As with IPD meta-analyses (see Chapter 18, Section 18.2.2) negotiating and establishing a strong collaboration with the participating trialists is essential for the success of a PMA. The focus of a PMA, however, is not primarily about locating and obtaining data from individual trials. As the collaboration needs to be formed prior to the results of any trial being known, the focus of a PMA’s collaborative efforts, at least initially, is on reaching agreement regarding study population, design and data collection methods for each of the participating studies. When members of a PMA collaborative group agree to participate in the project, they need to agree to a core common protocol and core common data items that will be collected across all trials. Individual trials can include local protocol amendments or additional data items but they need to ensure that these will not compromise the core common protocol elements.
In a PMA, efforts are made to identify all ongoing trials, both to maximize precision and to avoid bias that might be introduced by excluding studies based (at least in part) on knowledge of the results of those studies. To certify that an individual study is eligible for inclusion in the PMA there should be evidence to support the claim that, at the time of the agreement to be part of the PMA, trial results were not known outside the trial’s own data monitoring committee. This should ideally be in the form of evidence that the trial was prospectively registered (Laine 2007). It is also advisable for the collaborative group to obtain an explicit (and signed) agreement from each of the trial groups to collaborate. The idea is to encourage substantive contributions by the individual investigators and to get ‘buy-in’ to the concept of the PMA and the details of the protocol.