Report a systematic-review protocol amendment with its reason and timing

1. Start with the actual original commitment

Locate the original protocol and registration, their versions, dates and stated outcomes or methods. PRISMA 2020’s expanded checklist asks authors to describe amendments to registered or protocol information by identifying the change, its reason and the review stage at which it was implemented. It also asks for registration and protocol availability, including when those were not undertaken.

Do not create an apparent original protocol after seeing results and then describe it as prospective. If the team prepared no protocol, state that fact and explain the actual planning record. If several documents disagree, identify the authoritative version and retain the discrepancy. A later clarification may be useful without erasing what a reader could have known at the earlier date.

Reference: PRISMA 2020: expanded reporting checklist

2. Examine a fictional change in follow-up

In a wholly fictional review, a protocol dated 1 July 2026 specifies outcomes measured at twelve months. On 20 July, during extraction planning, the team learns that some reports describe six-month follow-up. The supplied premise is that it has seen which follow-up periods are available but has not viewed numerical outcome values or effect estimates.

On 22 July, the team adopts a documented amendment to add a separately reported six-month outcome category while retaining the twelve-month category. Analysis is still pending. This is an invented teaching record, not a completed review, registry update or approval. The amendment’s timing matters because the availability of follow-up information already influenced the decision.

On small screens, scroll the table horizontally, or focus it and use the arrow keys.

3. Explain the reason without making a stronger claim

A useful rationale identifies the problem the change addresses and its effect on interpretation. In this example the reason is the observed availability of reports at a different follow-up period. It is not “to confirm the intervention worked” or “to improve significance.” No direction or size of an effect is supplied, so neither can justify the decision in the teaching record.

Adding a six-month category does not make it equivalent to twelve-month follow-up. The team must define extraction and analysis for each and assess whether any synthesis is appropriate. Keep those methodological decisions open if they have not been made. A transparent amendment explains a choice; it does not automatically make the choice scientifically sound or remove the need for independent methods review.

4. Use the supported update process and keep the evidence

Check the actual registry’s amendment mechanism before changing its record. OSF documents a formal registration-update workflow with a justification and administrator review, alongside a preserved registration history. That process is different from silently replacing an attached original protocol. Its guidance also limits what can be changed in submitted registration files; check the current rules for the record concerned.

Keep the original plan, dated amendment, responsible contributors, required approvals and update receipt together. A draft update is not a completed registry event. For the fictional exercise, the team’s internal amendment is adopted, but no registry update or external approval is supplied. The log should retain those distinct states until actual evidence changes them.

Reference: OSF: registrations and preregistrations

5. Write the methods note from verified actions

Explain the change in the manuscript where the review method can be understood, with a link to the accessible protocol and amendment when available. Report which analyses were performed under the original plan and which followed later decisions. Identify any implications for extraction, grouping and interpretation. A link alone does not tell the reader what changed.

The example below states the adopted definition and keeps analysis pending. It does not claim that a registry accepted the update, that a committee approved it or that six-month results support a conclusion. Replace its training dates and actions with actual evidence before using a similar structure in a real manuscript.

6. Reconcile the complete report before submission

Compare the final protocol history with the extraction definitions, analysis files, tables, flow information and abstract. Check that a change described as separate has not been silently combined elsewhere. Confirm the actual timing from retained records, including whether outcome values were already known. If that knowledge state is uncertain, say so rather than reconstructing a more favourable chronology.

Ask an independent colleague to follow one amendment from the original clause through the decision and into the report. They should be able to identify the change, rationale, stage, execution status and remaining limitations. Retain withdrawals and superseded decisions as history. The goal is a report whose methods can be interpreted, not an unchanging protocol at the expense of an honest account.

Reference: PRISMA 2020: expanded reporting checklist

Questions for a practical final check

• Original protocol and registration status identified.
• Change, reason and review stage recorded.
• Prior outcome knowledge described honestly.
• Different follow-up periods not treated as equivalent.
• Internal adoption distinguished from registry completion.
• Final methods and outputs reconciled with the history.