Reliability of Assessing Hand Osteoarthritis on Digital Photographs and Associations With Radiographic and Clinical Findings

Abstract

Objective: To investigate the reliability and construct validity of an atlas for grading hand osteoarthritis (OA) on photographs in a separate younger community-dwelling population than the development cohort.

Methods: Participants were community-dwelling adults (ages ≥50 years) in North Staffordshire, UK with hand pain or hand problems in the last year who attended a research clinic. High-quality photographs were taken in a standardized position. A photographic atlas was used to score hand joints (second and third distal interphalangeal [DIP], second and third proximal interphalangeal [PIP], and first carpometacarpal [CMC] joints) and joint groups (DIP, PIP, and CMC joints) for OA on a 0-3 scale. Hand radiographs were graded for OA using the Kellgren/Lawrence (K/L) grading system. Clinical features (nodes, bony enlargement, and deformity) were determined by physical examination. Associations of photographic hand OA grades with radiographic OA and clinical features were determined to assess construct validity.

Results: In total, 558 participants (mean age 64 years, 62% women) were included in the analyses. Reliability for scoring OA on the photographs was good (mean intrarater intraclass correlation coefficient [ICC] 0.77 and mean interrater ICC 0.71). At the joint level, photographic hand OA grade was positively associated with radiographic OA grade (Spearman's ρ = 0.19-0.57, P < 0.001) and the number of clinical features (Spearman's ρ = 0.36-0.59, P < 0.001). At the person level, individuals with higher global photographic OA scores had higher summed K/L scores and higher percentages meeting the American College of Rheumatology clinical hand OA criteria.

Conclusion: This photographic scoring system was reliable and a good indicator of hand OA in a separate younger community-dwelling population than the development cohort. This method of data collection offers researchers a feasible alternative to physical examination and radiography.

Figure 1

In this study, the associations of photographic hand osteoarthritis (OA) with radiographic OA and clinical features were examined to explore the construct validity of hand photography as an indicator of hand OA. An example of a hand photograph and its corresponding radiographic image is shown.

Figure 2

Photographic hand osteoarthritis (OA) grades and the frequency of radiographic OA and clinical features. A, For each joint and joint group, the percentage of individuals with radiographic OA increased from grade 0 through to grade 3 of photographic hand OA scores. B, For each joint and joint group, the percentage of individuals with ≥1 clinical features as determined on the hand examination increased with photographic grade of hand OA. LDIP3 = left third distal interphalangeal joint; LPIP3 = left third proximal interphalangeal joint; LDIP2 = left second DIP joint; LPIP2 = left second PIP joint; LCMC1 = left first carpometacarpal joint; RCMC1 = right first CMC joint; RPIP2 = right second PIP joint; RDIP2 = right second DIP joint; RPIP3 = right third PIP joint; RDIP3 = right third DIP joint; K&L = Kellgren/Lawrence grade.

Figure 3

Global photographic hand osteoarthritis (OA) scores were compared with clinical OA at the person level. A higher percentage of those classified as having American College of Rheumatology (ACR) clinical hand OA were represented in those with grades 3 and ≥4 global hand OA scores (A). When using the relaxed ACR criteria (where hand pain was reported on some days or more in the last month rather than on most days or more) (B) and the ACR clinical hand OA features excluding the hand pain question (C), the percentage of individuals meeting the criteria increased as grade of global hand OA increased.