Could a vaping review be missing whole studies? Read its publication-bias assessment

Identify what is missing

Missing participants within a known trial, a study excluded for a stated eligibility reason, and an eligible study whose results are unavailable are different problems. Start by naming the one the review discusses. A date-limited search is another issue; its cutoff does not by itself demonstrate selective publication.

Cochrane describes non-reporting bias as arising when decisions about reporting are influenced by the results. Entire reports or particular results can be unavailable selectively. The concept concerns how the accessible evidence may differ from what is missing. It does not allow you to determine the motives of named authors from a gap alone.

Cochrane: Cochrane Handbook, Chapter 13: Risk of bias due to missing evidence

Look for an account of the search beyond published papers

Find the review’s search methods and limitations. Did it consult trial registries or other relevant sources, and did it identify projects with unavailable results? Read the authors’ explanation of what they could obtain. A large reference list is not enough to answer whether the search covered missing evidence.

In an original hypothetical example, a review lists a completed eligible project whose results could not be obtained. You can record “results unavailable” and the authors’ stated implications. Do not enter zero benefit for that project or invent participants’ outcomes. A completed registry entry is not itself the missing results report.

Do not treat a funnel plot as a detector of dishonesty

A funnel plot can help assess differences between smaller and larger studies. Cochrane cautions that asymmetry has several possible explanations and that related tests have limitations. A visually uneven plot does not diagnose publication bias by itself.

If a review reports that a test was not appropriate or could not be performed, preserve that limit. “Not assessed” is not “proved absent.” Equally, lack of a test does not prove that missing results exist. Do not count dots or calculate an improvised correction from a screenshot; interpreting these analyses needs the methods and context.

Keep the uncertainty beside the conclusion you share

A useful reading note has four lines: evidence the review obtained, relevant evidence it could not obtain, the authors’ assessment, and what remains uncertain. Link the methods or assessment section. If all you have is a headline, look for that account before describing the synthesis as complete.

This check does not determine whether a particular quitting intervention works or which care is suitable for you. Nor does it require assuming all unavailable reports would change the conclusion in one direction. Its purpose is to prevent “these are the published results” from silently becoming “these are all the results.” Bring consequential individual questions to an appropriate qualified professional.