Where is the evidence? A systematic review of shared decision making and patient outcomes.

BACKGROUND: Despite widespread advocacy for shared decision making (SDM), the empirical evidence regarding its effectiveness to improve patient outcomes has not been systematically reviewed. The purpose of this study was to systematically review the empirical evidence linking patient outcomes and SDM, when the decision-making process has been explicitly measured, and to identify under what measurement perspectives SDM is associated with which types of patient outcomes (affective-cognitive, behavioral, and health). DATA SOURCES: PubMed (through December 2012) and hand search of article bibliographies. STUDY SELECTION: Studies were included if they empirically 1) measured SDM in the context of a patient-clinician interaction and 2) evaluated the relationship between SDM and at least 1 patient outcome. DATA EXTRACTION: Study results were categorized by SDM measurement perspective (patient-reported, clinician-reported, or observer-rated) and outcome type (affective-cognitive, behavioral, or health). DATA SYNTHESIS: Thirty-nine studies met inclusion criteria. Thirty-three used patient-reported measures of SDM, 6 used observer-rated measures, and 2 used clinician-reported measures. Ninety-seven unique patient outcomes were assessed; 51% affective-cognitive, 28% behavioral, and 21% health. Only 43% of assessments (n = 42) found a significant and positive relationship between SDM and the patient outcome. This proportion varied by SDM measurement perspective and outcome category. It was found that 52% of outcomes assessed with patient-reported SDM were significant and positive, compared with 21% with observer-rated and 0% with clinician-reported SDM. Regardless of measurement perspective, SDM was most likely to be associated with affective-cognitive patient outcomes (54%), compared with 37% of behavioral and 25% of health outcomes. LIMITATIONS: The relatively small number of studies precludes meta-analysis. Because the study inclusion and exclusion criteria required both an empirical measure of SDM and an assessment of the association between that measure and a patient outcome, most included studies were observational in design.
CONCLUSIONS: SDM, when perceived by patients as occurring, tends to result in improved affective-cognitive outcomes. Evidence is lacking for the association between empirical measures of SDM and patient behavioral and health outcomes.