Few papers examine the pecuniary and non-pecuniary determinants of doctors' labour supply despite substantial predicted shortages in many OECD countries. We contribute to the literature by applying both a structural discrete choice and a reduced-form approach. Using detailed survey data for Australian physicians, we examine how these different modelling approaches affect estimated wage elasticities at the intensive margin. We show that all modelling approaches predict small negative wage elasticities for male and female General Practitioners (GPs) and specialists. Our detailed subgroup analysis does not reveal particularly strong responses to wage increases by any specific group. We show that the translog and Box-Cox utility functions outperform the quadratic utility function. Exploiting the advantages of the structural discrete choice model, we examine short-term effects at the intensive margin by calculating labour supply changes in response to 5 and 10% wage increases. The results show that such wage increases substantially reduce the full-time equivalent supply of male GPs, and to a lesser extent of male specialists and female GPs, but not of female specialists.