Does spellcheck catch wrong words in a CSR, protocol or investigator's brochure?
No. Spellcheck passes a correctly spelled wrong word in a clinical study report (CSR), protocol or investigator's brochure (IB) for the same reason it passes one on a drug label: the word is real.
ContextRx documents this failure on live FDA DailyMed labels, not in clinical documents. This page applies the same lesson to the long documents medical writers produce. The examples below are ones we made up to illustrate the point; none was found in a real document.
Same failure, longer documents
Our view: clinical documents are long, reuse text from earlier documents and templates, and pass through many hands. Each copy-forward is a chance for a wrong word to travel. Reviewers reading for consistency with source data are mostly looking at numbers, so a real word in the wrong place can slip by.
Where it bites in a protocol
Eligibility criteria and dosing instructions are where one word decides who is enrolled or what they receive. Made-up example: an exclusion line that should say "not" but says "now" can flip who is excluded. Another made-up example: a dose-hold rule tied to "hypotension" means something else if it says "hypertension".
Where it bites in a CSR
Results and safety narratives turn on direction and condition words. Made-up example: "hypoglycemia" and "hyperglycemia" name opposite events. A swap in a narrative or a summary line misreports what happened, and the table may not be checked against that sentence.
Where it bites in an IB
Our view: IB risk wording often feeds later labeling, so an error that survives in an IB can be carried into a label draft.
AI-drafted first drafts
AI tools can turn tables and outlines into fluent prose quickly. Our view: fluent text is harder to doubt, so a read for meaning matters more, not less. See does AI drafting make wrong-word errors moot and AI-drafted labeling and human review.
How writers catch it
Practice advice: add a short key-word pass on eligibility, dosing and safety text. Read each line for the one word it depends on, and ask whether it makes sense with that word and not its near twin. Pair it with your usual QC against source data, which checks numbers, not meaning. More in self-editing tips for regulatory writers and is a tool enough for labeling QC.
Sources
This page doesn't rely on outside sources: it gives practice advice, and its examples are made up.
Last reviewed: October 14, 2026
More on contextual errors in FDA DailyMed labels.
Questions about wording on your own labels? Talk to us.
We scan freshly published FDA DailyMed labels every day and pull out these high-stakes wrong words.
Every Tuesday and Friday, a new case like this lands in your inbox if you subscribe.
FAQ
Does spellcheck catch wrong words in a clinical study report?
No. It only checks spelling, so a correctly spelled wrong word passes.
Where do wrong words matter most in a protocol?
In our view, eligibility criteria and dosing instructions, where one word can change who is enrolled or what they receive.
Does QC against source data catch wrong words?
It mainly checks that numbers match the source. A wrong word in the surrounding text can still pass.
Does AI drafting of CSRs make this worse?
In our view, it adds fluent text quickly, so a read for meaning matters more, not less.
Has ContextRx found errors in CSRs or protocols?
No. ContextRx studies FDA DailyMed labels; this page applies the same lesson to clinical documents.