Background/Purpose: American rheumatology practice patterns vary considerably. One challenge is practitioner shortage, thus a patient-self-administered screening tool would fill an unmet need if it identified patients with the highest likelihood of having an inflammatory disease (ID).

Objective: To assess the appropriateness of referral patterns and concurrently to evaluate a previously published screening tool, Early Inflammatory Arthritis-3 (EIA–3)¹.

Methods: The US rheumatology database, JointMan® captures inflammatory/non-inflammatory diagnoses of referred patients at a large Washington State rheumatology practice. All patients at the practice fill in an electronic version of the EIA-3 tool. This published screening questionnaire estimates the likelihood of having an inflammatory disease [JIA, RA–fulfilling ACR criteria, Unspecified inflammatory arthritis, SpA, CTDs, SLE] vs non-inflammatory disease [OA, FM, CPPD]. The evaluating clinician was not aware of the EIA-3 results. The results were later correlated to the clinician’s diagnosis. Based on the confusion matrix, effect sizes (sensitivity, specificity and likelihood ratios) were calculated and compared to published results².

Results: There were 956 patients with a mean age of 55; 66.32% female; 77.82% Caucasian. Of those,726 patients had inflammatory diseases [137 also had non-inflammatory disease]; 230 had non-autoimmune disease. The EIA-3 suggested referrals for 725 patients and no referral for 231.

Conclusion: In this study, primary care physicians adeptly referred appropriate ID patients to less accessible rheumatologists (EIA-3 use only modestly increased preselection of ID patients).  It is possible that an electronic patient generated screening tool may help primary providers referring ID patients in less preselected populations (as in the first EIA-3 trial where non-inflammatory diseases were more common).

1.       Bell et al. BMC MSK Disorders 2010, 11:50

2.       Tavares, R. et al. J. Rheumatol 40(4):417-424