The Melbourne Assessment of Unilateral Upper Limb Function (Melbourne Assessment) measures quality of movement for children with congenital or acquired neurological conditions. It predominantly contains items at the level of body function and structure according to the International Classification of Functioning, Disability and Health[1] and measures capacity or best ability, rather than actual performance. Randall et al.[2] developed the original Melbourne Assessment following rigorous test development processes based on classical test development principles, as was the norm at the time of publication. Further development of this measure extended the age group to include children aged 2 years 6 months to 15 years and provided valuable information about psychometric properties. The principal value of the Melbourne Assessment has been as an outcome measure and it has attracted the interest of researchers who have added to our knowledge on reliability and validity.[3] The work presented in the current publication[2] used Rasch modelling to evaluate and revise the Melbourne Assessment. Rasch analysis of total scores indicated that the Melbourne Assessment was not unidimensional. Further Rasch analysis of the four subscales, which were compiled as part of the original test development, resulted in deletion of misfitting or redundant items and rescaling of disordered rating scales. Differential item function (DIF) was absent for age and sex, as expected, thus providing evidence for construct validity. DIF did exist between raters, adding to evidence suggesting interrater reliability may be a relative weakness.[3] Randall et al. have responded by providing access to online training and reliability resources for the revised Melbourne Assessment 2.

Revision of the Melbourne Assessment is akin to a double-edged sword. On the downside, the Melbourne Assessment 2 now needs to be evaluated with data collected using the revisions to items and rating scales in order to validate the new version. We also await knowledge regarding sensitivity to change and the amount of change following intervention which would be considered a minimally clinically important. Furthermore, we will need to consider carefully how, and if, we pool data collected with the Melbourne Assessment and Melbourne Assessment 2 in meta-analyses of the effects of interventions. The positive edge is that we have a clear understanding of important features of construct validity of the Melbourne Assessment 2 such as dimensionality and rating scale structure. We know it is inappropriate to use the total score for interpretation and scores derived from individual subscales are more robust. Understanding that the subscales are relatively discrete may assist in planning intervention. The solid basis for the Melbourne Assessment 2 is a launching platform for more research, which will aid our confidence in its use as a discriminative tool and outcome measure.

Decisions as to whether to use the Melbourne Assessment 2 for clinical and research purposes depend on balancing the requirements of the assessment, knowledge of psychometric properties, and clinical utility. The Melbourne Assessment is clearly feasible to use, considering frequency of use in intervention research, despite the need to videotape the assessment and time taken for scoring. The Melbourne Assessment 2 has a solid foundation to its psychometric properties based on Rasch modelling and extrapolating evidence from the existing Melbourne Assessment. The Melbourne Assessment 2 is more useful in outcome evaluation than treatment planning, although understanding that the four subscales are discrete highlights relative areas of strength and weakness to exploit during treatment planning. It is probably the assessment of choice when a measure of unilateral, as opposed to bimanual, upper limb capacity is required. The Melbourne Assessment 2 may be used to assess both upper limbs, but the assessment is then undertaken and scored for each upper limb separately. Evaluation of bimanual performance in naturalistic environments would complement information obtained from the Melbourne Assessment 2. One option is observation of spontaneous use of the affected hand in bimanual activities using the Assisting Hand Assessments (AHA)[4] for children with unilateral disability. The AHA is suitable for children aged 18 months to 12 years. Another option is the parent-rated ABILHAND-Kids[5] for older children aged 6–15 years and which is validated for use with unilateral or bilateral cerebral palsy.


  1. Top of page
  2. References