More follow-up measurements do not mean more study participants

Identify the person behind the observations

Read whether the report follows the same adults or recruits a fresh sample at each time. A table with three follow-up columns might describe one cohort repeatedly, not three separate cohorts. Also distinguish the number enrolled from the number contributing information at each measurement.

Cochrane’s unit-of-analysis guidance discusses repeated observations and the need to account for their structure. This guide helps identify that structure without performing a statistical analysis. It does not prescribe a model or establish that a named study handled repeated data correctly.

Cochrane: Cochrane Handbook, Chapter 6: Choosing effect measures

Count a fictional example at both levels

Suppose, for a hypothetical reading exercise, 80 adults are enrolled and each is invited to three follow-ups. There could be up to 240 follow-up observations, but that does not make 240 independently enrolled adults. Actual completion may be lower, and the report should describe those missing observations rather than hiding them inside a large total.

If the outcome is repeated use of a support action, one adult may also contribute several events. Events, visits, observations, and participants are different units. A headline emphasizing a large number of “records analyzed” cannot be read as a participant count without the original definition. None of the example’s numbers establishes efficacy.

Check how the analysis uses the repeated information

A report may analyze each follow-up separately, summarize a participant’s trajectory, or use a method accounting for relationships within a participant. Locate its actual approach and the effect question. Do not automatically add all columns together or pool estimates from the same people as though they were independent samples.

Also check whether a review includes more than one time point from the same study in the same pooled result. Cochrane explains that standard pooling can have a unit-of-analysis problem when repeated results are treated as independent. This introductory check identifies a question for the review; it does not prove an error from the mere presence of several follow-up graphs.

Cochrane: Cochrane Handbook, Chapter 6: Choosing effect measures

Preserve the study identity when describing confirmation

A bounded statement is: “This cohort was measured at these times, with this handling of repeated observations and these completion limits.” Independent replication requires independent evidence, not simply another measurement of the same adults. Repeated data can still be valuable when described and analyzed appropriately.

For an individual health decision, seek suitable professional interpretation of the full report. Your own daily observations likewise remain observations of one person, even when the file contains many rows. They can support practical reflection without becoming a large trial or demonstrating causation. The task ends when both the participant unit and the observation unit are visible in your summary.