Fugl-Meyer Assessment of Motor Recovery after Stroke

Stroke: 

(Sanford et al, 1993; n = 12; mean age = 66 years; stroke onset < 6 months, Acute Stroke)

• FMA total scores: 9.4

Stroke: 

(Wagner et al, 2008, n = 14, mean age = 59.9 (14.6) years, assessed on average 14 (6.5) months post stroke, Chronic Stroke)

• FMA = 5.2 points for the Upper Extremity portion of the assessment

Stroke: 

(Shelton et al, 2001; n = 171; mean age 70 (11) years; assessed within 17 (12) days of stroke, Acute Stroke)

FMA Motor Scores from Admission to Discharge 

• 10 point increase in FMA Upper Extremity = 1.5 change in discharge FIM 
• 10 point increase in FMA Lower Extremity = 1.9 change in discharge FIM

(Page SJ et al, 2012; n = 146; mean age 57 (11) years; chronic ischemic stroke, minimum-to-moderate UE hemiparesis)

Stroke: 

(Duncan et al, 2000; n = 459; mean age = 70 (11.4) years; stroke onset within 14 days, Acute Stroke)

Stroke:

(Duncan et al, 2000, Acute Stroke)

Stroke:

(Duncan et al, 1983; n = 19; mean age = 56 (13) years; same PT rating on 3 occasions each 3 weeks apart; VA sample, Chronic Stroke)

Sullivan et al, 2011; n=15 patients (10-71 days post-stroke; mean age 62.8 yrs. One expert PT rater compared to 17 trained PT's using standardized procedures developed for LEAPS trial.

Stroke:

(Lin et al, 2004; n = 176; mean age = 67.9 (10.9) years; assessed 14, 30, 90 and 180 days after stroke, Acute Stroke)

• Excellent internal consistency: alpha = 0.94 to 0.98 across 4 administrations

Stroke:

(Malouin et al, 1994; n = 32; mean age = 60; mean time since stroke = 64.5 days, Acute Stroke)

• Excellent FMA & Motor Assessment Scale (MAS) total score correlations (r = 0.96)  
• Poor FMA & MAS sitting balance item correlations (r = -0.10) 
• Motor and sensory FMA scores 5 days post-stroke were the strongest predictor of motor recovery 6 months post-stroke (Duncan et al, 1992)

Gait Speed (Nadeau et al, 1999; n = 16; mean age = 47.9 (15.6) years; mean number of months post-stroke = 43.9, Chronic Stroke)

• Poor correlations between FMA Sensation and Gait speed (r = 0.05) and comfort (r = 0.14)
• Excellent correlations between FMA total motor scores and gait speed (m = 1.09 meter per second; r = 0.61); and comfort (m = 0.76 meters per second; r = 0.61)

(Page SJ et al, 2015; n=32, mean age=57(10) years, chronic stroke patients with moderately impaired UE function)

• Excellent correlations between Wrist and Hand Upper Extremity Fugl-Meyer (w/h UE FM) and the Action Research Arm Test across two trials: (r = 0.74) and (r = 0.67)

See et al, 2013; n=12 .These exams were performed on 12 patients (of the cohort of 31 described under Reliability testing above), 4 separate visits across a treatment period, for a total of 48 exams. Testing was conducted on the UE FMA only. Performance for the total UE FMA, proximal arm subsection, and distal arm (wrist/hand) were examined separately.

ARAT = Action Research Arm Test; SIS = Stroke Impact Scale. Baseline values are mean ± SD except for Box & Blocks and 9-hole peg scores, which are median. Correlation values are Spearman r values which in all cases were significant with p < 0.0001.

(Wei et al, 2011; n=27 with chronic stroke (4.92±0.45 yrs post-stroke)

Excellent correlations were observed both pre-training and post-training among the Upper Extremity Fugl-Meyer Assessment , the Motor Status Scale and the Action Research Arm Test (ρ=0.91-0.93). Modified Ashworth Scale and the UE FMA were fairly to moderately correlated (ρ=0.40-0.62)

Acute Stroke:

• Excellent correlation: modified Balance Subscale on FMA and the Barthel Index; r = 0.86 - 0.89 (Mao et al, 2002)
• Excellent correlation: FMA and Functional Independence Measures (FIM) administered to 172 inpatients who had recently had a stoke; r = 0.63 (Shelton et al, 2000)
• FMA effectively distinguished between three levels of self care (Independent, Partly Dependant, and Dependant) in a sample of 109 recent (< 90 days) stroke survivors (Bernspang et al, 1987).
• FMA was a better measure of higher-level recovery than the MAS (Malouin, et al, 1994)

Chronic Stroke

 (Dettmann et al, 1987; n = 15; mean age = 64 years; mean time since stroke, 2 years, Chronic Stroke)

• The FMA and the Barthel Index were used to assess a group of 15 participants at an average of 2 years post stroke.  Correlations between the measures were excellent (r = 0.67). The strongest correlations were observed in the Balance subscore (r = 0.76) the Upper Extremitysubscore of the motor domain (r = 0.75) and FMA Motor total score (r = 0.74) 

(Hsieh et al, 2009; onset > 6 months, Chronic Stroke)

FMA Construct Validity (Spearman's rho followed by 95% CI)

• Excellent pre-treatment correlation between FMA and Action Research Arm Test: 0.73** (0.58, 0.83)
• Excellent pre-treatment correlation between FMA and Wolf Motor Function Test-TIME: 0.76** (0.63, 0.86)
• Excellent pre-treatment correlation between FMA and Wolf Motor Function Test-Functional Ability Scale: 0.71** (0.56, 0.82)
• Adequate pre-treatment correlation between the FMA and FIM-motor: 0.49** (0.27, 0.66)

*P < 0.05
**P < 0.01

• Excellent post-treatment correlation between FMA and Action Research Arm Test: 0.74** (0.60, 0.84)
• Excellent post -treatment correlation between FMA and Wolf Motor Function Test-TIME: 0.71** (0.56, 0.82)
• Excellent post -treatment correlation between FMA and Wolf Motor Function Test-Functional Ability Scale: 0.51** (0.29, 0.68)
• Adequate post -treatment correlation between the FMA and FIM-motor: 0.42** (0.18, 0.62)

*P < 0.05
**P < 0.01

Stroke:

(Woodbury et al, 2008; n = 377; men age = 69.2 (11.2) years, Acute Stroke) Upper extremity (modified version, 3 reflex items removed) 

• Administered at admission and 6-months post stroke
• Rasch Analysis demonstrated adequate fit for each of the 30 items except the hook grasp item
• Results suggest items contained in modified FMA, (except the hook grasp item) were assessing the same underlying construct

(Crow et al, 2008; n = 62; retrospective analysis, Chronic Stroke)  Upper and lower-extremity (excluding balance) portions of FMA: 

• Items within a scale are valid, unidimensional and cumulative
• Results suggest that if a patient is able to successfully complete an item of a certain difficulty, that same patient should be able to complete less difficult items,  implying that a shortened administration of the FMA may produce valid results

Stroke:

(Lin et al, 2004, Acute Stroke)

• Ceiling effects have been observed with the Sensation subscore. The following percentages of acute stroke patients reaching the highest possible score:

• Floor effects have been reported for the modified Balance domain.  The strongest effects were observed 14 days post stroke with 29.3% of participants not able to achieve the lowest possible score on the measure (Mao et al, 2002)Possible ceiling effects on hand and lower extremity items (Gladstone, et al, 2002).

Stroke:

(Mao et al, 2002, Acute Stroke)

• Excellent on the modified version of the FMA Balance score
  • Between assessments at 14, 30, 90 and 180 days post-stroke
  • Responsiveness decreased as the time between stroke and assessments increased

Moderate to Low: (Lin et al, 2004) responsiveness was found for the Sensation subscale of the FMA as assessed by Standardized Response Means (SRM). 

• Means were low at:
  • 14-30 days: SRM = 0.42
  • 30-90 days: SRM = 0.43
  • 90-180 days: SRM = 0.27  
• Moderate responsiveness when assessed from 14 to 180 days (SRM = 0.67)

(Hsueh et al, 2009, Chronic Stroke)

• Small to moderate  effect sizes were observed on the FMA, the Stroke Rehabilitation Assessment of Movement instrument (STREAM) and each of the measures shortened versions.  
• Moderate effect sizes on the shortened version of both measures (0.53 and 0.51)
• Small effect sizes on the long version of the measure  (0.045 and 0.38)

(Wei et al, 2011; n=27 with chronic stroke (4.92±0.45 yrs post-stroke

Pre- and post-training which consisted of 20 sessions of UE robotic training that could be completed in 4-7 weeks.

*GRI has no recommended value. Bigger GRI values mean larger effect size and more responsiveness. SRM = Standardized Response Mean; GRI= Guyatt’s Responsiveness Index[

General Rehab Sample: 

(Platz et al, 2005; n = 37 stroke, 14 MS, and 5 TBI patients; assessed twice within a 7 day interval, General Rehab Sample)

• Excellent Total Motor Score (ICC = 0.97)
• Excellent Sensation (ICC = 0.81)
• Excellent Passive Joint motion (ICC = 0.95)

Excellent Interrater Reliability:

(Duncan et al, 1983; 4 PT's using the above sample)

Excellent interrater reliability, FMA Total Score: (ICC = 0.96)(Sanford et al, 1993; n = 12; mean age = 66 (11.5) years;  0 to 6 months post stroke, Acute Stroke)

Wrist and Hand Upper Extremity Fugl-Meyer (w/h UE FM):

Study of the wrist stability and mobility hand scales (the 12 most distal items of the UE FM).

(Page SJ et al, 2015; n=32, mean age=57(10) years, chronic stroke patients with moderately impaired UE function)

• Excellent intrarater reliability (ICC = 0.95)

Subsection to FMA Total Score Correlations:

(Wood-Dauphinee et al, 1990; n = 167; assessed at admission and 5 weeks)

• Excellent correlation : Upper Extremity & FMA Total (r = 0.97)
• Excellent correlation: Mower Extremity & FMA Total (r = 0.90)
• Adequate correlation: Balance & FMA Total (r = 0.88)

(Page SJ et al, 2015; n=32, mean age=57(10) years, chronic stroke patients with moderately impaired UE function)

• Excellent internal consistency for the Wrist and Hand Upper Extremity Fugl-Meyer (w/h UE FM) across two trials, Cronbach's alpha = 0.82 and 0.84)

Gladstone et al, 2002: 

• Face and content validity for the motor domain are "very good"
• Scaling is heavily weighted for the upper extremity 
• Reflexes may be overrepresented in the scoring system