the following study aimed to: (1) compose a french version ofanderson et al.'s (1989) original diabetes attitude scale (das) by sampling 5 groups of quebec health care professionals; (2) increase the reliability and validity ofthe das (anderson et al., 1989) through the creation ofa revised version ofthis instrument. this scale, entitled the revised diabetes attitude scale (das-r), was developed through the efforts of a panel of 6 diabetes educators from notre dame hospital, montreal, quebec. thirty-eight items, thirty-one of these with cronbach alphas greater than 0.37, were selected from the original das (1989) and combined with 27 items that had been created by members of the panel to form a 65-item scale. the items were revised by the panel and those deemed to be redundant or to increase scale variability were dropped. a total of fifteen items were dropped, and the final version of the scale contained 50 questions. the 50 -item scale was then mailed to 478 health care professionals, nurses, dietitians, physicians, pharmacists, and psychologists (specialists and nonspecialists in diabetes care) in all regions of quebec. four-hundred additional surveys were distributed to health care professionals through quebec diabetes association conferences. the surveys were returned by 62 nurses, 49 dietitians, 149 physicians, 55 pharmacists, and 5 psychologists, totaling 320 returns (a return rate of 36%). the returned surveys were analyzed, and a 37-item das-r composed of 8 subscales resulted. evidence for the reliability and validity ofthe 37-item das-r are included in this study. the diabetes behaviour scale (dbs) was created to provide evidence for the validity of the das-r. this 13-item scale was developed specifically to accompany the das-r and is a measure ofdiabetes management-related behaviours applicable to physicians. the development ofthe dbs was based on information obtained from specialized texts in diabetes care and opinions from a specialist in diabetes care and education (notre dame hospital, montreal). the scale is comprised of questions based on various behavioural issues appropriate to diabetes care (i.e. are patients referred to cardiologists, ophthalmologists when required? does the physician attend diabetes workshops, seminars? are patients referred to outpatient education clinics?). an item analysis ofthe scale revealed that 7 items had poor item-total correlations. these were dropped from further analyses. the resulting 6-item scale had a satisfactory cronbach alpha value (alpha= .48). analyses of the results showed that there were significant correlations between two ofthe das-r subscales and the dbs. this is the first time in diabetes care research that an attempt at discovering a correlation between attitudes and behaviours has been accomplished using scales derived specifically for their measurement. the relationship between attitudes and behaviours found in this study supports ajzen and fishbein's theory of reasoned action.