What do GPs value about their job ?

In the mid 1990's, there was a shift in policy emphasis in the UK from policies that tried to influence the behaviour of GPs through financial incentives (e.g. the 1990 GP Contract), to policies that tried to improve the professional working lives of GPs to improve recruitment, retention and morale (e.g. changes to out of hours arrangements). This highlights the role non-pecuniary job characteristics in the GPs utility function. The aim of this paper is to examine the relative value GPs place on the pecuniary and non-pecuniary characteristics of their job. A model of GPs' choice of practice is presented. Information about job characteristics from studies of GPs' job satisfaction and factors influencing practice choice, and several interviews with GPs, were incorporated into a theoretical model of practice choice. This was tested using a discrete choice experiment (conjoint analysis) based in random utility theory. A self-administered questionnaire was posted to a stratified random sample of 1200 English and Scottish GPs. This presented GPs with a series of pairs of scenarios describing different practices. Each scenario differed with respect to a range of attributes as specified in the theoretical model. From each pair, GPs were asked which they preferred. A 65% response rate was achieved. Data were analysed using a random effects probit model. The results from the estimated utility function indicate the relative importance of pecuniary and non-pecuniary job characteristics; the monetary valuation of job characteristics; and the influence of GP characteristics and current work circumstances on preferences. The monetary valuations estimate GPs' strength of preference for specific job characteristics, and can be interpreted as GPs willingness to pay / willingness to accept for specific job characteristics. The most important attribute was out of hours arrangements, followed by the use of guidelines, daytime hours at work, and list size per GP. The number of hours of administration per week and whether the practice had opportunities to develop special interests/academic work, were not statistically significant in explaining practice choice. There were also variations in strength of preference across subgroups of GPs. The results are theoretically valid (showing a positive marginal utility of income, and negative marginal utility of time at work) suggesting the use of stated preference methods were appropriate in this context. However, the results do depend on the assumptions used in the model, including the assumption that GPs were trading-off. The results have important implications for the further development …