Background Story :
One day, my friend 99 discussed with me if a patient is randomized but not treated has some lab results besides screening, how should we classify the EPOCH variable for him/her in SE domain?
So today we will go over a basic example for SE domain: Subject Element.
The SE stores treatment data linked together with demographic data and Trial Design data, which is possible to standardize and create unique data-driven mapping for a high level view of study.
SE (Subject Element) contains subject level data about the timings of each subject transition from one element period to the other (screening, treatment, follow-up) and the treatment information (active, non-active) from multiple sources.
STUDYID, DOMAIN, USUBJID, SESEQ, ETCD, ELEMENT, SESTDTC, SEENDTC,TAETORD,EPOCH
Variable Derivation Rules:
*1.Seseq Sequential number is unique for the records within the domain by USUBJID. *2.ETCD Element Code is assigned (SCREEN, Treatment, Placebo, FUP). *3.ELEMENT Description of Element is assigned.(Screening, Treatment info, Follow up). *4.SESTDTC Start date of the element depends on Planned Arm relates to Informed consent date, Reference start date, Treatment Exposure Date. *5.SEENDTC End date of the elements equal to the next SESTDTC of next element. *6.TAETORD Planned Order of Elements within Arm is directly from TA *7.EPOCH Epoch is directly from TE.EPOCH it originated from to reference start/end date.
SE Mini Sample Data Set:
Note: The dataset is a dummy sample, real data will have different SESTDTC, SEENDTC dates
To end the story we started, SE would only keep the Screening record for a patient without treatment even he/she has follow up records.
Thanks 99 for sharing a case study with me!