ADaM BDS for average glucose measurements

Illustrates Domains

Illustrates Variables

LBSPEC
LBANMETH
LBTEST
LBTESTCD

Illustrates Data Structures

BDS

Content

This example illustrates raw glucose levels based on plasma equivalents obtained from a subject's CGM. These values are used in the derivation of analysis endpoints. In this study, the sponsor captured the raw glucose levels for a 2-week baseline period at the start of the study, and a 2-week period at the end of the study. Only example rows of the data are shown below. Glucose values were typically measured every 5 minutes. The sponsor decided not to populate VISITNUM for these data. However, VISITNUM is included in the dataset because it is an expected variable.

lb.xpt

Rows 1-4:	Show the plasma-equivalent glucose values that were obtained from the subject's CGM device about every 5 minutes. Only the first 4 records are shown for the baseline period. SPDEVID is used to indicate the device from which data were obtained. LBMETHOD was represented as "CALCULATION" and LBANMETH was represented as "DEVICE ALGORITHM". The sponsor can provide more details on the algorithm in the Device in Use (DU) domain.
Rows 5-7:	Show the plasma-equivalent glucose values that were obtained from the subject's CGM on the last day of the collection period at week 52. Only a few records are shown.

Rows 1-4:

Show the plasma-equivalent glucose values that were obtained from the subject's CGM device about every 5 minutes. Only the first 4 records are shown for the baseline period. SPDEVID is used to indicate the device from which data were obtained. LBMETHOD was represented as "CALCULATION" and LBANMETH was represented as "DEVICE ALGORITHM". The sponsor can provide more details on the algorithm in the Device in Use (DU) domain.

Rows 5-7:

Show the plasma-equivalent glucose values that were obtained from the subject's CGM on the last day of the collection period at week 52. Only a few records are shown.

lb.xpt

Row	STUDYID	DOMAIN	USUBJID	SPDEVID	LBSEQ	LBTESTCD	LBTEST	LBORRES	LBORRESU	LBSTRESC	LBSTRESN	LBSTRESU	LBSPEC	LBMETHOD	LBANMETH	LBDTC	LBDY
1	TD011	LB	001-001	CGM-001	1	GLUCPE	Plasma Equivalent Glucose	118	mg/dL	118	118	mg/dL	INTERSTITIAL FLUID	CALCULATION	DEVICE ALGORITHM	2016-09-08 T10:09:33	7
2	TD011	LB	001-001	CGM-001	2	GLUCPE	Plasma Equivalent Glucose	117	mg/dL	117	117	mg/dL	INTERSTITIAL FLUID	CALCULATION	DEVICE ALGORITHM	2016-09-08 T10:14:36	7
3	TD011	LB	001-001	CGM-001	3	GLUCPE	Plasma Equivalent Glucose	110	mg/dL	110	110	mg/dL	INTERSTITIAL FLUID	CALCULATION	DEVICE ALGORITHM	2016-09-08 T10:19:36	7
4	TD011	LB	001-001	CGM-001	4	GLUCPE	Plasma Equivalent Glucose	90	mg/dL	90	90	mg/dL	INTERSTITIAL FLUID	CALCULATION	DEVICE ALGORITHM	2016-09-08 T10:24:52	7
5	TD011	LB	001-001	CGM-001	4033	GLUCPE	Plasma Equivalent Glucose	107	mg/dL	107	107	mg/dL	INTERSTITIAL FLUID	CALCULATION	DEVICE ALGORITHM	2017-09-06 T10:09:33	369
6	TD011	LB	001-001	CGM-001	4034	GLUCPE	Plasma Equivalent Glucose	106	mg/dL	106	106	mg/dL	INTERSTITIAL FLUID	CALCULATION	DEVICE ALGORITHM	2017-09-06 T10:14:20	369
7	TD011	LB	001-001	CGM-001	4035	GLUCPE	Plasma Equivalent Glucose	100	mg/dL	100	100	mg/dL	INTERSTITIAL FLUID	CALCULATION	DEVICE ALGORITHM	2017-09-06 T10:19:03	369

This example uses ADaM BDS to model parameters for average glucose measurements using 3 specific time spans relative to the 2 study timepoints. It is not intended to illustrate every possible variable that might be included in datasets created for statistical analysis of CGM data, nor is it intended to serve as a recommendation for which time spans a sponsor should select when performing CGM base change from baseline analysis.

The example illustrates how some variables are derived from the glucose values downloaded from CGMs during a study. In this study, the sponsor was interested in comparing the change from the initial 2-week baseline period to the last 2-week post-baseline period at the end of the 52-week study. The 2-week periods are anchored by the LBDTC value chosen to anchor the period. There is an LBDTC value chosen to anchor the baseline period and and an LBDTC value chosen to anchor the week 52 period. This LBDTC value is used to determine the AEENDTM value used to define the analysis interval end; ASTDTM is derived by calculating the interval value described in PARAM prior to AENDTM.

The metadata and derivations presented are for illustrative purposes only. Refer to ADaM v2.1 and Version 1.1 of the ADaM Implementation Guide (ADaMIG) for required background about ADaM and the ADaM data structures.

This is an example ADCGM dataset. It is based on an SDTM LB dataset containing the observations per subject per parameter per visit per assessment.

adcgm.xpt

adcgm.xpt

Row	STUDYID	USUBJID	SPDEVID	TRT01P	PARAMN	PARAMCD	PARAM	AVISITN	AVISIT	AVAL	BASE	CHG	ABLFL	ASTDTM	AENDTM
1	TD011	001-001	CGM-001	DRUG A	1	AVGGL24H	Average Glucose (mg/dL) 24 Hours Prior to the End of the Analysis Interval	0	Baseline	113.7	113.7	.	Y	07OCT2016:13:10:33	08OCT2016:13:10:33
2	TD011	001-001	CGM-001	DRUG A	1	AVGGL24H	Average Glucose (mg/dL) 24 Hours Prior to the End of the Analysis Interval	52	Week 52	107.7	113.7	-6.2		07OCT2017:13:15:00	08OCT2017:13:1500
3	TD011	001-001	CGM-001	DRUG A	2	AVGGL4H	Average Glucose (mg/dL) 24 to 20 Hours Prior to the End of the Analysis Interval	0	Baseline	112.5	112.5	.	Y	07OCT2016:13:10:33	07OCT2016:17:10:33
4	TD011	001-001	CGM-001	DRUG A	2	AVGGL4H	Average Glucose (mg/dL) 24 to 20 Hours Prior to the End of the Analysis Interval	52	Week 52	107.8	112.5	-4.7		07OCT2017:13:15:00	07OCT2017:17:15:00
5	TD011	001-001	CGM-001	DRUG A	3	AVGGL2W	Average Glucose (mg/dL) 2 Weeks Prior to the End of the Analysis Interval	0	Baseline	104.0	104.0	.	Y	24SEP2016:13:10:33	08OCT2016:13:10:33
6	TD011	001-001	CGM-001	DRUG A	3	AVGGL2W	Average Glucose (mg/dL) 2 Weeks Prior to the End of the Analysis Interval	52	Week 52	101.6	104.0	-2.4		24SEP2017:13:10:33	08OCT2017:13:15:00

The following is an example of the ADCGM metadata for the analysis of the change from baseline for the endpoints defined in the study protocol. A BDS structure was used. The BDS dataset contains 1 or more records per subject, per analysis parameter, per analysis timepoint.

ADCGM Dataset Metadata

Dataset	Description	Class	Structure	Purpose	Keys	Location	Documentation
ADCGM	Analysis of Average Glucose using CGM Data	BASIC DATA STRUCTURE	One record per subject per parameter per visit.	Analysis	STUDYID, USUBJID, PARAMCD, AVISITN	ADCGM.xpt	ADCGM.SAS/SAP

ADCGM Variable Metadata

Variable Name	Variable Label	Type	Codelist/Controlled Terms	Source/Derivation/Comment
STUDYID	Study Identifier	text		ADSL.STUDYID
USUBJID	Unique Subject Identifier	text		ADSL.USUBJID
SPDEVID	Device Identifier	text		ADSL.SPDEVID
TRT01P	Planned Treatment for Period 01	text		ADSL.TRT01P
PARAM	Parameter	text	Average Glucose (mg/dL) 24 Hours Prior to the End of the Analysis Interval Average Glucose (mg/dL) 24 to 20 Hours Prior to the End of the Analysis Interval Average Glucose (mg/dL) 2 Weeks Prior to the End of the Analysis Interval	Assigned
PARAMCD	Parameter Code	text	AVGGL24H (= Average Glucose (mg/dL) 24 Hours Prior to the End of the Analysis Interval) AVGGL4H (= Average Glucose (mg/dL) 24 to 20 Hours Prior to the End of the Analysis Interval) AVGGL2W (= Average Glucose (mg/dL) 2 Weeks Prior to the End of the Analysis Interval)	Assigned
PARAMN	Parameter (N)	num	1 (= Average Glucose (mg/dL) 24 Hours Prior to the End of the Analysis Interval) 2 (= Average Glucose (mg/dL) 24 to 20 Hours Prior to the End of the Analysis Interval) 3 (= Average Glucose (mg/dL) 2 Weeks Prior to the End of the Analysis Interval)	Assigned
AVISITN	Analysis Visit (N)	num	0 (= Baseline) 52 (= Week 52)	Assigned
AVISIT	Analysis Visit	text	Baseline Week 52	Assigned: Set to "Baseline" for the PARAM values based on the baseline LBDTC value used to determine AENDTM and set to "Week 52" for the PARAM values based on the week 52 LBDTC value used to determine AENDTM.
AVAL	Analysis Value	float		Derived: See Parameter-Level Metadata.
BASE	Baseline Value	float		Derived: Equals to AVAL when ABLFL="Y" retained across all observations within a given USUBJID/PARAMCD.
CHG	Change from Baseline	float		Derived: If AVAL and BASE are not missing and ABLFL NE "Y", then set to AVAL-BASE.
ABLFL	Baseline Record Flag	text	Y	Set to "Y" for the AVISIT="Baseline" value
ASTDTM	Analysis Start Datetime	datetime		Derived: ASTDTM is the numeric equivalent of LBDTC at the end of the interval as indicated in the PARAM description, minus the assigned prior period. For PARAM description with "24 Hours Prior to the End of the Analysis Interval" and "24 to 20 Hours Prior to the End of the Analysis Interval", the interval will start 24 hours before the assigned endpoint (i.e., the numeric equivalent of LBDTC at the end of the interval minus 24 hours).
AENDTM	Analysis End Datetime	datetime		Derived: AENDTM is the numeric equivalent of LBDTC at the end of the interval as indicated in the PARAM description. When the PARAM description contains "24 Hours Prior to the End of the Analysis Interval" or "2 Weeks Prior to the End of the Analysis Interval", the numeric equivalent of LBDTC at the visit is used. If the PARAM description contains "24 to 20 Hours Prior to the End of the Analysis Interval", the numeric equivalent of LBDTC at the visit, minus 20 hours, is used.

Parameter Value List - ADCGM [AVAL]

Variable Name

Where

Type

Codelist/Controlled Terms

Source/Derivation/Comment

AVAL

PARAMCD="AVGGL24"

float

AVAL is the average of the glucose measurements collected by the CGM during the last 24 hours prior to the assigned LBDTC datetime endpoint. The measurements fall entirely within the 24-hour period ending within 5 minutes of the analysis timepoint. This is calculated programatically based on the CGM values.

AVAL

PARAMCD="AVGGL4H"

float

AVAL is the average of the glucose measurements collected by the CGM device during the first 4 hours of the 24-hour period prior to the assigned LBDTC datetime endpoint (i.e., 24 hours to 20 hours prior to the endpoint). The glucose measurements fall entirely within the first 4 hours of the 24-hour period ending within 5 minutes of the analysis timepoint. This is calculated programatically based on the CGM values.

AVAL

PARAMCD="AVGGL2W"

float

AVAL is the average of the glucose measurements collected during the 2-week period prior to the assigned LBDTC datetime endpoint. The glucose measurements fall entirely within the 2 weeks period ending within 5 minutes of the analysis timepoint. This is calculated programatically based on the CGM values.