Core concepts
Randomisation: Assigning participants to treatment groups by chance to reduce bias and improve validity.
Randomised subject: A participant who has been assigned to a group.
Randomisation ID: Unique, incremental code assigned after randomisation (e.g., R‑0001). Combined with the Subject ID, it uniquely identifies the participant in the study.
Randomisation allocation: The action of assigning a specific subject to a group once eligibility is confirmed.
Randomisation group: The treatment/intervention arm a subject is allocated to.
Allocation ratio / Group weight: Target proportion for group assignment (e.g., 1:1, 2:1). In Datacapt, set via Group Weight (e.g., entering
1-1for two groups).
Methods & lists
Randomisation list — The sequence of assignments.
Static list: Pre-generated at setup (used by Block method or Import).
Dynamic list: Generated on the fly per subject (used by Dynamic method).
Static randomisation (Block method) — Uses fixed block sizes (e.g., 4, 6) to keep groups balanced at regular intervals. Datacapt generates the list after you save the configuration.
Block size — Number of consecutive allocations treated as a unit to maintain balance.
Dynamic randomisation (Pocock & Simon): Recalculates the best next assignment for each subject to maintain balance across chosen factors (strata).
Distance methods (how imbalance is measured):
Range: Minimises the spread between the largest and smallest group counts within strata.
Variance: Minimises variance of group counts.
Max: Minimises the single worst imbalance across strata.
Stratification
Stratification: Balancing groups within predefined categories to control important prognostic factors.
Stratum (plural: strata): One category level within a stratification factor (e.g., Age <65, Age ≥65).
Stratification factor: The characteristic you balance on (e.g., centre, baseline severity).
Stratify by centre: Maintains balance per site.
Stratify by question: Uses a defined eCRF question/variable as a factor (must exist in the eCRF builder).
Operational terms
Randomisation slot: The next position in the sequence to be allocated.
Randomisation kit: Non‑identifiable code tied to product/pack used to keep assignments blinded.
Randomisation configuration: The saved method, groups, weights, block size (if applicable), and stratification options that govern allocation.
List import: Uploading an external randomisation list (contact your CSM).
Double‑blind import: Importing an external list that includes blinded kit IDs (contact your CSM).
Blinding
Blind study: At least one party (participant and/or investigator) does not know the assignment.
Single blind (simple blind): Participants are blinded; investigators are aware.
Double blind: Both participants and investigators are blinded.
Unblind action: Revealing treatment assignments under predefined conditions (e.g., study end or safety trigger).