US 20050027181A1 (19) United States (12) Patent Application Publication (10) Pub. No.: US 2005/0027181 A1 Goode, JR. et al. (43) Pub. Date: Feb. 3, 2005

(54) SYSTEM AND METHODS FOR PROCESSING (52) U.S. Cl...... 600/365; 128/920; 600/309 ANALYTE SENSOR DATA (76) Inventors: Paul V. Goode JR., Murrieta, CA (US); (57) ABSTRACT James H. Brauker, San Diego, CA (US); Apurv U. Kamath, San Diego, Systems and methods for processing Sensor analyte data, CA (US) including initiating calibration, updating calibration, evalu Correspondence Address: ating clinical acceptability of reference and Sensor analyte KNOBBE MARTENS OLSON & BEAR LLP data, and evaluating the quality of Sensor calibration. During 2040 MAIN STREET initial calibration, the analyte Sensor data is evaluated over FOURTEENTH FLOOR a period of time to determine stability of the sensor. The Sensor may be calibrated using a calibration Set of one or IRVINE, CA 92614 (US) more matched Sensor and reference analyte data pairs. The (21) Appl. No.: 10/633,404 calibration may be updated after evaluating the calibration set for best calibration based on inclusion criteria with newly (22) Filed: Aug. 1, 2003 received reference analyte data. Fail-safe mechanisms are provided based on clinical acceptability of reference and Publication Classification analyte data and quality of Sensor calibration. Algorithms provide for optimized prospective and retrospective analysis (51) Int. Cl...... A61B 5/00 of estimated blood analyte data from an analyte Sensor.

10 19 - 23 32(s 17

16

14 12 Patent Application Publication Feb. 3, 2005 Sheet 1 of 13 US 2005/0027181 A1

FIG. 1 Patent Application Publication Feb. 3, 2005 Sheet 2 of 13 US 2005/0027181 A1

20

22 POTENTIOSTAT

21 AVD CONVERTER

24

EEPROM

25 BATTERY RF TRANSCEIVER 26 CRYSTAL 28 CRYSTAL

FIG. 2 Patent Application Publication Feb. 3, 2005 Sheet 3 of 13 US 2005/0027181 A1

8000

7500 P s 7000 al SENSOR i DATA 6500 (counts) 6000 s i SN; : i.r i 5500

2300 2400 2500 2600 TIME (minutes)

FIG. 3 Patent Application Publication Feb. 3, 2005 Sheet 4 of 13 US 2005/0027181 A1

11:05AM 12:05 PM 282

F.G. 4B Patent Application Publication Feb. 3, 2005 Sheet 5 of 13 US 2005/0027181 A1

40

F.G. 4D Patent Application Publication Feb. 3, 2005 Sheet 6 of 13 US 2005/0027181 A1

50 53 CRYSTAL

51 52 54 RF TRANSCEIVER SRAM

MICROPROCESSOR

55 58

56 59 CRYSTAL COMPORT USER INTERFACE (REFERENCE ANALYTE KEYBOARD MONITOR)

57 SPEAKER

WIBRATOR

BACKLIGHT

CD

BUTTONS

FIG. 5 Patent Application Publication Feb. 3, 2005 Sheet 7 of 13 US 2005/0027181 A1

INITIAL CALIBRATION

61 RECEIVE SENSOR DATA FROM CONTINUOUSANALYTE SENSOR 62 RECEIVE REFERENCE DATA FROM REFERENCE ANALYTE SOURCE 63

MATCH TIME CORRESPONDING SENSOR DATA AND REFERENCE DATA TO PROVIDE MATCHED DATA PARS

FORMA CALIBRATION SET FROM MATCHED PAIR(S)

EVALUATE STABILITY OF CONTINUOUSANALYTE SENSOR OVER A PERIOD OF TIME

66

YES 67 CALCULATE CONVERSION FUNCTION USING CALIBRATION SET

CONVERT SENSOR DATA USNG CONVERSION FUNCTION FIG. 6

PROVIDE OUTPUT TO USER Patent Application Publication Feb. 3, 2005 Sheet 8 of 13 US 2005/0027181 A1

RAW SENSOR DATA (counts)

REFERENCE ANALYTE VALUE (mg/dL)

FG. 7 Patent Application Publication Feb. 3, 2005 Sheet 9 of 13 US 2005/0027181 A1

EVALUATE CLINICAL ACCEPTABILITY OF REFERENCE AND SENSOR DATA

RECEIVE REFERENCE DATA FROM REFERENCE ANALYTE SOURCE

EVALUATE CLNICAL ACCEPTABILITY OF REFERENCE DATA TO SUBSTANTIALLY TIME CORRESPONDING SENSOR DATA

CLINICALLY ACCEPTABLET)

OPTIONALLY RE-CALCULATE NO CONVERSION FUNCTION USING NEW REFERENCE DATA (FIG. 10)

CONTINUOUSLY CONVERT SENSOR DATA USING CONVERSION FUNCTION

CONTROL USER INTERFACE BASED ON CLNICAL ACCEPTABILITY

FIG. 8

Patent Application Publication Feb. 3, 2005 Sheet 11 of 13 US 2005/0027181 A1

VALUATE CALIBRATION SET FOR BEST CALIBRATION BASED ON INCLUSION CRITERA

RECEIVE NEW REFERENCE DATA FROM REFERENCE ANALYTE SOURCE

MATCH TIME CORRESPONDING SENSOR DATA TO UPDATED REFERENCE DATA TO PROVIDE NEWMATCHED DATA PAR

EVALUATE CALIBRATION SET WITH NO NEWMATCHED DATA PAR BASED ON INCLUSION CRITERA

UPDATE CALIBRATION SET2

YES

UPDATE CONVERSION FUNCTION USING NEW CALIBRATION SET BASED ON RESULTS OF EVALUATION

CONVERT SENSOR DATAUSING UPDATED CONVERSION FUNCTION

FIG. 10 Patent Application Publication Feb. 3, 2005 Sheet 12 of 13 US 2005/0027181 A1

EVALUATE OUALITY OF CALIBRATION SET DATA ASSOCATION

111 RECEIVE SENSOR DATA FROM CONTINUOUS SENSOR 112 RECEIVE REFERENCE DATA FROM REFERENCE GLUCOSE SOURCE 113

MATCH TIME CORRESPONDING SENSOR DATA AND REFERENCE DATA TO PROVIDE MATCHED DATA PAIRS

114 FORM CALIBRATION SET USING MATCHED DATA PAIR(S)

115 CALCULATE CONVERSION FUNCTION USING CALIBRATION SET

116 CONVERT SENSOR DATA USING CONVERSION FUNCTION

117 EVALUATE THE OUALITY OF CALIBRATION SET DATA ASSOCATION

118 CONTROL USER INTERFACE BASED ON OUALITY OF DATA ASSOCATION

FIG. 11 Patent Application Publication Feb. 3, 2005 Sheet 13 of 13 US 2005/0027181 A1

2500

2000 SENSOR DATA 1500 (Counts) 1OOO O

500

O O 50 1OO 150 200 250 300 350 REFERENCE DATA (mg/dL) FIG. 12A

2500

2000

1500 SENSOR DATA (Counts) 1000

O 50 100 150 200 250 300 350 REFERENCE DATA (mg/dL)

FIG. 12B US 2005/0027181 A1 Feb. 3, 2005

SYSTEMAND METHODS FOR PROCESSING 0007. In a first embodiment a method is provided for ANALYTE SENSOR DATA initializing a Substantially continuous analyte Sensor, the method including: receiving a data Stream from an analyte FIELD OF THE INVENTION Sensor, including one or more Sensor data points, receiving reference data from a reference analyte monitor, including 0001. The present invention relates generally to systems two or more reference data points, providing at least two and methods for analyte Sensor data processing. Particularly, matched data pairs by matching reference analyte data to the present invention relates to retrospectively and/or pro Substantially time corresponding Sensor data; forming a Spectively initiating a calibration, converting Sensor data, calibration Set including the at least two matching data pairs, updating the calibration, evaluating received reference and and determining a Stability of the continuous analyte Sensor. Sensor data, and evaluating the calibration for the analyte SCSO. 0008. In an aspect of the first embodiment, the step of determining the Stability of the Substantially continuous BACKGROUND OF THE INVENTION analyte Sensor includes waiting a predetermined time period between about one minute and about Six weeks. 0002 Diabetes mellitus is a disorder in which the pan creas cannot create Sufficient insulin (Type I or insulin 0009. In an aspect of the first embodiment, the step of dependent) and/or in which insulin is not effective (Type 2 determining the Stability of the Substantially continuous or non-insulin dependent). In the diabetic State, the victim analyte Sensor includes evaluating at least two matched data Suffers from high blood Sugar, which may cause an array of pairs. physiological derangements (e.g., kidney failure, skin ulcers, or bleeding into the vitreous of the eye) associated 0010. In an aspect of the first embodiment, the step of with the deterioration of Small blood vessels. A hypoglyce determining the Stability of the Substantially continuous mic reaction (low blood Sugar) may be induced by an analyte Sensor includes evaluating one of pH, oxygen, inadvertent overdose of insulin, or after a normal dose of hypochlorite, interfering Species, correlation of matched insulin or glucose-lowering agent accompanied by extraor pairs, R-value, baseline drift, baseline offset, and amplitude. dinary exercise or insufficient food intake. 0011. In an aspect of the first embodiment, the method 0.003 Conventionally, a diabetic person carries a self further includes providing one of an audible, Visual, or monitoring blood glucose (SMBG) monitor, which typically tactile output to a user based on the Stability of the Sensor. comprises uncomfortable finger pricking methods. Due to 0012. In an aspect of the first embodiment, the step of the lack of comfort and convenience, a diabetic will nor providing output based on the Stability of the Sensor includes mally only measure his or her glucose level two to four times indicating at least one of a numeric estimated analyte value, per day. Unfortunately, these time intervals are So far spread a directional trend of analyte concentration, and a graphical apart that the diabetic will likely find out too late, Sometimes representation of an estimated analyte value. incurring dangerous Side effects, of a hyper- or hypo glycemic condition. In fact, it is not only unlikely that a 0013 In an aspect of the first embodiment, the step of diabetic will take a timely SMBG value, but the diabetic will receiving Sensor data includes receiving Sensor data from a not know if their blood glucose value is going up (higher) or Substantially continuous glucose Sensor. down (lower) based on conventional methods, inhibiting their ability to make educated insulin therapy decisions. 0014. In an aspect of the first embodiment, the step of receiving Sensor data includes receiving Sensor data from an SUMMARY OF THE INVENTION implantable glucose Sensor. 0004 Systems and methods are needed that accurately 0015. In an aspect of the first embodiment, the step of provide estimated glucose measurements to a diabetic receiving Sensor data includes receiving Sensor data from patient continuously and/or in real time So that they may Subcutaneously implantable glucose Sensor. proactively care for their condition to safely avoid hyper 0016. In an aspect of the first embodiment, the step of and hypo-glycemic conditions. Real time and retrospective receiving reference data includes receiving reference data estimated glucose measurements require reliable data pro from a Self-monitoring blood glucose test. cessing in order to provide accurate and useful output to a patient and/or doctor. 0017. In an aspect of the first embodiment, the step of 0005 Similarly, systems and methods are needed that receiving reference data includes downloading reference accurately provide Substantially continuous estimated ana data via a cabled connection. lyte measurements for a variety of known analytes (e.g., 0018. In an aspect of the first embodiment, the step of oxygen, Salts, protein, and Vitamins) to provide prospective receiving reference data includes downloading reference and/or retrospective data analysis and output to a user. data via a wireleSS connection. 0006 Accordingly, systems and methods are provided for 0019. In an aspect of the first embodiment, the step of retrospectively and/or prospectively calibrating a Sensor, receiving reference data from a reference analyte monitor initializing a Sensor, converting Sensor data into calibrated includes receiving within a receiver internal communication data, updating and maintaining a calibration over time, from a reference analyte monitor integral with the receiver. evaluating received reference and Sensor data for clinical acceptability, and evaluating the calibration Statistical 0020. In an aspect of the first embodiment, the step of acceptability, to ensure accurate and Safe data output to a forming a calibration Set includes evaluating at least one patient and/or doctor. matched data pair using inclusion criteria. US 2005/0027181 A1 Feb. 3, 2005

0021. In an aspect of the first embodiment, the step of 0035) In an aspect of the second embodiment, the sensor receiving Sensor data includes receiving Sensor data that has data module is configured to receive Sensor data from an been algorithmically Smoothed. implantable glucose Sensor. 0022. In an aspect of the first embodiment, the step of 0036). In an aspect of the second embodiment, the sensor receiving Sensor data includes algorithmically Smoothing data module is configured to receive Sensor data from the received Sensor data. Subcutaneously implantable glucose Sensor. 0023. In an aspect of the first embodiment, the step of 0037. In an aspect of the second embodiment, the refer forming a calibration Set includes including in the calibra ence input module is configured to receive reference data tion Set between one and Six matched data pairs. from a Self-monitoring blood glucose test. 0024. In an aspect of the first embodiment, the step of 0038. In an aspect of the second embodiment, the refer forming a calibration Set includes including six matched data ence input module is configured to download reference data pairs. via a cabled connection. 0.025 In an aspect of the first embodiment, the step of 0039. In an aspect of the second embodiment, the refer forming a calibration Set further includes determining a ence input module is configured to download reference data value for n, where n is greater than one and represents the via a wireleSS connection. number of matched data pairs in the calibration Set. 0040. In an aspect of the second embodiment, the system 0026. In an aspect of the first embodiment, the step of further includes a reference analyte monitor integral with the determining a value for n is determined as a function of the System and wherein the reference input module is configured frequency of the received reference data points and Signal to receive an internal communication from the reference Strength over time. analyte monitor. 0027. In a second embodiment, a system is provided for initializing a continuous analyte Sensor, including: a Sensor 0041. In an aspect of the second embodiment, the pro data module operatively connected to a continuous analyte ceSSor module includes programming to evaluate at least one Sensor that receives a data Stream including a plurality of matched data pair using inclusion criteria. time Spaced Sensor data points from the analyte Sensor; a 0042. In an aspect of the second embodiment, the refer reference input module adapted to obtain reference data ence input module is configured to receive Sensor data that from a reference analyte monitor, including one or more has been algorithmically Smoothed. reference data points, a processor module that forms one or more matched data pairs by matching reference data to 0043. In an aspect of the second embodiment, the refer Substantially time corresponding Sensor data and Subse ence input module is configured to algorithmically Smooth quently forms a calibration Set including the one or more the received Sensor data. matched data pairs, and a start-up module associated with the processor module programmed to determine the Stability 0044) In an aspect of the second embodiment, the cali of the continuous analyte Sensor. bration Set includes between one and Six matched data pairs. 0028. In an aspect of the second embodiment, the sensor 0045. In an aspect of the second embodiment, the cali data module is adapted to wirelessly receive Sensor data bration Set includes Six matched data pairs. points from the Sensor. 0046. In an aspect of the second embodiment, the cali 0029. In an aspect of the second embodiment, the start-up bration Set includes n matched data pairs, where n is greater module is programmed to wait a predetermined time period than one. between six hours and Six weeks. 0047. In an aspect of the second embodiment, n is a 0.030. In an aspect of the second embodiment, the start-up function of the frequency of the received reference data module is programmed to evaluate at least two matched data points and Signal Strength over time. pairs. 0048. In a third embodiment, a computer system is pro 0031. In an aspect of the second embodiment, the start-up Vided for initializing a continuous analyte Sensor, the com module is programmed to evaluate one of pH, Oxygen, puter System including: a Sensor data receiving module that hypochlorite, interfering Species, correlation of matched receives Sensor data from the Substantially continuous ana pairs, R-value, baseline drift, baseline offset, and amplitude. lyte Sensor via a receiver, including one or more Sensor data points, a reference data receiving module that receives 0032. In an aspect of the second embodiment, the system reference data from a reference analyte monitor, including further includes an output control module associated with one or more reference data points, a data matching module the processor module and programmed to control output of that forms one or more matched data pairs by matching Sensor data. reference data to Substantially time corresponding Sensor 0033. In an aspect of the second embodiment, the output data; a calibration Set module that forms a calibration Set control module indicates at least one of a numeric estimated including at least one matched data pair; and a Stability analyte value, a directional trend of analyte concentration, determination module that determines the stability of the and a graphical representation of an estimated analyte value. continuous analyte Sensor. 0034. In an aspect of the second embodiment, the sensor 0049. In an aspect of the third embodiment, the stability data module is configured to receive Sensor data from determination module includes a System for waiting a pre Substantially the continuous glucose Sensor. determined time period. US 2005/0027181 A1 Feb. 3, 2005

0050. In an aspect of the third embodiment, the stability 0067. In an aspect of the third embodiment, n is a determination module evaluates at least two matched data function of the frequency of the received reference data pairs. points and Signal Strength over time. 0051. In an aspect of the third embodiment, the stability 0068. In a fourth embodiment, method is provided for determination module evaluates one of pH, Oxygen, initializing a Substantially continuous analyte Sensor, the hypochlorite, interfering Species, correlation of matched method including: receiving Sensor data from a Substantially pairs, R-value, baseline drift, baseline offset, and amplitude. continuous analyte Sensor, including one or more Sensor data points, receiving reference data from a reference analyte 0.052 In an aspect of the third embodiment, the computer monitor, including one or more reference data points; form System further includes an interface control module that ing one or more matched data pairs by matching reference provides output to the user based on the stability of the data to Substantially time corresponding Sensor data; form SCSO. ing a calibration Set including at least one matched data pair; 0053. In an aspect of the third embodiment, the output determining Stability of continuous analyte Sensor; and from the interface control module includes at least one of a outputting information reflective of the Sensor data once a numeric estimated analyte value, an indication of directional predetermined level of stability has been determined. trend of analyte concentration, and a graphical representa 0069. In a fifth embodiment, a system is provided for tion of an estimated analyte value. initializing a continuous analyte Sensor, including: a Sensor 0054. In an aspect of the third embodiment, the reference data module operatively linked to a continuous analyte data receiving module is adapted to receive Sensor data from Sensor and configured to receive one or more Sensor data a Substantially continuous glucose Sensor. points from the Sensor; a reference input module adapted to obtain one or more reference data points, and a processor 0055. In an aspect of the third embodiment, the reference module associated with the Sensor data module and the input data receiving module is adapted to receive Sensor data from module and programmed to match reference data points with an implantable glucose Sensor. time-matched Sensor data points to form a calibration Set including at least one matched data pair; and a start-up 0056. In an aspect of the third embodiment, the reference module associated with the processor module programmed data receiving module is adapted to receive Sensor data from to determine the Stability of the continuous analyte Sensor a Subcutaneously implantable glucose Sensor. and output information reflective of the Sensor data once a 0057. In an aspect of the third embodiment, the reference predetermined level of stability has been determined. data receiving module is adapted to receive Sensor data from 0070. In a sixth embodiment, a computer system is pro a Self-monitoring blood glucose test. Vided for initializing a continuous analyte Sensor, the System 0.058. In an aspect of the third embodiment, the reference including: a Sensor data receiving module that receives data receiving module is adapted to receive Sensor data from Sensor data including one or more Sensor data points from a cabled connection. the Substantially continuous analyte Sensor via a receiver; a reference data receiving module for receiving reference data 0059. In an aspect of the third embodiment, the reference from a reference analyte monitor, including one or more data receiving module is adapted to download reference data reference data points, a data matching module for forming via a wireleSS connection. one or more matched data pairs by matching reference data 0060. In an aspect of the third embodiment, the reference to Substantially time corresponding Sensor data; a calibration data receiving module is adapted to receive reference data Set module for forming a calibration Set including at least from an internal reference analyte monitor that is housed one matched data pair; a Stability determination module for integrally the computer System. evaluating the Stability of the continuous analyte Sensor; and an interface control module that outputs information reflec 0061. In an aspect of the third embodiment, the calibra tive of the Sensor data once a predetermined level of Stability tion Set module evaluates at least one matched data pair has been determined. using inclusion criteria. 0071. In a seventh embodiment, a method for initializing 0.062. In an aspect of the third embodiment, the sensor a glucose Sensor, the method including: receiving Sensor data receiving module is adapted to receive Sensor data that data from the glucose Sensor, including one or more Sensor has been algorithmically Smoothed. data points, receiving reference data from a reference glu cose monitor, including one or more reference data points; 0.063. In an aspect of the third embodiment, the computer forming one or more matched data pairs by matching System further includes a data Smoothing module that reference data to Substantially time corresponding Sensor Smoothes the received Sensor data. data; determining whether the glucose Sensor has reached a 0064. In an aspect of the third embodiment, the calibra predetermined level of stability. tion Set module includes between one and Six matched data 0072. In an eighth embodiment, a system is provided for pairs. initializing a continuous analyte Sensor, including: a Sensor data module operatively linked to a continuous analyte 0065. In an aspect of the third embodiment, the calibra Sensor and configured to receive one or more Sensor data tion Set module includes Six matched data pairs. points from the Sensor; a reference input module adapted to 0.066. In an aspect of the third embodiment, the calibra obtain one or more reference data points, and a processor tion Set includes n number of matched data pairs, where n is module associated with the Sensor data module and the input greater than one. module and programmed to match reference data points with US 2005/0027181 A1 Feb. 3, 2005

time-matched Sensor data points to form a calibration Set from its Substantially time corresponding reference or Sensor including at least one matched data pair; and a Stability data and clinical risk associated with that deviation based on module associated with the processor module programmed the glucose value indicated by at least one of the Sensor and to determine the Stability of the continuous analyte Sensor. reference data. 0073. In a ninth embodiment, a method is provided for 0083. In an aspect of the tenth embodiment, the system evaluating clinical acceptability of at least one of reference further includes means for providing an output based and Sensor analyte data, the method including: receiving a through a user interface responsive to the clinical accept data Stream from an analyte Sensor, including one or more ability evaluation. Sensor data points, receiving reference data from a reference analyte monitor, including one or more reference data 0084. In an aspect of the tenth embodiment, the means for points, and evaluating the clinical acceptability at least one providing an output includes means for alerting the user of the reference and Sensor analyte data using Substantially based on the clinical acceptability evaluation. time corresponding reference or Sensor data, wherein the at 0085. In an aspect of the tenth embodiment, the means for least one of the reference and Sensor analyte data is evalu providing an output includes means for altering the user ated for deviation from its Substantially time corresponding interface based on the clinical acceptability evaluation. reference or Sensor data and clinical risk associated with that 0086. In an aspect of the tenth embodiment, the means for deviation based on the glucose value indicated by at least altering the user interface includes at least one of providing one of the Sensor and reference data. color-coded information, trend information, directional 0.074. In an aspect of the ninth embodiment, the method information (e.g., arrows or angled lines), and/or fail-safe further includes providing an output through a user interface information. responsive to the clinical acceptability evaluation. 0087. In an aspect of the tenth embodiment, the means for 0075. In an aspect of the ninth embodiment, the step of evaluating the clinical acceptability includes using one of a providing an output includes alerting the user based on the Clarke Error Grid, a mean absolute difference calculation, a clinical acceptability evaluation. rate of change calculation, a consensus grid, and a Standard clinical acceptance test. 0.076. In an aspect of the ninth embodiment, the step of providing an output includes altering the user interface 0088. In an aspect of the tenth embodiment, the system based on the clinical acceptability evaluation. further includes means for requesting additional reference data if the clinical acceptability evaluation determines clini 0077. In an aspect of the ninth embodiment, the step of cal unacceptability. altering the user interface includes at least one of providing color-coded information, trend information, directional 0089. In an aspect of the tenth embodiment, the system information (e.g., arrows or angled lines), and/or fail-safe further includes means for repeated the clinical acceptability information. evaluation for the additional reference data. 0078. In an aspect of the ninth embodiment, the step of 0090. In an aspect of the tenth embodiment, the system evaluating the clinical acceptability includes using one of a further includes means for matching reference data to Sub Clarke Error Grid, a mean absolute difference calculation, a Stantially time corresponding Sensor data to form a matched rate of change calculation, a consensus grid, and a Standard data pair after the clinical acceptability evaluation. clinical acceptance test. 0091. In an eleventh embodiment, a computer system is 0079. In an aspect of the ninth embodiment, the method provided for evaluating clinical acceptability of at least one further includes requesting additional reference data if the of reference and Sensor analyte data, the computer System clinical acceptability evaluation determines clinical unac including: a Sensor data receiving module that receives a ceptability. data Stream including a plurality of time Spaced Sensor data points from a Substantially continuous analyte Sensor, a 0080. In an aspect of the ninth embodiment, the method reference data receiving module that receives reference data further includes repeating the clinical acceptability evalua from a reference analyte monitor, including one or more tion Step for the additional reference data. reference data points, and a clinical acceptability evaluation module that evaluates at least one of the reference and Sensor 0081. In an aspect of the ninth embodiment, the method analyte data using Substantially time corresponding refer further includes a step of matching reference data to Sub ence and Sensor data, wherein the at least one of the Stantially time corresponding Sensor data to form a matched reference and Sensor analyte data is evaluated for deviation pair after the clinical acceptability evaluation Step. from its Substantially time corresponding reference or Sensor 0082 In a tenth embodiment, a system is provided for data and clinical risk associated with that deviation based on evaluating clinical acceptability of at least one of reference the glucose value indicated by at least one of the Sensor and and Sensor analyte data, the method including: means for reference data. receiving a data Stream from an analyte Sensor, a plurality of time-Spaced Sensor data points, means for receiving refer 0092. In an aspect of the eleventh embodiment, the ence data from a reference analyte monitor, including one or computer System further includes an interface control mod more reference data points, and means for evaluating the ule that controls the user interface based on the clinical clinical acceptability of at least one of the reference and acceptability evaluation. Sensor analyte data using Substantially time corresponding 0093. In an aspect of the eleventh embodiment, the reference and Sensor data, wherein the at least one of the interface control module alerts the user based on the clinical reference and Sensor analyte data is evaluated for deviation acceptability evaluation. US 2005/0027181 A1 Feb. 3, 2005

0094. In an aspect of the eleventh embodiment, the monitor, including one or more reference data points, a interface control module alters the user interface based on clinical module that evaluates at least one of the reference the clinical acceptability evaluation. and Sensor analyte data using Substantially time correspond ing reference and Sensor data, wherein the at least one of the 0.095. In an aspect of the eleventh embodiment, the reference and Sensor analyte data is evaluated for deviation interface control module alters the user interface by provid from its Substantially time corresponding reference or Sensor ing at least one of providing color-coded information, trend data and clinical risk associated with that deviation based on information, directional information (e.g., arrows or angled the glucose value indicated by at least one of the Sensor and lines), and/or fail-safe information. reference data; and an interface control module that controls 0096. In an aspect of the eleventh embodiment, the the user interface based on the clinical acceptability evalu clinical acceptability evaluation module uses one of a Clarke ation. Error Grid, a mean absolute difference calculation, a rate of 0103) In an fifteenth embodiment, a computer system is change calculation, a consensus grid, and a Standard clinical provided for evaluating clinical acceptability of at least one acceptance test to evaluate clinical acceptability. of reference and Sensor analyte data, the computer System 0097. In an aspect of the eleventh embodiment, the including: a Sensor data module that receives a data Stream interface control module that requests additional reference including a plurality of time Spaced Sensor data points from data if the clinical acceptability evaluation determines clini a Substantially continuous analyte Sensor; a reference input cal unacceptability. module that receives reference data from a reference analyte monitor, including one or more reference data points, and a 0098. In an aspect of the eleventh embodiment, the clinical module that evaluates at least one of the reference interface control module evaluates the additional reference and Sensor analyte data with Substantially time correspond data using clinical acceptability evaluation module. ing reference and Sensor data, wherein the clinical module 0099. In an aspect of the eleventh embodiment, the uses one of a Clarke Error Grid, a mean absolute difference computer System further includes a data matching module calculation, a rate of change calculation, a consensus grid, that matches clinically acceptable reference data to Substan and a Standard clinical acceptance test to evaluate clinical tially time corresponding clinically acceptable Sensor data to acceptability. form a matched pair. 0104. In an sixteenth embodiment, a computer system is 0100. In a twelfth embodiment, a method is provided for provided for evaluating clinical acceptability of at least one evaluating clinical acceptability of at least one of reference of reference and Sensor analyte data, the computer System and Sensor analyte data, the method including: receiving a including: a Sensor data module that receives a data Stream data Stream from an analyte Sensor, including one or more including a plurality of time Spaced Sensor data points from Sensor data points, receiving reference data from a reference a Substantially continuous analyte Sensor via a receiver, a analyte monitor, including one or more reference data reference input module that receives reference data from a points, evaluating the clinical acceptability at least one of reference analyte monitor, including one or more reference the reference and Sensor analyte data using Substantially data points, and a clinical module that uses a Clarke Error time corresponding reference and Sensor data, wherein the at Grid to evaluate the clinical acceptability at least one of the least one of the reference and Sensor analyte data is evalu reference and Sensor analyte data using Substantially time ated for deviation from its Substantially time corresponding corresponding reference and Sensor data, and a fail-safe reference or Sensor data and clinical risk associated with that module that controls the user interface responsive to the deviation based on the glucose value indicated by at least clinical module evaluating clinical unacceptability. one of the Sensor and reference data; and providing an output 0105. In an seventeenth embodiment, a method is pro through a user interface responsive to the clinical accept Vided for evaluating clinical acceptability of at least one of ability evaluation. reference and Sensor glucose data, the method including: 0101. In an thirteenth embodiment, a method is provided receiving a data Stream from an analyte Sensor, including for evaluating clinical acceptability of at least one of refer one or more Sensor data points, receiving reference data ence and Sensor analyte data, the method including: receiv from a reference glucose monitor, including one or more ing a data Stream from an analyte Sensor, including one or reference data points, evaluating the clinical acceptability at more Sensor data points, receiving reference data from a least one of the reference and Sensor glucose data using reference analyte monitor, including one or more reference Substantially time corresponding reference and Sensor data, data points, and evaluating the clinical acceptability at least wherein the at least one of the reference and Sensor analyte one of the reference and Sensor analyte data using Substan data is evaluated for deviation from its substantially time tially time corresponding reference and Sensor data, includ corresponding reference or Sensor data and clinical risk ing using one of a Clarke Error Grid, a mean absolute asSociated with that deviation based on the glucose value difference calculation, a rate of change calculation, and a indicated by at least one of the Sensor and reference data; consensus grid. and a fail-safe module that controls the user interface responsive to the clinical module evaluating clinical unac 0102) In an fourteenth embodiment, a computer system is ceptability. provided for evaluating clinical acceptability of at least one of reference and Sensor analyte data, the computer System 0106. In an eighteenth embodiment, a method is provided including: a Sensor data module that receives a data Stream for maintaining calibration of a Substantially continuous including a plurality of time Spaced Sensor data points from analyte Sensor, the method including: receiving a data a Substantially continuous analyte Sensor; a reference input Stream from an analyte Sensor, including one or more Sensor module that receives reference data from a reference analyte data points, receiving reference data from a reference ana US 2005/0027181 A1 Feb. 3, 2005 lyte monitor, including two or more reference data points; 0118. In an aspect of the eighteenth embodiment, the step providing at least two matched data pairs by matching of creating a conversion function includes linear regression. reference analyte data to Substantially time corresponding Sensor data; forming a calibration Set including the at least 0119). In an aspect of the eighteenth embodiment, the step two matching data pairs, creating a conversion function of creating a conversion function includes non-linear regreS based on the calibration Set, converting Sensor data into Sion. calibrated data using the conversion function; Subsequently 0120 In an aspect of the eighteenth embodiment, the step obtaining one or more additional reference data points and of forming a calibration Set includes including in the cali creating one or more new matched data pairs, evaluating the bration Set between one and six matched data pairs. calibration Set when the new matched data pair is created, wherein evaluating the calibration Set includes at least one of 0121. In an aspect of the eighteenth embodiment, the step 1) ensuring matched data pairs in the calibration Set span a of forming a calibration Set includes including six matched predetermined time range, 2) ensuring matched data pairs in data pairs. the calibration Set are no older than a predetermined value, 0122) In an aspect of the eighteenth embodiment, the step 3) ensuring the calibration set has substantially distributed of forming a calibration Set further includes determining a high and low matched data pairs over the predetermined value for n, where n is greater than one and represents the time range, and 4) allowing matched data pairs only within number of matched data pairs in the calibration Set. a predetermined range of analyte values, and Subsequently modifying the calibration Set if Such modification is required 0123. In an aspect of the eighteenth embodiment, the step by the evaluation. of determining a value for n is determined as a function of the frequency of the received reference data points and 0107. In an aspect of the eighteenth embodiment, the step Signal Strength over time. of evaluating the calibration Set further includes at least one of evaluating a rate of change of the analyte concentration, 0.124. In an aspect of the eighteenth embodiment, the evaluating a congruence of respective Sensor and reference method further includes determining a set of matching data data in the matched data pairs, and evaluating physiological pairs from the evaluation of the calibration Set and re changes. forming a calibration Set. 0108. In an aspect of the eighteenth embodiment, the step 0.125. In an aspect of the eighteenth embodiment, the of evaluating the calibration Set includes evaluating only the method further includes repeating the Step of re-creating the new matched data pair. conversion function using the re-formed calibration Set. 0109. In an aspect of the eighteenth embodiment, the step 0.126 In an aspect of the eighteenth embodiment, the of evaluating the calibration Set includes evaluating all of the method further includes converting Sensor data into cali matched data pairs in the calibration Set and the new brated data using the re-created conversion function. matched data pair. 0127. In a nineteenth embodiment, a system is provided 0110. In an aspect of the eighteenth embodiment, the step for maintaining calibration of a Substantially continuous of evaluating the calibration Set includes evaluating combi analyte Sensor, the System including: means for receiving a nations of matched data pairs from the calibration Set and the data Stream from an analyte Sensor, a plurality of time new matched data pair. Spaced Sensor data points, means for receiving reference data from a reference analyte monitor, including two or more 0111. In an aspect of the eighteenth embodiment, the step reference data points, means for providing two or more of receiving Sensor data includes receiving a data Stream matched data pairs by matching reference analyte data to from a long-term implantable analyte Sensor. Substantially time corresponding Sensor data, means for 0112) In an aspect of the eighteenth embodiment, the step forming a calibration Set including at least two matched data of receiving Sensor data includes receiving a data Stream that pair; means for creating a conversion function based on the has been algorithmically Smoothed. calibration Set, means for converting Sensor data into cali brated data using the conversion function; Subsequently 0113. In an aspect of the eighteenth embodiment, the step obtaining one or more additional reference data points and of receiving Sensor data Stream includes algorithmically creating one or more new matched data pairs, means for Smoothing the data Stream. evaluating the calibration Set when the new matched data 0114. In an aspect of the eighteenth embodiment, the step pair is created, wherein evaluating the calibration Set of receiving reference data includes downloading reference includes at least one of 1) ensuring matched data pairs in the data via a cabled connection. calibration Set span a predetermined time range, 2) ensuring matched data pairs in the calibration Set are no older than a 0115) In an aspect of the eighteenth embodiment, the step predetermined value, 3) ensuring the calibration set has of receiving reference data includes downloading reference Substantially distributed high and low matched data pairs data via a wireless connection. over the predetermined time range, and 4) allowing matched 0116. In an aspect of the eighteenth embodiment, the step data pairs only within a predetermined range of analyte of receiving reference data from a reference analyte monitor values, and means for modifying the calibration Set if Such includes receiving within a receiver internal communication modification is required by the evaluation. from a reference analyte monitor integral with the receiver. 0128. In an aspect of the nineteenth embodiment, the 0117. In an aspect of the eighteenth embodiment, the means for evaluating the calibration Set further includes at reference analyte monitor includes Self-monitoring of blood least one of means for evaluating a rate of change of the analyte. analyte concentration, means for evaluating a congruence of US 2005/0027181 A1 Feb. 3, 2005

respective Sensor and reference data in matched data pairs, function of the frequency of the received reference data and means for evaluating physiological changes. points and Signal Strength over time. 0129. In an aspect of the nineteenth embodiment, the 0145. In an aspect of the nineteenth embodiment, the means for evaluating the calibration Set includes means for System further includes means for determining a Set of evaluating only the one or more new matched data pairs. matching data pairs from the evaluation of the calibration Set and re-forming a calibration Set. 0130. In an aspect of the nineteenth embodiment, the 0146 In an aspect of the nineteenth embodiment, the means for evaluating the calibration Set includes means for System further includes the means for repeating the Set of evaluating all of the matched data pairs in the calibration Set creating the conversion function using the re-formed cali and the one or more new matched data pairs. bration Set. 0131. In an aspect of the nineteenth embodiment, the means for evaluating the calibration Set includes means for 0147 In an aspect of the nineteenth embodiment, the evaluating combinations of matched data pairs from the System further includes means for converting Sensor data calibration Set and the one or more new matched data pair. into calibrated data using the re-created conversion function. 0.148. In a twentieth embodiment, a computer system is 0.132. In an aspect of the nineteenth embodiment, the provided for maintaining calibration of a Substantially con means for receiving Sensor data includes means for receiv tinuous analyte Sensor, the computer System including: a ing Sensor data from a long-term implantable analyte Sensor. Sensor data receiving module that receives a data Stream including a plurality of time Spaced Sensor data points from 0133. In an aspect of the nineteenth embodiment, the a Substantially continuous analyte Sensor; a reference data means for receiving Sensor data includes means for receiv receiving module that receives reference data from a refer ing Sensor data that has been algorithmically Smoothed. ence analyte monitor, including two or more reference data 0134. In an aspect of the nineteenth embodiment, the points, a data matching module that forms two or more means for receiving Sensor data includes means for algo matched data pairs by matching reference data to Substan rithmically Smoothing the receiving Sensor data. tially time corresponding Sensor data; a calibration Set module that forms a calibration Set including at least two 0135) In an aspect of the nineteenth embodiment, the matched data pairs; a conversion function module that means for receiving reference data includes means for creates a conversion function using the calibration Set, a downloading reference data via a cabled connection. sensor data transformation module that converts sensor data 0136. In an aspect of the nineteenth embodiment, the into calibrated data using the conversion function; and a means for receiving reference data includes means for calibration evaluation module that evaluates the calibration downloading reference data via a wireleSS connection. Set when the new matched data pair is provided, wherein evaluating the calibration Set includes at least one of 1) 0.137 In an aspect of the nineteenth embodiment, the ensuring matched data pairs in the calibration Set span a means for receiving reference data from a reference analyte predetermined time period, 2) ensuring matched data pairs monitor includes means for receiving within a receiver in the calibration Set are no older than a predetermined internal communication from a reference analyte monitor value, 3) ensuring the calibration set has Substantially dis integral with the receiver. tributed high and low matched data pairs over a predeter 0.138. In an aspect of the nineteenth embodiment, the mined time range, and 4) allowing matched data pairs only means for receiving reference data includes means for within a predetermined range of analyte values, wherein the receiving from a Self-monitoring of blood analyte. conversion function module is programmed to re-create the conversion function of Such modification is required by the 0.139. In an aspect of the nineteenth embodiment, the calibration evaluation module. means for creating a conversion function includes means for 0149. In an aspect of the twentieth embodiment, the performing linear regression. evaluation calibration module further evaluates at least one 0140. In an aspect of the nineteenth embodiment, the of a rate of change of the analyte concentration, a congru means for creating a conversion function includes means for ence of respective Sensor and reference data in matched data performing non-linear regression. pairs; and physiological changes. 0.141. In an aspect of the nineteenth embodiment, the 0150. In an aspect of the twentieth embodiment, the means for forming a calibration Set includes including in the evaluation calibration module evaluates only the new calibration Set between one and Six matched data pairs. matched data pair. 0151. In an aspect of the twentieth embodiment, the 0142. In an aspect of the nineteenth embodiment, the evaluation calibration module evaluates all of the matched means for forming a calibration Set includes including in the data pairs in the calibration Set and the new matched data calibration Set Six matched data pairs. pair. 0143. In an aspect of the nineteenth embodiment, the 0152. In an aspect of the twentieth embodiment, the means for forming a calibration Set further includes deter evaluation calibration module evaluates combinations of mining a value for n, where n is greater than one and matched data pairs from the calibration Set and the new represents the number of matched data pairs in the calibra matched data pair. tion Set. 0153. In an aspect of the twentieth embodiment, the 0144. In an aspect of the nineteenth embodiment, the Sensor data receiving module receives the data Stream from means for determining a value for n is determined as a a long-term implantable analyte Sensor. US 2005/0027181 A1 Feb. 3, 2005

0154) In an aspect of the twentieth embodiment, the data pairs by matching reference analyte data to Substan Sensor data receiving module receives an algorithmically tially time corresponding Sensor data, forming a calibration Smoothed data Stream. Set including the at least two matching data pairs, creating a conversion function based on the calibration Set, Subse O155 In an aspect of the twentieth embodiment, the quently obtaining one or more additional reference data Sensor data receiving module includes programming to points and creating one or more new matched data pairs, and Smooth the data Stream. evaluating the calibration Set when the new matched data 0156. In an aspect of the twentieth embodiment, the pair is created, wherein evaluating the calibration Set reference data receiving module downloads reference data includes at least one of 1) ensuring matched data pairs in the via a cabled connection. calibration Set span a predetermined time range, 2) ensuring matched data pairs in the calibration Set are no older than a O157. In an aspect of the twentieth embodiment, the predetermined value, 3) ensuring the calibration set has reference data receiving module downloads reference data Substantially distributed high and low matched data pairs via a wireleSS connection. over the predetermined time range, and 4) allowing matched 0158. In an aspect of the twentieth embodiment, the data pairs only within a predetermined range of analyte reference data receiving module receives within a receiver values. internal communication from a reference analyte monitor 0170 In a twenty-second embodiment, a computer sys integral with the receiver. tem is provided for maintaining calibration of a glucose 0159. In an aspect of the twentieth embodiment, the Sensor, the computer System including: a Sensor data module reference data receiving module receives reference data that receives a data Stream including a plurality of time from a Self-monitoring of blood analyte. Spaced Sensor data points from a Substantially continuous analyte Sensor; a reference input module that receives ref 0160 In an aspect of the twentieth embodiment, the erence data from a reference analyte monitor, including two conversion function module includes programming that per or more reference data points, a processor module that forms forms linear regression. two or more matched data pairs by matching reference data 0.161 In an aspect of the twentieth embodiment, the to Substantially time corresponding Sensor data and Subse conversion function module includes programming that per quently forms a calibration Set including the two or more forms non-linear regression. matched data pairs, and a calibration evaluation module that evaluates the calibration set when the new matched data pair 0162. In an aspect of the twentieth embodiment, the is provided, wherein evaluating the calibration Set includes calibration Set module includes in the calibration Set at least one of 1) ensuring matched data pairs in the between one and six matched data pairs. calibration set span a predetermined time period, 2) ensuring 0163. In an aspect of the twentieth embodiment, the matched data pairs in the calibration Set are no older than a calibration Set module includes in the calibration Set Six predetermined value, 3) ensuring the calibration set has matched data pairs. Substantially distributed high and low matched data pairs over a predetermined time range, and 4) allowing matched 0164. In an aspect of the twentieth embodiment, the data pairs only within a predetermined range of analyte calibration Set module further includes programming for values, wherein the conversion function module is pro determining a value for n, where n is greater than one and grammed to re-create the conversion function of Such modi represents the number of matched data pairs in the calibra tion Set. fication is required by the calibration evaluation module. 0171 In a twenty-third embodiment, a method is pro 0.165. In an aspect of the twentieth embodiment, the Vided for evaluating the quality of a calibration of an analyte programming for determining a value for n determines n as Sensor, the method including: receiving a data Stream from a function of the frequency of the received reference data an analyte Sensor, including one or more Sensor data points; points and Signal Strength over time. receiving reference data from a reference analyte monitor, 0166 In an aspect of the twentieth embodiment, data including two or more reference data points, providing at matching module further includes programming to re-form least two matched data pairs by matching reference analyte the calibration Set based on the calibration evaluation. data to Substantially time corresponding Sensor data; form ing a calibration Set including the at least two matching data 0167. In an aspect of the twentieth embodiment, the pairs, creating a conversion function based on the calibration conversion function module further includes programming Set, receiving additional Sensor data from the analyte Sensor; to re-create the conversion function based on the re-formed converting Sensor data into calibrated data using the con calibration Set. version function; and evaluating the quality of the calibra 0.168. In an aspect of the twentieth embodiment, the tion Set using a data association function. Sensor data transformation module further including pro 0172 In an aspect of the twenty-third embodiment, the gramming for converting Sensor data into calibrated using Step of receiving Sensor data includes receiving a data the re-created conversion function. Stream that has been algorithmically Smoothed. 0169. In a twenty-first embodiment, a method is provided 0.173) In an aspect of the twenty-third embodiment, the for maintaining calibration of a glucose Sensor, the method Step of receiving Sensor data includes algorithmically including: receiving a data Stream from an analyte Sensor, Smoothing the data Stream. including one or more Sensor data points, receiving refer 0.174. In an aspect of the twenty-third embodiment, the ence data from a reference analyte monitor, including two or Step of receiving Sensor data includes receiving Sensor data more reference data points, providing at least two matched from a Substantially continuous glucose Sensor. US 2005/0027181 A1 Feb. 3, 2005

0.175. In an aspect of the twenty-third embodiment, the 0190. In a twenty-fourth embodiment, a system is pro Step of receiving Sensor data includes receiving Sensor data Vided for evaluating the quality of a calibration of an analyte from an implantable glucose Sensor. Sensor, the System including: means for receiving a data Stream from an analyte Sensor, a plurality of time-spaced 0176). In an aspect of the twenty-third embodiment, the Sensor data points, means for receiving reference data from Step of receiving Sensor data includes receiving Sensor data a reference analyte monitor, including two or more reference from a Subcutaneously implantable glucose Sensor. data points, means for providing two or more matched data 0177. In an aspect of the twenty-third embodiment, the pairs by matching reference analyte data to Substantially Step of receiving reference data includes receiving reference time corresponding Sensor data, means for forming a cali data from a Self-monitoring blood glucose test. bration Set including at least two matched data pair; means for creating a conversion function based on the calibration 0178. In an aspect of the twenty-third embodiment, the Set; means for converting Sensor data into calibrated data Step of receiving reference data includes downloading ref using the conversion function; means for evaluating the erence data via a cabled connection. quality of the calibration Set based on a data association 0179. In an aspect of the twenty-third embodiment, the function. Step of receiving reference data includes downloading ref 0191 In an aspect of the twenty-fourth embodiment, the erence data via a wireleSS connection. means for receiving Sensor data includes means for receiv 0180. In an aspect of the twenty-third embodiment, the ing Sensor data that has been algorithmically Smoothed. Step of receiving reference data from a reference analyte monitor includes receiving within a receiver internal com 0.192 In an aspect of the twenty-fourth embodiment, the munication from a reference analyte monitor integral with means for receiving Sensor data includes means for algo the receiver. rithmically Smoothing the receiving Sensor data. 0181. In an aspect of the twenty-third embodiment, the 0193 In an aspect of the twenty-fourth embodiment, the Step of evaluating the quality of the calibration Set based on means for receiving Sensor data includes means for receiv a data association function includes performing one of linear ing Sensor data from Substantially continuous glucose Sen regression, non-linear regression, rank correlation, least SO. mean Square fit, mean absolute deviation, and mean absolute 0194 In an aspect of the twenty-fourth embodiment, the relative difference. means for receiving Sensor data includes means for receiv 0182. In an aspect of the twenty-third embodiment, the ing Sensor data from an implantable glucose Sensor. Step of evaluating the quality of the calibration Set based on a data association function includes performing linear least 0.195. In an aspect of the twenty-fourth embodiment, the Squares regression. means for receiving Sensor data includes means for receiv ing Sensor data from Subcutaneously implantable glucose 0183 In an aspect of the twenty-third embodiment, the SCSO. Step of evaluating the quality of the calibration Set based on a data association function includes Setting a threshold of 0196. In an aspect of the twenty-fourth embodiment, the data association. means for receiving reference data includes means for receiving reference data from a Self-monitoring blood glu 0184. In an aspect of the twenty-third embodiment, the COSe teSt. Step of evaluating the quality of the calibration Set based on data association includes performing linear least Squares 0197). In an aspect of the twenty-fourth embodiment, the regression and wherein the Step of Setting a threshold hold means for receiving reference data includes means for includes an R-value threshold of 0.79. downloading reference data via a cabled connection. 0185. In an aspect of the twenty-third embodiment, the 0198 In an aspect of the twenty-fourth embodiment, the method further includes providing an output to a user means for receiving reference data includes means for interface responsive to the quality of the calibration Set. downloading reference data via a wireleSS connection. 0186. In an aspect of the twenty-third embodiment, the 0199. In an aspect of the twenty-fourth embodiment, the Step of providing an output includes displaying analyte means for receiving reference data from a reference analyte values to a user dependent upon the quality of the calibra monitor includes means for receiving within a receiver tion. internal communication from a reference analyte monitor 0187. In an aspect of the twenty-third embodiment, the integral with the receiver. Step of providing an output includes alerting the dependent 0200. In an aspect of the twenty-fourth embodiment, the upon the quality of the calibration. means for evaluating the quality of the calibration Set 0188 In an aspect of the twenty-third embodiment, the includes means for performing one of linear regression, Step of providing an output includes altering the user inter non-linear regression, rank correlation, least mean Square fit, face dependent upon the quality of the calibration. mean absolute deviation, and mean absolute relative differ 0189 In an aspect of the twenty-third embodiment, the CCC. Step of providing an output includes at least one of providing 0201 In an aspect of the twenty-fourth embodiment, the color-coded information, trend information, directional means for evaluating the quality of the calibration Set information (e.g., arrows or angled lines), and/or fail-safe includes means for performing linear least Squares regreS information. Sion. US 2005/0027181 A1 Feb. 3, 2005

0202) In an aspect of the twenty-fourth embodiment, the 0214. In an aspect of the twenty-fifth embodiment, the means for evaluating the quality of the calibration Set Sensor data receiving module is adapted to receive Sensor includes means for Setting a threshold of data association. data from Subcutaneously implantable glucose Sensor. 0203. In an aspect of the twenty-fourth embodiment, the 0215. In an aspect of the twenty-fifth embodiment, the means for evaluating the quality of the calibration Set reference data receiving module is adapted to receive ref includes means for performing linear least Squares regreS erence data from a Self-monitoring blood glucose test. Sion and wherein the means for Setting a threshold hold 0216) In an aspect of the twenty-fifth embodiment, the includes an R-value threshold of 0.71. reference data receiving module is adapted to download 0204. In an aspect of the twenty-fourth embodiment, the reference data via a cabled connection. System further includes means for providing an output to a 0217. In an aspect of the twenty-fifth embodiment, the user interface responsive to the quality of the calibration Set. reference data receiving module is adapted to download 0205. In an aspect of the twenty-fourth embodiment, the reference data via a wireleSS connection. means for providing an output includes means for displaying 0218. In an aspect of the twenty-fifth embodiment, the analyte values to a user dependent upon the quality of the reference data receiving module is adapted to receive ref calibration. erence data from a reference analyte monitor integral with 0206. In an aspect of the twenty-fourth embodiment, the the receiver. means for providing an output includes means for alerting 0219. In an aspect of the twenty-fifth embodiment, the the dependent upon the quality of the calibration. quality evaluation module performs one of linear regression, 0207. In an aspect of the twenty-fourth embodiment, the non-linear regression, rank correlation, least mean Square fit, means for providing an output includes means for altering mean absolute deviation, and mean absolute relative differ the user interface dependent upon the quality of the calibra ence to evaluate calibration Set quality. tion. 0220. In an aspect of the twenty-fifth embodiment, the 0208. In an aspect of the twenty-fourth embodiment, the quality evaluation module performs linear least Squares means for providing an output includes at least one of regression. providing color-coded information, trend information, direc 0221) In an aspect of the twenty-fifth embodiment, the tional information (e.g., arrows or angled lines), and/or quality evaluation module sets a threshold for the data fail-safe information. asSociation function. 0209. In a twenty-fifth embodiment, a computer system is 0222. In an aspect of the twenty-fifth embodiment, the provided for evaluating the quality of a calibration of an quality evaluation module performs linear least Squares analyte Sensor, the computer System including: a Sensor data regression and wherein the threshold of the data association receiving module that receives a data Stream including a function includes an R-value threshold of at least 0.79. plurality of time Spaced Sensor data points from a Substan tially continuous analyte Sensor; a reference data receiving 0223) In an aspect of the twenty-fifth embodiment, the module that receives reference data from a reference analyte computer System further includes an interface control mod monitor, including two or more reference data points, a data ule that controls the user interface based on the quality of the matching module that forms two or more matched data pairs calibration Set. by matching reference data to Substantially time correspond 0224. In an aspect of the twenty-fifth embodiment, the ing Sensor data, a calibration Set module that forms a interface control module displays analyte values to a user calibration Set including at least two matched data pairs, a dependent upon the quality of the calibration Set. conversion function module that creates a conversion func tion using the calibration Set, a Sensor data transformation 0225. In an aspect of the twenty-fifth embodiment, the module that converts Sensor data into calibrated data using interface control module alerts the user based upon the the conversion function; and a quality evaluation module quality of the calibration Set. that evaluates the quality of the calibration Set based on a 0226. In an aspect of the twenty-fifth embodiment, the data association function. interface control module alters the user interface based upon 0210. In an aspect of the twenty-fifth embodiment, the the quality of the calibration Set. Sensor data receiving module receives Sensor data that has 0227. In an aspect of the twenty-fifth embodiment, the been algorithmically Smoothed. interface control module provides at least one of color-coded 0211. In an aspect of the twenty-fifth embodiment, the information, trend information, directional information computer System further includes a data Smoothing module (e.g., arrows or angled lines), and/or fail-safe information. that algorithmically Smoothes Sensor data received from the 0228. In a twenty-sixth embodiment, a method is pro Sensor data receiving module. Vided for evaluating the quality of a calibration of an analyte Sensor, the method including: receiving a data Stream from 0212. In an aspect of the twenty-fifth embodiment, the an analyte Sensor, including one or more Sensor data points; Sensor data receiving module is adapted to receive Sensor receiving reference data from a reference analyte monitor, data from Substantially continuous glucose Sensor. including two or more reference data points, providing at 0213. In an aspect of the twenty-fifth embodiment, the least two matched data pairs by matching reference analyte Sensor data receiving module is adapted to receive Sensor data to Substantially time corresponding Sensor data; form data from an implantable glucose Sensor. ing a calibration Set including the at least two matching data US 2005/0027181 A1 Feb. 3, 2005 pairs, creating a conversion function based on the calibration 0232. In a thirtieth embodiment, a method is provided for Set, receiving additional Sensor data from the analyte Sensor; evaluating the quality of a calibration of a glucose Sensor, converting Sensor data into calibrated data using the con the method including: receiving Sensor data from a glucose version function; and evaluating the quality of the calibra Sensor, including one or more Sensor data points, receiving tion Set based on a data association function Selected from reference data from a reference glucose monitor, including the group consisting of linear regression, non-linear regreS one or more reference data points, providing one or more Sion, rank correlation, least mean Square fit, mean absolute matched data pairs by matched reference glucose data to deviation, and mean absolute relative difference. Substantially time corresponding Sensor data; forming a 0229. In a twenty-seventh embodiment, a method is calibration Set including at least one matched data pair; and provided for evaluating the quality of a calibration of an evaluating the quality of the calibration Set based on data analyte Sensor, the method including: receiving a data asSociation. Stream from an analyte Sensor, including one or more Sensor data points, receiving reference data from a reference ana BRIEF DESCRIPTION OF THE DRAWINGS lyte monitor, including two or more reference data points; providing at least two matched data pairs by matching 0233 FIG. 1 is an exploded perspective view of a reference analyte data to Substantially time corresponding glucose Sensor in one embodiment. Sensor data; forming a calibration Set including the at least 0234 FIG. 2 is a block diagram that illustrates the sensor two matching data pairs, creating a conversion function electronics in one embodiment. based on the calibration Set, receiving additional Sensor data from the analyte Sensor; converting Sensor data into cali 0235 FIG. 3 is a graph that illustrates data Smoothing of brated data using the conversion function; evaluating the a raw data Signal in one embodiment. quality of the calibration Set using a data association func tion; and providing an output to a user interface responsive 0236 FIGS. 4A to 4D are schematic views of a receiver to the quality of the calibration Set. in first, Second, third, and fourth embodiments, respectively. 0230. In a twenty-eighth embodiment, a computer system 0237 FIG. 5 is a block diagram of the receiver electron is provided for evaluating the quality of a calibration of an ics in one embodiment. analyte Sensor, the computer System including: a Sensor data module that receives a data stream including a plurality of 0238 FIG. 6 is a flow chart that illustrates the initial time Spaced Sensor data points from a Substantially continu calibration and data output of the sensor data in one embodi ous analyte Sensor; a reference input module that receives ment. reference data from a reference analyte monitor, including 0239 FIG. 7 is a graph that illustrates a regression two or more reference data points, a processor module that performed on a calibration Set to obtain a conversion func forms two or more matched data pairs by matching reference tion in one exemplary embodiment. data to Substantially time corresponding Sensor data and Subsequently forms a calibration Set including the two or 0240 FIG. 8 is a flow chart that illustrates the process of more matched data pairs, and a conversion function module evaluating the clinical acceptability of reference and Sensor that creates a conversion function using the calibration Set, data in one embodiment. a Sensor data transformation module that converts Sensor data into calibrated data using the conversion function; a 0241 FIG. 9 is a graph of two data pairs on a Clarke quality evaluation module that evaluates the quality of the Error Grid to illustrate the evaluation of clinical acceptabil calibration Set based on a data association Selected from the ity in one exemplary embodiment. group consisting of linear regression, non-linear regression, 0242 FIG. 10 is a flow chart that illustrates the process rank correlation, least mean Square fit, mean absolute devia of evaluation of calibration data for best calibration based on tion, and mean absolute relative difference. inclusion criteria of matched data pairs in one embodiment. 0231. In a twenty-ninth embodiment, a computer system is provided for evaluating the quality of a calibration of an 0243 FIG. 11 is a flow chart that illustrates the process analyte Sensor, the computer System including: a Sensor data of evaluating the quality of the calibration in one embodi module that receives a data Stream including a plurality of ment. time Spaced Sensor data points from a Substantially continu 0244 FIGS. 12A and 12B are graphs that illustrate an ous analyte Sensor; a reference input module that receives evaluation of the quality of calibration based on data asso reference data from a reference analyte monitor, including ciation in one exemplary embodiment using a correlation two or more reference data points, a processor module that coefficient. forms two or more matched data pairs by matching reference data to Substantially time corresponding Sensor data and DETAILED DESCRIPTION OF CERTAIN Subsequently forms a calibration Set including the two or EMBODIMENTS more matched data pairs, and a conversion function module that creates a conversion function using the calibration Set, 0245. The following description and examples illustrate a Sensor data transformation module that converts Sensor Some exemplary embodiments of the disclosed invention in data into calibrated data using the conversion function; a detail. Those of skill in the art will recognize that there are quality evaluation module that evaluates the quality of the numerous variations and modifications of this invention that calibration Set based on data association; and a fail-safe are encompassed by its Scope. Accordingly, the description module that controls the user interface based on the quality of a certain exemplary embodiment should not be deemed to of the calibration set. limit the Scope of the present invention. US 2005/0027181 A1 Feb. 3, 2005

0246 Definitions roxine (T4); thyroxine-binding globulin; trace elements; 0247. In order to facilitate an understanding of the dis transferrin, UDP-galactose-4-epimerase, urea; uroporphy rinogen I Synthase; Vitamin A, white blood cells, and Zinc closed invention, a number of terms are defined below. protoporphyrin. Salts, Sugar, protein, fat, Vitamins and hor 0248. The term “analyte,” as used herein, is a broad term mones naturally occurring in blood or interstitial fluids may and is used in its ordinary Sense, including, without limita also constitute analytes in certain embodiments. The analyte tion, to refer to a Substance or chemical constituent in a may be naturally present in the biological fluid, for example, biological fluid (for example, blood, interstitial fluid, cere a metabolic product, a hormone, an antigen, an antibody, and bral spinal fluid, lymph fluid or urine) that can be analyzed. the like. Alternatively, the analyte may be introduced into the Analytes may include naturally occurring Substances, arti body, for example, a contrast agent for imaging, a radioiso ficial Substances, metabolites, and/or reaction products. In tope, a chemical agent, a fluorocarbon-based Synthetic Some embodiments, the analyte for measurement by the blood, or a drug or pharmaceutical composition, including Sensor heads, devices, and methods is analyte. However, but not limited to insulin; ethanol; (marijuana, other analytes are contemplated as well, including but not , hashish); inhalants (nitrous oxide, limited to acarboxyprothrombin; acylcarnitine, adenine amyl nitrite, butyl nitrite, chlorohydrocarbons, hydrocar phosphoribosyl transferase; adenosine deaminase, albumin; bons); cocaine (crack cocaine); stimulants (amphetamines, alpha-fetoprotein; amino acid profiles (arginine (Krebs methamphetamines, Ritalin, Cylert, Preludin, Didrex, cycle), histidine/urocanic acid, homocysteine, phenylala PreState, Voranil, Sandrex, Plegine); depressants (barbitu nine/tyrosine, tryptophan); andrenostenedione; antipyrine; ates, methaqualone, tranquilizerS Such as Valium, Librium, arabinitol enantiomers, arginase; benzoyle.cgonine Miltown, Serax, Equanil, Tranxene); hallucinogens (phen (cocaine); biotinidase; biopterin, c-reactive protein; cam cyclidine, lysergic acid, mescaline, peyote, psilocybin); nar itine, camosinase, CD4, ceruloplasmin, chenodeoxycholic cotics (heroin, codeine, morphine, opium, meperidine, Per acid; chloroquine, cholesterol, cholinesterase; conjugated cocet, Percodan, TuSSioneX, Fentanyl, Darvon, Talwin, 1-B hydroxy-cholic acid; cortisol, creatine kinase, creatine Lomotil); designer drugs (analogs of fentanyl, meperidine, kinase MM isoenzyme, cycloSporin A; d-penicillamine; amphetamines, methamphetamines, and phencyclidine, for de-ethylchloroquine; sulfate; DNA example, Ecstasy); anabolic Steroids; and . The (acetylator polymorphism, alcohol dehydrogenase, alpha metabolic products of drugs and pharmaceutical composi 1-antitrypsin, cystic fibrosis, Duchenne/Becker muscular tions are also contemplated analytes. Analytes Such as dystrophy, analyte-6-phosphate dehydrogenase, hemoglo neurochemicals and other chemicals generated within the binopathies, A, S, C, E, D-Punjab, beta-thalassemia, hepa body may also be analyzed, Such as, for example, ascorbic titis B virus, HCMV, HIV-1, HTLV-1, Leber hereditary optic acid, uric acid, dopamine, noradrenaline, 3-methoxy neuropathy, MCAD, RNA, PKU, Plasmodium vivax, sexual differentiation, 21-deoxycortisol); desbutylhalofantrine; tyramine (3MT), 3,4-Dihydroxyphenylacetic acid dihydropteridine reductase; diptheria/tetanus antitoxin; (DOPAC), Homovanillic acid (HVA), 5-Hydroxytryptamine erythrocyte arginase, erythrocyte protoporphyrin; esterase (5HT), and 5-Hydroxyindoleacetic acid (FHIAA). D; fatty acids/acylglycines, free B-human chorionic gona 0249. The terms “operably connected” and “operably dotropin; free erythrocyte porphyrin; free thyroxine (FT4); linked,” as used herein, are broad terms and are used in their free tri-iodothyronine (FT3); fumarylacetoacetase; galac ordinary Sense, including, without limitation, one or more tose/gal-1-phosphate, galactose-1-phosphate uridyltrans components being linked to another component(s) in a ferase, gentamicin; analyte-6-phosphate dehydrogenase; manner that allows transmission of Signals between the glutathione, glutathione perioxidase; glycocholic acid; gly components, e.g., wired or wirelessly. For example, one or cosylated hemoglobin, halofantrine, hemoglobin variants, more electrodes may be used to detect the amount of analyte hexosaminidase A, human erythrocyte carbonic anhydrase I; in a Sample and convert that information into a signal; the 17 alpha-hydroxyprogesterone; hypoxanthine phosphoribo Signal may then be transmitted to an electronic circuit Syl transferase; immunoreactive trypsin, lactate, lead; lipo means. In this case, the electrode is “operably linked' to the proteins ((a), B/A-1, f); lysozyme, mefloquine; netilmicin; electronic circuitry. phenobarbitone; phenyloin, phytanic/pristanic acid; proges 0250) The term “EEPROM,” as used herein, is a broad terone; prolactin; prolidase; purine nucleoside phosphory term and is used in its ordinary Sense, including, without lase; quinine; reverse tri-iodothyronine (rT3); Selenium; limitation, electrically erasable programmable read-only Serum pancreatic lipase, Sissomicin; Somatomedin C; Spe memory, which is user-modifiable read-only memory cific antibodies (adenovirus, anti-nuclear antibody, anti-Zeta (ROM) that can be erased and reprogrammed (e.g., written antibody, arbovirus, Aujeszky's disease virus, dengue virus, to) repeatedly through the application of higher than normal Dracunculus medinensis, EchinococcuS granulosus, Enta electrical Voltage. moeba histolytica, enterovirus, Giardia duodenalisa, Heli cobacter pylori, hepatitis B virus, herpes virus, HIV-1, IgE 0251) The term “SRAM,” as used herein, is a broad term (atopic disease), influenza virus, Leishmania donovani, lep and is used in its ordinary Sense, including, without limita toSpira, measles/mumpS/rubella, Mycobacterium leprae, tion, static random access memory (RAM) that retains data Mycoplasma pneumoniae, Myoglobin, Onchocerca volvu bits in its memory as long as power is being Supplied. lus, parainfluenza virus, Plasmodium falciparum, poliovirus, 0252) The term “A/D Converter,” as used herein, is a Pseudomonas aeruginosa, respiratory Syncytial virus, rick broad term and is used in its ordinary Sense, including, ettsia (Scrub typhus), Schistosoma man SOni, Toxoplasma without limitation, hardware that converts analog signals gondii, TrepenOma pallidium, TrypanoSoma Cruzi/rangeli, into digital Signals. vesicular stomatis virus, Wuchereria bancrofii, yellow fever 0253) The term “microprocessor,” as used herein, is a virus); Specific antigens (hepatitis B virus, HIV-1); Succiny broad term and is used in its ordinary Sense, including, lacetone; sulfadoxine; theophylline; thyrotropin (TSH); thy without limitation a computer System or processor designed US 2005/0027181 A1 Feb. 3, 2005 to perform arithmetic and logic operations using logic Number 5, September-October 1987, which is incorporated circuitry that responds to and processes the basic instruc by reference herein in its entirety. tions that drive a computer. 0263. The term “Consensus Error Grid', as used herein, 0254 The term “RF transceiver,” as used herein, is a is a broad term and is used in its ordinary Sense, including, broad term and is used in its ordinary Sense, including, without limitation, an error grid analysis that assigns a without limitation, a radio frequency transmitter and/or Specific level of clinical risk to any possible error between receiver for transmitting and/or receiving Signals. two time corresponding glucose measurements. The Con Sensus Error Grid is divided into Zones signifying the degree 0255 The term "jitter” as used herein, is a broad term and of risk posed by the deviation. See Parkes et al., “A New is used in its ordinary Sense, including, without limitation, Consensus Error Grid to Evaluate the Clinical Significance uncertainty or variability of waveform timing, which may be of Inaccuracies in the Measurement of Blood Glucose', cause by ubiquitous noise caused by a circuit and/or envi Diabetes Care, Volume 23, Number 8, August 2000, which ronmental effects, jitter can be seen in amplitude, phase is incorporated by reference herein in its entirety. timing, or the width of the Signal pulse. 0264. The term “clinical acceptability”, as used herein, is 0256 The term "raw data signal,” as used herein, is a a broad term and is used in its ordinary Sense, including, broad term and is used in its ordinary Sense, including, without limitation, determination of the risk of inaccuracies without limitation, an analog or digital signal directly related to a patient. Clinical acceptability considers a deviation to the measured analyte from the analyte Sensor. In one between time corresponding glucose measurements (e.g., example, the raw data Signal is digital data in “counts' data from a glucose Sensor and data from a reference glucose converted by an A/D converter from an analog signal (e.g., monitor) and the risk (e.g., to the decision making of a voltage or amps) representative of an analyte concentration. diabetic patient) associated with that deviation based on the glucose value indicated by the Sensor and/or reference data. 0257 The term “counts,” as used herein, is a broad term One example of clinical acceptability may be 85% of a given and is used in its ordinary Sense, including, without limita tion, a unit of measurement of a digital Signal. In one set of measured analyte values within the “A” and “B” example, a raw data Signal measured in counts is directly region of a standard Clarke Error Grid when the sensor related to a voltage (converted by an A/D converter), which measurements are compared to a Standard reference mea is directly related to current. Surement. 0265. The term “R-value,” as used herein, is a broad term 0258. The term “analyte sensor,” as used herein, is a and is used in its ordinary Sense, including, without limita broad term and is used in its ordinary Sense, including, tion, one conventional way of Summarizing the correlation without limitation, any mechanism (e.g., enzymatic or non of data; that is, a Statement of what residuals (e.g., root mean enzymatic) by which analyte can be quantified. For example, Square deviations) are to be expected if the data are fitted to Some embodiments utilize a membrane that contains glucose a Straight line by the a regression. oxidase that catalyzes the conversion of oxygen and glucose 0266 The term “data association” and “data association to and gluconate: function,” as used herein, are a broad terms and are used in Glucose--O->Gluconate--HO, their ordinary Sense, including, without limitation, a Statis tical analysis of data and particularly its correlation to, or 0259 Because for each glucose molecule metabolized, deviation from, from a particular curve. A data association there is a proportional change in the co-reactant O and the function is used to show data association. For example, the product HO, one can use an electrode to monitor the data that forms that calibration Set as described herein may current change in either the co-reactant or the product to be analyzed mathematically to determine its correlation to, determine glucose concentration. or deviation from, a curve (e.g., line or set of lines) that 0260 The term “host,” as used herein, is a broad term and defines the conversion function; this correlation or deviation is used in its ordinary Sense, including, without limitation, is the data association. A data association function is used to mammals, particularly humans. determine data association. Examples of data association functions include, but are not limited to, linear regression, 0261) The term “matched data pairs”, as used herein, is a non-linear mapping/regression, rank (e.g., non-parametric) broad term and is used in its ordinary Sense, including, correlation, least mean Square fit, mean absolute deviation without limitation, reference data (e.g., one or more refer (MAD), mean absolute relative difference. In one such ence analyte data points) matched with Substantially time example, the correlation coefficient of linear regression is corresponding sensor data (e.g., one or more Sensor data indicative of the amount of data association of the calibra points). tion Set that forms the conversion function, and thus the 0262 The term “Clarke Error Grid”, as used herein, is a quality of the calibration. broad term and is used in its ordinary Sense, including, 0267 The term “quality of calibration” as used herein, is without limitation, an error grid analysis, which evaluates a broad term and is used in its ordinary Sense, including, the clinical Significance of the difference between a refer without limitation, the Statistical association of matched data ence glucose value and a Sensor generated glucose value, pairs in the calibration Set used to create the conversion taking into account 1) the value of the reference glucose function. For example, an R-Value may be calculated for a measurement, 2) the value of the Sensor glucose measure calibration Set to determine its Statistical data association, ment, 3) the relative difference between the two values, and wherein an R-value greater than 0.79 determines a statisti 4) the clinical significance of this difference. See Clarke et cally acceptable calibration quality, while an R-Value leSS al., “Evaluating Clinical Accuracy of Systems for Self than 0.79 determines statistically unacceptable calibration Monitoring of Blood Glucose', Diabetes Care, Volume 10, quality. US 2005/0027181 A1 Feb. 3, 2005

0268. The term “substantially' as used herein, is a broad without limitation, any electronic connection known to those term and is used in its ordinary Sense, including, without in the art that may be utilized to interface the Sensor head limitation, being largely but not necessarily wholly that electrodes with the electronic circuitry of a device Such as which is Specified. mechanical (e.g., pin and Socket) or Soldered. 0269. The term “congruence” as used herein, is a broad 0275. The term “sensing membrane,” as used herein, is a term and is used in its ordinary Sense, including, without broad term and is used in its ordinary Sense, including, limitation, the quality or State of agreeing, coinciding, or without limitation, a permeable or Semi-permeable mem being concordant. In one example, congruence may be brane that may be comprised of two or more domains and determined using rank correlation. constructed of materials of a few microns thickneSS or more, which are permeable to oxygen and may or may not be 0270. The term “concordant’ as used herein, is a broad permeable to an analyte of interest. In one example, the term and is used in its ordinary Sense, including, without Sensing membrane comprises an immobilized glucose oxi limitation, being in agreement or harmony, and/or free from dase enzyme, which enables an electrochemical reaction to discord. occur to measure a concentration of glucose. 0271 The phrase “continuous (or continual) analyte 0276 The term “biointerface membrane,” as used herein, Sensing,” as used herein, is a broad term and is used in its is a broad term and is used in its ordinary Sense, including, ordinary Sense, including, without limitation, the period in without limitation, a permeable membrane that may be which monitoring of analyte concentration is continuously, comprised of two or more domains and constructed of continually, and or intermittently (but regularly) performed, materials of a few microns thickness or more, which may be for example, about every 5 to 10 minutes. placed over the Sensor body to keep host cells (e.g., mac 0272. The term “sensor head,” as used herein, is a broad rophages) from gaining proximity to, and thereby damaging, term and is used in its ordinary Sense, including, without the Sensing membrane or forming a barrier cell layer and limitation, the region of a monitoring device responsible for interfering with the transport of analyte acroSS the tissue the detection of a particular analyte. In one example, a device interface. Sensor head comprises a non-conductive body, a working 0277. In the disclosure which follows, the following electrode (anode), a reference electrode and a counter elec abbreviations apply: Eq and Eqs (equivalents); mEq (mil trode (cathode) passing through and Secured within the body liequivalents); M (molar); mM (millimolar) uM (micromo forming an electrochemically reactive Surface at one loca lar); N (Normal); mol (moles); mmol (millimoles); umol tion on the body and an electronic connective means at (micromoles); nmol (nanomoles); g (grams); mg (milli another location on the body, and a Sensing membrane grams); ug (micrograms); Kg (kilograms); L (liters); mL affixed to the body and covering the electrochemically (milliliters); dL (deciliters); ul (microliters); cm (centime reactive Surface. The counter electrode has a greater elec ters); mm (millimeters); um (micrometers); nm (nanom trochemically reactive Surface area than the working elec eters); hand hr (hours); min. (minutes); S and Sec. (Seconds); trode. During general operation of the Sensor a biological C. (degrees Centigrade). Sample (e.g., blood or interstitial fluid) or a portion thereof contacts (directly or after passage through one or more 0278). Overview membranes or domains) an enzyme (e.g., glucose oxidase); 0279 The preferred embodiments relate to the use of an the reaction of the biological Sample (or portion thereof) analyte Sensor that measures a concentration of analyte of results in the formation of reaction products that allow a interest or a Substance indicative of the concentration or determination of the analyte (e.g., glucose) level in the presence of the analyte. In Some embodiments, the Sensor is biological Sample. In Some embodiments, the Sensing mem a continuous device, for example a Subcutaneous, transder brane further comprises an enzyme domain (e.g., and mal, or intravascular device. In Some embodiments, the enzyme layer), and an electrolyte phase (e.g., a free-flowing device may analyze a plurality of intermittent blood liquid phase comprising an electrolyte-containing fluid Samples. The analyte Sensor may use any method of analyte described further below). Sensing, including enzymatic, chemical, physical, electro 0273. The term “electrochemically reactive surface,” as chemical, Spectrophotometric, polarimetric, calorimetric, used herein, is a broad term and is used in its ordinary Sense, radiometric, or the like. including, without limitation, the Surface of an electrode 0280 The analyte sensor uses any known method, where an electrochemical reaction takes place. In the case of including invasive, minimally invasive, and non-invasive the working electrode, the hydrogen peroxide produced by Sensing techniques, to provide an output Signal indicative of the enzyme catalyzed reaction of the analyte being detected the concentration of the analyte of interest. The output Signal creates a measurable electronic current (e.g., detection of is typically a raw signal that is used to provide a useful value analyte utilizing analyte oxidase produces H.O. peroxide as of the analyte of interest to a user, Such as a patient or a by product, HO reacts with the Surface of the working physician, who may be using the device. Accordingly, electrode producing two protons (2H), two electrons (2e) appropriate Smoothing, calibration, and evaluation methods and one molecule of oxygen (O) which produces the may be applied to the raw Signal and/or System as a whole electronic current being detected). In the case of the counter to provide relevant and acceptable estimated analyte data to electrode, a reducible species, e.g., O is reduced at the the user. electrode Surface in order to balance the current being generated by the working electrode. 0281) Sensor 0274 The term “electronic connection,” as used herein, is 0282. The analyte sensor useful with the preferred a broad term and is used in its ordinary Sense, including, embodiments may be any device capable of measuring the US 2005/0027181 A1 Feb. 3, 2005 concentration of an analyte of interest. One exemplary and therefore may be utilized to estimate a meaningful embodiment is described below, which utilizes an implant glucose value, Such as described elsewhere herein. able glucose Sensor. However, it should be understood that 0288 One problem of enzymatic glucose sensors such as the devices and methods described herein may be applied to described above is the non-glucose reaction rate-limiting any device capable of detecting a concentration of analyte of phenomenon. For example, if oxygen is deficient, relative to and providing an output Signal that represents the concen the amount of glucose, then the enzymatic reaction will be tration of the analyte. limited by oxygen rather than glucose. Consequently, the 0283 FIG. 1 is an exploded perspective view of a output signal will be indicative of the oxygen concentration glucose Sensor in one embodiment. The implantable glucose rather than the glucose concentration. Sensor 10 utilizes amperometric electrochemical Sensor 0289 FIG. 2 is a block diagram that illustrates the sensor technology to measure glucose. In this exemplary embodi electronics in one embodiment. In this embodiment, the ment, a body 12 and a head 14 house electrodes 16 and potentiostat 20 is shown, which is operatively connected to Sensor electronics, which are described in more detail with electrodes 16 (FIG. 1) to obtain a current value, and includes reference to FIG. 2. Three electrodes 16 are operably a resistor (not shown) that translates the current into Voltage. connected to the sensor electronics (FIG. 2) and are covered An A/D converter 21 digitizes the analog signal into counts by a sensing membrane 17 and a biointerface membrane 18, for processing. Accordingly, the resulting raw data Signal in which are attached by a clip 19. In alternative embodiments, counts is directly related to the current measured by the the number of electrodes may be less than or greater than potentiostat 20. three. 0290. A microprocessor 22 is the central control unit that 0284. The three electrodes 16, which protrude through houses EEPROM 23 and SRAM 24, and controls the pro the head 14, including a platinum working electrode, a cessing of the Sensor electronics. It may be noted that platinum counter electrode, and a Silver/silver chloride ref alternative embodiments utilize a computer System other erence electrode. The top ends of the electrodes are in than a microprocessor to process data as described herein. In contact with an electrolyte phase (not shown), which is a Some alternative embodiments, an application-specific inte free-flowing fluid phase disposed between the Sensing mem grated circuit (ASIC) may be used for some or all the brane and the electrodes. The sensing membrane 17 includes sensor's central processing. The EEPROM 23 provides an enzyme, e.g., glucose oxidase, which covers the electro Semi-permanent Storage of data, Storing data Such as Sensor lyte phase. In turn, the biointerface membrane 18 covers the ID and necessary programming to process data Signals (e.g., Sensing membrane 17 and Serves, at least in part, to protect programming for data Smoothing Such as described below). the Sensor from external forces that may result in environ The SRAM 24 is used for the system's cache memory, for mental StreSS cracking of the Sensing membrane 17. example for temporarily Storing recent Sensor data. 0285) In the illustrated embodiment, the counter elec 0291 Abattery 25 is operatively connected to the micro trode is provided to balance the current generated by the processor 22 and provides the necessary power for the Species being measured at the working electrode. In the case sensor. In one embodiment, the battery is a Lithium Man of a glucose oxidase based glucose Sensor, the Species being ganese Dioxide battery, however any appropriately sized measured at the working electrode is HO. Glucose oxidase and powered battery may be used (e.g., AAA, Nickel catalyzes the conversion of oxygen and glucose to hydrogen cadmium, Zinc-carbon, Alkaline, Lithium, Nickel-metal peroxide and gluconate according to the following reaction: hydride, Lithium-ion, Zinc-air, Zinc-mercury oxide, Silver Zinc, or hermetically-Sealed). In Some embodiments, a plu Glucose--O->Gluconate--HO, rality of batteries may be used to power the System. A Quartz 0286 The change in HO can be monitored to determine Crystal 26 is operatively connected to the microprocessor 22 glucose concentration because for each glucose molecule and maintains System time for the computer System as a metabolized, there is a proportional change in the product whole. HO. Oxidation of HO, by the working electrode is 0292 An RF Transceiver 27 is operably connected to the balanced by reduction of ambient oxygen, enzyme generated microprocessor 22 and transmits the Sensor data from the H2O, or other reducible Species at the counter electrode. sensor to a receiver (see FIGS. 4 and 5). Although an RF The HO produced from the glucose oxidase reaction transceiver is shown here, other embodiments include a further reacts at the Surface of working electrode and pro wired rather than wireleSS connection to the receiver. In yet duces two protons (2H), two electrons (2e), and one other embodiments, the receiver is transcutaneously pow oxygen molecule (O2) (See, e.g., Fraser, D. M. “An Intro ered via an inductive coupling, for example. A quartz crystal duction to In vivo Biosensing: Progress and problems.” In 28 provides the System time for Synchronizing the data “BioSensors and the Body,” D. M. Fraser, ed., 1997, pp. 1-56 transmissions from the RF transceiver. It may be noted that John Wiley and Sons, New York.) the transceiver 27 may be substituted for a transmitter in one 0287. In one embodiment, a potentiostat is used to mea embodiment. Sure the electrochemical reaction(s) at the electrode(s) (See FIG. 2). The potentiostat applies a constant potential 0293 Data Smoothing between the working and reference electrodes to produce a 0294 Typically, an analyte sensor produces a raw data current value. The current that is produced at the working Signal that is indicative of the analyte concentration of a electrode (and flows through the circuitry to the counter user, Such as described in more detail with reference to electrode) is proportional to the diffusional flux of H.O. FIGS. 1 and 2, above. However, it is well known that the Accordingly, a raw signal may be produced that is repre above described glucose Sensor is only one example of an Sentative of the concentration of glucose in the users body, abundance of analyte Sensors that are able to provide a raw US 2005/0027181 A1 Feb. 3, 2005

data Signal output indicative of the concentration of the In alternative embodiments, data Smoothing may be trans analyte of interest. Thus, it should be understood that the mitted from the Sensor to the receiver, and the data Smooth devices and methods of the preferred embodiments, includ ing performed at the receiver; it may be noted however that ing data Smoothing, calibration, evaluation, and other data there may be a risk of transmit-loSS in the radio transmission processing, may be applied to raw data obtained from any from the Sensor to the receiver when the transmission is analyte Sensor capable of producing a output signal. wireleSS. For example, in embodiments wherein a Sensor is 0295). It has been found that raw data signals received implemented in Vivo, the raw Sensor Signal may be more from an analyte Sensor include Signal noise, which degrades consistent within the Sensor (in vivo) than the raw signal the quality of the data. Thus, it has been known to use transmitted to a Source (e.g., receiver) outside the body (e.g., Smoothing algorithms help improve the Signal-to-noise ratio if a patient were to take the receiver off to shower, commu in the Sensor by reducing Signal jitter, for example. One nication between the Sensor and receiver may be lost and example of a conventional data Smoothing algorithms data Smoothing in the receiver would halt accordingly.) include finite impulse response filter (FIR), which is par Consequently, it may be noted that a multiple point data loSS ticularly Suited for reducing high-frequency noise (see Steil in the filter may take, for example, anywhere from 25 to 40 et al. U.S. Pat. No. 6,558,351). Other analyte sensors have minutes for the Smoothed data to recover to where it would utilized heuristic and moving average type algorithms to have been had there been no data loSS. accomplish data Smoothing of Signal jitter in data Signals, for 0302) Receiver example. 0303 FIGS. 4A to 4D are schematic views of a receiver 0296. It is advantageous to also reduce signal noise by in first, Second, third, and fourth embodiments, respectively. attenuating transient, low frequency, non-analyte related A receiver 40 comprises Systems necessary to receive, Signal fluctuations (e.g., transient ischemia and/or long process, and display Sensor data from an analyte Sensor, Such transient periods of postural effects that interfere with Sensor as described elsewhere herein. Particularly, the receiver 40 function due to lack of oxygen and/or other physiological may be a pager-sized device, for example, and comprise a effects). user interface that has a plurality of buttons 42 and a liquid 0297. In one embodiment, this attenuation of transient crystal display (LCD) Screen 44, and which may include a low frequency non-analyte related Signal noise is accom backlight. In Some embodiments the user interface may also plished using a recursive filter. In contrast to conventional include a keyboard, a speaker, and a vibrator Such as non-recursive (e.g., FIR) filters in which each computation described with reference to FIG. 5. uses new input data sets, a recursive filter is an equation that 0304 FIG. 4A illustrates a first embodiment wherein the uses moving averages as inputs; that is, a recursive filter receiver shows a numeric representation of the estimated includes previous averages as part of the next filtered output. analyte value on its user interface, which is described in Recursive filters are advantageous at least in part due to their more detail elsewhere herein. computational efficiency. 0305 FIG. 4B illustrates a second embodiment wherein 0298 FIG. 3 is a graph that illustrates data Smoothing of the receiver shows an estimated glucose value and one hour a raw data Signal in one embodiment. In this embodiment, of historical trend data on its user interface, which is the recursive filter is implemented as a digital infinite described in more detail elsewhere herein. impulse response filter (IIR) filter, wherein the output is computed using 6 additions and 7 multiplies as shown in the 0306 FIG.4C illustrates a third embodiment wherein the following equation: receiver shows an estimated glucose value and three hours of historical trend data on its user interface, which is described in more detail elsewhere herein. ao: x(n) -- a 3 x(n - 1) + a2 : x(n - 2) + as : x(n-3) - 0307 FIG. 4D illustrates a fourth embodiment wherein b : y (n-1)-bis y(n - 2) - bg : y (n-3) the receiver shows an estimated glucose value and nine bo hours of historical trend data on its user interface, which is described in more detail elsewhere herein. 0299. This polynomial equation includes coefficients that 0308. In some embodiments a user is able to toggle are dependent on Sample rate and frequency behavior of the through some or all of the screens shown in FIGS. 4A to 4D filter. In this exemplary embodiment, frequency behavior using a toggle button on the receiver. In Some embodiments, passes low frequencies up to cycle lengths of 40 minutes, the user is able to interactively Select the type of output displayed on their user interface. In Some embodiments, the and is based on a 30 Second Sample rate. Sensor output may have alternative configurations, Such as is 0300. In some embodiments, data Smoothing may be described with reference to FIG. 6, block 69, for example. implemented in the Sensor and the Smoothed data transmit ted to a receiver for additional processing. In other embodi 0309 FIG. 5 is a block diagram of the receiver electron ments, raw data may be sent from the Sensor to a receiver for ics in one embodiment. It may be noted that the receiver may data Smoothing and additional processing therein. In yet comprise a configuration Such as described with reference to FIGS. 4A to 4D, above. Alternatively, the receiver may other embodiments, the Sensor is integral with the receiver comprise any configuration, including a desktop computer, and therefore no transmission of data is required. laptop computer, a personal digital assistant (PDA), a server 0301 In one exemplary embodiment, wherein the sensor (local or remote to the receiver), or the like. In Some is an implantable glucose Sensor, data Smoothing is per embodiments, a receiver may be adapted to connect (via formed in the Sensor to ensure a continuous Stream of data. wired or wireless connection) to a desktop computer, laptop US 2005/0027181 A1 Feb. 3, 2005 computer, a PDA, a server (local or remote to the receiver), 0315 Communication ports, including a personal com or the like in order to download data from the receiver. In puter (PC) corn port 58 and a reference analyte monitor corn Some alternative embodiments, the receiver is housed within port 59 may be provided to enable communication with or directly connected to the Sensor in a manner that allows Systems that are separate from, or integral with, the receiver. Sensor and receiver electronics to work directly together The PC corn port 58 comprises means for communicating and/or share data processing resources. Accordingly, the with another computer system (e.g., PC, PDA, server, or the receiver, including its electronics, may be generally like). In one exemplary embodiment, the receiver is able to described as a “computer System.” download historic data to a physician’s PC for retrospective 0310. A quartz crystal 50 is operatively connected to an analysis by the physician. The reference analyte monitor RF transceiver 51 that together function to receive and corn port 59 comprises means for communicating with a Synchronize data Signals (e.g., raw data signals transmitted reference analyte monitor So that reference analyte values from the RF transceiver). Once received, the microprocessor may be automatically downloaded into the receiver. In one 52 processes the Signals, Such as described below. embodiment, the reference analyte monitor is integral with the receiver, and the reference analyte corn port 59 allows 0311. The microprocessor 52 is the central control unit internal communication between the two integral Systems. that provides the necessary processing, Such as calibration In another embodiment, the reference analyte monitor corn algorithms stored within an EEPROM 53. The EEPROM 53 port 59 allows a wireless or wired connection to the refer is operatively connected to the microprocessor 52 and ence analyte monitor Such as a Self-monitoring blood glu provides Semi-permanent Storage of data, Storing data Such cose monitor (e.g., for measuring finger Stick blood as receiver ID and necessary programming to proceSS data Samples). Signals (e.g., programming for performing calibration and other algorithms described elsewhere herein). In Some 0316 Algorithms embodiments, an application-Specific integrated circuit 0317 Reference is now made to FIG. 6, which is a flow (ASIC) may be used for some or all the receiver's central chart that illustrates the initial calibration and data output of processing. An SRAM 54 is used for the systems cache the Sensor data in one embodiment. memory and is helpful in data processing. 0318 Calibration of an analyte sensor comprises data 0312 The microprocessor 52, which is operatively con processing that converts Sensor data Signal into an estimated nected to EEPROM 53 and SRAM 54, controls the process analyte measurement that is meaningful to a user. Accord ing of the receiver electronics including, but not limited to, ingly, a reference analyte value is used to calibrate the data a Sensor data receiving module, a reference data receiving Signal from the analyte Sensor. module, a data matching module, a calibration Set module, 03.19. At block 61, a sensor data receiving module, also a conversion function module, a Sensor data transformation referred to as the Sensor data module, receives Sensor data module, a quality evaluation module, a interface control (e.g., a data stream), including one or more time-spaced module, and a Stability determination module, which are Sensor data points, from a Sensor via the receiver, which may described in more detail below. It may be noted that any of be in wired or wireless communication with the sensor. The the above processing may be programmed into and per Sensor data point(s) may be Smoothed, Such as described formed in the sensor electronics (FIG. 2) in place of, or in with reference to FIG.3, above. It may be noted that during complement with, the receiver electronics (FIG. 5). the initialization of the Sensor, prior to initial calibration, the 0313 A battery 55 is operatively connected to the micro receiver (e.g., computer System) receives and stores the processor 52 and provides the necessary power for the Sensor data, however may not display any data to the user receiver. In one embodiment, the battery is a AAA battery, until initial calibration and possibly stabilization of the however any appropriately sized and powered battery may Sensor has been determined. be used. In Some embodiments, a plurality of batteries may 0320 At block 62, a reference data receiving module, be used to power the System. A quartz crystal 56 is opera also referred to as the reference input module, receives tively connected to the microprocessor 52 and maintains reference data from a reference analyte monitor, including System time for the computer System as a whole. one or more reference data points. In one embodiment, the 0314. A user interface 57 comprises a keyboard, speaker, reference analyte points may comprise results from a Self vibrator, backlight, LCD, and a plurality of buttons. The monitored blood analyte test (e.g., from a finger Stick test). components that comprise the user interface 57 provide the In one Such embodiment, the user may administer a Self necessary controls to interact with the user. A keyboard may monitored blood analyte test to obtain an analyte value (e.g., allow, for example, input of user information about himself/ point) using any known analyte Sensor, and then enter the herself, Such as mealtime, exercise, insulin administration, numeric analyte value into the computer System. In another and reference analyte values. A Speaker may provide, for Such embodiment, a Self-monitored blood analyte test com example, audible signals or alerts for conditions Such as prises a wired or wireless connection to the receiver (e.g. present and/or predicted hyper- and hypoglycemic condi computer System) So that the user simply initiates a connec tions. A vibrator may provide, for example, tactile signals or tion between the two devices, and the reference analyte data alerts for reasons Such as described with reference to the is passed or downloaded between the self-monitored blood Speaker, above. A backlight may be provided, for example, analyte test and the receiver. In yet another Such embodi to aid the user in reading the LCD in low light conditions. ment, the Self-monitored analyte test is integral with the An LCD may be provided, for example, to provide the user receiver So that the user Simply provides a blood Sample to with Visual data output Such as described in more detail with the receiver, and the receiver runs the analyte test to deter reference to FIGS. 4A to 4D and FIG. 6. Buttons may mine a reference analyte value. provide toggle, menu Selection, option Selection, mode 0321. It may be noted that certain acceptability param Selection, and reset, for example. eters may be set for reference values received from the user. US 2005/0027181 A1 Feb. 3, 2005

For example, in one embodiment, the receiver may only 0326. At block 64, a calibration set module, also referred accept reference analyte values between about 40 and about to as the processor module, forms an initial calibration Set 400 mg/dL. Other examples of determining valid reference from a set of one or more matched data pairs, which are used analyte values are described in more detail with reference to to determine the relationship between the reference analyte FIG 8. data and the Sensor analyte data, Such as will be described 0322. At block 63, a data matching module, also referred in more detail with reference to block 67, below. to as the processor module, matches reference data (e.g., one 0327. The matched data pairs, which make up the initial or more reference analyte data points) with Substantially calibration Set, may be Selected according to predetermined time corresponding sensor data (e.g., one or more Sensor criteria. It may be noted that the criteria for the initial data points) to provide one or more matched data pairs. In calibration Set may be the same as, or different from, the one embodiment, one reference data point is matched to one criteria for the update calibration Set, which is described in time corresponding Sensor data point to form a matched data more detail with reference to FIG. 10. In some embodi pair. In another embodiment, a plurality of reference data ments, the number (n) of data pair(s) Selected for the initial points are averaged (e.g., equally or non-equally weighted calibration Set is one. In other embodiments, in data pairs are average, mean-value, median, or the like) and matched to Selected for the initial calibration Set wherein n is a function one time corresponding Sensor data point to form a matched of the frequency of the received reference data points. In one data pair. In another embodiment, one reference data point exemplary embodiment, Six data pairs make up the initial is matched to a plurality of time corresponding Sensor data calibration Set. points averaged to form a matched data pair. In yet another embodiment, a plurality of reference data points are aver 0328. In some embodiments, the data pairs are selected aged and matched to a plurality of time corresponding only within a certain analyte value threshold, for example Sensor data points averaged to form a matched data pair. wherein the reference analyte value is between about 40 and 0323 In one embodiment, a time corresponding sensor about 400 mg/dL. In some embodiments, the data pairs that data comprises one or more Sensor data points that occur form the initial calibration Set are Selected according to their 15-5 min after the reference analyte data timestamp (e.g., time Stamp. In Some embodiments, the calibration Set is the time that the reference analyte data is obtained). In this Selected Such as described with reference to FIG. 10 embodiment, the 15 minute time delay has been chosen to 0329. At block 65, a stability determination module, also account for an approximately 10 minute delay introduced by referred to as the Start-up module, determines the Stability of the filter used in data Smoothing and an approximately 5 the analyte Sensor over a period of time. It may be noted that minute physiological time-lag (e.g., the time necessary for Some analyte Sensors may have an initial instability time the analyte to diffusion through a membrane(s) of an analyte period during which the analyte Sensor is unstable for Sensor). In alternative embodiments, the time corresponding environmental, physiological, or other reasons. One Sensor value may be more or less than the above-described example of initial Sensor instability is an embodiment embodiment, for example +60 minutes. Variability in time wherein the analyte Sensor is implanted Subcutaneously; in correspondence of Sensor and reference data may be attrib this example embodiment, Stabilization of the analyte Sensor uted to, for example a longer or shorter time delay intro may be dependent upon the maturity of the tissue ingrowth duced by the data Smoothing filter, or if the configuration of around and within the Sensor. Another example of initial the analyte Sensor incurs a greater or lesser physiological Sensor instability is in an embodiment wherein the analyte time lag. Sensor is implemented transdermally, in this example 0324. It may be noted that in some practical implemen embodiment, Stabilization of the analyte Sensor may be tations of the Sensor, the reference analyte data may be dependent upon electrode Stabilization and/or Sweat, for obtained at a time that is different from the time that the data example. is input into the receiver. Accordingly, it should be noted that 0330. Accordingly, in some embodiments, determination the “time Stamp' of the reference analyte (e.g., the time at of Sensor Stability may include waiting a predetermined time which the reference analyte value was obtained) is not the period (e.g., an implantable Sensor is known to require a Same as the time at which the reference analyte data was time period for tissue, and a transdermal Sensor is known to obtained by receiver. Therefore, some embodiments include require time to equilibrate the Sensor with the user's skin); a time Stamp requirement that ensures that the receiver in Some embodiments, this predetermined waiting period is Stores the accurate time Stamp for each reference analyte between about one minute and about Six weeks. In Some value, that is, the time at which the reference value was embodiments, the Sensitivity (e.g., Sensor signal strength actually obtained from the user. with respect to analyte concentration) may be used to 0325 In some embodiments, tests are used to evaluate determine the Stability of the Sensor; for example, amplitude the best matched pair using a reference data point against and/or variability of Sensor Sensitivity may be evaluated to individual Sensor values over a predetermined time period determine the stability of the sensor. In alternative embodi (e.g., about 30 minutes). In one Such exemplary embodi ments, detection of pH levels, oxygen, hypochlorite, inter ment, the reference data point is matched with Sensor data fering Species (e.g., ascorbate, urea, and acetaminophen), points at 5-minute intervals and each matched pair is evalu correlation between sensor and reference values (e.g., ated. The matched pair with the best correlation may be R-value), baseline drift and/or offset, and the like may be Selected as the matched pair for data processing. In Some used to determine the Stability of the Sensor. In one exem alternative embodiments, matching a reference data point plary embodiment, wherein the Sensor is a glucose Sensor, it with an average of a plurality of Sensor data points over a is known to provide a Signal that is associated with inter predetermined time period may be used to form a matched fering Species (e.g., ascorbate, urea, acetaminophen), which pair. may be used to evaluate Sensor Stability. In another exem US 2005/0027181 A1 Feb. 3, 2005

plary embodiment, wherein the Sensor is a glucose Sensor forms of linear and non-linear regression, for example fuzzy Such as described with reference to FIGS. 1 and 2, the logic, neural networks, piece-wise linear regression, poly counter electrode can be monitored for oxygen deprivation, nomial fit, genetic algorithms, and other pattern recognition which may be used to evaluate Sensor Stability or function and Signal estimation techniques. ality. 0338. In yet other alternative embodiments, the conver 0331 At decision block 66, the system (e.g., micropro Sion function may comprise two or more different optimal cessor) determines whether the analyte Sensor is Sufficiently conversions because an optimal conversion at any time is Stable according to certain criteria, Such as described above. dependent on one or more parameters, Such as time of day, In one embodiment wherein the Sensor is an implantable calories consumed, exercise, or analyte concentration above glucose Sensor, the System waits a predetermined time or below a set threshold, for example. In one Such exemplary period believed necessary for Sufficient tissue ingrowth and embodiment, the conversion function is adapted for the evaluates the Sensor Sensitivity (e.g., between about one estimated glucose concentration (e.g., high VS. low). For minute and Six weeks). In another embodiment, the receiver example in an implantable glucose Sensor it has been determines Sufficient Stability based on oxygen concentra observed that the cells Surrounding the implant will consume tion near the Sensor head. In yet another embodiment, the at least a Small amount of glucose as it diffuses toward the Sensor determines Sufficient Stability based on a reassess glucose Sensor. ASSuming the cells consume Substantially ment of baseline drift and/or offset. In yet another alternative the Same amount of glucose whether the glucose concen embodiment, the System evaluates Stability by monitoring tration is low or high, this phenomenon will have a greater the frequency content of the Sensor data Stream over a effect on the concentration of glucose during low blood predetermined amount of time (e.g., 24 hours); in this Sugar episodes than the effect on the concentration of alternative embodiment, a template (or templates) are pro glucose during relatively higher blood Sugar episodes. Vided that reflect acceptable levels of glucose physiology Accordingly, the conversion function is adapted to compen and are compared with the actual Sensor data, wherein a sate for the sensitivity differences in blood sugar level. In predetermined amount of agreement between the template one implementation, the conversion function comprises two and the actual Sensor data is indicative of Sensor Stability. It different regression lines wherein a first regression line is may be noted that a few examples of determining Sufficient applied when the estimated blood glucose concentration is at Stability are given here, however a variety of known tests or below a certain threshold (e.g., 150 mg/dL) and a second and parameters may be used to determine Sensor Stability regression line is applied when the estimated blood glucose without departing from the spirit and Scope of the preferred concentration is at or above a certain threshold (e.g., 150 embodiments. mg/dL). In one alternative implementation, a predetermined 0332) If the receiver does not assess that the stability of pivot of the regression line that forms the conversion func the Sensor is Sufficient, then the processing returns to block tion may be applied when the estimated blood is above or 61, wherein the receiver receives Sensor data Such as below a set threshold (e.g., 150 mg/dL), wherein the pivot described in more detail above. The above-described steps and threshold are determined from a retrospective analysis are repeated until Sufficient Stability is determined. of the performance of a conversion function and its perfor mance at a range of glucose concentrations. In another 0333) If the receiver does assess that the stability of the implementation, the regression line that forms the conver Sensor is Sufficient, then processing continues to block 67 Sion function is pivoted about a point in order to comply and the calibration Set is used to calibrate the Sensor. with clinical acceptability standards (e.g., Clarke Error Grid, 0334. At block 67, the conversion function module uses Consensus Grid, mean absolute relative difference, or other the calibration Set to create a conversion function. The clinical cost function). Although only a few example imple conversion function Substantially defines the relationship mentations are described, the preferred embodiments con between the reference analyte data and the analyte Sensor template numerous implementations wherein the conversion data. function is adaptively applied based on one or more param 0335) A variety of known methods may be used with the eters that may affect the Sensitivity of the Sensor data over preferred embodiments to create the conversion function time. from the calibration Set. In one embodiment, wherein a 0339 Referring again to FIG. 6, at block 68, a sensor plurality of matched data points form the initial calibration data transformation module uses the conversion function to Set, a linear least Squares regression is performed on the transform Sensor data into Substantially real-time analyte initial calibration Set Such as described with reference to value estimates, also referred to as calibrated data, as Sensor FIG. 7. data is continuously (or intermittently) received from the 0336 FIG. 7 is a graph that illustrates a regression sensor. For example, in the embodiment of FIG. 7, the performed on a calibration Set to create a conversion func Sensor data, which may be provided to the receiver in tion in one exemplary embodiment. In this embodiment, a “counts”, is translated in to estimate analyte value(s) in linear least Squares regression is performed on the initial mg/dL. In other words, the offset value at any given point in calibration Set. The X-axis represents reference analyte data; time may be subtracted from the raw value (e.g., in counts) the y-axis represents Sensor data. The graph pictorially and divided by the Slope to obtain the estimate analyte value: illustrates regression of the matched pairs 76 in the calibra tion Set. Regression calculates a Slope 72 and an offset 74 (rawvalue - offset) (y=mx+b), which defines the conversion function. mgfdL = H slope . 0337. In alternative embodiments other algorithms could be used to determine the conversion function, for example US 2005/0027181 A1 Feb. 3, 2005

0340. In some alternative embodiments, the sensor and/or acceptability of reference and Sensor data in one embodi reference analyte values are Stored in a database for retro ment. Although Some clinical acceptability tests are dis Spective analysis. closed here, any known clinical Standards and methodolo 0341. At block 69, an output module provides output to gies may be applied to evaluate the clinical acceptability of the user via the user interface. The output is representative reference and analyte data herein. of the estimated analyte value, which is determined by 0348. It may be noted that the conventional analyte converting the Sensor data into a meaningful analyte value meters (e.g., Self-monitored blood analyte tests) are known Such as described in more detail with reference to block 68, to have a +-20% error in analyte values. For example, grOSS above. User output may be in the form of a numeric errors in analyte readings are known to occur due to patient estimated analyte value, an indication of directional trend of error in self-administration of the blood analyte test. In one analyte concentration, and/or a graphical representation of Such example, if the user has traces of Sugar on his/her finger the estimated analyte data over a period of time, for while obtaining a blood Sample for a glucose concentration example. Other representations of the estimated analyte test, then the measured glucose value will likely be much values are also possible, for example audio and tactile. higher than the actual glucose value in the blood. Addition 0342. In one exemplary embodiment, such as shown in ally, it is known that Self-monitored analyte tests (e.g., test FIG. 4A, the estimated analyte value is represented by a Strips) are occasionally Subject to manufacturing error. numeric value. In other exemplary embodiments, Such as 0349 Another cause for error includes infrequency and shown in FIGS. 4B to 4D, the user interface graphically time delay that may occur if a user does not Self-test represents the estimated analyte data trend over predeter regularly, or if a user Self-tests regularly but does not enter mined a time period (e.g., one, three, and nine hours, the reference value at the appropriate time or with the respectively). In alternative embodiments, other time peri appropriate time Stamp. Therefore, it may be advantageous ods may be represented. to validate the acceptability of reference analyte values prior 0343. In some embodiments, the user interface begins to accepting them as valid entries. Accordingly, the receiver displaying data to the user after the Sensor's Stability has evaluates the clinical acceptability of received reference been affirmed. In some alternative embodiments however, analyte data prior to their acceptance as a valid reference the user interface displayS data that is Somewhat unstable value. (e.g., does not have Sufficient stability at block 66); in these 0350. In one embodiment, the reference analyte data embodiments, the receiver may also include an indication of (and/or sensor analyte data) is evaluated with respect to instability of the Sensor data (e.g., flashing, faded, or another Substantially time corresponding Sensor data (and/or Sub indication of Sensor instability displayed on the user inter Stantially time corresponding reference analyte data) to face). In Some embodiments, the user interface informs the determine the clinical acceptability of the reference analyte user of the status of the stability of the sensor data. and/or Sensor analyte data. Clinical acceptability considers a 0344). Accordingly, after initial calibration of the sensor, deviation between time corresponding glucose measure and possibly determination of Stability of the Sensor data, ments (e.g., data from a glucose sensor and data from a real-time continuous analyte information may be displayed reference glucose monitor) and the risk (e.g., to the decision on the user interface So that the user may regularly and making of a diabetic patient) associated with that deviation proactively care for his/her diabetic condition within the based on the glucose value indicated by the Sensor and/or bounds Set by his/her physician. reference data. Evaluating the clinical acceptability of ref 0345. In alternative embodiments, the conversion func erence and Sensor analyte data, and controlling the user tion is used to predict analyte values at future points in time. interface dependent thereon, may minimize clinical risk. These predicted values may be used to alert the user of 0351. In one embodiment, the receiver evaluates clinical upcoming hypoglycemic or hyperglycemic events. Addi acceptability each time reference data is obtained. In another tionally, predicted values may be used to compensate for the embodiment, the receiver evaluates clinical acceptability time lag (e.g., 15 minute time lag. Such as described else after the initial calibration and Stabilization of the Sensor, where herein), So that an estimate analyte value displayed to Such as described with reference to FIG. 6, above. In some the user represents the instant time, rather than a time embodiments, the receiver evaluates clinical acceptability as delayed estimated value. an initial pre-Screen of reference analyte data, for example 0346. In some embodiments, the substantially real time after determining if the reference glucose measurement is estimated analyte value, a predicted future estimate analyte between about 40 and 400 mg/dL. In other embodiments, value, a rate of change, and/or a directional trend of the other methods of pre-Screening data may be used, for analyte concentration is used to control the administration of example by determining if a reference analyte data value is a constituent to the user, including an appropriate amount physiologically feasible based on previous reference analyte and time, in order to control an aspect of the user's biologi data values (e.g., below a maximum rate of change). cal System. One Such example is a closed loop glucose 0352. After initial calibration such as described in more Sensor and insulin pump, wherein the analyte data (e.g., detail with reference to FIG. 6, the sensor data receiving estimated glucose value, rate of change, and/or directional module 61 receives Substantially continuous Sensor data trend) from the glucose Sensor is used to determine the (e.g., a data stream) via a receiver and converts that data into amount of insulin, and time of administration, that may be estimated analyte values. AS used herein, “Substantially given to a diabetic user to evade hyper- and hypoglycemic continuous” is broad enough to include a data Stream of conditions. individual measurements taken at time intervals (e.g., time 0347 Reference is now made to FIG. 8, which is a flow Spaced) ranging from fractions of a Second up to, e.g., 1, 2, chart that illustrates the process of evaluating the clinical or 5 minutes. AS Sensor data is continuously converted, it US 2005/0027181 A1 Feb. 3, 2005 may be occasionally recalibrated Such as described in more 0357. In one exemplary embodiment, the clinical accept detail with reference FIG. 10. Initial calibration and re ability evaluation module 82 matches the reference data with calibration of the Sensor requires a reference analyte value. a Substantially time corresponding converted Sensor value Accordingly, the receiver may receive reference analyte data such as described with reference to FIG. 6 above, and plots at any time for appropriate processing. These reference the matched data on a Clarke Error Grid Such as described analyte values may be evaluated for clinical acceptability in more detail with reference to FIG. 9. Such as described below as a fail-safe against reference analyte test errors. 0358 FIG. 9 is a graph of two data pairs on a Clarke Error Grid to illustrate the evaluation of clinical acceptabil 0353 At block 81, the reference data receiving module, ity in one exemplary embodiment. The Clarke Error Grid also referred to as the reference input module, receives may be used by the clinical acceptability evaluation module reference analyte data from a reference analyte monitor. In to evaluate the clinical acceptability of the disparity between one embodiment, the reference data comprises one analyte a reference glucose value and a Sensor glucose (e.g., esti value obtained from a reference monitor. In Some alternative mated glucose) value, if any, in an embodiment wherein the embodiments however, the reference data includes a set of Sensor is a glucose Sensor. The X-axis represents glucose analyte values entered by a user into the interface and reference glucose data and the y-axis represents estimated averaged by known methods Such as described elsewhere glucose Sensor data. Matched data pairs are plotted accord herein. ingly to their reference and Sensor values, respectively. In 0354) In some embodiments, the reference data is pre this embodiment, matched pairs that fall within the A and B Screened according to environmental and physiological regions of the Clarke Error Grid are considered clinically issues, Such as time of day, oxygen concentration, postural acceptable, while matched pairs that fall within the C, D, and effects, and patient-entered environmental data. In one E regions of the Clarke Error Grid are not considered example embodiment, wherein the Sensor comprises an clinically acceptable. Particularly, FIG. 9 shows a first implantable glucose Sensor, an oxygen Sensor within the matched pair 92 is shown which falls within the A region of glucose Sensor is used to determine if Sufficient oxygen is the Clarke Error Grid, therefore is it considered clinically being provided to Successfully complete the necessary acceptable. A second matched pair 94 is shown which falls enzyme and electrochemical reactions for glucose Sensing. within the C region of the Clarke Error Grid, therefore it is In another example embodiment wherein the Sensor com not considered clinically acceptable. prises an implantable glucose Sensor, the counter electrode 0359. It may be noted that a variety of other known could be monitored for a “rail-effect', that is, when insuf methods of evaluation of clinical acceptability may be ficient oxygen is provided at the counter electrode causing utilized. In one alternative embodiment, the Consensus Grid the counter electrode to reach operational (e.g., circuitry) is used to evaluate the clinical acceptability of reference and limits. In yet another example embodiment, the patient is Sensor data. In another alternative embodiment, a mean prompted to enter data into the user interface, Such as meal absolute difference calculation may be used to evaluate the times and/or amount of exercise, which could be used to clinical acceptability of the reference data. In another alter determine likelihood of acceptable reference data. native embodiment, the clinical acceptability may be evalu 0355. It may be further noted that evaluation data, such as ated using any relevant clinical acceptability test, Such as a described in the paragraph above, may be used to evaluate known grid (e.g., Clarke Error or Consensus), and including an optimum time for reference analyte measurement. Cor additional parameterS Such as time of day and/or the increase respondingly, the user interface may then prompt the user to or decreasing trend of the analyte concentration. In another provide a reference data point for calibration within a given alternative embodiment, a rate of change calculation may be time period. Consequently, because the receiver proactively used to evaluate clinical acceptability. In yet another alter prompts the user during optimum calibration times, the native embodiment, wherein the received reference data is in likelihood of error due to environmental and physiological Substantially real time, the conversion function could be limitations may decrease and consistency and acceptability used to predict an estimated glucose value at a time corre of the calibration may increase. sponding to the time Stamp of the reference analyte value 0356. At block 82, the clinical acceptability evaluation (this may be required due to a time lag of the Sensor data module, also referred to as clinical module, evaluates the Such as described elsewhere herein). Accordingly, a thresh clinical acceptability of newly received reference data and/ old may be set for the predicted estimated glucose value and or time corresponding Sensor data. In Some embodiments of the reference analyte value disparity, if any. evaluating clinical acceptability, the rate of change of the 0360 Referring again to FIG. 8, the results of the clinical reference data as compared to previous data is assessed for acceptability evaluation are assessed. If clinical acceptabil clinical acceptability. That is, the rate of change and accel ity is determined with the received reference data, then eration (or deceleration) of many analytes has certain physi processing continues to block 84 to optionally recalculate ological limits within the body. Accordingly, a limit may be the conversion function using the received reference data in Set to determine if the new matched pair is within a physi the calibration Set. If, however, clinical acceptability is not ologically feasible range, indicated by a rate of change from determined, then the processing progresses to block 86 to the previous data that is within known physiological and/or control the user interface, Such as will be described with Statistical limits. Similarly, in Some embodiments any algo reference to block 86 below. rithm that predicts a future value of an analyte may be used to predict and then compare an actual value to a time 0361. At block 84, the conversion function module corresponding predicted value to determine if the actual optionally recreates the conversion function using the value falls within a clinically acceptable range based on the received reference data. In one embodiment, the conversion predictive algorithm, for example. function module adds the newly received reference data US 2005/0027181 A1 Feb. 3, 2005 22

(e.g., including the matched sensor data) into the calibration ing the calibration Set, and thus recalculating the conversion Set, displaces the oldest, and/or least concordant matched function, over time according to a set of inclusion criteria. data pair from the calibration Set, and recalculates the 0369. At block 101, the reference data receiving module, conversion function accordingly. In another embodiment, also referred to as the reference input module, receives a the conversion function module evaluates the calibration Set new reference analyte value (e.g., data point) from the for best calibration based on inclusion criteria, Such as reference analyte monitor. In Some embodiments, the refer described in more detail with reference to FIG. 10. ence analyte value may be pre-Screened according to criteria 0362 At 85, the sensor data transformation module uses Such as described in more detail with reference to FIG. 6, the conversion function to continually (or intermittently) block 62. In Some embodiments, the reference analyte value convert Sensor data into estimated analyte values, also may be evaluated for clinical acceptability Such as described referred to as calibrated data, Such as described in more in more detail with reference to FIG. 8. detail with reference to FIG. 6, block 68. 0370. At block 102, the data matching module, also 0363 At block 86, the interface control module, also referred to as the processor module, forms one or more referred to as the fail-safe module, controls the user interface updated matched data pairs by matching new reference data based upon the clinical acceptability of the reference data to Substantially time corresponding Sensor data, Such as received. If the evaluation (block 82) deems clinical accept described in more detail with reference to FIG. 6, block 63. ability, then the user interface may function as normal; that 0371. At block 103, a calibration evaluation module is, providing output for the user Such as described in more evaluates the new matched pair(s) inclusion into the cali detail with reference to FIG. 6, block 69. bration Set. In Some embodiments, the receiver simply adds 0364. If however the reference data is not considered the updated matched data pair into the calibration Set, clinically acceptable, then the fail-safe module begins the displaces the oldest and/or least concordant matched pair initial Stages of fail-safe mode. In Some embodiments, the from the calibration Set, and proceeds to recalculate the initial Stages of fail-Safe mode include altering the user conversion function accordingly (block 105). interface So that estimated Sensor data is not displayed to the 0372. In some embodiments, the calibration evaluation user. In Some embodiments, the initial Stages of fail-safe includes evaluating only the new matched data pair. In Some mode include prompting the user to repeat the reference embodiments, the calibration evaluation includes evaluating analyte test and provide another reference analyte value. The all of the matched data pairs in the existing calibration Set repeated analyte value is then evaluated for clinical accept and including the new matched data pair; in Such embodi ability such as described with reference to blocks 81 to 83, ments not only is the new matched data pair evaluated for above. inclusion (or exclusion), but additionally each of the data 0365. If the results of the repeated analyte test are deter pairs in the calibration Set are individually evaluated for mined to be clinically unacceptable, then fail-safe module inclusion (or exclusion). In Some alternative embodiments, may alter the user interface to reflect full fail-safe mode. In the calibration evaluation includes evaluating all possible one embodiment, full fail-Safe mode includes discontinuing combinations of matched data pairs from the existing cali Sensor analyte display output on the user interface. In other bration Set and including the new matched data pair to embodiments, color-coded information, trend information, determine which combination best meets the inclusion cri directional information (e.g., arrows or angled lines), teria. In Some additional alternative embodiments, the cali gauges, and/or fail-safe information may be displayed, for bration evaluation includes a combination of at least two of example. the above-described embodiments. 0366 If the results of the repeated analyte test are deter 0373) Inclusion criteria comprise one or more criteria that mined to be clinically acceptable, then the first analyte value define a set of matched data pairs that form a Substantially is discarded, and the repeated analyte value is accepted. The optimal calibration Set. One inclusion criterion comprises proceSS returns to block 84 to optionally recalculate the ensuring the time stamp of the matched data pairs (that make conversion function, Such as described in more detail with up the calibration Set) span at least a set time period (e.g., reference to block 84, above. three hours). Another inclusion criterion comprises ensuring that the time Stamps of the matched data pairs are not more 0367 Reference is now made to FIG. 10, which is a flow than a set age (e.g., one week old). Another inclusion chart that illustrates the process of evaluation of calibration criterion ensures that the matched pairs of the calibration Set data for best calibration based on inclusion criteria of have a substantially distributed amount of high and low raw matched data pairs in one embodiment. Sensor data, estimated Sensor analyte values, and/or refer 0368. It may be noted that calibration of analyte sensors ence analyte values. Another criterion comprises ensuring may be variable over time; that is, the conversion function all raw Sensor data, estimated Sensor analyte values, and/or Suitable for one point in time may not be Suitable for another reference analyte values are within a predetermined range point in time (e.g., hours, days, weeks, or months later). For (e.g., 40 to 400 mg/dL for glucose values). Another criterion example, in an embodiment wherein the analyte Sensor is comprises evaluating the rate of change of the analyte Subcutaneously implantable, the maturation of tissue concentration (e.g., from Sensor data) during the time stamp ingrowth over time may cause variability in the calibration of the matched pair(s). For example, Sensor and reference of the analyte Sensor. AS another example, physiological data obtained during the time when the analyte concentra changes in the user (e.g., metabolism, interfering blood tion is undergoing a slow rate of change may be leSS constituents, lifestyle changes) may cause variability in the Susceptible inaccuracies caused by time lag and other physi calibration of the Sensor. Accordingly, a continuously updat ological and non-physiological effects. Another criterion ing calibration algorithm is disclosed that includes reform comprises evaluating the congruence of respective Sensor US 2005/0027181 A1 Feb. 3, 2005 23 and reference data in each matched data pair; the matched from one or more matched data pairS Such as described in pairs with the most congruence may be chosen. Another more detail with reference to FIGS. 6, 8, and 10. criterion comprises evaluating physiological changes (e.g., low oxygen due to a user's posture that may effect the 0383 At block 115, the conversion function module function of a Subcutaneously implantable analyte Sensor, or calculates a conversion function using the calibration Set, other effects such as described with reference to FIG. 6) to such as described in more detail with reference to FIGS. 6, ascertain a likelihood of error in the Sensor value. It may be 8, and 10. noted that evaluation of calibration Set criteria may comprise 0384 At block 116, the sensor data transformation mod evaluating one, Some, or all of the above described inclusion ule continuously (or intermittently) converts received sensor criteria. It is contemplated that additional embodiments may data into estimated analyte values, also referred to as cali comprise additional inclusion criteria not explicitly brated data, Such as described in more detail with reference described herein. to FIGS. 6, 8, and 10. 0374. At block 104, the evaluation of the calibration set 0385 At block 117, a quality evaluation module evalu determines whether to maintain the previously established ates the quality of the calibration. In one embodiment, the calibration set, or if the calibration set should be updated quality of the calibration is based on the association of the (e.g., modified) with the new matched data pair. In Some calibration Set data using Statistical analysis. Statistical embodiments, the oldest matched data pair is simply dis analysis may comprise any known cost function Such as placed when a new matched data pair is included. It may be linear regression, non-linear mapping/regression, rank (e.g., noted however that a new calibration Set may include not non-parametric) correlation, least mean Square fit, mean only the determination to include the new matched data pair, absolute deviation (MAD), mean absolute relative differ but in Some embodiments, may also determine which of the ence, and the like. The result of the Statistical analysis previously matched data pairs should be displaced from the provides a measure of the association of data used in calibration Set. calibrating the System. A threshold of data association may 0375. At block 105, the conversion function module be set to determine if sufficient quality is exhibited in a recreates the conversion function using the modified cali calibration Set. bration Set. The calculation of the conversion function is 0386. In another embodiment, the quality of the calibra described in more detail with reference to FIG. 6. tion is determined by evaluating the calibration Set for clinical acceptability, Such as described with reference to 0376. At block 106, the sensor data transformation mod blocks 82 and 83 (e.g., Clarke Error Grid, Consensus Grid, ule converts Sensor data to calibrated data using the updated or clinical acceptability test). As an example, the matched conversion function. Conversion of raw Sensor data into data pairs that form the calibration Set may be plotted on a estimated analyte values is described in more detail with Clarke Error Grid, such that when all matched data pairs fall reference to FIG. 6. within the A and B regions of the Clarke Error Grid, then the 0377 Reference is now made to FIG. 11, which is a flow calibration is determined to be clinically acceptable. chart that illustrates the process of evaluating the quality of 0387. In yet another alternative embodiment, the quality the calibration in one embodiment. The calibration quality of the calibration is determined based initially on the asso may be evaluated by determining the Statistical association ciation of the calibration Set data using Statistical analysis, of data that forms the calibration set, which determines the and then by evaluating the calibration Set for clinical accept confidence associated with the conversion function used in ability. If the calibration set fails the statistical and/or the calibration and conversion of raw Sensor data into estimated clinical test, the processing returns to block 115 to recalcu analyte values. late the conversion function with a new (e.g., optimized) set 0378. In one embodiment calibration quality may be of matched data pairs. In this embodiment, the processing evaluated after initial or updated calculation of the conver loop (block 115 to block 117) iterates until the quality Sion function Such as described elsewhere herein. However evaluation module 1) determines clinical acceptability, 2) it may be noted that calibration quality may be performed at determines Sufficient Statistical data association, 3) deter any time during the data processing. mines both clinical acceptability and Sufficient Statistical 0379 At block 111, a sensor data receiving module, also data association, or 4) Surpasses a threshold of iterations; referred to as the Sensor data module, receives the Sensor after which the processing continues to block 118. data from the Sensor Such as described in more detail with 0388 FIGS. 12A and 12B illustrate one exemplary reference to FIG. 6. embodiment wherein the accuracy of the conversion func 0380 At block 112, a reference data receiving module, tion is determined by evaluating the correlation coefficient also referred to as the reference input module, receives from linear regression of the calibration Set that formed the reference data from a reference analyte monitor, Such as conversion function. In this exemplary embodiment, a described in more detail with reference to FIG. 6. threshold (e.g., 0.79) is set for the R-value obtained from the correlation coefficient. 0381 At block 113, the data matching module, also 0389 FIGS. 12A and 12B are graphs that illustrate an referred to as the processor module, matches received ref evaluation of the quality of calibration based on data asso erence data with Substantially time corresponding Sensor ciation in one exemplary embodiment using a correlation data to provide one or more matched data pairs, Such as coefficient. Particularly, FIGS. 12A and 12B pictorially described in more detail with reference to FIG. 6. illustrate the results of the linear least Squares regression 0382. At block 114, the calibration set module, also performed on a first and a second calibration set (FIGS. 12A referred to as the processor module, forms a calibration Set and 12B, respectively). The X-axis represents reference US 2005/0027181 A1 Feb. 3, 2005 24 analyte data; the y-axis represents Sensor data. The graph analyte value. The updated analyte value is then processed pictorially illustrates regression that determines the conver as described above and the updated conversion function that Sion function. results from the repeated reference analyte test, if any, is 0390. It may be noted that the regression line (and thus evaluated for Statistical accuracy. the conversion function) formed by the regression of the first 0397) If the results of the updated evaluation again calibration set of FIG. 12A is the same as the regression line exhibit insufficient quality, then the fail-safe module alters (and thus the conversion function) formed by the regression user interface to reflect full fail-safe mode, which is of the second calibration set of FIG. 12B. However, the described in more detail with reference to FIG. 8. If how correlation of the data in the calibration Set to the regression ever the results of the updated evaluation exhibit sufficient line in FIG. 12A is significantly different than the correla quality, then the first reference analyte value is discarded, tion of the data in the calibration Set to the regression line in and the repeated reference analyte value is accepted and the FIG. 12A. In other words, there is a noticeably greater process continues as described herein. deviation of the data from the regression line in FIG. 12B 0398. It may be noted that the initial stages of fail-safe than the deviation of the data from the regression line in mode and full fail safe mode may be similar to that described FIG. 12A. with reference to FIG. 8, including user interface control for 0391) In order to quantify this difference in correlation, example. Additionally, it is contemplated herein that a an R-Value may be used to Summarize the residuals (e.g., variety of difference modes between initial and full fail-safe root mean Square deviations) of the data when fitted to a mode may be provided depending on the relative quality of Straight line via least Squares method, in this exemplary the calibration. In other words, the confidence level of the embodiment. R-value may be calculated according to the calibration quality may control a plurality of different user following equation: interface Screens providing error bars, tvalues, and the like. Similar Screens may be implements in the clinical accept ability embodiments described with reference to FIG. 8. 0399. The above description discloses several methods R= and materials of the disclosed invention. This invention is Susceptible to modifications in the methods and materials, as well as alterations in the fabrication methods and equipment. Such modifications will become apparent to those skilled in the art from a consideration of this disclosure or practice of the invention disclosed herein. Consequently, it is not 0392) In the above equation: i is an index (1 to n), X is a intended that this invention be limited to the specific reference analyte value, y is a sensor analyte value, X is an embodiments disclosed herein, but that it cover all modifi average of 1/n reference analyte values, and y is an average cations and alternatives coming within the true Scope and of 1/n Sensor analyte values. Spirit of the invention as embodied in the attached claims. 0393. In the exemplary calibration set shown in FIG. All patents, applications, and other references cited herein 12A, the calculated R-value is about 0.99, which may also are hereby incorporated by reference in their entirety. be expressed as the correlation coefficient of regression. Accordingly, the calibration exhibits Sufficient data associa What is claimed is: tion (and thus insufficient quality) because it falls above the 1. A method for maintaining calibration of a Substantially 0.79 threshold set in this exemplary embodiment. continuous analyte Sensor, the method comprising: 0394. In the exemplary calibration set shown in FIG. receiving a data Stream from an analyte Sensor, including 12B, the calculated R-value is about 0.77, which may also one or more Sensor data points, be expressed as the correlation coefficient of regression. receiving reference data from a reference analyte monitor, Accordingly, the calibration exhibits insufficient data asso including two or more reference data points; ciation (and thus insufficient quality) because it falls below providing at least two matched data pairs by matching the 0.79 threshold set in this exemplary embodiment. reference analyte data to Substantially time correspond 0395 Reference is again made to FIG. 11, at block 118, ing Sensor data, the interface control module, also referred to as the fail-safe forming a calibration Set including Said at least two module, controls the user interface based upon the quality of matching data pairs; the calibration. If the calibration is exhibits sufficient quality, then the user interface may function as normal; that is creating a conversion function based on Said calibration providing output for the user Such as described in more Set, detail with reference to FIG. 6. converting Sensor data into calibrated data using Said 0396 If however the calibration is not deemed sufficient conversion function; in quality, then fail-safe module 118 begins the initial Stages Subsequently obtaining one or more additional reference of fail-safe mode, which are described in more detail with data points and creating one or more new matched data reference to FIG. 8. In some embodiments, the initial stages pairs, of fail-safe mode include altering the user interface So that estimated Sensor data is not displayed to the user. In Some evaluating Said calibration Set when Said new matched embodiments, the initial Stages of fail-safe mode also data pair is created, wherein evaluating Said calibration include prompting the user to provide an updated reference Set includes at least one of 1) ensuring matched data US 2005/0027181 A1 Feb. 3, 2005 25

pairs in Said calibration Set span a predetermined time 18. The method of claim 17, wherein the step of deter range, 2) ensuring matched data pairs in Said calibra mining a value for n is determined as a function of the tion set are no older than a predetermined value, 3) frequency of the received reference data points and Signal ensuring Said calibration Set has Substantially distrib Strength over time. uted high and low matched data pairs over Said prede 19. The method of claim 1, further comprising determin termined time range, and 4) allowing matched data ing a set of matching data pairs from Said evaluation of Said pairs only within a predetermined range of analyte calibration Set and re-forming a calibration Set. values, and 20. The method of claim 19, further comprising repeating the Step of re-creating Said conversion function using Said Subsequently modifying Said calibration Set if Such modi re-formed calibration Set. fication is required by Said evaluation. 21. The method of claim 20, further comprising convert 2. The method of claim 1, wherein the Step of evaluating ing Sensor data into calibrated data using Said re-created Said calibration Set further comprises at least one of evalu conversion function. ating a rate of change of the analyte concentration, evalu 22. A System for maintaining calibration of a Substantially ating a congruence of respective Sensor and reference data in continuous analyte Sensor, the System comprising: Said matched data pairs, and evaluating physiological changes. means for receiving a data Stream from an analyte Sensor, 3. The method of claim 1, wherein the step of evaluating a plurality of time-spaced Sensor data points; Said calibration Set includes evaluating only Said new means for receiving reference data from a reference matched data pair. analyte monitor, including two or more reference data 4. The method of claim 1, wherein the step of evaluating points; Said calibration Set includes evaluating all of the matched means for providing two or more matched data pairs by data pairs in Said calibration Set and Said new matched data matching reference analyte data to Substantially time pair. corresponding Sensor data; 5. The method of claim 1, wherein the step of evaluating Said calibration Set includes evaluating combinations of means for forming a calibration Set including at least two matched data pairs from the calibration Set and Said new matched data pair; matched data pair. means for creating a conversion function based on Said 6. The method of claim 1, wherein the step of receiving calibration set; Sensor data compriseS receiving a data Stream from a long term implantable analyte Sensor. means for converting Sensor data into calibrated data 7. The method of claim 1, wherein the step of receiving using Said conversion function; Sensor data compriseS receiving a data Stream that has been Subsequently obtaining one or more additional reference algorithmically Smoothed. data points and creating one or more new matched data 8. The method of claim 1, wherein the step of receiving pairs, Sensor data Stream comprises algorithmically Smoothing means for evaluating Said calibration Set when Said new Said data Stream. matched data pair is created, wherein evaluating Said 9. The method of claim 1, wherein the step of receiving calibration set includes at least one of 1) ensuring reference data comprises downloading reference data via a matched data pairs in Said calibration Set span a pre cabled connection. determined time range, 2) ensuring matched data pairs 10. The method of claim 1, wherein the step of receiving in Said calibration Set are no older than a predetermined reference data comprises downloading reference data via a value, 3) ensuring said calibration set has Substantially wireleSS connection. distributed high and low matched data pairs over Said 11. The method of claim 1, wherein the step of receiving predetermined time range, and 4) allowing matched reference data from a reference analyte monitor comprises data pairs only within a predetermined range of analyte receiving within a receiver internal communication from a values, and reference analyte monitor integral with Said receiver. means for modifying Said calibration Set if Such modifi 12. The method of claim 1, wherein the reference analyte cation is required by Said evaluation. monitor comprises Self-monitoring of blood analyte. 23. The system of claim 22, wherein said means for 13. The method of claim 1, wherein the step of creating evaluating Said calibration Set further comprises at least one a conversion function comprises linear regression. of means for evaluating a rate of change of the analyte concentration, means for evaluating a congruence of respec 14. The method of claim 1, wherein the step of creating tive Sensor and reference data in matched data pairs, and a conversion function comprises non-linear regression. means for evaluating physiological changes. 15. The method of claim 1, wherein the step of forming 24. The system of claim 22, wherein said means for a calibration Set comprises including in Said calibration Set evaluating Said calibration Set includes means for evaluating between one and six matched data pairs. only Said one or more new matched data pairs. 16. The method of claim 1, wherein the step of forming 25. The system of claim 22, wherein said means for a calibration Set comprises including six matched data pairs. evaluating Said calibration Set includes means for evaluating 17. The method of claim 1, wherein the step of forming all of the matched data pairs in Said calibration Set and Said a calibration Set further comprises determining a value for n, one or more new matched data pairs. where n is greater than one and represents the number of 26. The system of claim 22, wherein said means for matched data pairs in the calibration Set. evaluating Said calibration Set includes means for evaluating US 2005/0027181 A1 Feb. 3, 2005 26 combinations of matched data pairs from the calibration Set a data matching module that forms two or more matched and Said one or more new matched data pair. data pairs by matching reference data to Substantially 27. The system of claim 22, wherein said means for time corresponding Sensor data; receiving Sensor data comprises means for receiving Sensor data from a long-term implantable analyte Sensor. a calibration Set module that forms a calibration Set 28. The system of claim 22, wherein said means for including at least two matched data pairs, receiving Sensor data comprises means for receiving Sensor a conversion function module that creates a conversion data that has been algorithmically Smoothed. function using Said calibration Set; 29. The system of claim 22, wherein said means for receiving Sensor data comprises means for algorithmically a Sensor data transformation module that converts Sensor Smoothing Said receiving Sensor data. data into calibrated data using Said conversion function; 30. The system of claim 22, wherein said means for and receiving reference data comprises means for downloading a calibration evaluation module that evaluates Said cali reference data via a cabled connection. 31. The system of claim 22, wherein said means for bration Set when Said new matched data pair is pro receiving reference data comprises means for downloading vided, wherein evaluating Said calibration Set includes reference data via a wireleSS connection. at least one of 1) ensuring matched data pairs in Said 32. The system of claim 22, wherein said means for calibration Set span a predetermined time period, 2) receiving reference data from a reference analyte monitor ensuring matched data pairs in Said calibration Set are comprises means for receiving within a receiver internal no older than a predetermined value, 3) ensuring said communication from a reference analyte monitor integral calibration Set has Substantially distributed high and with said receiver. low matched data pairs over a predetermined time 33. The system of claim 22, wherein said means for range, and 4) allowing matched data pairs only within receiving reference data comprises means for receiving from a predetermined range of analyte values, a Self-monitoring of blood analyte. wherein Said conversion function module is programmed 34. The system of claim 22, wherein said means for to re-create Said conversion function of Such modifi creating a conversion function comprises means for per cation is required by Said calibration evaluation mod forming linear regression. ule. 35. The system of claim 22, wherein said means for 44. The computer System of claim 43, wherein said creating a conversion function comprises means for per evaluation calibration module further evaluates at least one forming non-linear regression. of a rate of change of the analyte concentration, a congru 36. The system of claim 22, wherein said means for ence of respective Sensor and reference data in matched data forming a calibration Set comprises including in Said cali pairs; and physiological changes. bration Set between one and Six matched data pairs. 45. The computer system of claim 43, wherein said 37. The system of claim 22, wherein said means for evaluation calibration module evaluates only Said new forming a calibration Set comprises including in Said cali matched data pair. bration Set Six matched data pairs. 46. The computer system of claim 43, wherein said 38. The system of claim 22, wherein the means for evaluation calibration module evaluates all of the matched forming a calibration Set further comprises determining a data pairs in Said calibration Set and Said new matched data value for n, where n is greater than one and represents the pair. number of matched data pairs in the calibration Set. 47. The computer system of claim 43, wherein said 39. The system of claim 38, wherein the means for evaluation calibration module evaluates combinations of determining a value for n is determined as a function of the matched data pairs from the calibration Set and Said new frequency of the received reference data points and Signal matched data pair. Strength over time. 48. The computer system of claim 43, wherein said sensor 40. The system of claim 22, further comprising means for data receiving module receives Said data Stream from a determining a Set of matching data pairs from Said evalua long-term implantable analyte Sensor. tion of Said calibration Set and re-forming a calibration Set. 41. The system of claim 40, further comprising said 49. The computer system of claim 43, wherein said sensor means for repeating the Set of creating Said conversion data receiving module receives an algorithmically Smoothed function using Said re-formed calibration Set. data Stream. 42. The System of claim 41, further comprising means for 50. The computer system of claim 43, wherein said sensor converting Sensor data into calibrated data using Said re data receiving module comprises programming to Smooth created conversion function. Said data Stream. 43. A computer System for maintaining calibration of a 51. The computer system of claim 43, wherein said Substantially continuous analyte Sensor, the computer Sys reference data receiving module downloads reference data tem comprising: via a cabled connection. 52. The computer system of claim 43, wherein said a Sensor data receiving module that receives a data Stream reference data receiving module downloads reference data comprising a plurality of time Spaced Sensor data points via a wireleSS connection. from a Substantially continuous analyte Sensor; 53. The computer system of claim 43, wherein said a reference data receiving module that receives reference reference data receiving module receives within a receiver data from a reference analyte monitor, including two or internal communication from a reference analyte monitor more reference data points; integral with Said receiver. US 2005/0027181 A1 Feb. 3, 2005 27

54. The computer system of claim 43, wherein said creating a conversion function based on Said calibration reference data receiving module receives reference data Set, from a Self-monitoring of blood analyte. Subsequently obtaining one or more additional reference 55. The computer system of claim 43, wherein said data points and creating one or more new matched data conversion function module comprises programming that performs linear regression. pairs; and 56. The computer system of claim 43, wherein said evaluating Said calibration Set when Said new matched conversion function module comprises programming that data pair is created, wherein evaluating Said calibration performs non-linear regression. Set includes at least one of 1) ensuring matched data 57. The computer system of claim 43, wherein said pairs in Said calibration Set span a predetermined time calibration Set module includes in Said calibration Set range, 2) ensuring matched data pairs in said calibra between one and six matched data pairs. tion set are no older than a predetermined value, 3) 58. The computer system of claim 43, wherein said ensuring Said calibration Set has Substantially distrib calibration Set module includes in Said calibration Set Six uted high and low matched data pairs over Said prede matched data pairs. termined time range, and 4) allowing matched data 59. The computer system of claim 43, wherein the cali pairs only within a predetermined range of analyte bration Set module further comprises programming for values. determining a value for n, where n is greater than one and 65. A computer System for maintaining calibration of a represents the number of matched data pairs in the calibra glucose Sensor, the computer System comprising: tion Set. a Sensor data module that receives a data Stream com 60. The computer system of claim 59, wherein said prising a plurality of time Spaced Sensor data points programming for determining a value for n determines n as from a Substantially continuous analyte Sensor, a function of the frequency of the received reference data points and Signal Strength over time. a reference input module that receives reference data from 61. The computer system of claim 43, wherein data a reference analyte monitor, including two or more matching module further comprises programming to re-form reference data points; Said calibration Set based on Said calibration evaluation. a processor module that forms two or more matched data 62. The computer system of claim 61, wherein said pairs by matching reference data to Substantially time conversion function module further comprises programming corresponding Sensor data and Subsequently forms a to re-create Said conversion function based on Said re calibration set including said two or more matched data formed calibration set. pairs; and 63. The computer System of claim 62, Sensor data trans a calibration evaluation module that evaluates Said cali formation module further comprising programming for con bration Set when Said new matched data pair is pro Verting Sensor data into calibrated using Said re-created vided, wherein evaluating Said calibration Set includes conversion function. at least one of 1) ensuring matched data pairs in Said 64. A method for maintaining calibration of a glucose calibration Set span a predetermined time period, 2) Sensor, the method comprising: ensuring matched data pairs in Said calibration Set are receiving a data Stream from an analyte Sensor, including no older than a predetermined value, 3) ensuring said calibration Set has Substantially distributed high and one or more Sensor data points, low matched data pairs over a predetermined time receiving reference data from a reference analyte monitor, range, and 4) allowing matched data pairs only within including two or more reference data points; a predetermined range of analyte values, providing at least two matched data pairs by matching wherein Said conversion function module is programmed reference analyte data to Substantially time correspond to re-create Said conversion function of Such modifi ing Sensor data; cation is required by Said calibration evaluation mod ule. forming a calibration Set including Said at least two matching data pairs;