Made with ScholarlyTurn your own files into documents like this
SAS, CDISC & SDTM Foundations — Part 1
Big picture
Clinical-trial data work moves from conducting the trial, to collecting, cleaning, and validating the data, and then to submitting standardized data to regulators. SDTM (Study Data Tabulation Model) provides a common structure for submission datasets. The workshop introduces raw Excel data, SDTM domains and variables, validation, and SAS Studio import steps.
Raw trial data and SDTM basics
CRF (Case Report Form): A form used to capture information collected about trial participants. The collected source fields are raw data; they are not yet SDTM datasets.
Domain: A subject-area dataset. Examples include DM (Demographics), VS (Vital Signs), LB (Laboratory), AE (Adverse Events), MH (Medical History), EX (Exposure), and DS (Disposition).
Variable: A field in a domain. Raw field names and values are mapped into standard SDTM variables and conventions.
USUBJID: The unique subject identifier used consistently to link a participant’s records across domains.
RFSTDTC / RFENDTC: Subject reference start and end date variables.
Standardized values matter too: categorical fields such as sex and race should follow controlled terminology, and planned versus actual treatment-arm information should be checked for consistency.
Mapping examples
DM — Demographics
A raw subject ID can be mapped to USUBJID; date of birth to BRTHDTC; and the dataset is identified as domain DM. The example derives age from birth date and a reference date, and removes names from the submission-style data.
AE — Adverse Events
Examples of raw-to-standard mappings include severity → AESEV, seriousness criteria → AESER, relationship to study treatment → AEREL, outcome → AEOUT, and event start/end dates → AESTDTC / AEENDTC. AESEQ provides a sequence number for records within a subject.
EX — Exposure
The EX domain records investigational-product administration. Example fields include drug, dose, dose units, route, and start/end dates; named variables include EXDOSE, EXDOSU, EXSTDTC, and EXENDTC.
VS — Vital Signs
A combined measurement such as blood pressure is represented as separate test records, such as systolic and diastolic pressure. Important variables include:
VSTESTCD — test code; VSTEST — test name
VSORRES — original result; VSORRESU — original-result units
VSSEQ — record sequence; VSPOS — position; VSDTC — measurement date/time
Examples include height, weight, and systolic/diastolic blood pressure. The speaker describes test codes as uppercase and test names as mixed case. Dates use ISO 8601 conventions; date/time variables commonly end in DTC.
Data validation and comparison
Before treating a dataset as submission-ready:
Sort or review records by subject identifier; look for unexpected duplicate records and missing values.
Check date formats and chronology—for example, whether exposure, consent, treatment, and vital-sign dates make sense relative to one another.
Compare the programmed output with the expected SDTM structure: variable names and order, types, labels, lengths, formats, and controlled terminology.
The emphasis is on checking metadata and standardized values as well as the data values themselves.
SAS Studio workflow: importing raw Excel data
Access SAS OnDemand, create/sign in to an account, and launch SAS Studio. The speaker describes it as free for students and academic users.
In SAS Studio, use Upload (the up-arrow) to select and upload the Excel workbook.
Use PROC IMPORT to read a worksheet into a SAS dataset. The demonstrated options include:
DATAFILE= — uploaded workbook path
OUT= — output SAS dataset
DBMS=XLSX — Excel workbook engine
REPLACE — replace an existing output dataset
SHEET= — worksheet to import
GETNAMES=YES — use the first row as variable names
Repeat for the raw DM, AE, and VS worksheets, using the correct file path and worksheet name.
Inspect the imported data and metadata (variable names, types, lengths, and formats) before mapping.
Key distinction: PROC IMPORT brings the raw worksheet into SAS; it does not by itself create SDTM domains or standard variables.
SAS Studio areas
Code: Write and run SAS programs.
Log: Review notes, warnings, and errors from program execution.
Results/output: View generated output.
The hands-on workshop then uses imported raw data to build DM and VS SDTM datasets.
Quick review
What is the difference between a raw worksheet and an SDTM domain?
Why should USUBJID remain consistent across domains?
How are systolic and diastolic blood pressure represented in VS?
Name three structural or metadata features to compare when validating SDTM output.
Which PROC IMPORT option specifies the worksheet, and which uses the first row as variable names?
Answers: Raw worksheets contain source fields; SDTM domains use standardized structures and variables. Consistent USUBJID links a subject’s data across domains. Systolic and diastolic are separate test records. Validation includes names/order, types, labels, lengths, formats, and controlled terminology. SHEET= chooses the worksheet; GETNAMES=YES uses first-row names.