“…In the 31 studies included in the block which studied doctors [10,11,15–22,24–30,32–35,37–39,41–44,47,48,51], the median percentage participation was 65%, and the data were collected by means of questionnaires in 27 cases [10,11,15–22,25–27,29,30,33–35,37–39,41–43,47,48,51] (with this procedure being combined with collection based on a review of clinical histories in four cases [18,19,42,43] and a review of prescriptions in one case) [37], and through observers/actors present at the medical visit in four cases [24,28,32,49]. Questionnaires were distributed by post in over 50% of cases [10,16,19,24–27,29,30,34,38,39,51]; nine were self‐administered [11,18,20,21,33,35,37,47,48], four made no mention of the method of distribution [15,22,28,41] and one took the form of a telephone survey [17].…”