The relationship between dual practice and physicians’ work behaviour in the public hospitals: Results from the Danish survey

In numerous healthcare systems, some physicians combine public hospital employment with a private practice - a phenomenon referred to as physician dual practice. Dual practice raises numerous concerns with regard to its possible negative effects on the work performance of physicians in public hospitals. The purpose of this study is to test the theoretical predictions with regards to the dual-practice effects on work inputs of public hospital physicians. We use a crosssectional data on physicians work inputs from a 2008 nation representative survey of public hospital physicians in Denmark. To tackle the limitations imposed by the cross-sectional character of the data and the self-(selection) problem we employ a multilevel mixed model. Multilevel modelling is a Bayesian alternative to the frequentist models of causal inference from observational data such as simultaneous equation or matching methods, including propensity score analysis. The multilevel mixed model gives the relationship between dual practice and the physicians’ behaviour in the public hospitals, controlling for confounders such as a hospital type, medical specialty, position, income, age, gender, and years of working experience, as well as potential confounders related to all these background characteristics. The results do not reveal any patterned relationship between dual practice and public hospital work hours, participation in voluntary tasks or activities that might conflict with the private-practice hours, or preferences for part-time employment. The results also do not support the general presumption that the physicians who work exclusively in public hospitals are more altruistic and hence, voluntarily provide more work inputs than dual-practitioners. The causal inference with the current data is challenging, especially in the view of the labour economics and personnel economics studies on dual job holding, which suggest that dual and single jobholders represent different types of personality, with the dual jobholders being the more proactive employees. The current analysis does not account for the possible differences in the personality, hence it cannot be excluded that the dual-practitioners change someway their work behaviour in public hospitals after the commencement of dual practice. Still, it can be concluded that physicians engaged in dual practice perform at least as good in the public hospitals as their counterparts who do not hold a second job.