Source document design for clinical research sites
Everything your trial knows lives in the source.
Protocol, EDC build, lab manual, visit schedule, training — consolidated into one document your coordinator opens instead of six tabs and a shared drive.
48–72 hour turnaround. Built by someone who has sat where your coordinator sits.
| Field | Record | Unit |
|---|---|---|
| Event term | Verbatim | — |
| Start date | DD-MON-YYYY | date |
| Severity grade | 1 2 3 4 5 | CTCAE v5.0 |
| Source | CRN / SPN / MD | code |
| Related to study drug | Y / N / UNK | — |
Draft — for qualified review and approval
What goes wrong
Four ways a source document stops working
These are the failure modes that turn into queries, rework at 10pm, and findings in a monitoring letter. Each one has a specific cause, and none of them is a formatting problem.
Field instructions nobody can interpret
Units missing, grading standard unstated, source of truth unspecified. Every coordinator fills it in differently, and the data proves it.
Timepoints that don't match the protocol
Visits that skip a required assessment, extra fields a study visit never collects, windows written differently from the Schedule of Assessments.
A data set that doesn't reach the EDC
Fields on paper that aren't collected electronically, EDC fields with no home, and manual reconciliation after every visit.
Structure that breaks in the binder
Grids that split across pages, no subject ID or page numbering, no correction convention, unusable once it's printed in black and white.
What you get
A document set, and the crosswalk that proves it reconciles
We take the inputs your trial already holds — protocol and Schedule of Assessments, EDC build, lab manual, visit schedule, training material — and consolidate them into the source document your site actually works from.
- Every document type the Schedule of Assessments calls for, in
.docxwith editable headers and footers - Field-level instructions: units, expected source, grading standard with version
- Header carrying site number, PI, protocol and version — footer carrying page x of y and effective date
- Stated correction convention: strikethrough, initial, date. Never overwrite
- Printed in black and white, tested on a real copier before it ships
- 48–72 hours from intake to delivery
| Protocol element | Source doc | EDC field | Type & unit |
|---|---|---|---|
| 4.2 Severity | AE Log | AE_SEV |
CTCAE v5.0 |
| 5.1 Hemoglobin | Labs | LB_HGB |
g/dL |
| 6.3 Tumor response | Efficacy | RS_EVAL |
RECIST 1.1 |
| 7.1 Conmed start | Conmeds | CMB_STDT |
date |
One row per required data element. The crosswalk is the artifact a monitoring visit checks first, and it is the one thing a site cannot rebuild from a document alone.
How every document is built
Every field, accounted for
The checklist we run before anything leaves us. It is the same checklist we would want on our own clipboard during a monitoring visit.
A Field instructions
- Units stated on every numeric field
- Graded fields cite the standard and its version
- Every field names its expected source: lab report, device printout, subject report, EDC
- “Record the highest grade” rule present wherever multiple toxicities can occur
- Date format, reference ranges and out-of-window handling specified
- Every field defines blank, unknown and not-done — none may be silently empty
B Timepoints against the protocol
- Visit names and windows match the Schedule of Assessments exactly
- Unscheduled and repeat visits accommodated
- Every visit carries exactly its required fields — no gaps, no extras
- Crosswalk shows every protocol-required element with a home
C Data set against the EDC build
- Every source field exists in the EDC, and every EDC field has a source
- Anything present only on paper is flagged as intentional source-only
- Repeated and parent-child structures modelled correctly: labs by visit, events by term
- Data-distribution-relevant fields present
D Structure in the binder
- Header: site number, PI, protocol number, version, subject ID
- Footer: page x of y, document version, effective date
- Correction convention stated: strikethrough, initial, date, never overwrite
- No grid row broken across a page break
- Legible printed in black and white, and complete in Word without reflowing
Services and pricing
Build your study the way you need it
Tick what you need and the total updates as you go. This is a scoping tool, not a checkout — you get a quote, and site teams rarely need more than a purchase order.
Additional sites beyond five are quoted. Maintenance runs for the length of the study and can be added or dropped when you buy. No rush surcharge — tell us your kickoff date and we will tell you honestly whether we can hit it.
Source maintenance
A source document that survives a three-year study
A study is not a one-time build. The protocol amends, the EDC gets rebuilt, a new coordinator joins in year two. Maintenance is what keeps the document true to the study instead of slowly drifting away from it.
- Unlimited protocol amendments, each issued within two weeks
- EDC field mapping updated whenever the build changes
- Version and effective-date control kept continuous
- Quarterly review against the current protocol
- Support for questions that come up around monitoring visits
- Amendments clustered at the start of a study included
| Change | We update | Turnaround |
|---|---|---|
| Protocol v3 → v4 | Visit structure, field set, crosswalk | 10 working days |
| EDC rebuild | Field mappings and data set check | 5 working days |
| New lab manual | Reference ranges and source notes | 5 working days |
| New coordinator | Onboarding pack and binder walkthrough | On request |
Maintenance is quoted per study, not per change. You are not counting amendments; we are counting what it costs to keep up.
A look at the work
What a finished document looks like
Redacted from real engagements. This is the level of instruction and document control you are paying for.
| Test | Result | Unit |
|---|---|---|
| Haemoglobin | value | g/dL |
| Platelet count | value | ×10ⁿ/L |
| Creatinine | value | mg/dL |
| Source | central lab | code |
- Units on every row, so nobody has to look them up
- Source of truth stated — and what not to use
- Out-of-range values are still recorded, not silently dropped
- Placeholder version and effective date in the footer
Start here
Four fields, then we take it from there
Tell us who you are and what is going on. Everything else happens in the reply, where you send the protocol anyway.
- Already know what you need?Email us directly and skip the form. Include the protocol number, the site, and the documents you want.
- What happens nextWe reply within one working day with a scoped quote, and countersign the NDA before opening anything.
- ConfidentialityProtocols are pre-publication material. NDA countersigned before any file is opened, and never used to train a model.
Not sure what you need yet? Book 15 minutes
Tell us the study and what is going wrong with the documents you have now. We will say what we would build and roughly what it costs. No deliverable, no obligation, no sales call.
Questions
What sites ask first
Is this a validated form?
No, and we do not claim it is. We draft to your protocol. Your PI or your sponsor's qualified person reviews and approves before anyone uses it. Every document ships as a draft, marked as such, with a version footer and a stated correction convention — because that is what your monitoring binder expects to find.
Why can't I attach the protocol here?
We take it as an email reply rather than a form upload, so it lands in the same thread as your confirmation and quote. Send the file as an attachment or a secure link, and we countersign the NDA before opening anything.
Why not just use the document our CRO sent?
Often you can, and we will say so. The problem is when it does not match your EDC build, your Schedule of Assessments, or the way your site actually runs a visit. We read all three and tell you plainly where they disagree.
How fast is delivery?
Tell us your kickoff date and we will say honestly whether we can hit it. The 48–72 hour figure you may have seen applies to a document audit. A full study engagement is scoped after we read the protocol.
What happens when the protocol amends?
With source maintenance, unlimited amendments are included and issued within two weeks. Studies typically amend most in the first months, and that clustering is covered. Without maintenance, a re-issue is quoted as a one-off.
What happens to our protocol?
Protocols are unpublished material and often your sponsor's core intellectual property. We sign an NDA before opening any file, we do not retain documents after delivery without a written request, and no protocol content is ever used to train a model.
Who owns the document?
You do. Edit it, extend it, reuse it at other sites. Up to five sites are included in the study engagement; beyond that we quote per site.
Can we use licensed assessment instruments?
We will flag any instrument with license terms when we build, and tell you to confirm your rights before printing. We do not supply proprietary instruments and we do not strip copyright notices off existing forms.
Do you do the EDC build too?
No. We read it, reconcile against it, and hand you the crosswalk that shows where everything lives. You keep your own builder.