Pos0476 quantification of referral bias in patient- and clinician-reported measures of rheumatoid arthritis severity by geographic distance from an academic rheumatology center

Background Systematic bias to greater severity of rheumatoid arthritis (RA) is often presumed to exist in patients referred for evaluation at academic centers. However, the magnitude of this potential bias has not been quantified to our knowledge. Objectives The objective was to test the hypothesis that patients in a referral population have greater RA severity based on both patient- and clinician-reported measures than local patients. Methods This study included eligible patients with RA who attended an in-person or virtual appointment in the outpatient rheumatology clinic at an academic center between 1/1/2020 and 10/28/2021. RA was defined by at least 2 ICD-10 diagnosis codes ≥30 days but <2 years apart plus use of a qualifying RA medication. Referral population was ascertained by geographic distance from patient residence to the clinic building and categorized as local, <50 miles; regional, ≥50 to <150 miles; or national/international, ≥150 miles. Data were collected from the electronic health records for patient-reported and clinician-reported measures of disease severity. Patient-reported measures included global pain, global arthritis, and Patient-Reported Outcomes Information System (PROMIS) computer adaptive tests for Pain Interference, Fatigue, Physical Function, and Ability to Participate in Social Roles and Activities. Chi-square or Kruskal-Wallis tests were used to analyze differences between groups, adjusting for age at the clinic visit and sex. Linear regression models were used to test for differences between groups in PROMIS measures and Clinical Disease Activity Index (CDAI), adjusting for age, sex, race/ethnicity, and appointment type (new vs. established). Results The study population included 3220 patients with RA, including 1631 local, 956 regional, and 633 national/international patients. Overall, mean (SD) age was 62.9 (13.7) yrs., 2312 (72%) were female, and 2947 (91.5%) were in-person visits. Proportions of new patients in the local, regional, and national/international populations were 5%, 9%, and 17%, respectively. All PROMIS measures were available within 7 days of appointment for 2677 (83%) patients, with no differences between groups (p = 0.394). Regional and national/international patients had significantly higher global pain, pain interference, and fatigue and significantly lower physical function and ability to participate than local patients. Regression analysis showed that regional and national/international patients had higher pain interference (on average: 1.3 and 1.6 units; p<0.001 and p<0.001, respectively) and worse physical function (on average: -1.0 and -2.2 units; p=0.008 and p<0.001, respectively) than local referent patients, adjusting for age, sex, race/ethnicity, and appointment type. In contrast, there were no significant differences between groups in CDAI (on average for regional and national/international: 0.3 and 1.1 units; p=0.73 and p=0.29, respectively). Variable Local (n = 1631) Regional (n = 956) National/ International (n = 633) Total (n = 3220) p Global pain score (0-100) 35.9 (29.2) 39.5 (29.0) 41.8 (30.6) 38.2 (29.5) 0.002 PROMIS T-scores Pain Interference 57.0 (8.3) 58.4 (8.3) 58.8 (8.4) 57.8 (8.4) <0.001 Fatigue 53.5 (9.6) 54.9 (9.8) 56.2 (10.1) 54.4 (9.8) <0.001 Physical Function 42.4 (8.8) 41.2 (8.2) 40.0 (8.6) 41.6 (8.6) <0.001 Ability to Participate 50.3 (9.1) 48.5 (9.4) 47.4 (9.2) 49.2 (9.3) <0.001 Clinical Disease Activity Index 12.2 (11.7) 12.2 (10.7) 13.4 (12.1) 12.4 (11.5) 0.44 C-reactive protein, mg/L 7.8 (11.5) 7.2 (10.9) 7.9 (11.7) 7.6 (11.4) 0.99 Values are mean (SD) or number (%). Conclusion Referral bias by geographic distance from the outpatient rheumatology clinic is more evident in patient-reported than clinician-reported measures of RA severity. The findings inform the interpretation of RA disease severity measures in clinical practice and research and highlight the importance of patient-reported outcome measures. REFERENCES: NIL. Acknowledgements: NIL. Disclosure of Interests John M Davis III Grant/research support from: Pfizer, Girihlet, Sara Achenbach: None declared, Courtney Arment: None declared, Delamo Bekele: None declared, Vanessa Kronzer: None declared, Thomas Mason: None declared, Elena Myasoedova: None declared, Lynne Peterson: None declared, Kerry Wright: None declared, Cynthia S. Crowson: None declared.