Research question 4 - How consistent are the findings across the trials?
A meta-analysis of quantitative and a meta-ethnography of qualitative data from all five studies will be conducted.
A qualitative meta-ethnography will be conducted using the data from the interviews. For this purpose a description of subcategories with significant quotations from all studies will be provided in English. This material will be analyzed according to the principles of Noblit and Hare’s reciprocal translational analysis, which is recommended to be used when analyzing studies with similar topics [17, 18]. The result of this analysis will be a set of key themes relevant in all studies. Interviews will be re-analyzed to explore how each single study contributes to the overall key themes. In a last step third order interpretation across the key themes will be considered if supported by a ‘line of argument’ across all studies.
A quantitative meta-analysis will be conducted, focused on the summary outcome measure in each trial, using the predictors that are identified as result from the analysis for question 3. While the items composing this aggregate outcome obviously differ among the trials, the assumption is that the resulting measures are reasonably comparable to allow pooling. Based on hypotheses derived from phase 1 and 2, individual data and/or clustered data from all trials will be included in a meta-regression analysis. This last step will efficiently make use of the total number of cases included across all studies and the shared instruments as described in this protocol.