder of this disclosure is described primarily in the context of glucose. However, the features described herein can be applied to other blood constituents or concentrations thereof and to other physiological parameters.
• FIG. 1 illustrates an embodiment of acalibration system100 that can adaptively adjust parameter measurements of a user. Thecalibration system100 can start with an empirically-derived model for measuring one or more parameters. The empirical model can be a calibration curve that is generated based on measurement data taken from a population. Advantageously, in certain embodiments, thecalibration system100 can adapt the empirical model to an individual. As a result, measurements taken for the individual can be more accurate.
• In the depicted embodiment, thecalibration system100 includes aparameter calculator101. Theparameter calculator101 can include hardware (such as one or more processors), software, and/or firmware for calculating a physiological parameter such as glucose, oxygen saturation, hemoglobin, or the like. Inputs to theparameter calculator101 can include, among others,measurement data105 andalternative data103 indicative of one or more parameters. Themeasurement data105 can be obtained from a physiological sensor (not shown), such as a noninvasive optical sensor. Examples of noninvasive optical sensors are described below (see, e.g.,FIGS. 4 and 12).
• Thealternative data103 can be obtained from an alternative source, such as another patient monitor or another sensor (not shown). The alternative patient monitor can be an invasive or minimally-invasive monitor or can even be another noninvasive monitor. For example, in the context of glucose, the alternative patient monitor could be a spot-check monitor (e.g., a finger-prick glucose meter) or a continuous glucose monitor.
• Theparameter calculator101 can calculate one or more physiological parameters based on themeasurement data105. If themeasurement data105 is provided by an optical sensor, themeasurement data105 can include light transmittance values after attenuation by tissue of a patient. Theparameter calculator101 can compare the transmittance values or a ratio derived from the transmittance values to a calibration curve to obtain a parameter value.
• As an example, a pulse oximetry sensor can shine red and infrared wavelengths of light into a tissue site. A photodetector can receive the red and infrared light attenuated by the tissue site, and in response, transmit ameasurement data105 signal to theparameter calculator101. Theparameter calculator101 can compute a ratio of the red to infrared values in themeasurement data105 and compare this ratio with an empirically-generated calibration curve to obtain a value for oxygen saturation (SpO₂).
• In certain embodiments, theparameter calculator101 advantageously uses thealternative data103 to adjust calculation of the one or more physiological parameters. For instance, theparameter calculator101 can use thealternative data105 to adjust an algorithm for calculating a physiological parameter. In one embodiment, adjusting the algorithm can include adjusting an empirically-derived calibration curve used by the algorithm. Advantageously, in certain embodiments, the calibration curve used by theparameter calculator101 can therefore account for characteristics of an individual under measurement. Embodiments of algorithms that can be used by theparameter calculator101 are described in greater detail below with respect toFIGS. 5 through 11B.
• Theparameter calculator101 canoutput parameter data107 indicative of the calculated parameters. Theparameter data107 can be displayed on adisplay device115. Theparameter calculator101 could also output thealternative data103. For example, theparameter calculator101 could output a finger-prick glucose measurement alongside or in place of a noninvasive glucose measurement. In another embodiment, theparameter calculator101 provides parameter values as anoutput106 to another device, for example, over a network.
• Theparameter calculator101 can also calculate trend data reflecting trend information for theparameter data107. Theparameter calculator101 can also synthesize or scale waveform data. In addition to outputting theparameter data107, theparameter calculator101 can output trenddata109, synthesized, scaled, oractual waveforms111,calibration data113, and alarms114. Thecalibration data113 can include information related to calibrations performed by the parameter calculator101 (see, e.g.,FIG. 7). Theparameter calculator101 can provide theoutputs107,109,111,113 to thedisplay115, to a separate patient monitoring device, or to another device configured to receive physiological parameter information.
• In an embodiment, theparameter calculator101 is implemented in a single monitoring device. In an embodiment, the features of theparameter calculator101 are distributed among separate devices. In an embodiment, theparameter calculator101 includes a processor, processor board, or an Original Equipment Manufacture (OEM) board. In an embodiment, theparameter calculator101 is portable. Data communicated between the various components of thecalibration system100 can be communicated through cables or wirelessly. Other inputs and/or outputs can be included with the system. For example, an error data output can be used to communicate an error calculated between the measureddata105 and thealternative data103.
• FIG. 2 illustrates anexample monitoring system200 that can implement thecalibration system100 ofFIG. 1. Themonitoring system200 includes anoninvasive monitor210, which is coupled to an individual202 via asensor201. Thesensor201 is an example of a noninvasive optical sensor. Thesensor201 can provide measurement data to thenoninvasive monitor210, which can calculate a physiological parameter based at least in part on the measurement data and display the physiological parameter. Thenoninvasive monitor210 can output glucose values derived from thesensor210. Thenoninvasive monitor210 can measure glucose values continuously or can be used for spot-checks.
• Thenoninvasive monitor210 can implement thecalibration system100 described above. For example, thenoninvasive monitor210 can adjust a glucose algorithm based on alternative data received from an alternative monitor. The alternative monitor in the depicted embodiment is aglucose meter220, sometimes referred to as a glucometer or a finger-prick meter. The individual202 can use theglucose meter220 to prick the individual's finger. The individual can then apply a blood sample to atest strip230. The individual can insert thetest strip230 into the glucose monitor220, which can determine a glucose measurement from the blood sample on thetest strip230.
• Theglucose meter220 can provide glucose measurements to thenoninvasive monitor210 wirelessly or through a wired connection. Alternatively, the individual202 can input measurement values calculated by theglucose meter220 into thenoninvasive monitor210. Thenoninvasive monitor210 can then use the features of thecalibration system100 described above and in greater detail below to adjust the calculation of noninvasive glucose.
• FIG. 3 illustrates anotherexample monitoring system300 that can implement thecalibration system100 ofFIG. 1. Themonitoring system300 includes amonitor310, which like thenoninvasive monitor210, obtains glucose measurements noninvasively from asensor301 coupled to an individual202. Additionally, themonitor300 includes the functionality of theglucose meter220 described above. Namely, themonitor300 can be used to obtain blood samples from atest strip330. Thus, in certain embodiments, themonitor310 can calculate both noninvasive glucose values and alternative glucose values. Themonitor310 can implement thecalibration system100 ofFIG. 1 to adjust calculation of the noninvasive glucose based at least in part on the alternative glucose measurements. Although not shown, in other embodiments themonitor310 can include the features of a continuous minimally-invasive glucose monitor instead of or in addition to the features of a finger-prick monitor.
• FIG. 4 illustrates anotherexample monitoring device400 in which thecalibration system100 can be housed. Advantageously, in certain embodiments, theexample monitoring device400 shown can have a shape and size that allows a user to operate it with a single hand or attach it, for example, to a user's body or limb.
• In the depicted embodiment, themonitoring device400 includes afinger clip sensor401 connected to amonitor409 via acable412. In the embodiment shown, themonitor409 includes adisplay410,control buttons408 and a power button. Moreover, themonitor409 can advantageously include electronic processing, signal processing, and data storage devices capable of receiving signal data from thesensor401, processing the signal data to determine one or more output measurement values indicative of one or more physiological parameters of a user, and displaying the measurement values, trends of the measurement values, combinations of measurement values, and the like.
• Thecable412 connecting thesensor401 and themonitor409 can be implemented using one or more wires, optical fiber, flex circuits, or the like. In some embodiments, thecable412 can employ twisted pairs of conductors in order to minimize or reduce cross-talk of data transmitted from thesensor401 to themonitor409. Various lengths of thecable412 can be employed to allow for separation between thesensor401 and themonitor409. Thecable412 can be fitted with a connector (male or female) on either end of thecable412 so that thesensor401 and themonitor409 can be connected and disconnected from each other. Alternatively, thesensor401 and themonitor409 can be coupled together via a wireless communication link, such as an infrared link, a radio frequency channel, or any other wireless communication protocol and channel. Thesensor401 could also be integrated with amonitor409 in other embodiments.
• Moreover, thesensor401 and/or monitor409 can include any of the features described in any of the following related applications, each of which is hereby incorporated by reference in its entirety: U.S. application Ser. No. 12/534,827, filed Aug. 3, 2009, titled “MULTI-STREAM DATA COLLECTION SYSTEM FOR NONINVASIVE MEASUREMENT OF BLOOD CONSTITUENTS”; U.S. application Ser. No. 12/534,812, filed Aug. 3, 2009, titled “MULTI-STREAM SENSOR FRONT ENDS FOR NONINVASIVE MEASUREMENT OF BLOOD CONSTITUENTS”; U.S. application Ser. No. 12/534,823, filed Aug. 3, 2009, titled “MULTI-STREAM SENSOR FOR NONINVASIVE MEASUREMENT OF BLOOD CONSTITUENTS”; U.S. application Ser. No. 12/534,825, filed Aug. 3, 2009, titled “MULTI-STREAM EMITTER FOR NONINVASIVE MEASUREMENT OF BLOOD CONSTITUENTS”; U.S. application Ser. No. 12/497,528, filed Jul. 2, 2009, titled “NOISE SHIELDING FOR A NONINVASIVE DEVICE”; U.S. application Ser. No. 12/497,523, filed Jul. 2, 2009, titled “CONTOURED PROTRUSION FOR IMPROVING SPECTROSCOPIC MEASUREMENT OF BLOOD CONSTITUENTS”; U.S. Provisional Application No. 61/239,741, filed Sep. 3, 2009, titled “EMITTER DRIVER FOR NONINVASIVE PATIENT MONITOR”; U.S. Design application No. 29/323,409, filed Aug. 25, 2008, titled “PATIENT MONITORING SENSOR”; U.S. Design application No. 29/323,408, filed Aug. 25, 2008, titled “PATIENT MONITOR”; U.S. application Ser. No. 12/497,506, filed Jul. 2, 2009, titled “HEAT SINK FOR NONINVASIVE MEDICAL SENSOR”; U.S. Provisional Application No. 61/243,507, filed Sep. 17, 2009, titled “IMPROVING ANALYTE MONITORING USING ONE OR MORE ACCELEROMETERS”; U.S. Provisional Application No. 61/177,971, filed May 13, 2009, titled “CALIBRATIONLESS SPECTROPHOTOMETER”; and U.S. Provisional Application No. 61/228,495, filed Jul. 24, 2009, titled “INTERFERENCE DETECTOR FOR PATIENT MONITOR.”
• Themonitor409 can be attached to the patient. For example, themonitor409 can include a belt clip or straps (not shown) that facilitate attachment to a patient's belt, arm, leg, or the like. Themonitor409 can also include a fitting, slot, magnet, LEMO snap-click connector, or other connecting mechanism to allow thecable412 andsensor401 to be attached to themonitor409.
• Themonitor409 can also include other components, such as a speaker, power button, removable storage or memory (e.g., a flash card slot), an AC power port, and one or more network interfaces, such as a universal serial bus interface or an Ethernet port. For example, themonitor409 can include adisplay410 that can indicate a measurement for glucose, for example, in mg/dL. Other analytes and forms of display can also appear on themonitor409.
• Thesensor401 can measure various blood constituents or analytes noninvasively using multi-stream spectroscopy. In an embodiment, the multi-stream spectroscopy can employ visible, infrared and near infrared wavelengths. Thesensor401 can include photocommunicative components, such as an emitter, a detector, and other components (not shown). The emitter can include a plurality of sets of optical sources that, in an embodiment, are arranged together as a point source. The various optical sources can emit a sequence of optical radiation pulses at different wavelengths towards a measurement site, such as a patient's finger. Detectors can then detect optical radiation from the measurement site. The optical sources and optical radiation detectors can operate at any appropriate wavelength, including, for example, infrared, near infrared, visible light, and ultraviolet. In addition, the optical sources and optical radiation detectors can operate at any appropriate wavelength, and modifications to the embodiments desirable to operate at any such wavelength can be used in certain embodiments. Thesensor401 or themonitor409 can also provide outputs to a storage device or network interface.
• In addition, although asingle sensor401 with asingle monitor409 is shown, different combinations of sensors and device pairings can be implemented. For example, multiple sensors can be provided for a plurality of differing patient types or measurement sites or even patient fingers. As described above with respect toFIG. 3, themonitor409 can also include noninvasive or minimally-invasive features for measuring analytes such as glucose.
• FIGS. 5A through 5C illustrate embodiments of glucose monitoring systems500. Each of the glucose monitoring systems500 can implement certain features of theparameter calculator101 described above.
• Referring toFIG. 5A, an embodiment of aglucose monitoring system500A is shown. In theglucose monitoring system500A, signal inputs are provided to a filtering andprocessing module510. The signal inputs can include signals received from a noninvasive optical sensor or the like. For example, the signal inputs could include optical signals, photoplethysmograph signals, transmittance signals, combinations of the same, and the like. Moreover, in some implementations, the input signals can include data regarding optical path length between one or more emitters and one or more detectors, a straight line distance between an emitter and a detector, and an angle between the straight line distance and an axis parallel to or substantially parallel to a detector shell or finger.
• The filtering andprocessing module510 can include analog and/or digital circuitry for filtering and processing one or more of the signal inputs. For instance, themodule510 can include analog conditioning circuitry, anti-alias filters, analog-to digital-converters, and the like. Themodule510 can pride one or more outputs to anadaptive glucose algorithm520. These outputs can include one or more ratios of wavelengths of light, transmittance values, coefficients of absorption, absorption values, combinations of the same, and the like.
• Theadaptive glucose algorithm520 can receive these outputs as well as an alternative measurement input, G_u. Theadaptive glucose algorithm520 can be implemented in hardware and/or software. The alternative measurement G_ucan be an invasive or minimally-invasive glucose value. In another embodiment, the alternative measurement input can include raw signal inputs from an invasive or minimally-invasive glucose monitor.
• In certain embodiments, theadaptive glucose algorithm520 initially starts noninvasive monitoring using an empirically-derived calibration curve. Theadaptive glucose algorithm520 can then use the alternative measurement to adjust the calibration curve to better match characteristics of a user. Theadaptive glucose algorithm520 can output a noninvasive measurement Ĝ_u, which reflects adjustment by the alternative measurement G_u. More detailed embodiments of theadaptive glucose algorithm520 are described below with respect toFIG. 6.
• FIG. 5B illustrates another embodiment of aglucose monitoring system500B. Thesystem500B also includes the filtering andprocessing module510, which receives the signal inputs and provides filtered and processed outputs to aglucose algorithm522. Theglucose algorithm522 in the depicted embodiment uses an empirically-derived calibration curve to generate noninvasive glucose measurements Ĝ_u. The measurements Ĝ_uare output to acalibration module530.
• Thecalibration module530 can be implemented in hardware and/or software. Thecalibration module530 can calibrate the noninvasive glucose measurements Ĝ_uusing one or more alternative glucose measurements G_u. For example, thecalibration module530 could perform regression or another statistical estimation technique to calibrate the noninvasive measurement. Theglucose monitoring system500B can therefore calibrate glucose measurements based on alternative measurements without adapting theglucose algorithm522.
• Examples of these calibration techniques are described below with respect toFIG. 7. Additional calibration embodiments that might be performed by the calibration module are described below with respect toFIGS. 10 and 11.
• FIG. 5C illustrates another embodiment of aglucose monitoring system500C. Thesystem500C includes the features of both theglucose monitoring systems500A and500B described above. For example, signal inputs are provided to a filtering andprocessing block510, which provides outputs to theadaptive glucose algorithm520. Theadaptive glucose algorithm520 can receive an alternative glucose measurement G_uand use this measurement to adjust the calibration curve. Theadaptive glucose algorithm520 can output a noninvasive measurement Ĝ_u, which can reflect adjustment by the alternative measurement G_u.
• The output Ĝ_uof theadaptive glucose algorithm520 is provided to thecalibration module530, which can further calibrate the noninvasive glucose measurement Ĝ_uwith the alternative measurement G_u. The resulting output of thecalibration module530 is a calibrated glucose value {circumflex over (Ĝ)}_u.
• FIG. 6 illustrates a more detailed embodiment of aglucose monitoring system600. In particular, theglucose monitoring system600 is a more detailed embodiment of theglucose monitoring system500A ofFIG. 5A. As such, theglucose monitoring system600 includes the filtering andprocessing module510 and anadaptive glucose algorithm620.
• As before, signal inputs are received by the filtering andprocessing module510, which outputs one or more ratios of wavelengths of light, transmittance values, coefficients of absorption, absorption values, and the like. In some embodiments, the noninvasive sensor used to obtain the signal inputs obtains n signals corresponding to n detected wavelengths of light, where n is an integer. The filtering andprocessing module510 can output ratios of a subset of these wavelengths. For two wavelengths, themodule510 could output a single ratio. For multiple (e.g., more than two) wavelengths, themodule510 could combine any of the wavelengths into ratios.
• Theadaptive glucose algorithm620 receives the ratios from themodule510 and processes the ratios or other outputs to determine an initial noninvasive glucose measurement Ĝ_u. In the depicted embodiment, theadaptive glucose algorithm620 includes a plurality of basis functions624 (ψ_i). Each of the basis functions624 can receive one or more ratios (or other outputs of the filtering and processing module510) as inputs. The basis functions624 can be blending functions or the like. The basis functions624 are multiplied by weights626 (w_i) to produce a noninvasive glucose measurement Ĝ_u. The combination of basis functions624 andweights626 can be expressed as follows, via sum block628:
• $\begin{array}{cc}{\hat{G}}_u=\sum _i^{}W_i{\psi }_i& \left(1\right)\end{array}$
• In one embodiment, the basis functions624 are polynomial functions, such as x, x², (x₁+x₂)/2, and the like, where x represents a ratio or some other output of the filtering andprocessing module510. The basis functions624 can also be logarithmic (e.g., ln(x)), trigonometric, Fourier basis functions, wavelet basis functions, combinations of the same, and the like. In another embodiment, at least some of the basis functions624 are radial basis functions. The radial basis functions are Gaussian in one embodiment, having the form
• 
  φ(r)=e^{−α(r-r0)T(r-r0)}  (2)
• where r represents a vector of ratios, r₀represents an initial vector of ratios, and α is a constant.
• In certain embodiments, the combination of the basis functions624 andweights626 approximates a calibration curve. Application of the basis functions624 andweights626 to the ratios (orother module510 outputs) effectively applies the calibration curve to the ratios (or other outputs). Advantageously, in some implementations, radial basis functions624 can approximate the calibration curve more accurately than other bases, such as polynomial bases.
• In one embodiment, the basis functions624 andweights626 are selected empirically. In addition, the number of basis functions624 andweights626 used can be determined empirically. In one embodiment, the more ratios that are provided as inputs to theadaptive glucose algorithm620, the fewer basis functions624 andweights626 that might be used, and vice versa.
• Because the basis functions624 andweights626 are selected empirically, they may not accurately represent a true calibration curve for a given individual. Thus, in certain embodiments, anadaptive algorithm625 can be used to adapt theweights626 of theglucose algorithm620. Theadaptive algorithm625 receives an error signal or cost function e(n), which can be the difference between the noninvasive measurement Ĝ_uand an alternative (e.g., invasive) measurement G_u(via sum block627). In one embodiment, theadaptive algorithm625 can minimize the cost function e(n) to obtain adjustment factors for theweights626 ornew weights626.
• Any of a variety ofadaptive algorithms625 may be used. For instance, theadaptive algorithm625 could implement one or more of the following: a least mean squares algorithm (LMS), a least squares algorithm, a recursive least squares (RLS) algorithm, a Kalman filter, a joint process estimator, an adaptive joint process estimator, a least-squares lattice joint process estimator, a least-squares lattice predictor, a correlation canceller, optimized or frequency domain implementations of any of the above, any other linear predictor, combinations of the same, and the like.
• Theadaptive glucose algorithm620 can adapt in other ways. For example, theadaptive glucose algorithm620 could use fewer, more, or different types ofbasis functions624 for certain individuals to adapt the calibration curve. In addition, in alternative embodiments, theadaptive glucose algorithm620 adapts the raw signal output from the filtering andprocessing module510 with the alternative measurement or a raw alternative measurement. For instance, in themonitoring system300 ofFIG. 3, where themonitor310 includes noninvasive and minimally-invasive features together, themonitor310 could adapt raw noninvasive signals with raw minimally-invasive signals.
• FIG. 7 illustrates another embodiment of aglucose monitoring system700. Theglucose monitoring system700 includes the features of theglucose monitoring system600. In addition, theglucose monitoring system700 supplies the noninvasive output Ĝ_uto thecalibration module530 described above with respect toFIG. 5B. Thecalibration module530 also receives an alternative measurement, for example, from an invasive or minimally-invasive monitor. Thecalibration module530 can advantageously further refine the noninvasive measurement to output a calibrated glucose value {circumflex over (Ĝ)}_u.
• Thecalibration module530 can perform calibration using various techniques, such as linear regression. In one embodiment, the regression can be performed by first determining an offset, for example, as follows:
• 
  offset=G_u−Ĝ_u  (3)
• where G_urepresents an alternative glucose measurement and where Ĝ_urepresents a noninvasive glucose measurement. The offset can be used to obtain a calibrated measurement, for example, as follows:
• 
  {circumflex over (Ĝ)}_u=Ĝ_u+offset  (4)
• where {circumflex over (Ĝ)}_uis the calibrated glucose value. Equation (4) can be refined in one embodiment by applying a multiplier to the noninvasive measurement. Equation (5) illustrates an example of using a multiplier:
• 
  {circumflex over (Ĝ)}_u=αĜ_u+offset  (5)
• where α is the multiplier.
• In addition, multiple alternative and noninvasive measurements can be combined together in a system of linear equations based on equation (5) to calculate or estimate α and the offset values. An example of a system of two equations is shown in equation (6):
• $\begin{array}{cc}\left[\begin{array}{c}{G}_{u1}\\ G_{u2}\end{array}\right]=\underset{\underset{A}{}}{\left[\begin{array}{cc}{\hat{\hat{G}}}_{u1}& 1\\ {\hat{\hat{G}}}_{u2}& 1\end{array}\right]}\left[\begin{array}{c}\alpha \\ \mathrm{offset}\end{array}\right]& \left(6\right)\end{array}$
• In equation (6), A is a matrix. Assuming A is nonsingular, equation (6) can be solved for α and the offset as follows:
• $\begin{array}{cc}\left[\begin{array}{c}\alpha \\ \mathrm{offset}\end{array}\right]=A^{-1}\left[\begin{array}{c}{G}_{u1}\\ G_{u2}\end{array}\right]& \left(7\right)\end{array}$
• If A is singular, or if more than two alternative measurements are being analyzed, in some embodiments the pseudoinverse of A can be used to calculate α and the offset. For example:
• $\begin{array}{cc}\left[\begin{array}{c}\alpha \\ \mathrm{offset}\end{array}\right]=P_{\mathrm{INV}}\left(A^{-1}\right)\left[\begin{array}{c}{G}_{u1}\\ G_{u2}\\ ⋮\\ G_{\mathrm{uN}}\end{array}\right]& \left(8\right)\end{array}$
• where P_INV( ) denotes the Moore-Penrose pseudoinverse (or another suitable pseudoinverse).
• In still other embodiments, polynomial regression may be used to calibrate the glucose measurements. Polynomial regression can take the form of a polynomial equation in Ĝ_u(a noninvasive measurement). One example of a polynomial function that may be used is shown in equation (9):
• 
  {circumflex over (Ĝ)}_u=P(Ĝ_u)=α_nĜ_uⁿ+α_n-1Ĝ_u^n-1+ . . . +offset  (9)
• where P( ) represents a polynomial operator, α_nrepresent the coefficients of the noninvasive measurement, and n is an integer that represents the order of the polynomial.
• To solve for the coefficients α_nand the offset, a system of polynomial equations can be generated from multiple noninvasive and alternative measurements. This system can be written in matrix form as follows:
• $\begin{array}{cc}\underset{\underset{y}{}}{\left[\begin{array}{c}{G}_{u1}\\ G_{u2}\\ ⋮\\ G_{\mathrm{un}+1}\end{array}\right]}=\underset{\underset{A}{}}{\left[\begin{array}{cccc}{\hat{G}}_{u1}^n& {\hat{G}}_{u1}^{n-1}& \dots & {\hat{G}}_{u1}\\ {\hat{G}}_{u2}^n& {\hat{G}}_{u2}^{n-1}& \dots & {\hat{G}}_{u2}\\ ⋮& ⋮& \ddots & ⋮\\ {\hat{G}}_{\mathrm{un}+1}^n& {\hat{G}}_{\mathrm{un}+1}^{n-1}& \dots & {\hat{G}}_{\mathrm{un}+1}\end{array}\right]}\underset{\underset{x}{}}{\left[\begin{array}{c}{\alpha }_n\\ {\alpha }_{n-1}\\ ⋮\\ {\mathrm{<figure-callout\; label="\alpha "\; filenames="US20140051953A1-20140220-D00009.png,US20140051953A1-20140220-D00010.png"\; state="\{\{state\}\}">\alpha </figure-callout>}}_1\\ \mathrm{offset}\end{array}\right]}& \left(10\right)\end{array}$
• A solution to this system of equations can be found as follows:
• 
  x=A⁻¹y  (11)
• Various estimation or approximation techniques could be used to avoid the computational inefficiency of calculating the inverse of A. In addition, other forms of regression or other approximation or estimation methods may be used to calibrate the noninvasive measurements.
• FIGS. 8 and 9 illustrate further embodiments ofglucose monitoring systems800,900 that implement a plurality of glucose algorithms. Referring toFIGS. 8 and 9 together, signal inputs are received by the filtering andprocessing module510, described above. The filtering andprocessing module510 provides outputs, such as ratios or the like, to a plurality ofglucose algorithms820,920. Each glucose algorithm outputs a glucose value Ĝ_ui. The plurality ofglucose algorithms820,920 can be considered parallel engines in certain embodiments.
• Some or all of theglucose algorithms820,920 can be adaptive glucose algorithms, though this need not be the case. At least some of thealgorithms820,920 can implement some or all of the features of theadaptive glucose algorithm620 described above with respect toFIG. 6. Some of theglucose algorithms820,920 could use basis functions to approximate a calibration curve, for instance. Different ones of thealgorithms820,920 could use different numbers or types of basis functions to provide different results. Advantageously, in certain embodiments, these results can be combined, compared, or otherwise selected to provide more accurate glucose values.
• Referring toFIG. 8, weights822 (w_i) can be applied to the glucose values from each algorithm and combined at acombiner824 to output an overall glucose value Ĝ_u. Theweights824 can average or otherwise blend the outputs of theglucose algorithms820. Although not shown, these weights may be adapted using any of the adaptive algorithms described above, e.g., based on alternative measurements. Fewer than all of the algorithms may be combined in some embodiments; a subset of thealgorithms920 may be selected to be combined for certain individuals.
• Referring toFIG. 9, the outputs of thealgorithms920 can be provided to aselector module924, which can select one or more of the outputs of thealgorithms920 to output as a final glucose value Ĝ_u. Theselector920 can selectdifferent algorithms920 for different types of patients (e.g., neonates versus adults). Alternatively, theselector920 can compare the outputs from theglucose algorithms920 to determine which, if any, of the algorithms provided outliers. Theselector924 could reject the outliers and combine (e.g., average) the outputs of the remainingalgorithms920.
• In yet another embodiment, theselector924 could determine which of the algorithm outputs are close to each other (e.g., within a tolerance) and output a combination of thoseoutputs920. For example, if there are fivealgorithms920 and three of the algorithms produce a similar output and two are outliers, theselector924 could average the three similar outputs or select one of the three outputs as the final glucose value. Moreover, theselector924 can learn over time and can select one ormore algorithms920 for a particular person based on past performance of that algorithm. Many other configurations and extensions of theselector924 are possible.
• FIGS. 10 and 11A illustrate embodiments ofprocesses1000,1100A for determining whether to suggest an alternative measurement of a physiological parameter. These processes can be implemented by theparameter calculator101 or by any of the other systems described herein. Advantageously, theprocesses1000,1100A can receive alternative and noninvasive measurements as inputs. Theprocesses1000,1100A can determine, based at least partly on the noninvasive measurements and/or based on an elapse of time, whether an alternative measurement should be taken. As a result, a user can be spared the discomfort of painful finger pricks or other invasive/minimally-invasive measurements while a noninvasive measurement is within a certain tolerance or while a certain amount of time has yet to elapse.
• Referring toFIG. 10, atblock1002 of theprocess1000, an alternative glucose measurement is obtained. The alternative glucose measurement can be obtained, for example, from an invasive or minimally-invasive device. Atblock1004, a first noninvasive glucose measurement is determined. This block can be implemented by a noninvasive glucose device.
• The alternative glucose measurement is output atblock1006, e.g., by the noninvasive device. Alternatively, the noninvasive glucose measurement can be output instead. Atblock1008, a new noninvasive glucose measurement is determined. If at decision block1010 a difference between the new and the first noninvasive measurements exceeds a threshold, an alarm or other indicator is output indicating or suggesting that a new alternative glucose measurement should be taken. Otherwise, theprocess1000 loops back to block1008, where a new noninvasive measurement is determined.
• Thus, theprocess1000 can loop fromblock1010 to block1008 until the noninvasive measurement changes enough to trigger an alarm or other indication. Advantageously, in certain embodiments, theprocess1000 can therefore allow a user to postpone painful finger prick or other invasive tests, while the noninvasive measurement is within a certain tolerance of a finger-prick measurement.
• In certain embodiments, theparameter calculator101 ofFIG. 1 performs theprocess1000 by continuously comparing the noninvasive measurement to a stored alternative measurement or measurements. If the alternative measurements come from a continuous glucose or other monitor, theparameter calculator101 can compare periodicity of the noninvasive and alternative measurements, peak value, low value, wave shape, or other factors. Additionally, in alternative embodiments, theparameter calculator101 can output the noninvasive value instead of or in addition to the alternative value.
• FIG. 11A illustrates anotherexample process1100A for determining whether to suggest an alternative measurement to be taken. As in theprocess1000, alternative and noninvasive glucose measurements are taken atblocks1102 and1104, and the alternative measurement is output atblock1106. Likewise, a new noninvasive glucose value is determined atblock1108.
• Atdecision block1110, it is determined whether a difference between new and first noninvasive values exceeds a threshold. If so, an alarm can be output atblock1112 indicating that a new alternative glucose measurement should be obtained, as in theprocess1000. If not, it is further determined atdecision block1114 whether an elapsed time has exceeded a threshold. If so, then the alarm is output atblock1112. Thus, in certain embodiments, an alarm is triggered by elapsed time, even when the noninvasive value has not changed beyond a threshold.
• In an alternative embodiment, instead of determining whether the noninvasive measurement differs more than a threshold, an elapse of time can be the main or only criteria used to determine whether to trigger an alarm. For example, a periodic time interval could be set for recommending an alternative measurement. Additionally, in alternative embodiments, the noninvasive value can be output instead of or in addition to the alternative value.
• FIG. 11B illustrates an embodiment of a process11008 for adjusting a glucose measurement. Test strips for finger-prick glucose meters can be sensitive to different concentrations of various hemoglobin species in an individual's blood. To compensate for these variations in hemoglobin species, in certain embodiments, the process11008 adjusts glucose measurements based on one or more hemoglobin measurements. These features can also be applied to adjust noninvasive glucose measurements. Theprocess1100B can be implemented by theparameter calculator101 or by another physiological monitor.
• Atblock1122, a glucose measurement is obtained. The glucose measurement can be invasive or noninvasive. Atblock1124, one or more hemoglobin measurements are obtained. The hemoglobin measurements may include any of total hemoglobin (Hbt), methemoglobin, carboxyhemoglobin, or other hemoglobin species. The noninvasive sensor described above with respect toFIG. 4 could obtain these measurements. A more detailed example of a sensor for obtaining these measurements is described in U.S. Publication No. 2006/0211924, filed Mar. 1, 2006, titled “Multiple Wavelength Sensor Emitters,” the disclosure of which is hereby incorporated by reference in its entirety.
• The glucose measurement is adjusted atblock1126 based at least partly on the one or more hemoglobin measurements. In certain embodiments, thisblock1126 can further include generating or modifying a calibration curve that relates glucose and one or more hemoglobin species. Like the calibration curve adjustments described above, an initial calibration curve relating glucose and hemoglobin species can be derived empirically. Then, atblock1126, the calibration curve can be adjusted based at least partly on the measurements of the hemoglobin species, thereby adjusting the glucose measurement. The adjustment of the glucose/hemoglobin calibration curve can be performed using any of the techniques described above, including adjusting weights of basis functions, using an adaptive algorithm, and so forth. Atblock1128, the adjusted glucose measurement is output. For example, the adjusted glucose measurement can be output on a display.
• In some embodiments, the process11006 can be used to adjust noninvasive glucose measurements based on hemoglobin species. Moreover, an invasive/minimally-invasive glucose measurement can be adjusted based on hemoglobin measurements, and the adjusted invasive/minimally-invasive measurement can be supplied to any of the adaptive glucose algorithms described above. In one embodiment, a test strip glucose reader can be made more accurate by accounting for changes in hemoglobin species. In addition, rather than displaying a noninvasive glucose measurement, an invasive or minimally-invasive glucose measurement adjusted for hemoglobin species can be displayed. Moreover, in yet another embodiment, the monitor includes a strip reader and a hemoglobin sensor but not a noninvasive glucose sensor. Many other configurations are possible.
• FIG. 12 illustrates an example of adata collection system1200. In certain embodiments, thedata collection system1200 noninvasively measures a blood analyte, such as oxygen, carbon monoxide, methemoglobin, total hemoglobin, glucose, proteins, glucose, lipids, a percentage thereof (e.g., saturation) or one or more other physiologically relevant patient characteristics. Thesystem1200 can also measure additional blood constituents or analytes and/or other physiological parameters useful in determining a state or trend of wellness of a patient.
• Thedata collection system1200 can be capable of measuring optical radiation from the measurement site. For example, in some embodiments, thedata collection system1200 can employ one or more photodiodes. In an embodiment, the photodiodes have an area from about 1 mm²-5 mm²(or higher) and are capable of detecting about 100 nanoamps (nA) or less of current resulting from measured light at full scale. In addition to having its ordinary meaning, the phrase “at full scale” can mean light saturation of a photodiode amplifier (not shown). Of course, other sizes and types of photodiodes can be used in various embodiments.
• Thedata collection system1200 can measure a range of approximately about 2 nA to about 100 nA or more full scale. Thedata collection system1200 can also include sensor front-ends that are capable of processing and amplifying current from the detector(s) at signal-to-noise ratios (SNRs) of about 100 decibels (dB) or more, such as about 120 dB in order to measure various desired analytes. Thedata collection system1200 can operate with a lower SNR if less accuracy is desired for an analyte like glucose.
• Thedata collection system1200 can measure analyte concentrations, including glucose, at least in part by detecting light attenuated by ameasurement site1202. Themeasurement site1202 can be any location on a patient's body, such as a finger, foot, ear lobe, or the like. For convenience, this disclosure is described primarily in the context of afinger measurement site1202. However, the features of the embodiments disclosed herein can be used withother measurement sites1202.
• In the depicted embodiment, thesystem1200 includes an optional tissue thickness adjuster ortissue shaper1205, which can include one or more protrusions, bumps, lenses, or other suitable tissue-shaping mechanisms. In certain embodiments, thetissue shaper1205 is a flat or substantially flat surface that can be positioned proximate themeasurement site1202 and that can apply sufficient pressure to cause the tissue of themeasurement site1202 to be flat or substantially flat. In other embodiments, thetissue shaper1205 is a convex or substantially convex surface with respect to themeasurement site1202. Many other configurations of thetissue shaper1205 are possible. Advantageously, in certain embodiments, thetissue shaper1205 reduces thickness of themeasurement site1202 while preventing or reducing occlusion at themeasurement site1202. Reducing thickness of the site can advantageously reduce the amount of attenuation of the light because there is less tissue through which the light must travel. Shaping the tissue in to a convex (or alternatively concave) surface can also provide more surface area from which light can be detected.
• The embodiment of thedata collection system1200 shown also includes anoptional noise shield1203. In an embodiment, thenoise shield1203 can be advantageously adapted to reduce electromagnetic noise while increasing the transmittance of light from themeasurement site1202 to one or more detectors1206 (described below). For example, thenoise shield1203 can advantageously include one or more layers of conductive coated glass or a metal grid electrically communicating with one or more other shields of thesensor1201 or electrically grounded. In an embodiment where thenoise shield1203 includes conductive coated glass, the coating can advantageously include indium tin oxide. In an embodiment, the indium tin oxide includes a surface resistivity ranging from approximately 30 ohms per square inch to about 500 ohms per square inch. In an embodiment, the resistivity is approximately 30, 200, or 500 ohms per square inch. Other resistivities can also be used which are less than about 30 ohms or more than about 500 ohms. Other conductive materials that are transparent or substantially transparent to light can be used instead.
• In some embodiments, themeasurement site1202 is located somewhere along a non-dominant arm or a non-dominant hand, e.g., a right-handed person's left arm or left hand. In some patients, the non-dominant arm or hand can have less musculature and higher fat content, which can result in less water content in that tissue of the patient. Tissue having less water content can provide less interference with the particular wavelengths that are absorbed in a useful manner by blood analytes like glucose. Accordingly, in some embodiments, thedata collection system1200 can be used on a person's non-dominant hand or arm.
• Thedata collection system1200 can include a sensor1201 (or multiple sensors) that is coupled to a processing device orphysiological monitor1209. In an embodiment, thesensor1201 and themonitor1209 are integrated together into a single unit. In another embodiment, thesensor1201 and themonitor1209 are separate from each other and communicate one with another in any suitable manner, such as via a wired or wireless connection. Thesensor1201 and monitor1209 can be attachable and detachable from each other for the convenience of the user or caregiver, for ease of storage, sterility issues, or the like. Thesensor1201 and themonitor1209 will now be further described.
• In the depicted embodiment shown inFIG. 12, thesensor1201 includes anemitter1204, anoptional tissue shaper1205, a set ofdetectors1206, and a front-end interface1208. Theemitter1204 can serve as the source of optical radiation transmitted towardsmeasurement site1202. As will be described in further detail below, theemitter1204 can include one or more sources of optical radiation, such as LEDs, laser diodes, incandescent bulbs with appropriate frequency-selective filters, combinations of the same, or the like. In an embodiment, theemitter1204 includes sets of optical sources that are capable of emitting visible and near-infrared optical radiation.
• In some embodiments, theemitter1204 is used as a point optical source, and thus, the one or more optical sources of theemitter1204 can be located within a close distance to each other, such as within about a 2 mm to about 4 mm. Theemitters1204 can be arranged in an array, such as is described in U.S. Publication No. 2006/0211924, filed Sep. 21, 2006, titled “Multiple Wavelength Sensor Emitters,” the disclosure of which is hereby incorporated by reference in its entirety. In particular, theemitters1204 can be arranged at least in part as described in paragraphs [0061] through [0068] of the aforementioned publication, which paragraphs are hereby incorporated specifically by reference. Other relative spatial relationships can be used to arrange theemitters1204.
• For analytes like glucose, currently available non-invasive techniques often attempt to employ light near the water absorbance minima at or about 1600 nm. Typically, these devices and methods employ a single wavelength or single band of wavelengths at or about 1600 nm. However, to date, these techniques have been unable to adequately consistently measure analytes like glucose based on spectroscopy.
• In contrast, theemitter1204 of thedata collection system1200 can emit, in certain embodiments, combinations of optical radiation in various bands of interest. For example, in some embodiments, for analytes like glucose, theemitter1204 can emit optical radiation at three (3) or more wavelengths between about 1600 nm to about 1700 nm. In particular, theemitter1204 can emit optical radiation at or about 1610 nm, about 1640 nm, and about 1665 nm. In some circumstances, the use of three wavelengths within about 1600 nm to about 1700 nm enable sufficient SNRs of about 100 dB, which can result in a measurement accuracy of about 20 mg/dL or better for analytes like glucose.
• In other embodiments, theemitter1204 can use two (2) wavelengths within about 1600 nm to about 1700 nm to advantageously enable SNRs of about 85 dB, which can result in a measurement accuracy of about 25-30 mg/dL or better for analytes like glucose. Furthermore, in some embodiments, theemitter1204 can emit light at wavelengths above about 1670 nm. Measurements at these wavelengths can be advantageously used to compensate or confirm the contribution of protein, water, and other non-hemoglobin species exhibited in measurements for analytes like glucose conducted between about 1600 nm and about 1700 nm. Of course, other wavelengths and combinations of wavelengths can be used to measure analytes and/or to distinguish other types of tissue, fluids, tissue properties, fluid properties, combinations of the same or the like.
• For example, theemitter1204 can emit optical radiation across other spectra for other analytes. In particular, theemitter1204 can employ light wavelengths to measure various blood analytes or percentages (e.g., saturation) thereof. For example, in one embodiment, theemitter1204 can emit optical radiation in the form of pulses at wavelengths of about 905 nm, about 1050 nm, about 1200 nm, about 1300 nm, about 1330 nm, about 1610 nm, about 1640 nm, and/or about 1665 nm. In another embodiment, theemitter1204 can emit optical radiation ranging from about 860 nm to about 950 nm, about 950 nm to about 1100 nm, about 1100 nm to about 1270 nm, about 1250 nm to about 1350 nm, about 1300 nm to about 1360 nm, and/or about 1590 nm to about 1700 nm. Of course, theemitter1204 can transmit any of a variety of wavelengths of visible or near-infrared optical radiation.
• Due to the different responses of analytes to the different wavelengths, certain embodiments of thedata collection system1200 can advantageously use the measurements at these different wavelengths to improve the accuracy of measurements. For example, the measurements of water from visible and infrared light can be used to compensate for water absorbance that is exhibited in the near-infrared wavelengths.
• As briefly described above, theemitter1204 can include sets of light-emitting diodes (LEDs) as its optical source. Theemitter1204 can use one or more top-emitting LEDs. In particular, in some embodiments, theemitter1204 can include top-emitting LEDs emitting light at about 850 nm to 1350 nm.
• Theemitter1204 can also use super luminescent LEDs (SLEDs) or side-emitting LEDs. In some embodiments, theemitter1204 can employ SLEDs or side-emitting LEDs to emit optical radiation at about 1600 nm to about 1800 nm.Emitter1204 can use SLEDs or side-emitting LEDs to transmit near infrared optical radiation because these types of sources can transmit at high power or relatively high power, e.g., about 40 mW to about 100 mW. This higher power capability can be useful to compensate or overcome the greater attenuation of these wavelengths of light in tissue and water. For example, the higher power emission can effectively compensate and/or normalize the absorption signal for light in the mentioned wavelengths to be similar in amplitude and/or effect as other wavelengths that can be detected by one or more photodetectors after absorption. However, certain the embodiments do not necessarily require the use of high power optical sources. For example, some embodiments may be configured to measure analytes, such as total hemoglobin (tHb), oxygen saturation (SpO₂), carboxyhemoglobin, methemoglobin, etc., without the use of high power optical sources like side emitting LEDs. Instead, such embodiments may employ other types of optical sources, such as top emitting LEDs. Alternatively, theemitter1204 can use other types of sources of optical radiation, such as a laser diode, to emit near-infrared light into themeasurement site1202.
• In addition, in some embodiments, in order to assist in achieving a comparative balance of desired power output between the LEDs, some of the LEDs in theemitter1204 can have a filter or covering that reduces and/or cleans the optical radiation from particular LEDs or groups of LEDs. For example, since some wavelengths of light can penetrate through tissue relatively well, LEDs, such as some or all of the top-emitting LEDs can use a filter or covering, such as a cap or painted dye. This can be useful in allowing theemitter1204 to use LEDs with a higher output and/or to equalize intensity of LEDs.
• Thedata collection system1200 also includes adriver1211 that drives theemitter1204. Thedriver1211 can be a circuit or the like that is controlled by themonitor1209. For example, thedriver1211 can provide pulses of current to theemitter1204. In an embodiment, thedriver1211 drives theemitter1204 in a progressive fashion, such as in an alternating manner. Thedriver1211 can drive theemitter1204 with a series of pulses of about 1 milliwatt (mW) for some wavelengths that can penetrate tissue relatively well and from about 40 mW to about 100 mW for other wavelengths that tend to be significantly absorbed in tissue. A wide variety of other driving powers and driving methodologies can be used in various embodiments.
• Thedriver1211 can be synchronized with other parts of thesensor1201 and can minimize or reduce jitter in the timing of pulses of optical radiation emitted from theemitter1204. In some embodiments, thedriver1211 is capable of driving theemitter1204 to emit optical radiation in a pattern that varies by less than about 10 parts-per-million.
• Thedetectors1206 capture and measure light from themeasurement site1202. For example, thedetectors1206 can capture and measure light transmitted from theemitter1204 that has been attenuated or reflected from the tissue in themeasurement site1202. Thedetectors1206 can output adetector signal1207 responsive to the light captured or measured. Thedetectors1206 can be implemented using one or more photodiodes, phototransistors, or the like.
• In addition, thedetectors1206 can be arranged with a spatial configuration to provide a variation of path lengths among at least some of thedetectors1206. That is, some of thedetectors1206 can have the substantially, or from the perspective of the processing algorithm, effectively, the same path length from theemitter1204. However, according to an embodiment, at least some of thedetectors1206 can have a different path length from theemitter1204 relative to other of thedetectors1206. Variations in path lengths can be helpful in allowing the use of a bulk signal stream from thedetectors1206. In some embodiments, thedetectors1206 may employ a linear spacing, a logarithmic spacing, or a two or three dimensional matrix of spacing, or any other spacing scheme in order to provide an appropriate variation in path lengths.
• Thefront end interface1208 provides an interface that adapts the output of thedetectors1206, which is responsive to desired physiological parameters. For example, thefront end interface1208 can adapt asignal1207 received from one or more of thedetectors1206 into a form that can be processed by themonitor1209, for example, by asignal processor1210 in themonitor1209. Thefront end interface1208 can have its components assembled in thesensor1201, in themonitor1209, in connecting cabling (if used), combinations of the same, or the like. The location of thefront end interface1208 can be chosen based on various factors including space desired for components, desired noise reductions or limits, desired heat reductions or limits, and the like.
• Thefront end interface1208 can be coupled to thedetectors1206 and to thesignal processor1210 using a bus, wire, electrical or optical cable, flex circuit, or some other form of signal connection. Thefront end interface1208 can also be at least partially integrated with various components, such as thedetectors1206. For example, thefront end interface1208 can include one or more integrated circuits that are on the same circuit board as thedetectors1206. Other configurations can also be used.
• Thefront end interface1208 can be implemented using one or more amplifiers, such as transimpedance amplifiers, that are coupled to one or more analog to digital converters (ADCs) (which can be in the monitor1209), such as a sigma-delta ADC. A transimpedance-basedfront end interface1208 can employ single-ended circuitry, differential circuitry, and/or a hybrid configuration. A transimpedance-basedfront end interface1208 can be useful for its sampling rate capability and freedom in modulation/demodulation algorithms. For example, this type offront end interface1208 can advantageously facilitate the sampling of the ADCs being synchronized with the pulses emitted from theemitter1204.
• The ADC or ADCs can provide one or more outputs into multiple channels of digital information for processing by thesignal processor1210 of themonitor1209. Each channel can correspond to a signal output from adetector1206.
• In some embodiments, a programmable gain amplifier (PGA) can be used in combination with a transimpedance-basedfront end interface1208. For example, the output of a transimpedance-basedfront end interface1208 can be output to a PGA that is coupled with an ADC in themonitor1209. A PGA can be useful in order to provide another level of amplification and control of the stream of signals from thedetectors1206. Alternatively, the PGA and ADC components can be integrated with the transimpedance-basedfront end interface1208 in thesensor1201.
• In another embodiment, thefront end interface1208 can be implemented using switched-capacitor circuits. A switched-capacitor-basedfront end interface1208 can be useful for, in certain embodiments, its resistor-free design and analog averaging properties. In addition, a switched-capacitor-basedfront end interface1208 can be useful because it can provide a digital signal to thesignal processor1210 in themonitor1209.
• As shown inFIG. 12, themonitor1209 can include thesignal processor1210 and a user interface, such as adisplay1212. Themonitor1209 can also include optional outputs alone or in combination with thedisplay1212, such as astorage device1214 and anetwork interface1216. In an embodiment, thesignal processor1210 includes processing logic that determines measurements for desired analytes, such as glucose, based on the signals received from thedetectors1206. Thesignal processor1210 can be implemented using one or more microprocessors or subprocessors (e.g., cores), digital signal processors, application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), combinations of the same, and the like.
• Thesignal processor1210 can provide various signals that control the operation of thesensor1201. For example, thesignal processor1210 can provide an emitter control signal to thedriver1211. This control signal can be useful in order to synchronize, minimize, or reduce jitter in the timing of pulses emitted from theemitter1204. Accordingly, this control signal can be useful in order to cause optical radiation pulses emitted from theemitter1204 to follow a precise timing and consistent pattern. For example, when a transimpedance-basedfront end interface1208 is used, the control signal from thesignal processor1210 can provide synchronization with the ADC in order to avoid aliasing, cross-talk, and the like. As also shown, anoptional memory1213 can be included in the front-end interface1208 and/or in thesignal processor1210. Thismemory1213 can serve as a buffer or storage location for the front-end interface1208 and/or thesignal processor1210, among other uses.
• Theuser interface1212 can provide an output, e.g., on a display, for presentation to a user of thedata collection system1200. Theuser interface1212 can be implemented as a touch-screen display, an LCD display, an organic LED display, or the like. In addition, theuser interface1212 can be manipulated to allow for measurement on the non-dominant side of patient. For example, theuser interface1212 can include a flip screen, a screen that can be moved from one side to another on themonitor1209, or can include an ability to reorient its display indicia responsive to user input or device orientation. In alternative embodiments, thedata collection system1200 can be provided without auser interface1212 and can simply provide an output signal to a separate display or system.
• Astorage device1214 and anetwork interface1216 represent other optional output connections that can be included in themonitor1209. Thestorage device1214 can include any computer-readable medium, such as a memory device, hard disk storage, EEPROM, flash drive, or the like. The various software and/or firmware applications can be stored in thestorage device1214, which can be executed by thesignal processor1210 or another processor of themonitor1209. Thenetwork interface1216 can be a serial bus port (RS-232/RS-485), a Universal Serial Bus (USB) port, an Ethernet port, a wireless interface (e.g., WiFi such as any 802.1x interface, including an internal wireless card), or other suitable communication device(s) that allows themonitor1209 to communicate and share data with other devices. Themonitor1209 can also include various other components not shown, such as a microprocessor, graphics processor, or controller to output theuser interface1212, to control data communications, to compute data trending, or to perform other operations.
• Although not shown in the depicted embodiment, thedata collection system1200 can include various other components or can be configured in different ways. For example, thesensor1201 can have both theemitter1204 anddetectors1206 on the same side of themeasurement site1202 and use reflectance to measure analytes. Thedata collection system1200 can also include a sensor that measures the power of light emitted from theemitter1204.
• Depending on the embodiment, certain acts, events, or functions of any of the algorithms described herein can be performed in a different sequence, can be added, merged, or left out all together (e.g., not all described acts or events are necessary for the practice of the algorithm). Moreover, in certain embodiments, acts or events can be performed concurrently, e.g., through multi-threaded processing, interrupt processing, or multiple processors or processor cores, rather than sequentially.
• The various illustrative logical blocks, modules, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, modules, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. The described functionality can be implemented in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the disclosure.
• The various illustrative logical blocks and modules described in connection with the embodiments disclosed herein can be implemented or performed by a machine, such as a general purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. A general purpose processor can be a microprocessor, but in the alternative, the processor can be a processor, controller, microcontroller, or state machine, combinations of the same, or the like. A processor can also be implemented as a combination of computing devices, e.g., a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration.
• The steps of a method or algorithm described in connection with the embodiments disclosed herein can be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module can reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, a removable disk, a CD-ROM, or any other form of processor-readable or computer-readable storage medium known in the art. An exemplary storage medium can be coupled to the processor such that the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium can be integral to the processor. The processor and the storage medium can reside in an ASIC. The ASIC can reside in a user terminal. In the alternative, the processor and the storage medium can reside as discrete components in a user terminal.
• Conditional language used herein, such as, among others, “can,” “could,” “might,” “may,” “e.g.,” and the like, unless specifically stated otherwise, or otherwise understood within the context as used, is generally intended to convey that certain embodiments include, while other embodiments do not include, certain features, elements and/or states. Thus, such conditional language is not generally intended to imply that features, elements and/or states are in any way required for one or more embodiments or that one or more embodiments necessarily include logic for deciding, with or without author input or prompting, whether these features, elements and/or states are included or are to be performed in any particular embodiment.
• While the above detailed description has shown, described, and pointed out novel features as applied to various embodiments, it will be understood that various omissions, substitutions, and changes in the form and details of the devices or algorithms illustrated can be made without departing from the spirit of the disclosure. As will be recognized, certain embodiments of the inventions described herein can be embodied within a form that does not provide all of the features and benefits set forth herein, as some features can be used or practiced separately from others. The scope of certain inventions disclosed herein is indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.

Claims (8)

What is claimed is:

1. A method of determining whether to recommend an alternative measurement of a physiological parameter, the method comprising:

obtaining a noninvasive measurement of a physiological parameter using an optical sensor;

receiving an alternative measurement of the physiological parameter, the alternative measurement being generated by an alternative sensor;

analyzing the noninvasive and alternative measurements by a processor to determine whether a condition has been met; and

in response to the condition being met, outputting an indication that a new measurement should be obtained from the alternative sensor.

2. The method ofclaim 1, wherein analyzing the noninvasive and alternative measurements to determine whether a condition has been met comprises determining whether a difference between the noninvasive and alternative measurements exceeds a threshold.

3. The method ofclaim 1, wherein analyzing the noninvasive and alternative measurements to determine whether a condition has been met comprises determining a predetermined amount of time has passed.

4. The method ofclaim 1, wherein the indication comprises an alarm.

5. The method ofclaim 1, further comprising displaying the alternative measurement.

6. The method ofclaim 1, wherein the alternative source comprises an invasive monitor.

7. The method ofclaim 1, wherein the alternative source comprises a minimally-invasive monitor.

8. The method ofclaim 1, wherein the physiological parameter is glucose.

Images