Ubel identified a recency effect whereby women at high risk of breast cancer who learned first about the risks of tamoxifen prophylaxis therapy remembered the benefits of tamoxifen better and thought more favourably of the drug in comparison to women who learned first about the benefits [15]. We speculate that the influence of order effects will be greater in PtDAs with greater numbers of attributes. We also predict that primacy vs recency effects will differ depending on list length and where in the PtDA the patient is asked their treatment preference. Future studies exploring this website different designs with both fewer and greater numbers of attributes should

further examine the influence of both primacy and recency effects. We found that younger people (≤35) were more influenced by the primacy effect, which could be because this group has preformed habits for

reading web pages. Studies of web browsing have found that older users are more likely to read all of the information NU7441 ic50 on a screen before committing themselves to move to the next screen [31]. Younger users are more likely to read less of the on-screen information on a web page, often reading the top line and then scanning vertically down the left of the page [31]. If this phenomenon is also present with web based PtDAs, it is plausible that younger people are more influenced by order effects. A specific strength of this study is the randomized experiment used to detect differences between PtDA designs. Despite over 86 randomized trials of PtDAs [32], few have used randomization to examine the influence of design issues. The majority of those few studies considered the influence of individualized risk estimates and found only limited impact [30]. This study contributes to the small literature researching the effect of information design on decision-making. Our results should be interpreted with caution given certain study limitations. First, the task was hypothetical and so we cannot be sure that the results observed would also be found among sleep apnea patients making actual treatment

decisions. If this experiment was not hypothetical, it is quite plausible that patients would spend more time studying the 17-DMAG (Alvespimycin) HCl information provided and be less influenced by order effects as a result. Consequently, it is possible the size of effects may be an overestimate of what would happen in clinical practice. The study by Ubel et al. did however find small order effects among women at a high risk of breast cancer who used a PtDA on preventative therapy options. Thus while the effects we observed might be reduced, they are unlikely to be eliminated if our study were replicated with a sample of sleep apnea patients [15]. Second, the results could have been confounded by the order in which we presented the value clarification exercise and treatment option information.