LAUSR

Nonlinear associations in rheumatoid arthritis (RA) prevalence and sociodemographic indices suggest social determinants of health affect RA.1 RA screening tools have lower sensitivity and specificity for Hispanic/Latinx individuals compared with white individuals (77% vs 85% sensitivity and 45% vs 87% specificity).2 Hispanic/Latinx people also present to physicians later and with more severe RA.3 There is little research in this population to explain these differences or how they can be addressed. Because it is known the Hispanic/Latinx population has lower English-language proficiency, we hypothesized this is in part because of the lack of validated Spanish-language educational materials and research tools for RA.To validate a Spanish-language patient-provider questionnaire (PPQ) for Hispanic/Latinx patients with RA that had previously been validated as concordant for primary care visits in Sweden.4To understand if rheumatologist and patient assessments for Hispanic/Latinx people with RA are concordant when using a Spanish-language digital PPQ.To determine if a Spanish-language rheumatologist-completed PPQ could be a proxy for patient impressions in a prospective clinical study.A Spanish-language PPQ for RA was created by translating 9 of 13 questions previously validated in the Swedish study,4 and adding a question about treating to target specific to RA. The survey was made available on tablet devices in 4 rheumatology clinics in the U.S. and Puerto Rico. Clinic staff obtained patient consent and gathered demographic information to generate a unique confidential identifier code for each visit, which was inputted into surveys on tablet devices. After the visit ended the patient and rheumatologist were each given the PPQ coded for that visit to complete independently of one another. The PPQs were submitted electronically to a secure database in which the visit code was the only identifier recorded.Across 114 clinical visits, 96.75% of possible answers were recorded and were almost invariably positive with scores of 5 (strongly agree; 88%), 4 (agree; 12%), or 3 (neither agree nor disagree; 0.09%). Physicians responded with 4 (agree) more often than patients (18% vs 6% of responses). Responses from both patient and physician were available for (96.64%) of answered questions. Within these paired answers, 80.67% were concordant (same answer from both patient and physician). Physicians answered 4 when patients answered 5 in 76.5% of discordant responses (different ratings from patient and physician). Most physician ratings of 4 came from 1 of the 4 physicians involved and only 12.5% of patients were responsible for 75% of the patient responses of 4.Hispanic/Latinx patients with RA and their rheumatologist rated their communication, goal setting, and relationships extremely positively, making it difficult to evaluate true concordance and not possible to use rheumatologist-completed PPQs as proxy for patient assessments. Notably, ratings were substantially different from what is typically seen on Likert scales, which normally skew positively but with a normal distribution. This finding may reflect social determinants of health or cultural differences such as a social-desirability bias toward positive statements about physician-patient interactions. Heterogeneity within the participants is also a plausible explanation, considering that a distinct subset of respondents account for almost all responses below the 5 rating. Further research is needed to identify best practices for measuring treatment to target and patient-rheumatologist interactions in the Hispanic/Latinx population with RA.[1]Safiri S, et al. Ann Rheum Dis. 2019;78(11):1463-71. doi:10.1136/annrheumdis-2019-215920[2]Potter J, et al. J Rheumatol. 2008;35(8):1545-9. PMID: 18597406[3]Riad M, et al. J Clin Rheumatol. 2019. doi:10.1097/RHU.0000000000001085[4]Ahlén GC, et al. Fam Pract. 2007;24(5):498503. doi.org/10.1093/fampra/cmm043We wish to acknowledge the important contributions of our HOPE CAPE RA participants, including the patients and providers and Anne Sydor, PhD for her invaluable advice and help throughout the project.This project was funded by the Bristol Myers Squibb Foundation.Daniel Hernandez Grant/research support from: This project is supported by the Bristol Myers Squibb Foundation, Julio Bravo Grant/research support from: This project is supported by the Bristol Myers Squibb Foundation, Juan Jose Maya Grant/research support from: This project is supported by the Bristol Myers Squibb Foundation, Oscar Soto-Raíces Grant/research support from: This project is supported by the Bristol Myers Squibb Foundation, Angel Tapia Grant/research support from: This project is supported by the Bristol Myers Squibb Foundation, Guillermo Valenzuela Grant/research support from: This project is supported by the Bristol Myers Squibb Foundation, W. Benjamin Nowell Grant/research support from: This research was supported by the Bristol Myers Squibb Foundation, Shilpa Venkatachalam Grant/research support from: This project is supported by the Bristol Myers Squibb Foundation