Background/Purpose:  

Early treatment of inflammatory arthritis (IA), and especially rheumatoid arthritis (RA), leads to improved outcomes, although diagnostic delays may impede timely treatment. We hypothesized that case finding in a community based health fair may improve the timing of diagnosis and treatment of IA. However, the diagnostic accuracy of this approach in a health fair where IA is likely rare is largely unknown. Therefore, we developed and tested the diagnostic accuracy of a case-finding strategy for IA in a Colorado-based health fair.

Methods:  

Volunteers were evaluated between 2009-2012 at 18 health fair sites using a self-reported joint symptom questionnaire and CCP3 (INOVA; ≥20 units positive) and RF (Roche RFII; >14 units positive) testing. IA was defined as synovitis identified by a trained clinician on examination of the wrists, elbows, and PIP and MCP joints performed at the same time as the questionnaire and blood testing.

Results:  

3260 volunteers completed a questionnaire, joint examination and blood testing; 164 (5%) were identified with active, previously undiagnosed IA, of whom 33/164 (20%) were classified as RA by 2010 ACR/EULAR criteria. After modifications of an initial questionnaire to improve its ease of use and sensitivity for IA, the optimal prediction model for IA (multiple regression; bootstrap validated AUC of 0.73) included 4 items that each contributed similarly to the model: CCP positivity and self-reported: a) morning stiffness ≥1 hour, b) nodules, and c) joint pain, stiffness or swelling present the day of questionnaire completion. The maximal sensitivity (SENS) for IA was with ≥1 item positive (96%; corresponding specificity (SPEC) of 24%); ≥2 positive items resulted in the highest combination of SENS and SPEC (85% and 52%, respectively). In 2012 we piloted a questionnaire with a figure for subjects to identify symptom location, and symptoms in PIPs and MCPs were 93% SENS and 65% SPEC for IA, although there were too few cases of IA (N=15 out of 338 subjects) to evaluate adequately the diagnostic accuracy of additional factors. CCP and/or RF were positive in 279/3260 (9%) of subjects, and both autoantibodies (Abs) were independently significantly associated with IA (p<0.01 and 0.03, respectively). In evaluation of the diagnostic accuracy of Abs alone for IA, CCP and/or RF positivity was 11% SENS and 92% SPEC for IA, reflecting that most IA was seronegative. Additionally, CCP and RF levels >3 times normal were each highly specific for IA (CCP: SENS 5%, SPEC 99%; RF: SENS 5%, SPEC 99%).

Conclusion:  

Assessing self-reported symptoms and Abs in a health fair may identify previously undiagnosed IA/RA. As such, these items may be used initially independently of a clinician thus broadening the applicability of this approach to settings where trained joint examiners are absent. Further testing is necessary to validate these findings and to determine whether case-finding accuracy may be improved by ascertaining specific locations of joint symptoms. Additionally, determining final formal diagnoses, timing of treatment for IA, and post-health fair costs (including costs of evaluating false-positive results) will help determine the overall impact of this approach in arthritis care.

---

« Back to 2013 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/assessment-of-the-diagnostic-accuracy-of-different-strategies-for-health-fair-case-finding-for-undiagnosed-inflammatory-arthritis/