A systematic review of the psychometric properties of the Action Research Arm Test in neurorehabilitation.

Abstract

The Action Research Arm Test (ARAT) measures upper limb activity limitations in people with acquired brain injuries. Evidence relating to the use of this test in neurorehabilitation is scattered. This review identifies, rates and synthesises evidence on the original 1981 ARAT use within neurorehabilitation. Psychometric properties are reviewed, including specific examination of participants with upper limb spasticity. Systematic review of published articles describing psychometric properties and/or use of the original version of the ARAT in neurorehabilitation. COnsensus-based Standards for the selection of health Measurement INstruments (COSMIN) search strategy, reporting and methodological checklist with criterion-based appraisal of quality criteria for good measurement properties were applied. A best evidence synthesis for each psychometric property was completed. In 28 included studies, participants had suffered a stroke or traumatic brain injury, with 46% >6months post-injury. Six studies identified participants with upper limb spasticity. Methodological quality of psychometric properties ranged from poor to excellent. Best evidence synthesis determined moderate positive evidence for using the ARAT with people without limb spasticity: intra-rater reliability (ICC 0.71 (95% CI 0.53-0.89) to 0.99 (95% CI 0.98, 0.99)); responsiveness (ROC curve 0.72-0.88, SRM 0.89); and regarding construct validity, it is a valid measure of activity limitation. Limited evidence for psychometric properties of the ARAT were found when used with people with upper limb spasticity for construct validity and responsiveness (ES 0.55-0.78). Gaps in evidence were found for inter and test-retest reliability, measurement error, content validity, structural validity, floor and ceiling effects. The ARAT is an appropriate measure of activity limitation post-stroke and should be considered for use with people with TBI; evidence for people with upper limb spasticity is limited. Gaps and mixed limited to moderate evidence for psychometric properties in neurorehabilitation mean further research is required.