Abstract Background: Breast cancer risk assessment and interventions for prevention, such as chemoprevention, are underutilized in the U.S. Reasons for low uptake include inability to routinely screen for high-risk women in the primary care setting, inadequate time for counseling, and insufficient knowledge about risk-reducing strategies. We developed an initial prototype of a web-based decision aid (DA), RealRisks, which incorporates experience-based dynamic interfaces to communicate risk aimed at reducing inaccurate risk perceptions, particularly in low-numerate populations. Methods: RealRisks is a patient DA that models patient-provider dialogue with modules on breast cancer risk, genetic testing, and chemoprevention. Embedded within the narrative are 2 games of experience-based risk interfaces, demonstrating average 5-year and lifetime breast cancer risk. Both games ask players to sample from a pictograph of 100 clickable women to better learn the meaning of a pre-set probability (i.e., 12 out of 100 women or 12%). We conducted four focus groups of 7-9 English-speaking women over the age of 18, recruited from the local community in Upper Manhattan in New York City. These recorded sessions lasted about 90 minutes and involved use of RealRisks on a laptop, questionnaire completion before and after interacting with the DA, and a semi-structured group discussion. Questionnaires included information about demographics, numeracy, internet access, breast cancer risk factors, perceived breast cancer risk, and evaluation of RealRisks on a 7-point Likert scale. Descriptive statistics were generated to document baseline characteristics and frequencies of positive and negative attitudes about RealRisks. Paired t-test and McNemar's test were used to compare within-individual changes in accuracy of perceived breast cancer risk. During the group discussion, verbal responses were condensed into themes using a qualitative approach. Results: From May to June 2013, 34 women were enrolled. Median age was 53.5 (range, 35-75); 85% were either black or Hispanic; 41% met criteria for low numeracy; and 88% had internet access. After removing 3 women with a history of breast cancer, 3 (9.7%) met high-risk criteria for breast cancer according to the Gail model (≥1.67% 5-year risk) and mean 5-year and lifetime breast cancer risk were 1.11% (±0.77) and 7.46% (±2.87), respectively. After interacting with RealRisks, the difference in perceived vs. actual breast cancer risk according to the Gail model significantly improved for 5-year risk (p=0.008), but not lifetime risk (p=0.20). Before exposure to RealRisks, 52% had accurate breast cancer risk perceptions (defined as within ±5% of actual lifetime risk according to the Gail model) compared to 70% after RealRisks (p=0.10). In particular, 4 out of 5 women who overestimated their lifetime breast cancer risk by >30% had accurate risk perceptions after exposure to RealRisks. We found a significant association between numeracy and accuracy of risk perception after interacting with RealRisks (p=0.05). Over 85% of the participants thought RealRisks was useful, easy to use, increased their knowledge about breast cancer and understanding of breast cancer risk factors. From the focus group discussions, we found that knowledge about breast cancer risk factors, apart from family history, was limited. Participants were interested in receiving a personalized breast cancer risk assessment and found the interactive games engaging. Discussion: In a multi-ethnic low-numerate population, we demonstrated a significant improvement in accuracy of perceived breast cancer risk after exposure to RealRisks. Based upon feedback from our focus groups, we were able to identify information needed to fully represent the important issues of breast cancer risk to further develop our prototype for testing in a randomized controlled trial. Citation Format: Tong Xiao, Katherine D. Crew, Parijatham S. Sivasubramanian, Alejandra N. Aguirre, Cindy K. Smalletz, Rita Kukafka. Development of a patient-centered decision aid to improve accuracy of breast cancer risk perception. [abstract]. In: Proceedings of the Twelfth Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2013 Oct 27-30; National Harbor, MD. Philadelphia (PA): AACR; Can Prev Res 2013;6(11 Suppl): Abstract nr C03.