Decision aids prepare people to participate in 'close call' decisions that involve weighing benefits, harms, and scientific uncertainty. To conduct a systematic review of randomised controlled trials (RCTs) evaluating the efficacy of decision aids for people facing difficult treatment or screening decisions. We searched MEDLINE (Ovid) (1966 to July 2006); Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library; 2006, Issue 2); CINAHL (Ovid) (1982 to July 2006); EMBASE (Ovid) (1980 to July 2006); and PsycINFO (Ovid) (1806 to July 2006). We contacted researchers active in the field up to December 2006. There were no language restrictions. We included published RCTs of interventions designed to aid patients' decision making by providing information about treatment or screening options and their associated outcomes, compared to no intervention, usual care, and alternate interventions. We excluded studies in which participants were not making an active treatment or screening decision, or if the study's intervention was not available to determine that it met the minimum criteria to qualify as a patient decision aid. Two review authors independently screened abstracts for inclusion, and extracted data from included studies using standardized forms. The primary outcomes focused on the effectiveness criteria of the International Patient Decision Aid Standards (IPDAS) Collaboration: attributes of the decision and attributes of the decision process. We considered other behavioural, health, and health system effects as secondary outcomes. We pooled results of RCTs using mean differences (MD) and relative risks (RR) using a random effects model. This update added 25 new RCTs, bringing the total to 55. Thirty-eight (69%) used at least one measure that mapped onto an IPDAS effectiveness criterion: decision attributes: knowledge scores (27 trials); accurate risk perceptions (11 trials); and value congruence with chosen option (4 trials); and decision process attributes: feeling informed (15 trials) and feeling clear about values (13 trials).This review confirmed the following findings from the previous (2003) review. Decision aids performed better than usual care interventions in terms of: a) greater knowledge (MD 15.2 out of 100; 95% CI 11.7 to 18.7); b) lower decisional conflict related to feeling uninformed (MD -8.3 of 100; 95% CI -11.9 to -4.8); c) lower decisional conflict related to feeling unclear about personal values (MD -6.4; 95% CI -10.0 to -2.7); d) reduced the proportion of people who were passive in decision making (RR 0.6; 95% CI 0.5 to 0.8); and e) reduced proportion of people who remained undecided post-intervention (RR 0.5; 95% CI 0.3 to 0.8). When simpler decision aids were compared to more detailed decision aids, the relative improvement was significant in knowledge (MD 4.6 out of 100; 95% CI 3.0 to 6.2) and there was some evidence of greater agreement between values and choice.In this review, we were able to explore the use of probabilities in decision aids. Exposure to a decision aid with probabilities resulted in a higher proportion of people with accurate risk perceptions (RR 1.6; 95% CI 1.4 to 1.9). The effect was stronger when probabilities were measured quantitatively (RR 1.8; 95% CI 1.4 to 2.3) versus qualitatively (RR 1.3; 95% CI 1.1 to 1.5).As in the previous review, exposure to decision aids continued to demonstrate reduced rates of: elective invasive surgery in favour of conservative options, decision aid versus usual care (RR 0.8; 95% CI 0.6 to 0.9); and use of menopausal hormones, detailed versus simple aid (RR 0.7; 95% CI 0.6 to 1.0). There is now evidence that exposure to decision aids results in reduced PSA screening, decision aid versus usual care (RR 0.8; 95% CI 0.7 to 1.0) . For other decisions, the effect on decisions remains variable.As in the previous review, decision aids are no better than comparisons in affecting satisfaction with decision making, anxiety, and health outcomes. The effects of decision aids on other outcomes (patient-practitioner communication, consultation length, continuance, resource use) were inconclusive.There were no trials evaluating the IPDAS decision process criteria relating to helping patients to recognize a decision needs to be made, understand that values affect the decision, or discuss values with the practitioner. Patient decision aids increase people's involvement and are more likely to lead to informed values-based decisions; however, the size of the effect varies across studies. Decision aids have a variable effect on decisions. They reduce the use of discretionary surgery without apparent adverse effects on health outcomes or satisfaction. The degree of detail patient decision aids require for positive effects on decision quality should be explored. The effects on continuance with chosen option, patient-practitioner communication, consultation length, and cost-effectiveness need further evaluation.