Multiattribute utility scores for predicting family physicians' decisions regarding sinusitis

To examine whether multiattribute utility (MAU) scores can be used to predict family physicians' decisions regarding patients suspected to have sinusitis and rhinitis, 100 randomly selected family physicians from the Leiden area (The Netherlands) were asked to rank a set of six attributes regarding their importance, yielding attribute weights. Next, the physicians were asked to rate the degrees to which five decision alternatives optimized each attribute, yielding utilities, regarding three case vignettes about a patient suspected to have acute maxillary sinusitis and rhinitis, with a brief clinical history. By combining attribute weights and utilities, a MAU score was calculated for each decision alternative regarding each case vignette. Finally, for each case vignette the physicians' treatment preferences were assessed by means of an open-ended question. For the clear-cut sinusitis case, management strategies and highest MAU scores were concordant for 80% of the physicians. Regarding the dubious sinusitis and rhinitis case, concordance was 50%. The latter was associated with small differences in values between the highest MAU scores. Because agreement among the physicians regarding the management strategies and weight ranks was high and there was little variation in the highest MAU scores, the kappas between the reported management strategies and weight ranks were poor (0.24 and lower). It is concluded that MAU scores may be used to predict family physicians' decisions regarding the management of patients suspected of having sinusitis where there are significant differences in values between the highest MAU scores.