Despite increasing interdependencies, national decision-makers have been reluctant to delegate healthcare competences to the supranational level in the European Union (EU). To overcome this impasse, EU institutions and member states have agreed on middle ground compromises by means of experimentalist governance. In this paper, we examine a tool of experimentalist governance in the making, i.e., the network formed by the cross-border healthcare expert group (CBHC) in the Patient Rights Directive. We ask whether interaction by means of transitive relations carrying trust, takes place and the extent to which domestic institutions, i.e., healthcare models, condition such interaction and thus learning. To examine network interactions, we use social network analysis on the basis of collected survey data on the exchange of information, advice and best practices within the CBHC network. We develop an Exponential Random Graph Model of the network to test the extent to which domestic institutions condition such interactions. For this, we conduct a cluster analysis and build a healthcare typology of EU27 plus the UK, Norway and Iceland, identifying five distinct healthcare types. We find that this type of networked governance brings EU healthcare cooperation forward, while domestic institutions greatly condition who interacts with and learns from whom.