Stroke is one of the main causes of acquired physical disability in adults [1]. Paresis affecting the upper extremity is a common motor impairment in patients with stoke [2]. According to a previous study, 38% of patients with a paretic upper extremity regained limited dexterity at 6 months after stroke, whereas only 11.6% demonstrated complete functional recovery [3]. Dexterity, i.e., coordination of the fingers and manipulation of the objects, is important for most activities of daily living. The loss of hand dexterity often leads to limitations in activities of daily living, and may decrease the quality of life [4]. Therefore, improvements in dexterity for functional recovery and better quality of life are the main goals of rehabilitation in stroke patients with a paretic hand.

The measurement of dexterity or upper limb function is an important component of rehabilitation programs. A valid outcome measurement leads to more effective rehabilitation by quantifying the severity of disability, assisting in the development of a care plan, comparing the functional status between examination periods, and improving the efficiency of practice [5]. Many outcome measurements have been examined for the evaluation of upper limb function in patients with stroke, including the Jebsen-Taylor Hand Function Test (JTHFT) [6,7,8,9], Fugl-Meyer Assessment (FMA) [10,11,12], Box and Block Test (BBT) [10,11,13,14,15], and Nine-Hole Peg Test [6,11,13,16].

Button fastening requires fine and sequential coordination of the fingers and thumb, which is a skilled motor task that requires little attention, effort, or conscious control [17]. Clawson et al. [18] recorded the time taken to open and fasten six buttons that measured 0.75 inch in diameter and were mounted on a board placed in front of patients with rheumatoid arthritis to evaluate hand disability. This measurement was later modified to a standard board with five buttons by Pincus et al. [19], and these authors found excellent interobserver and intraobserver reliabilities for this test. The button fastening test was performed to evaluate hand function impairment in patients with Parkinson disease who showed a poor performance during a highly skilled movement [17]. This assessment was conducted by recording the time taken to fasten five buttons measuring 2.5 cm in diameter on a cardigan. The modified button fastening test was used by Nitz et al. [20] to assess hand disability in patients with myotonic dystrophy. Patients were instructed to fasten the lower four buttons that measured 6 mm in diameter on a standard shirt. The number of buttons fastened within 25 seconds, which was considered normal performance, was counted. In the aforementioned study, the buttoning test was performed with both hands, however, in a few studies, only the dominant hand was used to perform the test [21]. The buttoning test has been used to assess the hand function of patients with hand disability, but no standardized method has yet been established.

Although various outcome measurements previously mentioned including the JTHFT, highlight a stroke patient's dexterity, these tests require a specific tool and a certain amount of time. However, the buttoning test does not require a specific tool except clothing with buttons, and it can be performed quickly compared with other tests (within 2 minutes). This motor task is acquired in childhood and practiced throughout daily living, often without visual control and attention [17]. Therefore, the buttoning test can be performed by patients without a detailed explanation of the task.

Although this skilled motor task provides important information on hand disability in patients with various diseases, little information is available about the validity of the buttoning test, especially in patients with stroke. There are no standard methods and times required by non-impaired subjects to fasten a button.

The purpose of this study was to compare the relationship between the buttoning test and JTHFT, and to investigate the role of buttoning test as a valid measurement of hand disability in patients with stroke.

The analysis included 151 ischemic stroke patients with the dominant hand affected. The average age of individuals was 61.8±11.5 years, and the average time since stroke was 8.0±15.2 days. The sample had a greater proportion of men than women, 94.7% of them were right-handed, and 92.1% of patients had a lesion in the left hemisphere. Characteristics of patients with the dominant hand affected and patients with the non-dominant hand affected are listed in Table 1. No statistically significant differences between the groups were found for the following factors: age; sex; time since stroke; hand dominance; side of the affected hemisphere; location of the lesion; JTHFT; scores of total, visual, motor arm, and language items from the NIHSS; Motricity Index; and Mini-Mental State Examination-Korean version score.

Our results showed that the button fastening test was negatively correlated with the JTHFT (p<0.001) (Table 2). The correlation between the button fastening time and JTHFT was considered moderate-to-good (r=-0.672). The correlation between the button fastening time and each subtest of the JTHFT was evaluated. A fair-to-good correlation with statistical significance was found between the button fastening time and each subtest of the JTHFT: writing letters (r=-0.590, p<0.001), card turning (r=-0.546, p<0.001), moving small objects (r=-0.574, p<0.001), stimulated feeding (r=-0.601, p<0.001), stacking checkers (r=-0.595, p<0.001), moving large light objects (r=-0.555, p<0.001), and moving large heavy objects (r=-0.470, p<0.002). However, the button opening time was not statistically correlated with the total score and subtests of the JTHFT.

The JTHFT scores among the buttoning test groups are shown in Table 3. Fifty-two of 151 subjects (group A, 34.4%) were unable to fasten the button with their paretic hand. Thirty patients (group C, 19.9%) were able to fasten the button inside the 18-second time constraint. The patients who had difficulty with the buttoning test showed lower mean JTHFT scores (p<0.0001, Jonckheere-Terpstra test). The Kruskal-Wallis test revealed significant differences in JTHFT scores among the three buttoning test groups (p<0.0001). The Mann-Whitney U-test indicated that all pairs of groups were significantly different from each other (p<0.017 level).

The distribution of the JTHFT scores in each buttoning test group is shown in Fig. 1. There were considerable overlaps in JTHFT scores between the buttoning test groups, especially in groups B and C.

There were also significant differences between groups A and B, and between groups A and C in all subtest scores of the JTHFT (p<0.0001). Significant differences were observed between groups B and C in writing letters (p=0.090), moving small objects (p=0.003), and stacking checkers (p=0.001), which involve fine hand motor function (Table 4).

Assessment of the ability to fasten buttons provides the physician or occupational therapist with an overview of a patient's hand function [17,20,25]. Hand disability is a common motor impairment in patients [2]. However, little information is available concerning the validity of the buttoning test for patients with stroke.

Button fastening involves a more skilled, complex motor task than button opening. In this study, the magnitude of correlation between the button fastening test and JTFHT was higher than between the button opening test and JTFHT. We examined for statistical differences in mean JTHFT scores among the button fastening test levels to identify the discriminant validity of the buttoning test as an assessment of hand disability. There were significant differences in JTHFT scores among the three study groups. Results of the Mann-Whitney U-test showed that all pairs of groups (A–B, A–C, and B–C) were significantly different from each other (p<0.017). Moreover, the more difficult it was to fasten the button, the lower the JTHFT score was (p<0.0001, Jonckheere-Terpstra test). These results support the validity of the button fastening test as a hand disability test after stroke.

Nitz et al. [20] reported that the button fastening ability was related to muscle strength of the proximal upper limb in patients with myotonic dystrophy. Therefore, upper limb posture is important to exclude the effect of proximal upper limb weakness and to perform the hand action. All patients sat upright with the shoulder joint adducted, elbow joint flexed, and forearm fully supported on the lapboard. We assessed only one button at the same level of the hand that was placed on the lapboard. In this study, the motor arm item that represented proximal upper limb strength in the NIHSS was not statistically correlated with the button fastening time (r=0.227, p=0.159), although it was correlated with the JTHFT score (r=-0.603, p<0.001). This different result may be attributed to the method of the buttoning test that minimizes the contribution of the proximal upper limb. Our results suggest that the buttoning test can be used as a measure of hand dysfunction rather than the entire upper limb in patients with stroke.

Subtests of the JTHFT requiring fine hand function, such as writing letters, moving small objects, and stacking checkers, showed statistically significant differences among the three buttoning groups. These results were consistent with the need for fine, sequential, and bimanual coordination skills of the fingers and thumb [17]. The button test can be considered as a test that more closely reflects the fine motor function than the gross motor function.

Cognitive dysfunction altered consciousness, severe aphasia, and visual disturbance may have a detrimental effect on functional performance in patients with stroke, and may affect the outcome of hand function measures. However, button testing has been identified as an automatic process that requires little conscious control, attention, and resources [17]. The timed button fastening task was not correlated with the degree of visual, spatial, or cognitive dysfunction [26]. Although our study excluded stroke patients with cognitive impairment, severe aphasia, complete hemianopsia, and inability to follow 2-step commands, further studies that include patients with these impairments may need to reinforce the utility of buttoning test as a measure of hand function in variable degrees of impairment in patients with stroke.

Advantages of the buttoning test are its relative simplicity and the shorter time required for administration compared with other outcome measurements of hand disability. If the goal is to obtain a more comprehensive assessment of dexterity, the JTHFT may be the preferred option. However, only 2 minutes and a hospital gown with buttons are needed to perform the buttoning test. Therefore, this measurement is advantageous in clinical settings such as clinical rounds and in outpatient clinics, where time and resources are limited. In this study, one standardized gown was used; however, under the clinical settings, various sizes of buttons and buttonholes can be used, and the difficulty of button fastening can vary depending on the nature of the buttonhole, which may affect the normality criterion and degree of reflection associated with hand disability. Further studies will be needed to confirm whether the buttoning test using non-standardized gowns is useful as a screening test to identify hand disability.

An easy evaluation of hand disability requires three levels of grading. However, the range of JTHFT scores in all buttoning test groups was wide, with considerable overlaps between the groups, possibly due to the concise grading system of the buttoning test. Additionally, although the buttoning test showed good reliability [19], one trial of button fastening was performed for each measurement causing the overlapped distribution. To avoid these issues, a grading system based on further information of the standard time required by non-impaired subjects to fasten a button and repeated button fastening may be a solution.

When the results of this study are interpreted, some potential limitations warrant consideration. First, our sample size was small (n=30) for group C. The small sample size reduced statistical power in the correlational analyses and validity test. Second, we did not compare the buttoning test with other tests such as the FMA and BBT for evaluating hand disability, and Modified Barthel Index for evaluating activities of daily living in patients with stroke. Third, this study did not include patients with hemorrhagic stroke that accounted for a large number of stroke patients. Forth, there was little follow-up data available and the role of button test in assessing hand function improvement during stroke recovery could not be confirmed.

Despite the findings supporting the merits of buttoning test, a further study involving a larger sample with a broader range of stroke-related impairments in dexterity should include patients with hemorrhagic stroke, and non-impaired subjects. A longitudinal follow-up evaluation is also needed.

In summary, the buttoning test was selected as an outcome measurement of hand disability in patients with stroke. This test was significantly and negatively correlated with the JTHFT and represents a valid measurement tool. Furthermore, the results of this study suggest that the buttoning test was more appropriate for the evaluation of fine hand motor function. The 2-minute time required to perform this assessment may be of considerable value in a clinical setting. This study did not account for the overlapped distribution between the buttoning groups and precise grading references, suggesting the need for a further study with a larger sample to validate the finding.