Variability in clinical assessment of clade IIb mpox lesions

Abstract

ObjectivesThe ongoing global outbreak of mpox, caused by clade IIb mpox virus, poses significant challenges in accurately categorizing and assessing the diversity of lesions. With lesion resolution being a key endpoint in clinical trials and observational studies, it is essential to evaluate the inter-rater reliability and agreement of clade IIb mpox lesion assessment among clinicians. MethodsClinicians experienced in clade IIb mpox disease were surveyed online with 20 lesion images. They categorized lesions into active, crusted, resolved, or unclassifiable groups. Reliability was assessed with Fleiss' kappa and agreement with proportion of exact agreement. ResultsFifty-three clinicians completed the survey, with a median self-reported confidence rating of 7 (on a scale of 1 to 10) in assessing mpox lesions. The inter-rater reliability was found to be moderate, with a Fleiss' kappa coefficient of 0.417 (P <0.05, 95% confidence interval: 0.409-0.425). The inter-rater agreement was 61%. ConclusionThis study demonstrates moderate inter-rater reliability and agreement in clade IIb mpox lesion assessment among clinicians. The findings emphasize the importance of standardizing lesion classification systems to facilitate clinical care (e.g., decision to start treatment) and public health (e.g. isolation) decisions and a need to explore alternative endpoints for clinical trials.