Head injury is the leading cause of death and serious morbidity in bicycle accidents. There is good evidence to recommend helmets, yet few children wear them. We evaluated helmet promotion in a randomized trial targeting children presenting to primary care settings for routine ambulatory care. The intervention consisted of physician counseling and take-home pamphlets. The study involved 339 families, 167 in the intervention group and 172 in the control group. In a follow-up telephone call, 2 to 3 weeks later, only 7.2% of the intervention group had purchased helmets, compared with 7.0% of the control group (chi 2 = 0.0056, P = .94). During the latter half of the study, bicycle safety received considerable media attention in Ottawa, and the provincial medical society sponsored a $5 discount campaign. Therefore both groups were subject to community "co-intervention." Nonetheless, we were surprised that physician counseling made no additional impact. Our results and the success of certain community programs suggest that physicians interested in helmet promotion would do better to participate in the design and implementation of multidisciplinary campaigns.