Using discrete choice experiments to evaluate alternative electronic prescribing systems

Abstract

Abstract Objective To assess the relative importance to pharmacists and general practitioners (GPs) of different characteristics of electronic prescribing systems. Methods A discrete choice experiment (DCE) was used to obtain preferences for the following attributes of an electronic prescribing scheme: “typical response time”, “frequency of slow responses”, “frequency of unscheduled downtime”, “length of unscheduled downtime”, “frequency of scheduled downtime”, and “frequency of lost/corrupted data”. Preferences of GPs and pharmacists were compared using the Likelihood ratio test and the Wald statistic. The rate at which respondents were willing to trade between attributes and benefit scores for different systems were estimated. Rationality of responses and theoretical validity of responses to the DCE were also assessed. Subjects and setting 199 pharmacists and 197 GPs in Scotland. Key findings The usable response rate for the DCE was 42.4 per cent. Preferences of pharmacists and GPs were significantly different. Given the units of measurement, the most important attribute for pharmacists was “length of unscheduled downtime” whereas for GPs it was “typical response time”. Evidence was found of rationality of responses and theoretical validity. Conclusions All technical attributes of an electronic prescribing system assessed in this study were considered by the respondents to be important. Discrete choice experiments provide useful information on the relative importance and trade-offs between attributes as well as benefit scores for different systems. The preferences of pharmacists and GPs are significantly different. Future research should investigate the nature of the benefit function, the assumption of trading across attributes and the external validity of the technique.