Background/Purpose: Vulnerable populations with rheumatoid arthritis (RA) have poorer outcomes and report suboptimal shared decision-making communication. Patient involvement in the choice of disease modifying anti-rheumatic drugs (DMARDs) for RA may result in improved adherence and better outcomes. Our objective was to test the efficacy of an RA decision aid tool, developed for low literacy and non-English-speaking patients, to improve knowledge and enhance patient involvement in treatment decisions.

Methods: We conducted a pilot trial to test the utility of the decision aid, a set of 5 issue cards which described 12 DMARDs by frequency and mode of administration, time to onset of action, cost, side effects and contraindications. Patients at 2 university-affiliated rheumatology clinics faced with a possible medication change were enrolled consecutively into one of three arms: Arm 1 – existing medication summary guide (usual care) provided pre-clinic visit; Arm 2 – low literacy RA guide provided pre-visit; Arm 3 – low literacy guide provided pre-visit and decision aid used during visit. Immediately post-visit and during a telephone interview 3 months post-visit, subjects completed an RA knowledge questionnaire and the low-literacy version of the Decisional Conflict Scale (DCS; range 0-100; higher scores reflect more conflict).  At 3 months, subjects were also asked about self-reported medication adherence. All materials were available in English, Spanish, and Chinese. DCS, knowledge, and adherence were compared by arm using linear (for DCS) or logistic regression with adjustment for gender immediately post-visit and 3 months.

Results: Of 166 patients enrolled, 88% were female, mean age 59±12; 87% were non-white; 54% spoke a language other than English; and 66% had not graduated high school. There were no statistically significant differences by study arm for education, language, or race/ethnicity; however, there were differences by gender (p=0.02). Compared with the existing guide (Arm 1, n=58), patients who received the low literacy guide plus decision aid (Arm 3, n=60), were more likely to have adequate RA knowledge (OR 3.1, 95% CI 1.4-7.0) immediately post-visit; a difference not seen at 3 months (Table). Among patients who reported a medication change during the visit, Arm 3 patients reported lower decisional conflict compared to those in Arm 1 post-visit (mean DCS 11 vs. 24, p =0.05) and at 3 months (17 vs. 32, p<0.05).  Self-reported adherence did not differ by study arm at 3 months.

Conclusion: An innovative, multi-lingual, low literacy decision aid tool effectively reduced decisional conflict among vulnerable patients with RA faced with choices about DMARDs. The decision aid and medication guide also enhanced RA-specific knowledge immediately post-visit. Future studies are needed to assess the impact of the decision aid on health outcomes in vulnerable populations.