Modelling entire disc galaxies is a difficult task for the most advanced computer system, resulting in the small scale physics of the interstellar medium being simplified, often to a single phase (or fixed two or three phase) gas. Simulations of local patches, however, reveal a complex ISM; a turbulent mix where many forces compete during star formation. This suggests global models are likely to be a poor approximation to the real thing and that we may be missing important clues in their evolution and star formation histories. That said, observations show simple relations between global star formation and gas density, indicating the opposite is true and the detail of the ISM is not important. So can we ignore the fine structure on large scales and if so why? In this work, I present some of the first global models that develop a multiphase ISM, comparing isothermal systems with galaxies that contain feedback and background heating to try and uncover the most important parameters in star formation in disc galaxies.