Abstract
Background/Objectives: The Mallet classification systems are widely used clinician-reported measures of shoulder function in children with brachial plexus birth injury (BPBI). However, “Mallet” encompasses multiple forms differing in task content, scoring, and interpretation. This review aimed to critically evaluate their measurement properties using COSMIN methodology. Methods: This review followed COSMIN 2.0, PRISMA 2020, and PRISMA-COSMIN for OMIs 2024 guidance and was registered in PROSPERO (CRD420261457308). Five databases were searched from inception to 23 July 2026, supplemented by grey literature and citation searches. Methodological quality was assessed using COSMIN Risk of Bias tools, and certainty was graded using the modified GRADE approach. Results: Thirteen studies met the eligibility criteria and represented seven review-defined Mallet operationalisations; no included study evaluated an eligible measurement property of the original patient-level Mallet grade. Task-level reliability varied by task, and task-level construct-validity findings more often met the review hypotheses than composite-score findings; however, certainty was low or very low. Clinical grades generally overestimated angle-defined movement, particularly external rotation. In one low-certainty study, reliability varied more across rater configurations than across assessment modes. No eligible evidence was identified for measurement error, responsiveness, content validity, cross-cultural validity, or minimal important change, and structural-validity evidence was absent for multi-item forms used to generate composite scores. Conclusions: Mallet operationalisations should not be treated as interchangeable instruments. Task-level findings require cautious, operationalisation-specific interpretation, while the dimensionality and measurement error of composite scores remain unestablished; any preference for V1 task-level reporting is provisional.