Compliance question · SpreadsheetsUpdated October 13, 2026

Is Excel compliant for clinical trial data? Part 11 and audit-trail gaps

Part 11 neither certifies nor bans Excel. A workbook can be used under validation and procedure, but its file-based design makes a trustworthy audit trail, access control and version control hard to evidence.

  • File control and macros
  • Validation burden
  • Switch plan

Free sandbox · No credit card · 21 CFR Part 11 aligned

Validating one study workbook · demo effort
123456789101112

weeks

Requirements and risk assessment
Build: protection, validation rules, macros
Test scripts and executed evidence
Summary report and approval
Re-test after any change

Repeats

Illustrative planning view with invented durations, not a benchmark

The direct answer

  • Excel is not "compliant" or "non-compliant" in itself. Part 11 is met by a system's controls, your validation and your procedures. A validated, access-controlled workbook can be defended for some uses, but few teams can show all the controls for a participant database.
  • The weak points are structural. A workbook is a file: copies multiply, anyone with the file can edit it, and there is no built-in, tamper-evident, per-field audit trail with a reason for change.
  • Change tracking is limited and depends on where the file lives. Per secondary sources we could check, Show Changes needs a file in OneDrive or SharePoint and covers a limited window, and Version History keeps whole-file snapshots. Confirm against Microsoft's own documentation.
  • Macros and formulas are software. A workbook with VBA, lookups or derived columns needs the same validation thinking as any computerised system, and every edit to the logic is a change to a validated state.
  • Capture is the form-based alternative: field-level audit trail, edit checks, role-based access and approved, locked forms, tested free in a sandbox. Part 11 is a shared responsibility either way.

The question behind the question

Why "is Excel compliant?" has no yes or no answer

Searching for a verdict on Excel usually ends in contradictory advice, because the right answer is conditional. 21 CFR Part 11 asks that an electronic records system is validated, limits access to authorised individuals, keeps secure computer-generated, time-stamped audit trails, and links signatures to records. ICH E6(R2) section 5.5.3 asks that electronic trial data systems are validated, have SOPs, maintain an audit trail and protect against unauthorised access. The EU's Annex 11 (a GMP annex that inspectors and the EMA's guideline on computerised systems in clinical trials use as a reference point) adds expectations on change and configuration management, security and archiving. None of these names Excel. All of them ask what you did with it.

Excel's design makes the "what you did" part heavy. The unit of control is the file. A data manager can protect a sheet, restrict editing and add data validation lists, but the workbook still travels by email, USB stick and shared drive, and each copy is a new object to control. For a pilot with a handful of non-identifiable values this may be acceptable if your sponsor agrees in writing. For a participant database it is a lot to evidence.

Teams considering the move can read how to migrate from Excel spreadsheets and the general Capture vs Excel comparison. This page focuses on the compliance reasoning and on what auditors will ask.

Control by control

Excel's specific gaps against Part 11, Annex 11 and ICH GCP

Regulatory references are general; Excel statements come from secondary sources checked on 13 October 2026 and should be confirmed in Microsoft's documentation for your version and licence.

ExpectationWhere Excel needs workCapture
Audit trail (11.10(e); Annex 11 s.9)Show Changes needs OneDrive or SharePoint storage and covers a limited recent window; Version History is file-level snapshots, not a per-field record with reasons. Local and e-mailed copies have none.Field-level, append-only, hash-chained; user, timestamp, old and new value, reason
Validation (11.10(a); Annex 11 s.4)Formulas, lookups, conditional formats and VBA macros all need specification and test evidence; Excel version updates may change behaviour.Sponsor validation and UAT documentation for enterprise customers; shared responsibility
Change control (Annex 11 s.10)Anyone who can unprotect a sheet can change logic. Is each change reviewed, tested and recorded?Draft to approved lifecycle; approved forms locked for live use
Access limitation (11.10(d))Passwords on sheets and workbooks are weak controls and are shared; no per-user rights inside the file.Role-based access, row-level security, two-factor authentication for staff
Accurate and complete copies (11.10(b))Exports are the file itself; copies drift. Which one is the record?One database; CSV and Excel exports with an automatic data dictionary
Data checks at entry (ICH E6 5.5.3)Data validation rules reject some inputs but raise no query to the site.Edit checks raise an auto-query
Signatures (11.50, 11.70, 11.200)No signature with meaning and re-authentication linked to the record.Password re-authentication and a meaning statement for staff signatures
Backup and retention (Annex 11 s.7, s.17)Depends on where people saved the file.Hosted in the EU (Frankfurt) or the USA (N. Virginia) with soft-delete only

General information, not legal or regulatory advice. Your quality unit decides what is adequate.

The hidden validation scope

A derived column is a function you have to validate

A study workbook usually contains logic: BMI from height and weight, a visit window from two dates, a flag when a value is out of range. Each is a function whose output people trust. In Excel you must specify it, test it across edge cases, protect the cells and re-test whenever the formula changes. In Capture, calculated fields such as BMI are built in, read-only during entry, and filled from mapped inputs.

  • BMI, BSA, MAP, eGFR (CKD-EPI 2021), Cockcroft-Gault, QTc (Bazett and Fridericia) and more are built-in calculators.
  • The value appears read-only, so a site cannot overtype it.
  • Changes to a form go through draft and approval, with an audit record.
ECG · Visit 4
Demo study · Subject 004-0009Draft

Heart rate

68bpm

QT

412ms

QTcF

Fridericia, calculated

404 ms Calculated

Overall interpretation

NormalAbnormal, not clinically significantAbnormal, clinically significant
Demo values. Calculation shown read-only in the form

See a calculated field that cannot be overtyped

Open the sandbox, add an ECG form and change the inputs. No credit card, every feature.

Try a form-based build free

What you would have to build around it

If you keep Excel: the minimum control set a reviewer expects

Teams that continue with Excel for regulated data usually combine technical settings with procedures. The list is not exhaustive and it must be tailored to your risk assessment.

Controlled storage and naming

One master file in a managed location, a naming and versioning convention, no local copies, and a procedure for what happens when someone saves elsewhere. Record who held the master at each point. This is the file-control layer that an EDC replaces with a single database.

Protected template, validated once

Build entry sheets as a locked template: protected cells, data validation lists, derived columns, no macros if you can avoid them. Validate the template against a written specification and re-validate after any change or after significant Excel updates. See computer system validation for clinical trial software and the data validation plan template.

Change log and periodic review

Because built-in history is partial, teams add a manual or scripted change log with a reason and an independent reviewer. That log is itself a record to protect. See audit trail software for clinical trials for what a native trail looks like.

Separate identifiers

Keep names and contact details out of the analysis workbook, and control the key that links them. See role-based access control and ALCOA+ data integrity.

Switch plan

Moving a study workbook to Capture

The sequence below keeps your existing workbook as the archived record and avoids any claim about automated data import.

  1. 1

    Lock and archive the workbook

    Make a read-only master, record the date, the file hash if your procedure uses one, and the person responsible. File it with the study documents.

  2. 2

    Turn each sheet into a form

    A sheet of repeated measurements becomes a form with a table section. A sheet of one-off baseline values becomes a baseline form. Upload the XLSX or headers and let AI draft the questions; review each one.

  3. 3

    Replace formulas with calculated fields

    Use the built-in calculators where they exist and rebuild any other logic as edit checks, reviewed before approval.

  4. 4

    Define visits and roles

    Add the schedule your protocol requires and give site staff the access they need. The templates library has standard forms to copy.

  5. 5

    Test in the sandbox and keep the evidence

    Run mock participants end to end: break a rule, edit a value with a reason, read the audit trail. File the evidence with your validation documentation.

  6. 6

    Go live and document the boundary

    State in your data management plan which records lived in Excel and which live in Capture, and how they are combined at analysis.

FAQ

Questions teams ask before they switch

Something not covered here? Ask us directly.

Is Excel compliant for clinical trial data?

Excel is not certified or prohibited by Part 11. Compliance depends on validation, access control, audit trail and procedures around your use. File-based design makes this hard to evidence for a participant database.

Does Excel have an audit trail?

Per secondary sources, change tracking for shared workbooks needs OneDrive or SharePoint and covers a limited window, and Version History keeps file snapshots. Neither is a per-field audit trail with a reason for change. Confirm with Microsoft documentation.

Do I need to validate an Excel spreadsheet used for a trial?

If the spreadsheet supports regulated records, expect to specify, test and document its formulas, macros and protection, and re-test after changes. Your quality unit sets the level, based on risk.

Can I use Excel for a small academic study?

Possibly, if the data are not subject to Part 11 or GCP expectations, or your sponsor and ethics committee accept it. Agree this in writing before you start.

What replaces Excel in Capture?

Form-based eCRFs with typed fields, edit checks, calculated fields, roles and a field-level audit trail. Exports to CSV or Excel are still available for analysis.

Can Capture import my existing Excel data?

We do not claim automated data import. AI drafting can build forms from an uploaded XLSX; archive the workbook and document how earlier data are combined at analysis.

Is Capture suitable for Phase 1, 2 and 3?

Yes. Capture is suitable for Phase 1, 2 and 3 studies.

How do I try it?

Open the free sandbox, with every feature and no credit card. You pay only when you go live.

Replace formulas with controlled forms

Free sandbox, every feature. Build the form, break a rule and read the audit trail. No credit card.

Try a form-based build free