Individual-based network models are increasingly being applied to understand the transmission dynamics of infectious diseases. Research in this area has mostly focused on networks defined under a limited set of rules (e.g., preferential attachment, sexual partner formation and dissolution) that are supposed to mimic the real world but are often defined heuristically due to lack of empirical knowledge. Here, two different mechanisms (M- and lambda(2)-rules) were used to generate a wide range of networks and to show the extent to which microstructures such as the mean component size, the size of the giant component and the cumulative nomination centrality index may vary between networks with fixed predetermined macrostructure characteristics (size, node degree distribution and mixing pattern) and influence disease transmission. It is important to carefully consider the limitations of network models and to appreciate the extent to which a given degree distribution and mixing pattern will be consistent with a wide range of underlying network microstructures.