PT - JOURNAL ARTICLE AU - Rachel Thompson AU - Gabrielle Stevens AU - Hillary Washburn AU - Paul Barr TI - 49 Is disclosure of financial conflicts of interest in patient decision aids effective? Insights from a simulated pilot AID - 10.1136/bmjebm-2019-EBMLive.57 DP - 2019 Jul 01 TA - BMJ Evidence-Based Medicine PG - A30--A30 VI - 24 IP - Suppl 1 4099 - http://ebm.bmj.com/content/24/Suppl_1/A30.1.short 4100 - http://ebm.bmj.com/content/24/Suppl_1/A30.1.full SO - BMJ EBM2019 Jul 01; 24 AB - Objectives Patient decision aids are regarded as a strategy for engaging patients in decision-making, facilitating evidence translation, and reducing low-value health care. However, financial conflicts of interest among authors of these resources have the potential to bias their content and undermine their usefulness. The prevailing quality framework in this field – the International Patient Decision Aids Standards – requires that decision aids disclose author financial conflicts of interest. However, to our knowledge, no research has examined whether such disclosure has the desired effect. Our objective was to conduct a simulated pilot study of how members of the public perceive and respond to financial conflicts of interest disclosure in patient decision aids.Methods We conducted an Internet-based between-subjects fractional factorial experiment. Participants were English-speaking adults in the United States invited by a research recruitment service. Quotas were imposed so the sample was composed equally of participants with adequate (AHL) and limited health literacy (LHL). Participants were asked to imagine they had been diagnosed with an illness, were presented with a fictional one-page decision aid on three treatment options (No Treatment, Abbatide, Benedent), and were advised either that they found it online or received it from a doctor. The decision aid featured a financial conflicts of interest disclosure either negative or positive in content and either basic or enhanced in detail and design. Participants completed several measures. Two items administered to verify the effectiveness of our experimental manipulation are the primary focus of this analysis.Results After excluding people ineligible for the study or with missing or unreliable data, our sample comprised 395 participants. When asked if they read the conflicts of interest disclosure when they viewed the decision aid, 197 participants (49.9%) responded ‘Yes’ (AHL: 55.3%; LHL: 44.4%). Of participants who reported reading the disclosure, 72 (36.5%) gave a correct answer when asked, ‘Are any of the authors of the decision aid paid by the company that sells Abbatide or Benedent?’ (AHL 34.9%; LHL 38.6%). Overall, 18.2% of participants reported reading the disclosure and correctly recalled whether a conflict of interest was disclosed, with the lowest proportion (5.3%) observed for the basic, negative disclosure.Conclusions A minority of study participants reported reading and correctly recalled the content of a financial conflicts of interest disclosure in a fictional decision aid. Whether these findings can be generalised to actual decision aid users is uncertain. Our recruitment of members of the public and use of a simulated illness scenario, although motivated by ethical considerations, may have weakened the salience of both the decision aid and the disclosure. On the other hand, knowledge of study participation along with the relative prominence of the disclosure in the fictional decision aid may have amplified participants’ attention to it. Altogether, we tentatively conclude that disclosure of financial conflicts of interest in patient decision aids may not have the desired effect and advocate further research attention to optimal methods of addressing this issue.