Study build and CRF design

Review status

Included in the proposed model: EDC Study, Study Design Version, Design Publication, Site Design Assignment, Casebook Design Adoption, CRF Definition, CRF Placement, Item Group Definition, Item Group Placement, Item Definition, Item Placement, Codelist, Unit List, Condition, Derivation, Edit Check, Design Validation, and the initial references to independently versioned review, coding, and signature plans.

Open: formal approval roles, localization depth, and whether reusable standards libraries are required in the first release.

Clinical purpose

Study build turns the protocol, schedule of activities, data-management plan, monitoring plan, coding conventions, and statistical requirements into the exact collection design that sites will use. The design must be understandable before first-subject-first-visit and reconstructable years later.

The design is not merely a set of screens. It determines what data are expected, when they are expected, what values are allowed, which discrepancies are raised, what requires review, and what the investigator will eventually sign.

Records and fields

EDC Study

Field	Clinical meaning
Study reference	Link to the canonical clinical study
EDC study name and number	Recognizable EDC identity used by study teams
Blinding model	Which roles and transfers may see restricted data
Subject-numbering policy	Site prefix, sequence rules, correction policy
Active design versions	Versions currently assigned to one or more sites or casebooks
Study status	Build, testing, live, suspended, closing, archived

The EDC Study is responsible for EDC participation. It does not redefine the protocol, sponsor, countries, or sites owned by the clinical-suite platform.

Study Design Version

Field	Clinical meaning
Version number	Human-recognizable casebook/design version
Protocol basis	Protocol version and amendment references addressed by the design
Predecessor	Earlier published design from which an amendment was prepared
Change summary	Clinically meaningful account of what changed and why
Build owner	Accountable clinical programmer or study designer
Authoring/approval status	Draft, in review, validation failed, validation passed, approved, withdrawn
Approval evidence	Approver, role, date, decision, and any conditions

A design version contains the complete schedule, CRFs, terminology, and collection rules needed to run it. It records the exact initial versions of independently governed review, coding, and signature plans. Subject collection never looks up “the latest” library component or plan.

Design Publication, Site Design Assignment, and Casebook Design Adoption

Record	Clinical decision and fields
Design Publication	Makes one approved design available for production; publication number, approved version, publisher/time, validation evidence, content inventory and integrity evidence
Site Design Assignment	Selects the publication used for new casebooks at one EDC site; site, publication, effective time, assigner, reason and prior assignment
Casebook Design Adoption	Moves one existing casebook to a later publication after compatibility review; casebook, from/to publication, assessment, conflicts, resolutions and adoption time

Publication does not assign a site. Site assignment does not update existing casebooks. Casebook adoption does not alter the published design.

CRF Definition

Field	Clinical meaning
Stable CRF identifier	Identity retained across versions where clinical meaning remains the same
Name, short label, and purpose	Terms visible in specifications and casebooks
Completion instructions	Study-specific guidance for site users
Item groups	Ordered sections and repeating logs
Completion policy	What must be answered or accounted for before submission
Signature inclusion	Whether the form is included in investigator endorsement
Standard origin	Adopted standard and declared study-specific differences

CRF Placement

A placement is the use of one CRF at one visit or event. It carries:

• visit/event reference;
• display order;
• expected, optional, or conditional status;
• placement-specific instructions;
• condition controlling applicability;
• repeat policy; and
• any placement-specific review or signature treatment.

Changing the reusable CRF changes every linked placement in the same draft. Copying a CRF creates an independent definition. Detaching one placement creates an independent copy with recorded source history.

Item Group Definition and Item Group Placement

The definition describes a reusable clinical section or log. The placement says where that section is used in a CRF and how it behaves there.

Definition field	Clinical meaning
Name and label	Recognizable CRF section or log
Row label fields	Values used to distinguish adverse events, medications, samples and similar rows
Item placements	Ordered uses of reusable item definitions

Placement field	Clinical meaning
Item-group definition	Exact reusable section used
CRF definition	Parent CRF in which it appears
Order and display label	Position and contextual wording
Repeatability and limit	Single section or repeating rows; maximum when clinically bounded
Applicability	Condition controlling whether the section is expected
Completion rule	Whether created rows must be accounted for before form submission

Item Definition and Item Placement

The Item Definition carries stable clinical meaning. The Item Placement is the use of that item in a particular item group. This prevents a reused standard question from gaining one global requiredness, order, or contextual instruction.

Item Definition field	Clinical meaning
Stable item identifier	Study-local identity retained across compatible versions
Clinical definition	Observation or activity being collected
Datatype	Text, integer, decimal, date, datetime, coded response, boolean, or other approved type
Length and precision	Valid representational bounds
Codelist	Allowed submitted codes and displayed labels
Unit list	Allowed units and standard/reporting unit policy
Date precision	Whether unknown day, month, or time is permitted
Collection standard mapping	CDASH meaning and terminology where applicable
Submission mapping	Intended SDTM target or transformation guidance
Criticality	Importance to participant safety or reliability of trial results

Item Placement field	Clinical meaning
Item definition	Exact reusable clinical question used
Item-group placement	Parent section in the CRF design
Question text and prompt	Contextual wording shown to the site
Help and completion instructions	What to report and from which source/timepoint
Order	Position within the item group
Requiredness	Always, conditional, or optional in this placement
Applicability/visibility	Conditions controlling whether the item applies or displays
Source-data designation	Whether direct eCRF entry is intended to be source here

One Item Definition can have several Item Placements. Each placement has one definition and one parent item-group placement. Subject datapoints are created for Item Placements, not directly for the reusable Item Definition.

Question wording, datatype, terminology, and clinical definition must agree. A field called “Date of onset” cannot secretly collect a text narrative because storage convenience allows it.

Codelist and Unit List

A codelist retains stable list identity, exact version, submitted code, displayed label, order, enabled/retired status, and external terminology reference. Label correction does not change the submitted meaning. Removing a value from future entry does not erase its historical use.

A unit list retains allowed units, preferred entry unit, standard reporting unit, and conversion policy. Unit conversion is a declared clinical calculation, not display formatting.

Plans owned by study build

Plan	What it decides	What it does not prove
SDV Plan version	Which items/forms/subjects require source data verification	That SDV was performed
Data-Management Review Plan version	Which records require clinical/logical review	That the data are clean
Coding Configuration version	Which verbatims require which dictionary and context	That a term was coded correctly
Signature Plan version	Which milestones and data the investigator must endorse	That a signature is due or completed
Export Specification	Required raw and standards mappings	That a particular extract was complete

Separate status journeys

Design authoring and approval

Draft
→ In clinical/programming review
→ Validation failed → Draft
→ Validation passed
→ Approved

Draft / In review / Approved → Withdrawn, with reason

Design Publication availability

Prepared from one approved Design Version
→ Published and available for assignment
→ Retired from new assignment

Retirement never changes casebooks already using the publication.

Site Design Assignment

Planned → Effective → Superseded by a later effective assignment
                     ↘ Ended, with reason

Casebook Design Adoption

Planned → Compatibility assessed
        → Conflicts requiring resolution → Reassessed
        → Authorized → Applied
                     ↘ Failed, with retained reason and safe prior assignment

Approval, publication, site assignment, and casebook adoption are independent. None is a status of the other record.

Actions

Action	Required information	Result
Start initial design	study, protocol basis, accountable builder	Complete editable draft
Start amendment design	predecessor, amendment or defect reason	New draft with recorded history
Add or revise CRF	clinical purpose, groups, items, instructions	Draft CRF revision
Place CRF	CRF, visit/event, order, applicability	Contextual use without copying definition
Copy CRF	source and reason	Independent identifiers and retained source history
Adopt standard component	exact approved source revision, adoption mode	Exact study-owned content
Define terminology	codes, labels, version, external reference	Valid controlled responses
Define condition/derivation/check	clinical intent, inputs, result/effect	Data-type-checked, testable study rule
Validate design	full design package	Findings or passed validation report
Approve design	review evidence and approver authority	Approved version
Publish design	approved version and release evidence	Production package that cannot be changed
Assign publication to site	site, publication, effective time, authority	New casebooks use that publication
Adopt publication for casebook	compatibility assessment and resolved conflicts	Existing casebook moves with retained adoption history

Business rules

1. A published design cannot be changed; correction requires a new version and publication.
2. Every amendment names its predecessor and reason.
3. Every placed CRF resolves to an exact CRF definition in the same publication.
4. Every item-group and item identifier is unique within the published design.
5. Every coded item resolves to one exact codelist version.
6. Every unit-bearing item declares allowed units and the meaning of conversion.
7. Conditions, derivations, checks, schedules, and mappings reference valid, data-type-compatible Item Placements.
8. Circular rule dependencies prevent publication.
9. Requiredness, visibility, and applicability remain separate decisions.
10. A design cannot be assigned to a production site until validation and required approvals pass.
11. Retirement prevents new use but never breaks historical casebooks.
12. Collection standards guide meaning; exchange and submission structures do not dictate ownership of operational clinical records.
13. Publication, site assignment, and casebook adoption are separate decisions with separate evidence.
14. Independently governed plan versions never change a published design silently; assignments retain effective dates and prior versions.

Validation evidence

A publishable design retains:

• protocol and amendment traceability;
• annotated CRFs and design specification;
• all definitions, placements, terminology, units, and rules;
• rule dependency and cycle checks;
• schedule and repeat-limit checks;
• sample/expected test cases and results;
• unresolved finding disposition;
• reviewer and approver evidence;
• publication package identity and integrity value; and
• rule-processing version needed to reproduce behavior.

Exceptions that require deliberate handling

• A translation changes the apparent clinical meaning of a question.
• A standard CRF is adopted but the study needs a justified deviation.
• Two items share a label but collect different clinical concepts.
• A codelist value is retired after it was used in existing studies.
• A derivation needs more than one occurrence of a repeating row and no selector is defined.
• An item is required while its containing form is conditional.
• A form is linked at several visits but one visit requires different wording.
• A critical endpoint item lacks a source designation or completion instruction.

Handoffs

• Platform supplies Study and Study Site identities and evidence coordinates.
• Study Startup or CTMS may supply approval milestones; EDC makes its own site-activation decision.
• RTSM supplies randomization/treatment facts without giving EDC ownership of allocation.
• eCOA and laboratories supply data with origin and transfer evidence.
• Statistical programming consumes versioned extracts and mappings, not the mutable live casebook.