[Preferences for work in primary care among medical students in Minas Gerais State, Brazil: evidence from a discrete choice experiment].

This article presents the results of a discrete choice experiment (DCE) conducted in 2012 with 277 final-year medical students from Minas Gerais State, Brazil. The experiment tested students' preferences concerning future work as physicians in primary health care, based on hypothetical job scenarios aimed at measuring the likelihood of placement in areas with a shortage of doctors. Application of DCE involved (i) a qualitative stage to define the attributes and their respective levels to comprise the job scenarios, (ii) construction and application of the instrument, and (iii) analysis with application of multinomial logit with conditional probability to estimate the weight of attributes and to construct scenarios for choice probability. The results indicate that the job attribute that most impacted students' choice was location, followed by job conditions, pay, access to medical residency, type of employment relationship, and workload. Students from private medical schools, with higher family income, and females were generally more likely to resist job assignments in unsafe urban areas and remote areas of the countryside. The job scenarios that proved most plausible in terms of public intervention were those that combined middle-level wages, good working conditions, and 10 to 20 bonus points on medical residency exams.