The objective of this study, undertaken in the USA, was to investigate the consequences of autonomy-related companion behaviors on patient decision-making activity during geriatric primary care visits. Videotapes were analyzed to characterize patient and companion decision-making activity and related companion behaviors. These behaviors were coded throughout the visit using an autonomy-based framework that included both autonomy enhancing (i.e. facilitating patient understanding, patient involvement, and doctor understanding) and detracting behaviors, (i.e. controlling the patient and building alliances with the physician). Patients (N = 93) in this cross-sectional sample range in age from 65 to 95 years and are mostly white (n = 73, 79%) and female (n = 67, 72%). Companions are spouses (n = 42, 46%), adult children (n = 33, 36%), or other relatives and friends (n = 15, 16%) of patients. Companions are active participants in medical visits and engage in more autonomy enhancing than detracting behaviors. Companions of sicker (compared with less sick) patients were more likely to facilitate patient understanding, p < .05; doctor understanding, p < .01; and patient involvement, p = .06, in care. Patients whose companions facilitated their involvement in the medical visit by asking the patient questions, prompting the patient to talk, and asking for the patient's opinion were more than four times as likely to be active in decision-making as patients whose companions did not assist in this manner (unadjusted OR 3.5, CI 1.4-8.7, p < .01; adjusted OR 4.5, CI 1.6-12.4, p < .01). Companions can play an important role in the visits of geriatric patients by facilitating communication throughout the visit as well as patient activity in decision-making.