Eliciting primary healthcare physicians' preferences for job characteristics in rural China: a discrete choice experiment

The outflow and scarcity of physicians in rural areas can adversely affect universal health coverage and population health outcomes, which are critical concerns in China. This study explored primary healthcare physicians' job preferences using a discrete choice experiment to identify appropriate incentives for retention. Eight job characteristics were identified through a literature review and qualitative studies as the attributes relevant to designing the discrete choice experiment, with levels varying between two hypothetical jobs. The data were analysed using conditional logit model, mixed logit model and latent class model. A total of 1781 licensed physicians (including licensed assistant physicians) from township health centres in rural areas were surveyed. Policy simulation suggested that they were sensitive to both monetary and non-monetary policy incentives. As for non-monetary job characteristics, a highly intense doctor-patient relationship, bianzhi (the number of personnel allocated to each employer by the government) and educational opportunities were highly valued by the respondents. The latent class model could identify distinct groups with different job preferences according to their memberships. Urban jobs were much preferred to rural ones. However, policy incentives can lend themselves to effective retention strategies. It is also important to tailor policy incentives to different subgroups.