This is a commentary on an article by Andrew Peterson, Emily Largent and Jason Karlawish on supported decision-making (SDM) in people with dynamic impairments of decision-making capacity (DMC). We argue that Peterson and colleagues fail to address convincingly the concern that supported decision-making may render people with impaired DMC vulnerable to undue influence. We propose our “combined supported decision-making model” as a way to address this concern. The combined SDM model employs a functional and time- and task-indexed DMC assessment as an allocation and success criterion for SDM: SDM is necessary when a person’s level of DMC falls below the threshold at which she can make meaningful treatment decisions, and successful when it raises the person’s level of DMC above this threshold. The model reduces the risk of undue influence by maintaining a clear distribution of responsibility: when SDM is successful, responsibility for the treatment decision is accorded to the person who is supported; and when a person’s DMC remains below the threshold despite SDM, responsibility is accorded to a substitute decision-maker.