Can These Studies Be Pooled? A Feasibility Walkthrough

A shared topic is a starting point

Imagine three fictional papers about recovery after an intervention. One measures symptom score at two weeks, another records readmission by 30 days, and the third reports a functional scale at six months. The papers concern the same broad topic, but their outcomes do not represent one interchangeable endpoint.

This worked example illustrates a planning decision; it is not an assessment of a named clinical evidence base.

Build the comparison before calculating anything

Make one row per eligible study and record population, intervention or exposure, comparator, design, outcome definition, measurement scale, time point, and reported effect measure. Add a field for possible participant overlap.

Check the exact question each estimate answers. Adjusted and unadjusted associations may not target the same quantity. A review containing two reports from the same cohort cannot treat their participants as independent simply because the reports have different titles.

Choose a defensible synthesis plan

Some differences may be handled through a justified transformation or separate analyses. Others may mean that a structured narrative synthesis is more appropriate. A standardized mean difference does not make distinct clinical constructs equivalent.

An I² threshold alone cannot decide whether studies belong together. Clinical and methodological compatibility must be assessed, and with few studies statistical heterogeneity is difficult to characterize precisely.

Record a stopping rule

Write down what information would make you proceed, narrow the question, seek author clarification, or stop quantitative pooling. This protects the project from drifting toward a convenient analysis merely because the software can run it.

Use the pooling planner and download the study comparison worksheet. If you already have a proposed question, request a feasibility check.

Methodological source: Cochrane Handbook Chapter 10. Apply it alongside a documented search and study-level assessment.