Choose clinical article keywords from the study, then check the vocabulary

1. Describe the actual study in one sentence

Begin with population, setting, concept and design. The fictional manuscript in this exercise reviews documentation of inhaler-technique assessment in adult asthma outpatient records from two clinics. It describes whether an assessment was recorded. It does not measure symptom control, compare treatments or establish whether undocumented assessments were never performed. No sample size or result has been supplied, so keyword selection cannot add either.

Ask a clinical co-author to read the sentence without the proposed keywords. If they infer a treatment trial, rewrite the description before choosing terms. This is a meaning check, not a claim that one colleague represents all readers. Keep the manuscript’s limits close to the selection sheet so an attractive but unsupported phrase does not enter the list during a later copyedit.

2. Separate author keywords from assigned indexing

NLM describes MeSH as a controlled, hierarchical vocabulary. PubMed distinguishes author keywords from MeSH indexing; author keywords can be searched through specified fields such as title/abstract or other term. Submitting a MeSH-like keyword to a journal does not establish that NLM has assigned that heading to the published article. Record which kind of term you are discussing.

Read the journal’s keyword instructions, including any required vocabulary and number of terms. If it requests free-text terms, an accurate phrase may be useful even when it is not a preferred MeSH heading. If it specifically requires controlled headings, verify each selected heading and its scope. Do not label an entire list “MeSH terms” merely because the first word appears in the browser.

Reference: NLM: Medical Subject Headings · NLM: PubMed search help

3. Build a concept-to-term worksheet

List why each candidate belongs and what would make it misleading. In this example, “asthma” identifies the condition, while “inhaler technique” captures the assessment concept in ordinary clinical language. “Clinical documentation” reflects the source being examined. These are proposed author keywords, not a completed search strategy or an assertion about the indexing of a real article.

Remove terms that introduce a stronger claim than the paper supports. “Treatment effectiveness” suggests an outcome comparison absent from the study. “GCC-wide practice” extends two unspecified clinics to a region that the supplied facts do not establish. Avoid adding fashionable technology terms when the manuscript does not evaluate that technology. The best justification is a clear connection to the actual question, not a guessed number of searches.

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

4. Inspect an actual vocabulary record carefully

The current NLM browser has an Asthma heading, and Administration, Inhalation describes administration of drugs through the respiratory route. Those are real vocabulary records linked below. The second is broader than this fictional paper’s focus on documentation of technique assessment. Its existence does not automatically make it the best keyword, or establish that “inhaler technique” is its exact synonym.

A useful review note contains the vocabulary version, heading, scope examined and decision. Preserve the distinction between a broad related concept and a precise description of the paper. Ask a librarian or appropriately experienced co-author about unresolved choices. The Arabic equivalent should convey the same concept, while an English controlled heading can be retained explicitly where the journal requests its exact form. Do not invent an official Arabic MeSH translation.

Reference: NLM MeSH 2026: Asthma · NLM MeSH 2026: Administration, Inhalation

5. Treat a search query as a separate experiment

A short keyword list is not a reproducible literature search. If the team wants to inspect how terms retrieve relevant material, specify the database, interface, date, exact query and any filters. Review the returned records against a stated purpose. A search used to explore terminology should not be reported as a systematic review search unless it was designed and documented for that purpose.

PubMed documents automatic term mapping and Search Details, as well as field-tag behavior. Inspect the actual translation of a query rather than assuming the typed words were searched literally. For this guide, no PubMed result query was run: there is no retrieval count, relevance estimate or coverage comparison to report. A future test record should retain those observations only after execution, together with its limitations.

Reference: NLM: PubMed search help

6. Check the final list in both languages

Read the title, abstract and final keyword list together. Each term should still be supported after manuscript revisions. If the study’s setting or outcome changes during analysis, revisit the list instead of preserving a polished but outdated phrase. Confirm the number, separators and vocabulary format required by the target journal, and distinguish an author-approved list from one merely drafted by a writing assistant.

In Arabic, distinguish “توثيق تقييم تقنية الاستنشاق” from “تحسين نتائج علاج الربو”: the first concerns the recorded assessment, while the second claims improvement in treatment outcomes. The fictional manuscript supports only the former. After publication, read back the displayed and exported keywords. That confirms what the publisher supplied; any database-assigned headings or discovery effects must be checked separately on the actual record.

Questions for a practical final check

• Study question and limits written before keyword selection
• Author keywords separated from MeSH indexing
• Candidate terms justified by supplied concepts
• Unsupported treatment and regional claims removed
• Vocabulary records and unresolved choices documented
• No unperformed search count or ranking outcome reported