There is growing interest in shared decision making (SDM) in the United States and globally, at both the clinical and policy levels. SDM is typically employed during "preference-sensitive" decisions, where there is equipoise between treatment options with equal or similar outcomes from a medical standpoint. In these situations, patients' preferences for the possible risks, benefits, and trade-offs between options are central to the decision. However, SDM also may be appropriate in clinical situations besides those in which data demonstrate equipoise. In situations of low evidence, where evidence is conflicting, unavailable or not applicable to an individual patient, supporting SDM can present unique challenges, above and beyond the challenges faced during more standard preference-sensitive decisions. This article discusses challenges in supporting shared decisions when clinical evidence is low, describes strategies that can facilitate SDM despite low evidence, and suggests avenues for future research to explore further these proposed strategies.