Inter-rater reliability and validity of the Action Research arm test in stroke patients

Age and Ageing, March, 1998 by Ching-Lin Hsieh, I-Ping Hsueh, Fu-Mei Chiang, Po-Hsin Lin

The results showed a strong relation between performance on the ARAT and the UEMAS (r = 0.96). The extreme association between the ARAT and the UEMAS implies a similarity in the construct (i.e. arm disability) being evaluated. The results support the validity of the ARAT as a measure of upper extremity function in patients who have a stroke. The high r-value does not indicate agreement. Exact agreement between scores on the ARAT and the UEMAS would not be expected because of the different methods employed in scoring the two assessments.

The results also showed that the scores of the ARAT were highly associated with those of the AMI and the UEMMAC (impairment measurements). This finding of the present study is similar to that of DeWeerdt and Harrison's study [7], in that the scores of the ARAT were closely correlated with those of the Fugl-Meyer assessment (a test of impairment) in stroke patients. The results indicate that the scores of the ARAT may reflect not only arm function but also upper extremity motor impairment that represents the exteriorization of neurophysiological states due to cerebrovascular diseases. Thus, the scores of the ARAT may also represent the degree of upper extremity motor impairment.

The ARAT is designed for evaluation of both sides of patients with cortical injuries, helping to obtain a more total description of the upper extremity function than investigation of only the hemiplegic side. From the point of view of dependency it is necessary to know whether the patient has unlimited function on the non-affected side. In particular, some patients (e.g. individuals with brain stem lesions) have both sides affected, although generally one side to a lesser extent than the other. Furthermore, a number of studies have reported that the ipsilateral non-affected side of patients with a single focal hemispheric infarct showed slowed sensory-motor responses [26- 28].

The evaluation of the non-paretic side is not time-consuming because of the hierarchical design of the ARAT. In fact, the evaluation of the non-affected side can also serve as a kind of demonstration in which the rater will determine whether the patient understands his/her commands. The evaluation will be hence performed more smoothly.

Other commonly used instruments for assessing arm-hand function in stroke patients which might be used instead of the ARAT include: the grip strength test [29], the nine-hole peg test [30] and the Frenchay arm test [31]. Although grip strength is probably a sensitive measure of recovery from stroke [17], the test measures distal strength rather than proximal strength and dexterity as in the ARAT. The nine-hole peg test mainly focuses on finger dexterity, but it cannot detect deficits of proximal strength and is not useful when impairment is severe, especially when the patient's upper extremity motor function is limited [30]. The Frenchay arm test assesses proximal control and dexterity. It contains only five subtests and is simple and quick to administer. Patients tend to either pass or fail all subtests [30], which suggests this test might not be sensitive enough to distinguish patients with minor difference in motor control and dexterity. The ARAT assesses not only proximal and distal strength but also dexterity. It is easy and quick to use. Therefore, the ARAT may be preferable among the instruments which are used to evaluate arm-hand function in stroke patients. Although it requires construction of a special table, the materials needed for the ARAT are not expensive or difficult to obtain.