Manual Review

Manual review lets you work through the checklist yourself, item by item, and have the module compute a recommendation from your answers.


The review flow

After the study type is determined and you select manual review, the module:

  1. Fetches the checklist for that study type from the data service
  2. Presents the items one set at a time, each as a yes/no judgement
  3. Lets you answer every item
  4. Computes the result when you click Calculate Results

For each checklist item, you record whether the manuscript adequately addresses it -- yes (addressed) or no (not adequately addressed).

The Calculate Results button is disabled until every item has been answered. This ensures the recommendation is based on a fully-completed checklist.


How answers map to issues

When you calculate results, the module sorts every unmet item (answered "no") into a severity bucket based on its category:

  • If category is minor -- added to minor issues
  • If category is major -- added to major issues

Items answered "yes" are considered satisfied and do not generate issues.


The escalation-to-reject rule

After sorting, the module applies an escalation rule:

If the number of major issues is 3 or more, OR any major issue concerns "bias" or "design", then all major issues are escalated to REJECT.

A manuscript with three or more major gaps, or with a major gap relating to bias or study design, is escalated from "major revisions" to "reject" -- reflecting that fundamental design or bias problems are not fixable by revision alone.


The recommendation

The final recommendation follows from which buckets contain issues, in priority order:

Condition Recommendation
Any reject issues Reject -- not acceptable in current form
Any major issues (no reject) Major Revisions
Any minor issues only Minor Revisions
No issues at all Accept

The logic is strictly hierarchical: a single reject-level finding outweighs any number of major or minor ones.


The results display

The results panel shows:

  • The overall decision (Accept / Minor / Major / Reject)
  • The major issues (or reject issues) -- specific checklist items not adequately addressed
  • The minor issues -- shown when there are no reject-level findings

Each listed issue is the actual checklist question that failed.


Follow-up chat

When the decision is anything other than Accept, the module automatically opens the Review Chat so you can discuss findings -- ask why an item matters, how to address a gap, or what the guideline expects. For an accept decision, no follow-up chat is opened.


Manual vs AI review

Manual review AI review
Who judges each item You The AI
Checklist used Study type's checklist Same checklist
Issue severity From item's major/minor category From AI evaluation + item category
Recommendation scale Accept / Minor / Major / Reject Accept / Minor / Major / Reject
Best for Careful, reviewer-controlled assessment Fast automated first pass

Both modes assess the same items and use the same four-level recommendation scale.