Fig 1.
Patient selection process. Diagram showing the selection process of participants through the study. After allocation, 33 and 31 participants were analyzed for reliability and validity testing, respectively.
Table 1.
Clinical characteristics of the analyzed patients.
Table 2.
Raw data for Action Research Arm Test total scores and subscale scores.
Table 3.
Reliability of the total score and the subscale scores of the Action Research Arm Test.
Table 4.
Intra-rater reliability of ARAT subscales assessed using weighted kappa coefficients.
Table 5.
Inter-rater reliability of ARAT subscales assessed using weighted kappa coefficients.
Table 6.
Bland–Altman analysis of total and subscale scores between the new and original Action Research Arm Test.
Fig 2.
Bland–Altman plots for intra-rater agreement of the Action Research Arm Test (ARAT) subscales and total score. The Bland–Altman plots show (a) ARAT grasp score, (b) ARAT grip score, (c) ARAT pinch score, (d) ARAT gross movement score, and (e) ARAT total score. The solid blue line indicates bias, and the red dashed lines indicate the upper and lower LOA, which were drawn using exact calculated values before rounding. A systematic error is suggested when the bias shifts upward or downward, and a proportional bias is suggested when the difference changes linearly with the mean. Points outside the LOA indicate outliers. Agreement is considered clinically acceptable when most points lie within the LOA, and the LOA width falls within the acceptable range. ARAT, Action Research Arm Test; LOA, Limits of Agreement.
Fig 3.
Bland–Altman plots for inter-rater agreement of the Action Research Arm Test subscales and total score. The Bland–Altman plots show (a) ARAT grasp score, (b) ARAT grip score, (c) ARAT pinch score, (d) ARAT gross movement score, and (e) ARAT total score. The solid blue line indicates bias, and the red dashed lines indicate the upper and lower LOA, which were drawn using exact calculated values before rounding. A systematic error is suggested when the bias shifts upward or downward, and a proportional bias is suggested when the difference changes linearly with the mean. Points outside the LOA indicate outliers. Agreement is considered clinically acceptable when most points lie within the LOA, and the LOA width falls within the acceptable range. ARAT, Action Research Arm Test; LOA, Limits of Agreement.
Fig 4.
Correlation between the Action Research Arm Test and upper limb function assessment. The scatter plot shows the relationship between the ARAT total score and (a) the FMA-UE total score, (b) the BBT on the paretic side, and (c) grip strength on the paretic side. The strength of the correlation was interpreted based on the absolute value of the correlation coefficient as follows: < 0.4 = weak, 0.4–0.74 = moderate, 0.75–0.9 = strong, and ≥0.9 = very strong. ARAT, Action Research Arm Test; BBT, Box and Block Test; FMA-UE, Fugl-Meyer Assessment of the Upper Extremity.
Fig 5.
Correlation between the Action Research Arm Test and self-reported outcomes. The scatter plot shows the relationship between the ARAT total score and (a) MAL Amount of Use, (b) MAL Quality of Movement, (c) JASMID Quantity, and (d) JASMID Quality. The strength of the correlation was interpreted based on the absolute value of the correlation coefficient as follows: < 0.4 = weak, 0.4–0.74 = moderate, 0.75–0.9 = strong, and ≥0.9 = very strong. ARAT, Action Research Arm Test; JASMID, Jikei Assessment Scale for Motor Impairment in Daily Living; MAL, Motor Activity Log.