Fig. 3

Part 2 survey respondents’ ordering of potential harmful outcome domains that would affect their decision to use a treatment. Twelve groups of survey participants were presented with 7/21 potential harms. Each participant (N = 363) ranked the importance of the potential harms in making a decision about whether to use a treatment for pain, and we used BWS to order them. For each harm in the set, we subtracted the proportion of participants who ranked each potential harm “1” from the proportion of participants who ranked each harm “7” to obtain a BWS score. Thus, scores greater than 0 were ranked 1 (“most affects your decision”) more often than they were ranked 7 (“least affects your decision”)