A cohort study investigating a simple, early assessment to predict upper extremity function after stroke - a part of the SALGOT study

BACKGROUND: For early prediction of upper extremity function, there is a need for short clinical measurements suitable for acute settings. Previous studies demonstrate correct prediction of function, but have ether included a complex assessment procedure or have an outcome that does not automatically correspond to motor function required to be useful in daily activity. The purpose of this study was to investigate whether a sub-set of items from the Action Research Arm Test (ARAT) at 3 days and 1 month post-stroke could predict the level of upper extremity motor function required for a drinking task at three later stages during the first year post-stroke. METHODS: The level of motor function required for a drinking task was identified with the Fugl-Meyer Assessment for Upper Extremity (FMA-UE). A structured process was used to select ARAT items not requiring special equipment and to find a cut-off level of the items’ sum score. The early prognostic values of the selected items, aimed to determine the level of motor function required for a drinking task at 10 days and 1 and 12 months, were investigated in a cohort of 112 patients. The patients had a first time stroke and impaired upper extremity function at day 3 after stroke onset, were ≥18 years and received care in a stroke unit. RESULTS: Two items, “Pour water from glass to glass” and “Place hand on top of head”, called ARAT-2, met the requirements to predict upper extremity motor function. ARAT-2 is a sum score (0-6) with a cut-off at 2 points, where >2 is considered an improvement. At the different time points, the sensitivity varied between 98 % and 100 %, specificity between 73 % and 94 %. Correctly classified patients varied between 81 % and 96 %. CONCLUSIONS: Using ARAT-2, 3 days post-stroke could predict the level of motor function (assessed with FMA-UE) required for a drinking task during the first year after a stroke. ARAT-2 demonstrates high predictive values, is easily performed and has the potential to be clinically feasible. TRAIL REGISTRATION: ClinicalTrials.gov: NCT01115348