BACKGROUND: Medical decision making has long been in focus, but little is known of the preferences and conditions for elderly people with co-morbidities to participate in medical decision making. The main objective of the present study was to investigate the preferred and the actual degree of control, i.e. the role elderly people with co-morbidities wish to assume and actually had with regard to information and participation in medical decision making during their last stay in hospital.This study was a cross-sectional survey including three Swedish hospitals with acute admittance. The participants were patients aged 75 years and above with three or more diagnoses according to the International Classification of Diseases (ICD-10) and three or more hospitalisations during the last year. METHODS: We used a questionnaire combined with a telephone interview, using the Control Preference Scale to measure each participant's preferred and actual role in medical decision making during their last stay in hospital. Additional questions were asked about barriers to participation in decision making and preferred information seeking role. The results are presented with descriptive statistics with kappa weights. RESULTS: Of the 297 elderly patients identified, 52.5% responded (n = 156, 46.5% male). Mean age was 83.1 years. Of the respondents, 42 of 153 patients said that they were not asked for their opinion (i.e. no shared decision making). Among the other 111 patients, 49 had their exact preferred level of participation, 37 had less participation than they would have preferred, and 23 had more responsibility than they would have preferred. Kappa statistics showed a moderate agreement between preferred and actual role (κw = 0.57; 95% CI: 0.45-0.69). Most patients wanted to be given more information without having to ask. There was no correlation between age, gender, or education and preferred role. 35% of the patients agreed that they experienced some of the various barriers to decision making that they were asked about: 1) the severity of their illness, 2) doctors with different treatment strategies, 3) difficulty understanding the medical information, and 4) difficulty understanding doctors who did not speak the patient's own language. CONCLUSIONS: Physicians are not fully responsive to patient preferences regarding either the degree of communication or the patient's participation in decision making. Barriers to participation can be a problem, and should be taken into account more often when dealing with hospitalised elderly people.