Providers who referred patients to Reach to Recovery (Reach), an American Cancer Society breast cancer support group, were compared with those who did not to evaluate whether providers who identified problems with the program were less likely to make referrals. Also considered were contact with a Reach volunteer, having a Reach program in the area, years since residency or medical training, perceived value of the Reach program, belief that one's peers refer patients to Reach, urban or rural location of practice, and size of practice. When other variables were considered, physicians who identified problems with Reach were no less likely to refer patients to the program than those who did not. The factors most associated with referral (P < .001) were "having a Reach program in the area" and "having had contact with a Reach volunteer." Implications of these findings and the experience of conducting the evaluation through the Collaborative Evaluations Fellows Project are discussed.