A quitting network meta-analysis: was that comparison direct?

Separate a network from one multi-group trial

A multi-group trial has several groups within one study. A network connects comparisons across studies. Cochrane explains how direct and indirect evidence can contribute to the same pairwise estimate, subject to important comparability assumptions. Recognizing that route helps you understand what the published number represents.

Locate the network diagram and the actual comparison table. Identify the intervention descriptions rather than reducing them to brand names or delivery labels. Two support arrangements called digital may differ substantially. A graph node is not evidence that any app sharing one feature was included in the analysis.

Cochrane: Cochrane Handbook, Chapter 11: Undertaking network meta-analyses

Draw an invented three-node route

Imagine fictional studies comparing arrangement A with B, and separate studies comparing A with C. There is no direct B-versus-C study in this example. Draw two solid links for the comparisons that exist, then a differently labeled indirect route between B and C through A. Do not calculate an effect from made-up labels.

Now inspect what A meant in both sets of studies. If one version provides frequent conversations and the other only one note, the common name does not guarantee a comparable bridge. The invented map makes a reporting question visible; it provides no evidence that one arrangement helps quitting more than another.

Read comparability and disagreement questions

The linked comparisons need sufficiently similar factors that could affect relative effects. Check what the review says about that assumption and about disagreement between direct and indirect evidence where both exist. A connected drawing alone cannot establish validity; this requires informed methodological assessment.

If the paper gives a ranking, find the outcome and the uncertainty behind it. Rankings can differ across outcomes and depend on which interventions entered the network. Do not convert first place into a guaranteed best choice for every person, or assume that small differences in rank describe a large practical difference in effect.

Cochrane: Cochrane Handbook, Chapter 11: Undertaking network meta-analyses

Preserve the route in a short summary

A useful statement identifies whether the quoted comparison was direct, indirect, or combined, then retains the outcome and limitations. If you cannot determine the route, say the source of the comparison remains unverified. This is more informative than calling the entire network a single head-to-head trial.

This guide does not evaluate a named network or choose treatment. A qualified professional can help assess the whole evidence for individual care. Your own use of multiple supports is not a network analysis, and comparing personal tracker screens cannot recreate randomized study links. The practical output is a clearly labeled evidence route and a focused question about its assumptions.