Shared decision making (SDM) is a process by which a healthcare choice is made jointly by the practitioner and the patient and is said to be the crux of patient-centred care. Policy makers perceive SDM as desirable because of its potential to a) reduce overuse of options not clearly associated with benefits for all (e.g., prostate cancer screening); b) enhance the use of options clearly associated with benefits for the vast majority (e.g., cardiovascular risk factor management); c) reduce unwarranted healthcare practice variations; d) foster the sustainability of the healthcare system; and e) promote the right of patients to be involved in decisions concerning their health. Despite this potential, SDM has not yet been widely adopted in clinical practice. To determine the effectiveness of interventions to improve healthcare professionals' adoption of SDM. We searched the following electronic databases up to 18 March 2009: Cochrane Library (1970-), MEDLINE (1966-), EMBASE (1976-), CINAHL (1982-) and PsycINFO (1965-). We found additional studies by reviewing a) the bibliographies of studies and reviews found in the electronic databases; b) the clinicaltrials.gov registry; and c) proceedings of the International Shared Decision Making Conference and the conferences of the Society for Medical Decision Making. We included all languages of publication. We included randomised controlled trials (RCTs) or well-designed quasi-experimental studies (controlled clinical trials, controlled before and after studies, and interrupted time series analyses) that evaluated any type of intervention that aimed to improve healthcare professionals' adoption of shared decision making. We defined adoption as the extent to which healthcare professionals intended to or actually engaged in SDM in clinical practice or/and used interventions known to facilitate SDM. We deemed studies eligible if the primary outcomes were evaluated with an objective measure of the adoption of SDM by healthcare professionals (e.g., a third-observer instrument). At least two reviewers independently screened each abstract for inclusion and abstracted data independently using a modified version of the EPOC data collection checklist. We resolved disagreements by discussion. Statistical analysis considered categorical and continuous primary outcomes. We computed the standard effect size for each outcome separately with a 95% confidence interval. We evaluated global effects by calculating the median effect size and the range of effect sizes across studies. The reviewers identified 6764 potentially relevant documents, of which we excluded 6582 by reviewing titles and abstracts. Of the remainder, we retrieved 182 full publications for more detailed screening. From these, we excluded 176 publications based on our inclusion criteria. This left in five studies, all RCTs. All five were conducted in ambulatory care: three in primary clinical care and two in specialised care. Four of the studies targeted physicians only and one targeted nurses only. In only two of the five RCTs was a statistically significant effect size associated with the intervention to have healthcare professionals adopt SDM. The first of these two studies compared a single intervention (a patient-mediated intervention: the Statin Choice decision aid) to another single intervention (also patient-mediated: a standard Mayo patient education pamphlet). In this study, the Statin Choice decision aid group performed better than the standard Mayo patient education pamphlet group (standard effect size = 1.06; 95% CI = 0.62 to 1.50). The other study compared a multifaceted intervention (distribution of educational material, educational meeting and audit and feedback) to usual care (control group) (standard effect size = 2.11; 95% CI = 1.30 to 2.90). This study was the only one to report an assessment of barriers prior to the elaboration of its multifaceted intervention. The results of this Cochrane review do not allow us to draw firm conclusions about the most effective types of intervention for increasing healthcare professionals' adoption of SDM. Healthcare professional training may be important, as may the implementation of patient-mediated interventions such as decision aids. Given the paucity of evidence, however, those motivated by the ethical impetus to increase SDM in clinical practice will need to weigh the costs and potential benefits of interventions. Subsequent research should involve well-designed studies with adequate power and procedures to minimise bias so that they may improve estimates of the effects of interventions on healthcare professionals' adoption of SDM. From a measurement perspective, consensus on how to assess professionals' adoption of SDM is desirable to facilitate cross-study comparisons.
