Mutual influences of general practitioners in partnerships

The aim of this study was to find out whether or not general practitioners (GPs) within the same partnership show more similarities in attitudes and behaviour than GPs in different partnerships, and what the causes of these similarities might be. Knowledge of the causes of patterns of similarities within medical teams contributes to understanding medical practice variation, which is crucial in developing effective health care policies. Data were used from the Dutch National Survey of General Practice ('87/'88), consisting of a stratified sample of 161 Dutch GPs, who served 335,000 patients in total. To find out whether GPs in the same partnership are indeed more similar than GPs randomly chosen from different partnerships, we constructed two kinds of pairs: all possible pairs of GPs working in the same partnership (actual pairs), and randomly constructed pairs of GPs who are not working in the same partnership (random pairs). For each pair difference scores were computed for a variety of attitudes and behaviour. Difference scores for actual and random pairs were analysed using multi-level analysis. Most differences in attitudes and behaviour were smaller for actual pairs than for random pairs. Furthermore, in the majority of the cases differences were no longer statistically significant after explanatory variables indicating selection, gradual adaptation and rapid adaptation through shared circumstances were taken into account. It was found that Dutch GPs working in the same partnership showed more resemblance in attitudes and behaviour than GPs not working in the same partnership. Most indications point towards circumstances, and to a lesser extent towards adaptation, as an explanation of similarities within partnerships. The implication of this study is that medical practice variations are not merely individual differences in preferred practice style, but are patterned by social processes in partnerships and local circumstances.