Evaluation of the Diagnostic Accuracy of Teledermatology for Skin Lesions in Patients with Inflammatory Rheumatic Diseases

Abstract

This study aimed to investigate the diagnostic accuracy and validity of teledermatology in patients with rheumatological diseases. The study included patients with inflammatory rheumatologic diseases who reported active skin and nail lesions during follow-up. A rheumatologist first evaluated each patient, and a preliminary diagnosis of the skin/nail lesion was recorded. Following, the images of the lesions were shared with two dermatologists and preliminary diagnoses were documented. The patients were then referred to a dermatologist for a face-to-face visit and a final diagnosis was made by using additional tests or skin biopsies if needed. After six months, the same images of the lesions were evaluated by one teledermatologist again, blind to the final diagnosis. The validity and reliability of teledermatology were assessed. Also the positive predictive values were calculated for the teledermatologists and rheumatologists. The study included 113 patients (female 75.2%, mean age: 44.9 years). The two teledermatologists were in perfect agreement; the rheumatologists were in substantial agreement with the final diagnosis (λ: 0.95, 0.82, and 0.61 respectively). The inter-rater agreements of two teledermatologists were as 0.80 (p = 0. 38); the intra-rater agreement of the one teledermatologist was as 0.90 (p = 0.29). Rheumatologists' positive predictive values for the most common lesions (psoriasis, leukocytoclastic vasculitis, and erythema nodosum) were shown to be lower than teledermatologists'. Teledermatology could be used in rheumatological patients with skin/nail lesions as a viable and reliable approach. When compared to teledermatologists, rheumatologists slightly underperformed in diagnosing the most common skin/nail lesions encountered in clinical practice.