Importance of the methods used to support the node-making process in network meta-analysis

      We thank Shi et al. for their interest [
      • Shi C.
      • Westby M.
      • Norman G.
      • Dumville J.
      • Cullum N.
      Node-making processes in network meta-analysis of non-pharmacological interventions should be well planned and reported.
      ] in our article “Node-making process in network meta-analysis of non-pharmacological interventions should be well planned and reported” [
      • James A.
      • Yavchitz A.
      • Ravaud P.
      • Boutron I.
      Node-making process in network meta-analysis of nonpharmacological treatment are poorly reported.
      ].
      We totally agree that methodological choices might alter the final network and consequently the results of the study. The four proposals made by Shi et al. concerning the possible approach to lumping and splitting all seem to be feasible options although we are not sure to understand the difference between the first and the second one. Furthermore, we would like to ask Shi et al. if they can elaborate on their methodological vision of these options and especially if they see them as complementary or exclusive.
      The four amendments made to our proposals in the memory box are also interesting. They nevertheless raise some issues that we would like to discuss. First, concerning “a pre-specification of lumping and/or splitting approach to node-making,” we think that this point is already encompassed by the proposal “pre-specify inclusion criteria for each node.” Indeed, inclusion criteria will modulate interventions' grouping granularity. Second, we agree that the node-making process should respond to the objectives of the network meta-analysis. However, we are not sure that basing the node-making process on sensitivity analysis is a good recommendation to make. Obviously, sensitivity analyses raise issues if they are not planned in the protocol. Moreover, we think that they should remain exploratory analysis as recommended by the PRISMA extension statement [
      • Hutton B.
      • Salanti G.
      • Caldwell D.M.
      • Chaimani A.
      • Schmid C.H.
      • Cameron C.
      • et al.
      The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations.
      ] and not condition the node-making process. Otherwise, it might lead to a significantly biased network that would fit the authors' vision of the results. Third, the “description of the sources of information” did not seem specific to network meta-analyses reporting. Furthermore, a proposal in the published memory box is already asking for “a description of each intervention” which takes into account this problematic. Fourth, we don't think that “a description of the methods used to avoid arbitrary node-making” requires a specific point because the whole memory box finally aimed at this objective.
      Finally, we think that these four amendments might request further discussions before being included in the memory box.
      Sincerely,
      AmélieYavchitz

      References

        • Shi C.
        • Westby M.
        • Norman G.
        • Dumville J.
        • Cullum N.
        Node-making processes in network meta-analysis of non-pharmacological interventions should be well planned and reported.
        J Clin Epidemiol. 2018; (http://dx.doi.org/10.1016/j.jclinepi.2018.04.009 [Epub ahead of print])
        • James A.
        • Yavchitz A.
        • Ravaud P.
        • Boutron I.
        Node-making process in network meta-analysis of nonpharmacological treatment are poorly reported.
        J Clin Epidemiol. 2017; 97: 95-102
        • Hutton B.
        • Salanti G.
        • Caldwell D.M.
        • Chaimani A.
        • Schmid C.H.
        • Cameron C.
        • et al.
        The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: checklist and explanations.
        Ann Intern Med. 2015; 162: 777-784

      Linked Article