how to design a custom sdtm domain for nonclinical data ... · how to design a custom sdtm domain...
TRANSCRIPT
1 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
How to Design a Custom SDTM Domain for Nonclinical Data
PhUSE Standards Roadmap Team
Contents Abstract: 1
Using the “How to Design a Domain” resource 2
DOMAIN TABLE SKELETON – XX 5
Appendix 1. Table of SDTM Variable Categories and Subcategories 13
Appendix 2. Use Cases for the Custom Domain Resource 14
Abstract:
Data Standards are being developed to convey complex clinical and nonclinical data in a standard form
across organizations and for regulatory submissions. CDISC SEND is a nonclinical standard for general
toxicology studies and carcinogenicity studies. The SEND standard is currently being expanded to include
reproductive and developmental toxicology studies and safety pharmacology studies. Though study
designs for many nonclinical studies have uniform design and often capture similar endpoints, some design
elements and study endpoints may have a more unique quality that has not been specifically considered in
the standard. The science of study conduct and data interpretation allows data to be captured in various
ways across studies. Providing a dataset standard for a given data type will end up being a compromise of
the different ways of collecting this data. This built in flexibility in the dataset standard will now in every
instance be subject to interpretations of the tenets of the standard, and approaches to its use may be so
creative that similar data across studies may be tabulated in multiple ways. In such cases, data loses a
standard quality. The Standards Roadmap Team created a decision tree to guide a data-driven decision-
making for capturing not-yet standardized data types in a standard dataset. The intent is to streamline and
improve uniformity in the interpretation of the tenets of the SDTM standard and thereby providing a
consistent way of creating a standard domain structure specifically suited for the data belonging to a
findings general observation class.
2 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Using the “How to Design a Domain” resource
The purpose of this resource is to guide users toward appropriate variables that will accommodate data
types that are not specified in the standard. The primary reference is the decision tree which will guide
users on which domain variables are most appropriate for their data types. The decision tree is directly
below this section (Figure 1) and is also provided as a Visio file that accompanies this resource. Users
should begin at the top left of the decision tree at the category “Core Dataset Structure” and follow the
“swim lane” going right across the decision tree for this category of variables. The categories on the
decision tree have corresponding color-coding on the Domain Table, which is a supplemental reference
document that provides more detail on the nature of variables. The questions in the decision tree will
guide the user as to when groups of variables are needed for their particular data. When users encounter a
grouping of variables on the decision tree, they should then refer to the Domain Table “Core” column to
determine which variables are considered required, expected or permissible for their data type. After
completing the decisions for a category (i.e. traveling to the end of a lane) the user should go to the lane
directly below the one they completed and follow the decision tree (left to right) across that lane. The user
should consider each category (i.e. Core Dataset Structure, Object Identifiers etc.) and it is essential to use
each category in a dataset. In other words, all categories of variables are essential to a dataset but the
decision tree will guide the user to select the relevant groups of variables within these categories. Some
guidance for selecting variables within a group is provided in the “Core (within a variable group)” column,
however the fine tuning of the domain structure should reflect the data need.
If you answer both yes and no to some of the questions, you are likely dealing with more than one data
type in mind. Have each of the data types go through the decision tree and decide in the end if the record
structure is similar enough to allow for collapsing the two resulting domain structures into one. This is not
recommended if the domain structures have different Object Identifiers, Test Variables or Timing Variables.
The CDISC SDTM model allows for addition of an extensive number of variables to a findings observation
class domain. The Standards Roadmap Team has assessed these variables from a nonclinical perspective
and selected the ones considered most applicable for nonclinical studies. These are the variables presented
in the domain table below and those that you are guided through in the decision tree. If you find that there
is an additional need for further variables to accurately reflect the data, it is to the user’s discretion to add
SDTM variables allowed for a Findings General Observation Class domain from the SDTM while maintaining
the proper sequence of all variables as described in SDTM.
The data in scope for this resource is collected data from nonclinical studies that would be expected to be
modelled as findings data and which can be supplemented to a data package based on CDISC SEND version
3.0. This resource has not been applied to clinical data and such use is not recommended. Further
development is necessary to address clinical needs.
The domain table below should not be read without the decision tree. From a familiarity point of view, the
“Core” column in the domain table below is not referring to a requirement on a domain level, but a
requirement on a variable grouping level. Example: If the object being assessed is a group of subjects, then
you are prompted in the decision tree to add the “OI-1” variables. You should not also add “OI-2” or “OI-3”
variables if they are indicated as required in the “Core” column. Only when the decision tree warrants you
to add a group of variables, should you then refer to the “Core” column for further guidance on expectancy
3 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
within the group of variables. Once you have settled with the right groups of variables, the sequence must
match that of the domain table below. Permissible variables that you have gained as part of a group can be
freely deleted if there is no use for them for your particular data, however required and expected variables
must be included due to the answers you have provided in the decision tree. The required and some of the
expected variables will be the natural keys in your new domain. The natural keys are those variables that
collectively provide a unique identification of every record in the dataset and they define the record
structure. Note that the expectancy of the variables is not intended to reflect any published SEND or SDTM
domain, but rather serve as a guide towards common scenarios for how to tabulate nonclinical data.
Appendices are provided as general reference material and as confirmation that the output from the
resource follows basic SDTM concepts. Appendix 1 is a tabular view of the categories, subcategories and
variables that are used in this resource. The list in Appendix 1 serves as a general reference to ensure that
users considered all variables within each subcategory. Several use cases are provided in Appendix 2 to
demonstrate how unmodeled data can be modeled using this resource
As part of developing this resource the PhUSE Nonclinical Roadmap team looked into current published
validation rules to see if some guidance could be provided on how to verify that fundamentals of the
standard are maintained in the new custom domain. The analysis of validation rules resulted in a non-
exhaustive list of checks from the public domain which can be considered key elements for any SDTM
findings dataset. The analysis outcome is presented as a separate Excel file and can be found at
http://www.phusewiki.org/wiki/index.php?title=WG6_Nonclinical_-_Standardization_Roadmap.
Additional conformance rules for SDTM datasets can be found at:
http://www.phusewiki.org/wiki/index.php?title=Representing_CDISC_Conformance_Checks.
4 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Figure 1 – Decision Tree
5 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
DOMAIN TABLE SKELETON – XX
xx.xpt, - Findings. Tabulation, Record structure will depend on the outcome of the decision tree.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Core Dataset
structure
Mandatory Req STUDYID Study Identifier Char Identifier Unique identifier for a study.
Core Dataset
structure
Mandatory Req DOMAIN Domain
Abbreviation
Char XX Identifier Two-character abbreviation for the domain. Refer to
CDISC SDTMIG and SENDIG for reserved domain
codes. Domain codes starting with “X”, “Y” and “Z”
are always considered
Object Identifier OI-2 Req USUBJID Unique Subject
Identifier
Char Identifier Identifier used to uniquely identify a subject across all
studies for all application or submissions involving the
product.
Object Identifier OI-1 Req POOLID Pool Identifier Char Identifier Pooling Identifier for Samples. Either USUBJID or
POOLID must be populated.
Core Dataset
structure
Mandatory Req --SEQ Sequence
Number
Num Identifier Sequence number given to ensure uniqueness of
subject records within a domain. May be any valid
number.
Object Identifier OI-3 Perm --REFID Reference
Identifier
Char Identifier Optional internal or external identifier such as lab
specimen ID, or UUID for an ECG waveform or a
medical image.
Core Dataset
structure
CD-1 Perm --SPID Sponsor-
Defined
Identifier
Char Identifier Sponsor-defined reference number. Perhaps preprinted
as an explicit line identifier or defined in the sponsor’s
operational database. Example: Line number on the
Lab page.
6 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Test Variables Mandatory Req --TESTCD Measurement,
Test or
Examination
Short Name
Char Controlled
terminology
is expected
Topic Short name of the measurement, test, or examination
described in --TEST. It can be used as a column name
when converting a dataset from a vertical to a
horizontal format. The value in --TESTCD cannot be
longer than 8 characters, nor can it start with a number
(e.g., “1TEST” is not valid). LBTESTCD cannot
contain characters other than letters, numbers, or
underscores.
Test Variables Mandatory Req --TEST Measurement,
Test or
Examination
Name
Char Controlled
terminology
is expected
Synonym
Qualifier
Long name for --TESTCD. The value in LBTEST
cannot be longer than 40 characters.
Test Variables TV-1 Exp
--CAT Measurement,
Test or
Examination
Category
Char Controlled
terminology
is expected
Grouping
Qualifier
Used to define a category of related tests across
subjects.
Test Variables TV-1 Perm --SCAT Measurement,
Test or
Examination
Subcategory
Char Grouping
Qualifier
A further categorization of a test category
Object Identifier OI-4 Perm --POS Position of
Subject during
Observation
Char (POSITIO
N)
Record
Qualifier
Position of the subject during a measurement or
examination. Examples: SUPINE, STANDING,
SITTING.
Result Variables Mandatory Exp --ORRES Result or
Finding as
Collected
Char Result
Qualifier
Result of the measurement or finding as originally
received or collected.
Results Variables RV-3 Exp --ORRESU Unit of the
Original Result
Char (UNIT) Variable
Qualifier
The unit for the original result. The unit of the original
result should be mapped to a synonymous unit on the
Controlled Terminology list.
Results Variables RV-4 Perm
--ORNRLO Reference
Range Lower
Limit-Orig Unit
Char Variable
Qualifier
Lower end of reference range used at the time of
collection, in original units.
7 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Results Variables RV-4 Perm --ORNRHI Reference
Range Upper
Limit-Orig Unit
Char Variable
Qualifier
Upper end of reference range used at the time of
collection, in original units.
Results Variables Mandatory Exp --STRESC Standardized
Result in
Character
Format
Char Controlled
terminology
is expected
Result
Qualifier
Contains the result value for all findings, copied or
derived from --ORRES in a standard format or
standard units. --STRESC should store all results or
findings in character format; if results are numeric,
they should also be stored in numeric format in --
STRESN.
Results Variables RV-3 Exp --STRESN Standardized
Result in
Numeric Format
Num Result
Qualifier
Used for continuous or numeric results or findings in
standard format; contains the numeric form of --
STRESC.
Results Variables RV-3 Exp --STRESU Unit of the
Standardized
Result
Char (UNIT) Variable
Qualifier
Standardized unit used for --STRESC and --STRESN.
Results Variables RV-4 Perm --STNRLO Reference
Range Lower
Limit-Std Unit
Num Variable
Qualifier
Lower end of reference range for continuous or
numeric standardized results (--STRESN) represented
in standardized units.
Results Variables RV-4 Perm --STNRHI Reference
Range Upper
Limit-Std Unit
Num Variable
Qualifier
Upper end of reference range for continuous or
numeric standardized results (--STRESN) represented
in standardized units.
Results Variables RV-4 Perm --STNRC Reference
Range for Char
Rslt-Std Unit
Char Variable
Qualifier
Reference range for results stored in --STRESC that
are character in ordinal or categorical scale. Example:
Negative to Trace.
8 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Results Variables RV-4 Perm --RIND Reference
Range Indicator
Char Variable
Qualifier
Indicates where value falls with respect to reference
range defined by --ORNRLO and --ORNRHI, by --
STRNRLO and --STNRHI, or by --STNRC. Examples:
NORMAL, ABNORMAL, HIGH, LOW. This should
not be used to indicate biological significance.
Sponsors should specify in the study metadata
(Comments column in the data definition file) whether
LBNRIND refers to the original or standard reference
ranges and results.
Results Variables RV-2 Exp --RESCAT Result Category Char Controlled
terminology
is expected
Variable
Qualifier
Used to categorize the result of a finding post
collection.
Result Variables Mandatory Perm --STAT Completion
Status
Char (ND) Record
Qualifier
Used to indicate when a test is not done or result is
missing. Should be null if a results exists in --ORRES.
Result Variables Mandatory Perm --REASND Reason Not
Done
Char Record
Qualifier
Describes why --STAT is NOT DONE, such as
BROKEN EQUIPMENT or SICK ANIMAL.
Core Dataset
structure
CD-1 Perm --XFN External File
Name
Char Record
Qualifier
Filename for an external file associated with the result
Test Variables TV-3 Perm --NAM Vendor Name Char Record
Qualifier
Name or identifier of the laboratory or vendor that
provided the test results.
Object Identifier OI-3 Exp --SPEC Specimen
Material Type
Char (SPEC) Record
Qualifier
Defines the type of tissue, organ, or fluid specimen
used as the object for the finding. Examples: GLAND,
ADRENAL; KIDNEY; VESSEL, LYMPHATIC.
9 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Object Identifier OI-3 Exp --ANTREG Anatomical
Region of
Specimen
Char Variable
Qualifier
Defines the specific anatomical or biological region of
a tissue, organ specimen or the region from which the
specimen was obtained, such as a section or part of
what is defined in the --SPEC variable. If the
anatomical region is not included in the specimen
description --SPEC, it may be included in this variable.
This field can be a combination of terms where needed.
This field can be null if not applicable. Examples:
CORTEX, MEDULLA, MUCOSA, SEROSA, ISLET,
ZONA FASICULATA, ZONA RETICULARIS,
CRANIAL, MEDIAN, ACCESSORY, SPINAL,
LUMBAR, FRONTAL.
Object Identifier OI-3 Perm --SPCCND Specimen
Condition
Char Variable
Qualifier
Free or standardized text describing the condition of
the specimen. Example: AUTOLYZED,
HEMOLYZED, ICTERIC, LIPEMIC, etc.
Object Identifier OI-3 Perm --SPCUFL Specimen
Usability for the
Test
Char (NY) Variable
Qualifier
Describes the usability of the specimen for the test.
Example: N = the specimen is not usable; otherwise
null.
Object Identifier OI-5 Perm --LOC Location
associated with
a result or
finding
Char Record
Qualifier
Anatomical location associated with the test. Example
rectal for temperature.
Object Identifier OI-3 Perm --LAT Specimen
Laterality
within Subject
Char (LAT) Variable
Qualifier
Qualifier for laterality of the specimen within the
subject for paired specimens. Examples: LEFT,
RIGHT, BILATERAL.
Object Identifier OI-3 Perm --DIR Specimen
Directionality
within Subject
Char (DIR) Variable
Qualifier
Qualifier for directionality of the specimen within the
subject. Examples: DORSAL, PROXIMAL.
Object Identifier OI-3 Perm --PORTOT Portion or
Totality
Char (PORTOT) Variable
Qualifier
Qualifier for anatomical location or specimen further
detailing the portion or totality which means
arrangement of, or apportioning of, for example,
ENTIRE, SINGLE, SEGMENT, MANY.
10 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Test Variables TV-2 Perm --
METHOD
Method of Test
or Examination
Char Record
Qualifier
Method of the test or examination.
Object Identifier OI-4 Perm --CSTATE Consciousness
State
Char (CSTATE) Record
Qualifier
The consciousness state of the subject at the time of
measurement.
Examples: CONSCIOUS, SEMI-CONSCIOUS,
UNCONSCIOUS.
Result Variables RV-3 Exp --BLFL Baseline Flag Char (NY) Record
Qualifier
A baseline indicator may be used to calculate
differences or changes from baseline. Value should be
Y or null. The baseline flag is sponsor-defined.
Object Identifier OI-4 Perm --FAST Fasting Status Char (NY) Record
Qualifier
Indicator used to identify fasting status. Null if not
relevant.
Test Variables TV-3 Perm --EVAL Evaluator Char Record
Qualifier
Role of the person who provided the evaluation. Used
only for results that are subjective (e.g., assigned by a
person or a group). Should be null for records that
contain collected or derived data. Examples:
RESPONSIBLE SCIENTIST, PRINCIPAL
INVESTIGATOR, PEER
REVIEWER.
Results Variables RV-1 Perm --SEV Severity Char (SEV) Record
Qualifier
Describes the severity or intensity of a particular
finding.
Test Variables TV-2 Perm --LLOQ Lower Limit of
Quantitation
Num Variable
Qualifier
Indicates the lower limit of quantitation for an assay.
Units will be those used in --STRESU.
Test Variables TV-2 Perm --ULOQ Upper Limit of
Quantitation
Num Variable
Qualifier
Indicates the upper limit of quantitation for an assay.
Units will be those used in --STRESU.
Results Variables Mandatory Perm --EXCLFL Exclusion Flag Char (NY) Record
Qualifier
Y if the result should be excluded from all calculations,
otherwise null.
Results Variable Mandatory Perm --REASEX Reason for
Exclusion
Char Record
Qualifier
The reason the result should be excluded from all
calculations. Used only when --EXCLFL is Y.
Timing Variables Mandatory Exp VISITDY Planned Study
Day
Num Timing This is the planned study day of collection.
11 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Timing Variables TM-1 Exp
--DTC Date/Time of
Test
Char ISO 8601 Timing Date/time of the measurement, test or observation in
IS0 8601 format.
Timing Variables TM-3 Exp --STDTC Start Date/Time
of Test
Char ISO 8601 Timing Date/time of the start of the measurement, test, or
observation in ISO 8601 format. Should be populated
only for continuous sample collection.
Timing Variables TM-3 Perm --ENDTC End Date/Time
of Test
Char ISO 8601 Timing Date/time of the end of the measurement, test, or
observation in ISO 8601 format. Should be populated
only for continuous sample collection.
Timing Variables TM-1 Exp --DY Study Day of
Test
Num Timing Study day of the measurement, test or observation, in
integer days. The algorithm for calculations must be
relative to the sponsor-defined RFSTDTC variable in
the Demographics (DM) domain.
Timing Variables TM-3 Exp --STDY Study Day of
Start of Test
Num Timing Study day of the start of the measurement, test, or
observation, in integer days. The algorithm for
calculations must be relative to the sponsor-defined
RFSTDTC variable in the Demographics (DM) domain.
Timing Variables TM-3 Perm --ENDY Study Day of End
of Test
Num Timing Study day of the end of the measurement, test, or
observation, in integer days. The algorithm for
calculations must be relative to the sponsor-defined
RFSTDTC variable in the Demographics (DM) domain.
Timing Variables TM-3 Perm --DUR Duration of Test Char ISO 8601 Timing Collected duration of a measurement, test, or
observation, represented in ISO8601 format. This
should not be used to report a derived duration.
Timing Variables TM-2 Exp --TPT Planned Time
Point Name
Char Timing Text description of the planned time of the
measurement, test or observation.
This may be represented as an elapsed time relative to
a fixed reference point, such as time of last dose. See
LBTPTNUM and LBTPTREF. Examples: Start, 5 min
post.
Timing Variables TM-2 Exp --TPTNUM Planned Time
Point Number
Num Timing Numerical version of --TPT to aid in sorting.
12 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables (see decision
tree)
SDTM / SEND definitions and rules for the population of the variables
Variable category Variable group Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role CDISC Notes
Timing Variables TM-2 Exp --ELTM Planned Elapsed
Time from Time
Point Ref
Char ISO 8601 Timing Planned elapsed time (in ISO 8601 format) relative to a
planned fixed reference (--TPTREF). This variable is
useful where there are repetitive measures. Not a clock
time or a date time variable. Represented as an ISO
8601 duration. Examples: “- P15M” to represent the
period of 15 minutes prior to the reference point
indicated by LBTPTREF, or “P8H” to represent the
period of 8 hours after the reference point indicated by
LBTPTREF.
Timing Variables TM-2 Exp --TPTREF Time Point
Reference
Char Timing Name of the fixed reference point referred to by --
ELTM, if used for LBTPTNUM, and LBTPT.
Examples: PREVIOUS DOSE, PREVIOUS MEAL.
Timing Variables TM-2 Exp --RFTDTC Date/Time of
Reference Time
Point
Char ISO 8601 Timing Date/time of the reference time point, --TPTREF
Timing Variables TM-4 Perm --EVLINT Evaluation
Interval
Char ISO 8601 Timing Duration of interval associated with an observation
such as a finding --TESTCD, represented in ISO 8601
character format. Example: -P2M to represent a period
of the past 2 months as the evaluation interval for a
question from a questionnaire such as SF-36.
Timing Variables TM-4 Perm --STINT Planned Start of
Assessment
Interval
Char ISO 8601 Timing The start of a planned evaluation or assessment interval
relative to the Time Point Reference
Timing Variables TM-4 Perm --ENINT Planned End of
Assessment
Interval
Char ISO 8601 Timing The end of a planned evaluation or assessment interval
relative to the Time Point Reference
13 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Appendix 1. Table of SDTM Variable Categories and Subcategories
(cross-check each variable before finalizing)
Core Dataset
Structure
Object
Identifier
(OI)
Results
Variables
(RV)
Test
Variable
(TV)
Timing
Variables
(TM)
Mandatory STUDYID ORRES TESTCD VISITDY
Variables DOMAIN STRESC TEST
in Bold SEQ STAT
REASND
EXCLFL
REASEX
OI-1 RV-1 TV-1 TM-1
POOLID SEV CAT DTC
SCAT DY
OI-2 RV-2
USUBJID RESCAT TV-2 TM-2
METHOD TPT
OI-3 RV-3 LLOQ TPTNUM
REFID ORRESU ULOQ ELTM
SPEC STRESN
TPTREF
ANTREG STRESU TV-3 RFTDTC
SPCCND BLFL NAM
SPCUFL EVAL TM3
LAT RV-4
STDTC
DIR ORNRLO
ENDTC
PORTOT ORNRHI
STDY
STNRLO
ENDY
OI-4 STNRHI
DUR
POS STNRC
CSTATE NRIND
TM-4
FAST
EVLINT
STINT
OI-5 ENINT
LOC
14 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Appendix 2. Use Cases for the Custom Domain Resource
Example 1: Mapping micronucleus end-points
A published in vivo mammalian cytogenetics study satisfying the OPPTS 870.5395; OECD 474
guideline was used to model endpoints from an erythrocyte micronucleus assay in mice.
Brief overview of the study (No. 221): Male mice in groups of 6 were administered either test
article, vehicle or positive control (cyclophosphamide). Bone marrow was harvested from the
hind limb (tibia) 24 and 48 hours after dosing. The following endpoints were assessed:
Endpoint Result type Unit
Number of micronucleated erythrocytes per
2000 polychromatic erythrocytes (PCE)
examined per animal
Number (integer) /2000 PCEs/animal
Ratio of polychromatic erythrocytes (PCE) to
monochromatic erythrocytes (NCE)
determined in at least 500 erytrocytes per
animal
Number (float) ratio
Data from the study:
Study day Animal Number # MN PCEs
/2000 PCEs
/animal
Ratio PCE:NCE
24 2602 1 0.47
24 2606 0 0.78
48 2632 2 0.56
The endpoints in the table above are sent through the decision tree, starting from the top left
side. Ideally each of the two endpoints above should be processed through the decision tree
separately; however the outcome through the decision will be the same for both, so only one
run through the decision tree is described below.
The figure below illustrates the steps that this particular data will go through and the
categories of variables selected by answering the questions.
15 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
16 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision steps:
1. The Core Dataset Structure mandatory variables (STUDYID, DOMAIN, --SEQ) are added
2. There is no need to provide record level references or identifiers, so CD-1 variables are
omitted.
3. The object assessed is not a group of subjects, so the answer to 3 is no.
4. The object assessed is not an entire subject, so the answer to 4 is no.
5. The object assessed is a specimen from a subject (i.e. bone marrow from the hind
limb), so the answer to 5 is yes. Therefore OI-2 (USUBJID), and OI-3 (--REFID, --SPEC,
--ANTREG, --SPCCND, --LAT, --DIR, --PORTOT) variables are added. Note: The answer
to either 3, 4 or 5 must be yes, in this case it was no. 5.
6. The Test mandatory variables (--TESTCD, --TEST) are added.
7. There is no need to further specify the test to distinguish between records, so all other
test variables (TV-1, TV-2 and TV-3) are omitted. The need to further specify a test
could be if the same test was measured using two different methods (i.e. dipstick and
mass spectrometry) or the same test performed by two different labs.
8. The Result mandatory variables (--ORRES, --STRESC, --STAT, --REASND, --EXCLFL, --
REASEX) are added.
9. The results are not character or categorical data, so the answer to 9 is no.
10. The results are numeric (i.e. integer and float), so the answer to 10 is yes. The RV-3
variables (--ORRESU, --STRESN, --STRESU, --BLFL) are added.
11. The results are not categorized for further analysis, so the answer to 11 is no and RV-2
variables are omitted.
12. There are no references ranges provided for the results, so the answer to 12 is no and
RV-4 variables are omitted.
13. The Timing mandatory variable (VISITDY) is added.
14. The test does not have a duration, so the answer to 14 is no. The test is performed at a
single point in time, TM-1 variables (--DTC, --DY) are added.
15. The test is performed relative to a fixed reference (i.e. 24 and 48 hours post dose), so
the answer to 15 is yes and TM-2 variables (--TPT, --TPTNUM, --ELTM, --TPTREF, --
RFTDTC) are added.
The resulting domain structure is given below with all omitted variables removed.
17 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables Custom SDTM dataset specification – Micronucleus data
Decision
steps
Variable
category
Variable
group
Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role
Step 1 Core Dataset
structure
Mandatory Req STUDYID Study Identifier Char Identifier
Step 1 Core Dataset
structure
Mandatory Req DOMAIN Domain
Abbreviation
Char XX Identifier
Step 5 Object Identifier OI-2 Req USUBJID Unique Subject
Identifier
Char Identifier
Step 1 Core Dataset
structure
Mandatory Req --SEQ Sequence
Number
Num Identifier
Step 5 Object Identifier OI-3 Perm --REFID Reference
Identifier
Char Identifier
Step 6 Test Variables Mandatory Req --TESTCD Measurement,
Test or
Examination
Short Name
Char Controlled
terminology is
expected
Topic
Step 6 Test Variables Mandatory Req --TEST Measurement,
Test or
Examination
Name
Char Controlled
terminology is
expected
Synonym
Qualifier
Step 8 Result Variables Mandatory Exp --ORRES Result or
Finding as
Collected
Char Result
Qualifier
Step 10 Results
Variables
RV-3 Exp --ORRESU Unit of the
Original Result
Char (UNIT) Variable
Qualifier
Step 8 Results
Variables
Mandatory Exp --STRESC Standardized
Result in
Character
Format
Char Controlled
terminology is
expected
Result
Qualifier
Step 10 Results
Variables
RV-3 Exp --STRESN Standardized
Result in
Numeric Format
Num Result
Qualifier
Step 10 Results
Variables
RV-3 Exp --STRESU Unit of the
Standardized
Result
Char (UNIT) Variable
Qualifier
Step 8 Result Variables Mandatory Perm --STAT Completion
Status
Char (ND) Record
Qualifier
Step 8 Result Variables Mandatory Perm --REASND Reason Not
Done
Char Record
Qualifier
Step 5 Object Identifier OI-3 Exp --SPEC Specimen
Material Type
Char (SPEC) Record
Qualifier
Step 5 Object Identifier OI-3 Exp --ANTREG Anatomical
Region of
Specimen
Char Variable
Qualifier
Step 5 Object Identifier OI-3 Perm --SPCCND Specimen
Condition
Char Variable
Qualifier
Step 5 Object Identifier OI-3 Perm --SPCUFL Specimen
Usability for the
Test
Char (NY) Variable
Qualifier
Step 5 Object Identifier OI-3 Perm --LAT Specimen
Laterality within
Subject
Char (LAT) Variable
Qualifier
Step 5 Object Identifier OI-3 Perm --DIR Specimen
Directionality
within Subject
Char (DIR) Variable
Qualifier
18 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
Decision based adding of variables Custom SDTM dataset specification – Micronucleus data Decision
steps
Variable
category
Variable
group
Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role
Step 5 Object Identifier OI-3 Perm --PORTOT Portion or
Totality
Char (PORTOT) Variable
Qualifier
Step 10 Result Variables RV-3 Exp --BLFL Baseline Flag Char (NY) Record
Qualifier
Step 8 Results
Variables
Mandatory Perm --EXCLFL Exclusion Flag Char (NY) Record
Qualifier
Step 8 Results Variable Mandatory Perm --REASEX Reason for
Exclusion
Char Record
Qualifier
Step 13 Timing
Variables
Mandatory Exp VISITDY Planned Study
Day
Num Timing
Step 14 Timing
Variables
TM-1 Exp
--DTC Date/Time of
Test
Char ISO 8601 Timing
Step 14 Timing
Variables
TM-1 Exp --DY Study Day of
Test
Num Timing
Step 15 Timing
Variables
TM-2 Exp --TPT Planned Time
Point Name
Char Timing
Step 15 Timing
Variables
TM-2 Exp --TPTNUM Planned Time
Point Number
Num Timing
Step 15 Timing
Variables
TM-2 Exp --ELTM Planned Elapsed
Time from Time
Point Ref
Char ISO 8601 Timing
Step 15 Timing
Variables
TM-2 Exp --TPTREF Time Point
Reference
Char Timing
Step 15 Timing
Variables
TM-2 Exp --RFTDTC Date/Time of
Reference Time
Point
Char ISO 8601 Timing
19 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
The dataset outline based on the domain structure above including example data is shown below. I have decided to name it
“MN” (for Micronucleus), which is populated in the DOMAIN variable and also used as prefix for all variables with -- in their
name in the domain structure.
The permissibility (Core within each variable group) is included. Once the data has been successfully mapped into the dataset
structure and you make sure that you have been able to map all information about each result, then a last exercise to remove
any non-used permissible variables (highlighted in blue) should be performed.
Note that as part of this final step we are removing four mandatory Result Variables because they are permissible and were not
used in the dataset. The expected but un-used variables are still kept in the dataset structure.
The actual date/time information about the results is likely available for data collected in today’s data capture systems, but in
this example that information cannot be populated because of the legacy nature of the used data.
Req Req Req Req Perm Req Req Exp Exp Exp Exp Exp
Row
no.
STUDYID DOMAIN USUBJID MNSEQ MNREFID MNTESTCD MNTEST MNORRES MNORRESU MNSTRESC MNSTRESN MNSTRESU
1 221 MN 221-2602 1 MN_PCE Micronucleated
Polychromatic
Erythrocyte
1 /2000 PCEs
/animal
1 1 /2000 PCEs
/animal
2 221 MN 221-2602 2 PCE/NCE Polychromatic
Monochromatic
Ratio
0.47 Ratio 0.47 0.47 Ratio
3 221 MN 221-2606 1 MN_PCE Micronucleated
Polychromatic
Erythrocyte
0 /2000 PCEs
/animal
0 0 /2000 PCEs
/animal
4 221 MN 221-2606 2 PCE/NCE Polychromatic
Monochromatic
Ratio
0.78 Ratio 0.78 0.78 Ratio
5 221 MN 221-2632 1 MN_PCE Micronucleated
Polychromatic
Erythrocyte
2 /2000 PCEs
/animal
2 2 /2000 PCEs
/animal
6 221 MN 221-2632 2 PCE/NCE Polychromatic
Monochromatic
Ratio
0.56 Ratio 0.56 0.56 Ratio
20 of 35 Authored by the PhUSE Nonclinical Roadmap team, December 2014.
(Cont.) Perm Perm Exp Exp Perm Perm Perm Perm Perm Exp Perm Perm
Row no. MNSTAT MNREASND MNSPEC MNANTREG MNSPCCND MNSPCUFL MNLAT MNDIR MNPORTOT MNBLFL MNEXCLFL MNREASEX
1 BONE
MARROW
2 BONE
MARROW
3 BONE
MARROW
4 BONE
MARROW
5 BONE
MARROW
6 BONE
MARROW
(Cont.) Exp Exp Exp Exp Exp Exp Exp Exp
Row no. VISITDY MNDTC MNDY MNTPT MNTPTNUM MNELTM MNTPTREF MNRFTDTC
1 2 2 24 hours after dose 1 P24H Day 1 dose
2 2 2 24 hours after dose 1 P24H Day 1 dose
3 2 2 24 hours after dose 1 P24H Day 1 dose
4 2 2 24 hours after dose 1 P24H Day 1 dose
5 3 3 48 hours after dose 2 P48H Day 1 dose
6 3 3 48 hours after dose 2 P48H Day 1 dose
Reference: http://www.epa.gov/opp00001/chem_search/cleared_reviews/csr_PC-072506_27-Jul-05_a.pdf
Example 2: Mapping Motor Activity
The source of the study data is a mocked study from a live data capture system.
Brief overview of the study (ABC) data:
Individual animals were placed in a cage divided into two boxes with photocells measuring the activity of
the animal based on the crossed photocells in X and Y axis in each box.
Endpoint Result type Unit
Number crossed photocell per axis, time
interval and defined space
Number (integer) count
Data from the study:
Study day Animal
Number
Space Time XT YT
5 M1 Box1 5min 862 502
5 M1 Box1 10min 699 465
5 M1 Box1 15min 383 263
5 M1 Box2 5min 484 351
5 M1 Box2 10min 429 259
5 M1 Box2 15min 187 115
Study day 5 was March 10th 2009.
The endpoints in the table above are sent through the decision tree, starting from the top left side. Ideally
each of the two endpoints above should be processed through the decision tree separately; however the
outcome through the decision will be the same for both, so only one run through the decision tree is
described below.
The figure below illustrates the steps that this particular data will go through and the categories of
variables selected by answering the questions.
Decision steps:
1. The Core Dataset Structure mandatory variables (STUDYID, DOMAIN, --SEQ) are added
2. There is no need to provide record level references or identifiers, so CD-1 variables are omitted.
3. The object assessed is not a group of subjects, so the answer to 3 is no.
4. The object assessed is entire subject, so the answer to 4 is yes.
5. The object assessed is not a specimen from a subject (i.e. bone marrow from the hind limb), so
the answer to 5 is no.
6. The assessment is not specific to a location on the animal, so OI-5 variables are omitted.
7. The Test mandatory variables (--TESTCD, --TEST) are added.
8. There is a need to further specify the test to distinguish between records, so the answer to 8 is
yes.
9. The test do not include category, so the answer to 9 is no.
10. The test does include method, so the answer to 10 is yes.
11. The responsability of the test is not important, so the answer to 11 is no.
12. The Result mandatory variables (--ORRES, --STRESC, --STAT, --REASND, --EXCLFL, --REASEX)
are added.
13. The results are not character or categorical data, so the answer to 13 is no (RV-1 and RV-2
variables are omitted).
14. The results are numeric, so the answer to 14 is yes. The RV-3 variables (--ORRESU, --STRESN, --
STRESU, --BLFL) are added.
15. There are no references ranges provided for the results, so the answer to 15 is no and RV-4
variables are omitted.
16. The Timing mandatory variable (VISITDY) is added.
17. The measure have a duration (it is a count per time interval), so the answer to 17 is yes, TM-3
variables (--STDTC, --ENDTC, --STDY, --ENDY, --DUR) are added.
18. The test is performed relative to a fixed reference, so the answer to 18 is yes, TM-2 (--TPT, --
TPTNUM, --ELTM, --TPTREF, --RFTDTC) and TM-4 variables (--EVLINT, --STINT, --ENINT) are
added.
The resulting domain structure is given below with all omitted variables removed.
Decision based adding of variables Custom SDTM dataset specification – Motor Activity data
Decision
steps
Variable
category
Variable
group
Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role
Step 1 Core Dataset
structure
Mandatory Req STUDYID Study Identifier Char Identifier
Step 1 Core Dataset
structure
Mandatory Req DOMAIN Domain
Abbreviation
Char XX Identifier
Step 4 Object
Identifier
OI-2 Req USUBJID Unique Subject
Identifier
Char Identifier
Step 1 Core Dataset
structure
Mandatory Req --SEQ Sequence
Number
Num Identifier
Step 7 Test Variables Mandatory Req --TESTCD Measurement,
Test or
Examination
Short Name
Char Controlled
terminology is
expected
Topic
Step 7 Test Variables Mandatory Req --TEST Measurement,
Test or
Examination
Name
Char Controlled
terminology is
expected
Synonym
Qualifier
Step 10 Test Variables TV-2 Perm --LOC Location
associated with
a result or
finding
Char Record
Qualifier
Step 10 Test Variables TV-2 Perm --METHOD Method of Test
or Examination
Char Record
Qualifier
Step 12 Result
Variables
Mandatory Exp --ORRES Result or
Finding as
Collected
Char Result
Qualifier
Step 14 Results
Variables
RV-3 Exp --ORRESU Unit of the
Original Result
Char (UNIT) Variable
Qualifier
Step 12 Results
Variables
Mandatory Exp --STRESC Standardized
Result in
Character
Format
Char Controlled
terminology is
expected
Result
Qualifier
Step 14 Results
Variables
RV-3 Exp --STRESN Standardized
Result in
Numeric Format
Num Result
Qualifier
Step 14 Results
Variables
RV-3 Exp --STRESU Unit of the
Standardized
Result
Char (UNIT) Variable
Qualifier
Step 12 Result
Variables
Mandatory Perm --STAT Completion
Status
Char (ND) Record
Qualifier
Step 12 Result
Variables
Mandatory Perm --REASND Reason Not
Done
Char Record
Qualifier
Step 14 Result
Variables
RV-3 Exp --BLFL Baseline Flag Char (NY) Record
Qualifier
Step 12 Results
Variables
Mandatory Perm --EXCLFL Exclusion Flag Char (NY) Record
Qualifier
Step 12 Results
Variable
Mandatory Perm --REASEX Reason for
Exclusion
Char Record
Qualifier
Step 16 Timing
Variables
Mandatory Exp VISITDY Planned Study
Day
Num Timing
Step 17 Timing
Variables
TM-3 Exp --STDTC Start Date/Time
of Test
Char ISO 8601 Timing
Step 17 Timing
Variables
TM-3 Perm --ENDTC End Date/Time
of Test
Char ISO 8601 Timing
Decision based adding of variables Custom SDTM dataset specification – Motor Activity data
Decision
steps
Variable
category
Variable
group
Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role
Step 17 Timing
Variables
TM-3 Exp --STDY Study Day of
Start of Test
Num Timing
Step 17 Timing
Variables
TM-3 Perm --ENDY Study Day of
End of Test
Num Timing
Step 17 Timing
Variables
TM-3 Perm --DUR Duration of Test Char ISO 8601 Timing
Step 18 Timing
Variables
TM-2 Exp --TPT Planned Time
Point Name
Char Timing
Step 18 Timing
Variables
TM-2 Exp --TPTNUM Planned Time
Point Number
Num Timing
Step 18 Timing
Variables
TM-2 Exp --ELTM Planned Elapsed
Time from Time
Point Ref
Char ISO 8601 Timing
Step 18 Timing
Variables
TM-2 Exp --TPTREF Time Point
Reference
Char Timing
Step 18 Timing
Variables
TM-2 Exp --RFTDTC Date/Time of
Reference Time
Point
Char ISO 8601 Timing
Step 18 Timing
Variables
TM-4 Perm --EVLINT Evaluation
Interval
Char ISO 8601 Timing
Step 18 Timing
Variables
TM-4 Perm --STINT Planned Start of
Assessment
Interval
Char ISO 8601 Timing
Step 18 Timing
Variables
TM-4 Perm --ENINT Planned End of
Assessment
Interval
Char ISO 8601 Timing
The dataset outline based on the domain structure above including example data is shown below. I have decided to name it “MC”
(for Motor Activity - MA is already used), which is populated in the DOMAIN variable and also used as prefix for all variables with --
in their name in the domain structure.
The results were collected over a planned collection interval of 5 minutes each, which is reflected in the ISO8601 duration format in
MCEVLINT (PT5M). Each time point described in TPT is labelled with the end of the interval, and the planned start and end of the
collection interval is represented in MCSTINT and MCENINT, respectively.
Any permissible variables not used to model the data (indicated in red) can be omitted from the dataset.
Req Req Req Req Req Req Exp Exp Exp Exp Exp
Row
no. STUDYID DOMAIN USUBJID MCSEQ MCTESTCD MCTEST MCORRES MCORRESU MCSTRESC MCSTRESN MCSTRESU
1 ABC MC ABC-M1 1 MC_XT Motor Activity Cross X Line 862 count 862 862 count 2 ABC MC ABC-M1 2 MC_YT Motor Activity Cross Y Line 502 count 502 502 count 3 ABC MC ABC-M1 3 MC_XT Motor Activity Cross X Line 699 count 699 699 count 4 ABC MC ABC-M1 4 MC_YT Motor Activity Cross Y Line 465 count 465 465 count 5 ABC MC ABC-M1 5 MC_XT Motor Activity Cross X Line 383 count 383 383 count 6 ABC MC ABC-M1 6 MC_YT Motor Activity Cross Y Line 263 count 263 263 count 7 ABC MC ABC-M1 7 MC_XT Motor Activity Cross X Line 484 count 484 484 count 8 ABC MC ABC-M1 8 MC_YT Motor Activity Cross Y Line 351 count 351 351 count 9 ABC MC ABC-M1 9 MC_XT Motor Activity Cross X Line 429 count 429 429 count
10 ABC MC ABC-M1 10 MC_YT Motor Activity Cross Y Line 259 count 259 259 count 11 ABC MC ABC-M1 11 MC_XT Motor Activity Cross X Line 187 count 187 187 count 12 ABC MC ABC-M1 12 MC_YT Motor Activity Cross Y Line 115 count 115 115 count
(Cont.) Perm Perm Perm Exp Perm Perm
Row no. MCMETHOD MCSTAT MCREASND MCBLFL MCEXCLFL MCREASEX 1 Box1 2 Box1 3 Box1 4 Box1 5 Box1 6 Box1 7 Box2 8 Box2 9 Box2
10 Box2 11 Box2 12 Box2
(Cont.) Exp Exp Perm Exp Perm Perm Exp Exp Exp Exp Exp Row no. VISITDY MCSTDTC MDENDTC MCSTDY MCENDY MCDUR MCTPT MCTPTNUM MCELTM MCTPTREF MCRFTDTC 1 5 2009-03-10 2009-03-10 5 5 PT5M 5 minutes after dose 1 Day 5 dose 2009-03-10
2 5 2009-03-10 2009-03-10 5 5 PT5M 5 minutes after dose 1 Day 5 dose 2009-03-10
3 5 2009-03-10 2009-03-10 5 5 PT5M 10 minutes after dose 2 Day 5 dose 2009-03-10
4 5 2009-03-10 2009-03-10 5 5 PT5M 10 minutes after dose 2 Day 5 dose 2009-03-10
5 3 2009-03-10 2009-03-10 5 5 PT5M 15 minutes after dose 3 Day 5 dose 2009-03-10
6 3 2009-03-10 2009-03-10 5 5 PT5M 15 minutes after dose 3 Day 5 dose 2009-03-10
7 5 2009-03-10 2009-03-10 5 5 PT5M 5 minutes after dose 1 Day 5 dose 2009-03-10
8 5 2009-03-10 2009-03-10 5 5 PT5M 5 minutes after dose 1 Day 5 dose 2009-03-10
9 5 2009-03-10 2009-03-10 5 5 PT5M 10 minutes after dose 2 Day 5 dose 2009-03-10
10 5 2009-03-10 2009-03-10 5 5 PT5M 10 minutes after dose 2 Day 5 dose 2009-03-10
11 3 2009-03-10 2009-03-10 5 5 PT5M 15 minutes after dose 3 Day 5 dose 2009-03-10
12 3 2009-03-10 2009-03-10 5 5 PT5M 15 minutes after dose 3 Day 5 dose 2009-03-10
(Cont.) Perm Perm Perm
Row no. MCEVLINT MCSTINT MCENINT 1 PT5M PT0M PT5M
2 PT5M PT0M PT5M
3 PT5M PT5M PT10M
4 PT5M PT5M PT10M
5 PT5M PT10M PT15M
6 PT5M PT10M PT15M
7 PT5M PT0M PT5M
8 PT5M PT0M PT5M
9 PT5M PT5M PT10M
10 PT5M PT5M PT10M
11 PT5M PT10M PT15M
12 PT5M PT10M PT15M
Example 3: Ocular Irritation
Study Number: 8289B
Title: “Primary Eye Irritation Study of Alipal CO-436 in New Zealand White Rabbits”
Study Date: 9/24/84
Accessed at: http://www.epa.gov/oppt/tsca8e/pubs/8ehq/2011/apr11/8ehq-0411-01708b.pdf
Summary:
New Zealand White Rabbits each received a single ocular dose of 0.1 mL of Alipal CO-436 with either no
washout (n=6) or a 30-second washout (n=3) with physiological saline. After administration of the test
article, the eyes were examined at 24, 48, and 72 hours, then at 7, 10, 13, 16, 19 and 21 days. For each
animal, the cornea, iris and conjunctiva were scored for irritation. The test article was administered to
one eye for each animal, with the second eye serving as an untreated control. Irritation was scored on a
classification system (i.e. Draize Scale) where numeric values are used to describe the level of irritation
(Appendix I). The sponsor recorded ocular irritation for each individual animal and then presented group
means for “no washout group” and “30-second washout group” at each time point.
Endpoint Result type Unit
Draize scoring (various tissues) Number (integer) N/A (rank)
Draize totals (derived values) Number (integer) N/A
Data from the study:
Animal
Number
Draize scoring Time
Tissue assessment 24 h 48 h 72 h Day 7 Day 10 84-3240 A - Cornea Opacity 1 1 1 1 1
B - Cornea Area Involved 4 4 2 1 1 Dose: 0.1 mL C - Iris 1 1 1 1 1
D - Conjuctiva redness 2 3 2 2 2 Rinse 30 sec E - Conjuctive Chemosis 3 3 4 3 3
After F - Conjuctiva Discharge 3 3 2 3 2 1 - Cornea Total ((A+B) x 5) 20 20 10 5 5 2 - Iris Total (C x 5) 5 5 5 5 5 3 - Conjuctivae Total ((D+E+F) x 2) 16 18 16 16 14 Total (1+2+3) 41 43 31 26 24
84-3240 A - Cornea Opacity 0 0 0 0 0 B - Cornea Area Involved 0 0 0 0 0
Dose: N/A C - Iris 0 0 0 0 0 Untreated D - Conjuctiva redness 0 0 0 0 0
eye E - Conjuctive Chemosis 0 0 0 0 0 F - Conjuctiva Discharge 0 0 0 0 0 1 - Cornea Total ((A+B) x 5) 0 0 0 0 0 2 - Iris Total (C x 5) 0 0 0 0 0 3 - Conjuctivae Total ((D+E+F) x 2) 0 0 0 0 0 Total (1+2+3) 0 0 0 0 0
The data highlighted in green is the data that is modelled as an example dataset.
The endpoints in the table above are sent through the decision tree, starting from the top left side. Ideally
each of the two endpoints above should be processed through the decision tree separately; however the
outcome through the decision will be the same for both as they are both integers without units, so only
one run through the decision tree is described below.
The figure below illustrates the steps that this particular data will go through and the categories of
variables selected by answering the questions.
Decision steps:
Step 1. Add Mandatory variable from Core Dataset Structure (STUDYID, DOMAIN, --SEQ).
Step 2. Reference level identifiers will not be added, so the answer is “No” to the question in the
Core Domain structure. Variables in CD-1 will not be added.
Step 3. The object being assessed is not a group of subjects (“No”), so the OI-2 variables are added
(USUBJID)
Step 4. The object beings assessed is an intact (entire) subject (“Yes”), because no specimen have
been excised from the subject.
Step 5. The state of the subject is not important for the interpretation of the result (“No”), so OI-4
variables are not added.
Step 6. The test is specific to a location on the intact subject (it is an eye assessment). The answer
is “Yes”, so OI-5 variables are added (--LOC).
Step 7. Add Mandatory variables for “Test” lane (--TESTCD, --TEST)
Step 8. There is no need to further specify the test to distinguish records (“No”), so the TV-1, TV-2
and TV-3 variables are omitted. Though it is tempting to think of “rinse” vs . “No-rinse” as a
methjod, it is more appropriate to place this aspect of the treatment in the EX (Exposure)
Domain (see the SEND Implementation Guide for further information).
Step 9. Add Mandatory variables for “Results” lane (--ORRES, --STRESC, --STAT, --REASND, --
EXCLFL, --REASEX).
Step 10. The results are not character format or categories (“No”).
Step 11. The results are quantitative (numerical scoring), so RV-3 variables are added (--ORRESU, --
STRESN, --STRESU, --BLFL).
Step 12. There are no reference ranges for numerical scoring. RV-4 variables are not added.
Step 13. Add Mandatory Timing Variables (VISITDY).
Step 14. The individual measurement does not have a duration (i.e. “No”). Because it is a
measurement taken at several single points in time, TM-1 variables are added (--DTC, --
DY).
Step 15. The individual measurement is planned relative to a fixed reference (“Yes”), so we add
relevant variables from TM-2 (--TPT, --TPTNUM, --ELTM, --TPTREF, --RFTDTC).
Finally all variables must be sorted so that they are in the same sequence as the Domain Table. The
resulting ocular Irritation domain structure is given below.
Decision based adding of variables Custom SDTM dataset specification – Ocular Irritation data
Decision
Steps
Variable category Variable
group
Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role
Step 1 Core Dataset
structure
Mandatory Req STUDYID Study Identifier Char Identifier
Step 1 Core Dataset
structure
Mandatory Req DOMAIN Domain
Abbreviation
Char XX Identifier
Step 3 Object Identifier OI-2 Req USUBJID Unique Subject
Identifier
Char Identifier
Step 1 Core Dataset
structure
Mandatory Req --SEQ Sequence Number Num Identifier
Step 7 Test Variables Mandatory Req --TESTCD Measurement, Test
or Examination
Short Name
Char Controlled
terminology
is expected
Topic
Step 7 Test Variables Mandatory Req --TEST Measurement, Test
or Examination
Name
Char Controlled
terminology
is expected
Synonym
Qualifier
Step 9 Result Variables Mandatory Exp --ORRES Result or Finding as
Collected
Char Result
Qualifier
Step 11 Results Variables RV-3 Exp --ORRESU Unit of the Original
Result
Char (UNIT) Variable
Qualifier
Step 9 Results Variables Mandatory Exp --STRESC Standardized Result
in Character Format
Char Controlled
terminology
is expected
Result
Qualifier
Step 11 Results Variables RV-3 Exp --STRESN Standardized Result
in Numeric Format
Num Result
Qualifier
Step 11 Results Variables RV-3 Exp --STRESU Unit of the
Standardized Result
Char (UNIT) Variable
Qualifier
Step 9 Result Variables Mandatory Perm --STAT Completion Status Char (ND) Record
Qualifier
Step 9 Result Variables Mandatory Perm --REASND Reason Not Done Char Record
Qualifier
Step 6 Object Identifier OI-5 Perm --LOC Location associated
with a result or
finding
Char Record
Qualifier
Step 11 Result Variables RV-3 Exp --BLFL Baseline Flag Char (NY) Record
Qualifier
Step 9 Results Variables Mandatory Perm --EXCLFL Exclusion Flag Char (NY) Record
Qualifier
Step 9 Results Variable Mandatory Perm --REASEX Reason for
Exclusion
Char Record
Qualifier
Step 13 Timing Variables Mandatory Exp VISITDY Planned Study Day Num Timing
Step 14 Timing Variables TM-1 Exp
--DTC Date/Time of Test Char ISO 8601 Timing
Step 14 Timing Variables TM-1 Exp --DY Study Day of Test Num Timing
Decision based adding of variables Custom SDTM dataset specification – Ocular Irritation data
Decision
Steps
Variable category Variable
group
Core
(within
variable
group)
Variable
Name
Variable Label Type Controlled
Terms,
Codelist, or
Format
Role
Step 15 Timing Variables TM-2 Exp --TPT Planned Time Point
Name
Char Timing
Step 15 Timing Variables TM-2 Exp --TPTNUM Planned Time Point
Number
Num Timing
Step 15 Timing Variables TM-2 Exp --ELTM Planned Elapsed
Time from Time
Point Ref
Char ISO 8601 Timing
Step 15 Timing Variables TM-2 Exp --TPTREF Time Point
Reference
Char Timing
Step 15 Timing Variables TM-2 Exp --RFTDTC Date/Time of
Reference Time
Point
Char ISO 8601 Timing
The dataset outline based on the domain structure above including example data is shown below. I have decided to name it “OC”
(for Ocular Irritation), which is populated in the DOMAIN variable and also used as prefix for all variables with -- in their name in the
domain structure.
The results are numeric, but they are a ranking, so the unit variables (OCORRESU and OCSTRESU) are not used for this data, but
they are left in the dataset as they are both Expected.
Typically the report would reflect which eye that was administered test article and which that was not (left and right). However this
data unless indicate that it was one and the other, so the location of the observation (indicated in OCLOC) is equally unspecific (Eye,
treated and Eye, untreated).
The data that is used in the example below is from 24 hours after test article administration; therefore the planned study day of
assessment is study day 2.
Dosing occurred on November 9th 1984 according to the report, so we can populate the date/time variables (OCDTC and OCRFTDTC)
based on this information.
For this study the following tests and test codes could be
--TESTCD* --TEST*
DSCOOP Draize Score: Cornea Opacity
DSCOAR Draize Score: Cornea Area Involved
DSI Draize Score: Iris
DCCOJRD Draize Score: Conjuctiva redness
DSCOJCM Draize Score: Conjuctive Chemosis
DSCOJDC Draize Score: Conjuctiva Discharge
DSCOTOT Draize Score: Cornea Total
DSITOT Draize Score: Iris Total
DSCOJTOT Draize Score: Conjuctivae Total
DSTOT Driaze Score: Total
*Note: These codes are not published controlled terminology.
Any permissible variables not used to model the data (indicated in red) can be omitted from the dataset.
Req Req Req Req Req Req Exp Exp Exp Exp Exp
Row
no. STUDYID DOMAIN USUBJID OCSEQ OCTESTCD OCTEST OCORRES OCORRESU OCSTRESC OCSTRESN OCSTRESU
1 8289B OC 8289B-
843240
1 DSCOOP Draize Score: Cornea Opacity 1 1 1
2 8289B OC 8289B-
843240
2 DSCOAR Draize Score: Cornea Area
Involved 4 4 4
3 8289B OC 8289B-
843240
3 DSI Draize Score: Iris 1 1 1
4 8289B OC 8289B-
843240
4 DCCOJRD Draize Score: Conjuctiva redness 2 2 2
5 8289B OC 8289B-
843240
5 DSCOJCM Draize Score: Conjuctive Chemosis 3 3 3
6 8289B OC 8289B-843240
6 DSCOJDC Draize Score: Conjuctiva Discharge 3 3 3
7 8289B OC 8289B-
843240
7 DSCOTOT Draize Score: Cornea Total 20 20 20
8 8289B OC 8289B-843240
8 DSITOT Draize Score: Iris Total 5 5 5
9 8289B OC 8289B-
843240
9 DSCOJTOT Draize Score: Conjuctivae Total 16 16 16
10 8289B OC 8289B-843240
10 DSTOT Driaze Score: Total 41 41 41
11 8289B OC 8289B-
843240
11 DSCOOP Draize Score: Cornea Opacity 0 0 0
12 8289B OC 8289B-843240
12 DSCOAR Draize Score: Cornea Area Involved
0 0 0
(Cont.) Perm Perm Perm Exp Perm Perm
Row no. OCSTAT OCREASND OCLOC OCBLFL OCEXCLFL OCREASEX 1 Eye, Treated 2 Eye, Treated 3 Eye, Treated 4 Eye, Treated 5 Eye, Treated 6 Eye, Treated 7 Eye, Treated 8 Eye, Treated 9 Eye, Treated
10 Eye, Treated 11 Eye, Untreated 12 Eye, Untreated
(Cont.) Exp Exp Exp Exp Exp Exp Exp Exp
Row no. VISITDY OCDTC OCDY OCTPT OCTPTNUM OCELTM OCTPTREF OCRFTDTC
1 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03
2 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03
3 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03
4 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03
5 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03
6 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03 7 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03 8 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03 9 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03 10 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03 11 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03 12 2 1984-09-04 2 24 hours after dose 1 PT24H Dosing of eye 1984-09-03