Rural practice preferences among medical students in Ghana: a discrete choice experiment.

OBJECTIVE To determine how specific job attributes influenced fourth year medical students' stated preference for hypothetical rural job postings in Ghana. METHODS Based on discussions with medical student focus groups and physicians in practice and in the Ministry of Health, we created a discrete choice experiment (DCE) that assessed how students' stated preference for certain rural postings was influenced by various job attributes: a higher salary, free superior housing, an educational allowance for children, improved equipment, supportive management, shorter contracts before study leave and a car. We conducted the DCE among all fourth year medical students in Ghana using a brief structured questionnaire and used mixed logit models to estimate the utility of each job attribute. FINDINGS Complete data for DCE analysis were available for 302 of 310 (97%) students. All attribute parameter estimates differed significantly from zero and had the expected signs. In the main effects mixed logit model, improved equipment and supportive management were most strongly associated with job preference (beta = 1.42; 95% confidence interval, CI: 1.17 to 1.66, and beta = 1.17; 95% CI: 0.96 to 1.39, respectively), although shorter contracts and salary bonuses were also associated. Discontinuing the provision of basic housing had a large negative influence (beta = -1.59; 95% CI: -1.88 to -1.31). In models including gender interaction terms, women's preferences were more influenced by supportive management and men's preferences by superior housing. CONCLUSION Better working conditions were strongly associated with the stated choice of hypothetical rural postings among fourth year Ghanaian medical students. Studies are needed to find out whether job attributes determine the actual uptake of rural jobs by graduating physicians.