160 likes | 181 Views
This presentation discusses the setup steps for lipid ADaM data analysis, including the creation of the ADaM Basic Data Structure (BDS) and traceability between analysis data and source data.
E N D
A Taste of ADaM Beilei Xu Accenture Changhong Shi Merck Sharp & Dohme Corp. Presented by: Peng/Zik Liu MSD (Shanghai) Pharma Co.
Outline • Background • ADaM • Setup Steps for Lipid ADaM data • Summary
Background • CDISC -Clinical Data Interchange Standards Consortium • SDTM - Study Data Tabulation Model • Standard for interchange of collected data • ADaM - Analysis Data Model • Standard for analysis data http://www.cdisc.org/adam
ADaM Standard Data Structure • ADSL: Subject level analysis data • one record per subject • subject-level population flags, planned and actual treatment variables, demographic information, randomization factors, sub-grouping variables, and important dates • BDS: Basic Data Structure • Long and skinny structure: contains one or more records per subject, per analysis parameter, and per analysis time point • “One-proc” away readiness for analysis • Traceability
BDS Variables • A central set of variables: • The analysis parameter: e.g., PARAM • The value being analyzed: e.g., AVAL and AVALC • Other variables: • Provide more information about the value being analyzed (e.g., the subject identification) • Describe and trace the derivation of the variable (e.g., DTYPE) • Enable the analyses (e.g., treatment variables, covariates)
Implementation Consideration • Number of ADaM datasets needed • Derivation of analysis endpoints, analysis windows, analysis values, and imputation of missing values • Setup of analysis flags and population flags
Lipid Analysis Data - ADLP • Lipid endpoints: • LDL - C • HDL - C • LDL/HDL ratio • Analysis time points: • SCREENING • BASELINE • WEEK 2 • WEEK 4 • Analysis population- Full Analysis Set
Lipid Analysis Data - ADLP • Subject identifiers: STUDYID, USUBJID, SUBJID, and SITEID • Treatment variables: TRTP, TRPA, TRTPN, and TRTAN • Analysis parameter variables: PARAM, PARAMCD, PARAMN, and PARAMTYP • Analysis timing variables: ADT and ADY • The analysis value variables: AVAL, BASE, and CHG • The analysis flag variable - ANL01FL • The parameter population flag - FASPFL • The traceability variables: SRCDOM, SRCVAR, and SRCSEQ
ADLP Setup Steps • Obtain Variables from Source SDTM LB Domain • Derive New Analysis Endpoints (PARAMTYP) • Handle Negatives (or under detection) and Multiple Records on the Same Date (DTYPE) • Set Analysis Flag Variables (ANLzzFL) • Compute Change, Percent Change from Baseline (BASE, CHG, PCHG) • Set Population Flag Variables
Obtain Data from SDTM LB Domain SDTM LB Domain: ADLP: 10
One-Proc Away procmixed data=adlp; where paramcd=‘LDL’ and anl01fl=‘Y’ and faspfl=‘Y’ and avisitn in (0,2,4); class subjid avisit trta; model chg=avisit trta trta*avisit; repeated avisit/subject=subjid type=un; run;
Summary • The setup steps shown above enable: • the creation of the ADaM Basic Data Structure (BDS) • traceability between analysis data and source data • "one-proc" away readiness for analysis • Further development can be made to standardize the programs for analysis data setup