Archives of Physical Medicine and Rehabilitation

journal homepage: www.archives-pmr.org

Archives of Physical Medicine and Rehabilitation 2023;104: 1596−605

ORIGINAL RESEARCH

Reliability of the Modified Frenchay Scale for the

Assessment of Upper Limb Function in Adults
With Hemiparesis
Zo
e Laclergue, MD,a Mouna Gh
edira, PT, PhD,b Caroline Gault-Colas, MD,a
Laurene Billy, OT,c Jean-Michel Gracies, MD, PhD,a,b Marjolaine Baude, MDa,b
From the aService de Re
e
ducation Neurolocomotrice, Ho^pitaux Universitaires Henri Mondor, Assistance Publique-Ho^pitaux de Paris, Cre
teil,
France; bUR 7377 BIOTN, Universite
Paris Est Cre
teil (UPEC), Cre
teil, France; and cPo^le de Me
decine Physique et Re
adaptation, Fondation
Mallet, Richebourg, France.

Abstract
Objectives: To investigate the reliability of the Modified Frenchay Scale (MFS) in adults with hemiparesis.
Design: Prospective analysis of videos.
Setting: Study conducted in a Neurorehabilitation Unit of a University Hospital.
Participants: Fifty-one patients (17 women [33%], age 46§15, time since injury 5.2§6.7 years) with hemiparesis secondary to stroke (N=47),
tumor (N=3), or spinal cord injury (N=1) were enrolled.
Intervention: The MFS measures active upper limb function in spastic hemiparesis based on a video recording of 10 daily living tasks, each rated
from 0 to 10. Six tasks are bimanual and 4 are unimanual with the paretic hand. MFS videos performed in routine care of patients with hemiparesis
between 2015 and 2021 were collected. After a 3-hour group training session, each MFS video was assessed twice, 1 week apart by 4 rehabilitation
professionals with various levels of experience in using the scale.
Main Outcome Measures: Internal consistency was determined using Cronbach’s alpha. Intra- and inter-rater reliability was measured using
intraclass correlation coefficients (ICC, mean [95% CI]), mean differences between ratings and minimal detectable change (MDC). Bland-Altman
plots were also performed for inter-rater assessments.
Results: The mean overall MFS score was 4.95§1.20 with no floor or ceiling effect. Cronbach’s a was 0.97. For the overall MFS score, intra- and
inter-rater ICCs were 0.99[0.99;1.00] and 0.97[0.95;0.98], respectively; mean intra- and inter-rater differences were 0.10§0.04 and 0.24§0.12,
respectively; and MDC were 0.17 and 0.37, respectively.
Conclusions: The MFS is an internally consistent and reliable scale to assess upper limb function in adults with hemiparesis.
Archives of Physical Medicine and Rehabilitation 2023;104:1596−605
Ó 2023 by the American Congress of Rehabilitation Medicine. Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/)

Spastic paresis is a syndrome caused by a lesion involving the cen- months),5,6 which represents a major challenge for the rehabilita-
tral motor pathways that process the execution of voluntary motor tion community.
command.1,2 The main cause is stroke, accounting for more than Meaningful assessment of motor function in the upper extrem-
13.7 million new cases and 80.1 million (74.1-86.3) prevalent ity remains difficult. While a number of scales have been devel-
cases worldwide in 2016.3 Approximately 70% of stroke survivors oped, none has achieved consensus and each presents with
experience an initial motor deficit, with recovery of walking drawbacks such as the lack of procedure and rater training stan-
proving possible in most cases.4 In contrast, only 10%-20% dardization.7 Currently, the most widely used upper limb motor
recover normal upper limb function in the chronic phase (>6 scale in clinical research is the Fugl-Meyer Assessment.8,9 How-
ever, this evaluation assesses upper limb motor impairment rather
than function, and its time consumption makes it cumbersome in
Disclosures: None.
clinical routine. Other more targeted measures, such as the Box

0003-9993/$36 - see front matter Ó 2023 by the American Congress of Rehabilitation Medicine. Published by Elsevier Inc. This is an open access article
under the CC BY license (http://creativecommons.org/licenses/by/4.0/)
https://doi.org/10.1016/j.apmr.2023.04.003
Reliability of Modified Frenchay Scale in hemiparesis 1597

and Block test,10 the Nine Hole Peg Test,11 or the Purdue Peg- to participate in the study and (2) cognitive impairment likely to
board Test12 also fall short of testing daily living tasks, in addition interfere with the ability to understand and follow verbal instruc-
to focusing mostly on distal dexterity rather than on overall upper tions to complete the assessments.
limb function. Other scales test performance closer to daily living
tasks, such as the Action Research Arm Test13 and the Wolf Motor Raters
Function Test14; however, these are not easily applicable in rou- All raters were rehabilitation professionals involved in the evalua-
tine practice because they require expensive equipment or are tion and rehabilitation of the upper limb of hemiparetic patients.
time consuming. Finally, assessments such as the Motor Activity They underwent a 3-hour group training session prior to the begin-
Log,15 ABILHAND,16 and the more recently validated DextQ- ning of the study, involving rating and discussion of 6 MFS videos
2417 are merely questionnaires on perceived function without together as a group.
actually testing active tasks, making them prone to capture com-
pensations rather than true upper limb function.
Among all existing scales of actual upper limb function, the
Procedure
Frenchay Arm Test18 assesses upper limb function on 7 daily tasks
(including 5 unimanual tasks), while combining the advantages of Study design
being simple, fast, low cost and representative of everyday life. We conducted a study of patients with spastic hemiparesis who
However, the original Frenchay Arm Test is hampered by major were assessed using the MFS in a Neurorehabilitation Unit of a
limitations that were noted by the authors themselves, as the test University Hospital. Any subject examined using the MFS
relies on a binary pass/fail rating with poor sensitivity to between January 2015 and January 2021 was included in the study
change19,20 and propensity to floor or ceiling effects. In addition, if he/she met the inclusion criteria and if 1 MFS video of the sub-
the Frenchay Arm Test may overemphasize unimanual tasks, con- ject could be identified. Demographic and clinical characteristics
trasting with the high number of bimanual activities usually were collected from medical records and each MFS video was
required in real life. rated twice a week apart by 4 rehabilitation professionals with var-
The Modified Frenchay Scale (MFS), based on the original ious levels of experience in using the scale. Each of the 4 investi-
Frenchay Arm Test, was first published in 2002.21 Three key mod- gators was blinded as to the assessments of others.
ifications from the original Frenchay Arm Test were proposed: (1)
increase in the number of tasks from 7 to 10, including 6 bimanual Modified Frenchay Scale
and 4 unimanual tasks with the paretic hand, to better reflect the The MFS consists in 10 daily life tasks to be performed by the
use of the paretic upper limb in daily life; (2) conversion from the patient using standardized daily life tools, arranged along a semi-
original binary rating to a quantitative rating ranging from 0 (no circle shape at arm length in front of the patient. The clinician
movement) to 10 (normal performance) in 0.5 steps, to improve asks the patient to perform the 6 bimanual and 4 unimanual tasks
sensitivity and minimize the risk of floor or ceiling effects; (3) sys- of the scale. Tasks are completed always in the same order, from
tematic videotaping of the performance as an intrinsic part of the left to right; unimanual tasks must be performed with the paretic
scale, so as to keep documents in the electronic chart and to moni- hand only. Each task is rated from 0 (no movement) to 10 (normal)
tor progress. Today, this scale is part of the Five-Step Assessment in half-point steps, with 5 corresponding to a task barely accom-
(FSA) in spastic paresis.22 plished. The overall score is the average of all 10 scores. The scale
The present study focused on investigating the reliability of the is systematically video-recorded. The patient is encouraged to
MFS in adult patients. self-stretch any overactive upper limb muscles between 2 tasks, to
transiently reduce spastic dystonia and spastic cocontraction.
To optimize reproducibility, the complete protocol with equip-
ment, instructions, and scoring aid is available in appendix 1, and
Methods the scoring sheet with task titles is displayed in appendix 2. The
MFS is freely available on the public website https://neuroloco.
This study followed guidelines for reporting reliability and agree-
ment studies.23 wixsite.com/mfs-scale, where some videos are also available for
demonstration. The scale has been initially created in English and
later translated into French (see website). The total cost of the
Participants daily life utensils used should not exceed 50 euros.

Patients
Sample size
Patients were followed for hemiparesis as outpatients or inpatients
at the time of enrolment in the protocol. Inclusion criteria were (1) Based on the COSMIN checklist24 and prior literature25-27 on reli-
age >18, (2) subacute (15 days-6 months) or chronic (>6 months) ability studies of upper limb motor scales, the sample size was
hemiparesis with upper limb involvement, secondary to vascular prospectively estimated to require at least 50 patient videos evalu-
or traumatic brain lesion or non-evolutive brain tumor, diagnosed ated by the 4 raters.
by a neurologist. Exclusion criteria were as follows: (1) no consent
Ethics
List of abbreviations:
This study was conducted in accordance with the ethical principles
ICC intraclass correlation coefficient
LOA limits of agreement of the Declaration of Helsinki. The study protocol was reviewed
MDC minimal detectable change and approved by the local Institutional Review Board at Henri
MFS Modified Frenchay Scale Mondor University Hospitals on October 1st, 2020. All patients
were informed about the study and attested in writing to their non-

www.archives-pmr.org
1598 Z. Laclergue et al

opposition to the use of their data for this research after reading Mean absolute differences
the study information note provided by the study physician. For clinical relevance, we determined mean absolute intra- and
inter-rater differences between ratings.
Statistical analysis
Standard error of measurement and minimal detectable change
The main objective of the study was to evaluate the reliability of The measurement error was determined by calculating the stan-
the MFS, for each task and for the overall score (average of the 10 dard error of measurement (SEM=SDdifference/x2). It provides
tasks). We chose to define the MFS scoring variables as quantita- information on the accuracy of measurement. Based on the SEM,
tive by considering that the 0.5 interval between each step was of the smallest real difference not due to random measurement error
steady clinical dimension, with 21 possible graduations between 0 was determined by calculating the minimal detectable change
and 10, which behaved in a linear way. We used descriptive statis- (MDC=SEM £ 1.96 £ x2), with a confidence level of 95%.
tics (mean, SD and frequency) to analyze the characteristics of the MDC corresponds to the minimum amount of change outside of
population as well as score distribution of the MFS. Floor and error that reflects true change between 2 measures rather than a
ceiling effects are defined as the proportion of patients scoring the variation in measurement.
lowest (floor) or highest (ceiling) possible score and a threshold of
15% was considered significant for the determination of overall Bland-Altman plot and 95% limits of agreement
MFS score floor or ceiling effects. Bland-Altman plots and 95% limits of agreement (LOA) were
used to visualize the agreement between assessments from various
raters. In the plot, the absolute difference between assessments of
Internal consistency the raters for the overall MFS score was plotted against the mean
Internal consistency refers to the general agreement between the scores for each patient. Assuming that differences follow a normal
multiple items that make up a composite score for a given concept, distribution, the LOA represents 95% of the differences between
here the upper limb function. This agreement is generally mea- the mean difference§1.96 £ SD of the differences.
sured by the correlation between items, so we first measured the Data were analyzed using Stata software 15.0 (Stata Statistical
correlation between each task independently using Pearson corre- Software: Release 15.0 College Station, TX: StataCorp LP).
lation coefficients, and then designed a correlation matrix. In this
matrix, each correlation is represented by a color with a gradient
from purple to yellow depending on the correlation coefficient’s Results
value. We also calculated Cronbach’s alpha coefficients.28 Internal
consistency was considered good if a was above 0.80 and excel-
lent if a was above 0.90. Participants
Based on the inclusion and exclusion criteria, 51 patients (women
Intra- and inter-rater reliability 17 [33%]; mean age 46§15 years) with hemiparesis secondary to
single or multiple strokes (ischemic, n=33; hemorrhagic, n=14),
Intra- and inter-rater reliability was assessed to examine the repro- tumor (n=3), or spinal cord injury (n=1) were included. The mean
ducibility between repeated measurements obtained from the time since the injury was 5.2§6.7 years. Regarding raters experi-
same patient by different raters or by the same rater at different ence, Rater-1 was a Physical Medicine and Rehabilitation special-
times. ist with more than 10 years of experience in MFS rating; Rater-2
was a Physical Medicine and Rehabilitation resident with 2 years
Intraclass correlation coefficient of experience in Neurorehabilitation, who started using the MFS
We conducted intraclass correlation coefficients (ICCs) with 95% scale 2 months only before beginning the study; Rater-3 was a
confidence interval [CI] to examine the extent of agreement physiotherapist with 9 years of experience in MFS rating; and
between intra- and inter-rater assessments in our study. If we had Rater-4 was an occupational therapist practicing the MFS for
considered variables as ordinal or categorical, we would have 5 years in another center. Table 1 summarizes the characteristics
used weighted Fleiss’ kappa but as Fleiss and Cohen showed in of patients and raters.
1973,29 ICCs are more relevant for scales with 5 or more catego-
ries (21 here) and we thus have considered MFS rating values as Description and score distribution
quantitative variables. Intra-rater absolute agreement ICCs were
computed using a 2-way random-effects model, specifying Figure 1 displays the mean scores obtained on each task of the
together the rating, the target and the judge. Both individual (to scale and the overall MFS score, which was 4.9§1.2, for a
study each of the 4 intra-rater reliabilities) and average ICCs (to [2.3;8.4] range. As expected, unimanual tasks were characterized
study the mean of the 4 intra-rater reliabilities) were analyzed. For by lower scores (mostly <5) than bimanual tasks but there was no
inter-rater measurements, the same model was used, but only aver- overall significant difference in mean values among tasks. As for
age ICCs were taken into account (to study the mean inter-rater the distribution of the overall MFS score, no floor (0%) or ceiling
reliability at each visit). The assumption was that raters were ran- (0%) effect was found for each rater.
domly selected among rehabilitation professionals and that the
results could be generalized to any similar raters in the field. Reli- Internal consistency
ability was considered excellent for ICCs greater than 0.90; good
between 0.76 and 0.90; moderate between 0.51 and 0.75; and poor Figure 2 represents the correlation matrix between tasks. The
below 0.50.30 The 95% confidence interval was calculated for range of values of correlation coefficients (showed by the color
each ICC to take sampling variations into account. gradient on the right of the figure) was between 0.55 and 0.96.

www.archives-pmr.org
Reliability of Modified Frenchay Scale in hemiparesis 1599

Table 1 Demographic characteristics of patients and raters ICC was 0.97[0.95-0.98] for the overall MFS score, indicating
excellent inter-rater reliability. The mean inter-rater difference
Characteristics Values*
for the overall MFS score was 0.24§0.12, ranging from 0.34§
Patients (n=51) 0.25 to 0.58§0.27 for individual tasks (table 3). Greater inter-
Age, years 46 (15) rater differences were observed for the broom and the clothes-
Sex, n women 17 (33) pins tasks.
Right-handed 44 (86) The inter-rater SEM and MDC for the overall MFS score were
Side of hemiparesis, n right 25 (49) 0.13 and 0.37, respectively.
Time since lesion (years) 5.2 (6.7), Using Bland and Altman statistics, the mean inter-rater differ-
median (IQR) 3.3 (1.1-7.7) ence between 2 overall scores was similar at 0.23 with LOA of
MFS overall score 4.9 (1.2) [-0.15;0.61] (fig 3), indicating that 95% of the differences in over-
Cause of hemiparesis all MFS scores between raters were between -0.15 and 0.61. Some
Ischemic stroke 33 (65) patients were characterized by differences between 0.61 and 0.80
Hemorrhagic stroke 14 (27) outside the LOA, indicating that the random measurement error
Tumor 3 (6) may reach 0.80/10 when the MFS is administered by different
Spinal cord injury 1 (2) raters.
Vascular risk factors
Hypertension 23 (45)
Dyslipidemia 13 (25)
Reliability and time to complete the MFS as a
Smoking 16 (31)
Alcohol 7 (14)
function of the overall MFS score
Diabetes 3 (6) Reliability may be optimal in the low and mid-range of MFS
Atrial fibrillation 4 (8) scores (fig 3). Mean time taken to complete the scale was 10§3
Raters (n=4) minutes. Relation between the overall MFS score and the time to
Age, years 36.8 (12) complete all tasks is displayed in Supplemental Figure 1. Time
Sex, n women 4 (100) was longest for patients with moderate overall MFS score
Experience in the rehabilitation of 12.3 (9.5) (between 4 and 5.5).
spastic paresis (years)
MFS experience (years) 7 (5.4)
Profession
PM&R specialists 2 (50)
Physiotherapist 1 (25)
Discussion
Occupational therapist 1 (25) This reliability study of the MFS to evaluate active upper limb
* Values indicated are mean (SD) for quantitative variables and the function in spastic paresis yielded excellent intra- and inter-
number with percentage n (%) for qualitative variables. rater reliability, for both the overall MFS score and individual
tasks scores, as well as excellent internal consistency. No floor
or ceiling effect was found from the sample of participants
The unimanual hand opening tasks in particular, small bottle, studied.
large bottle and cup, were very closely correlated with each other
with correlation coefficients close to 1. On the other hand, the Robustness of the present reliability data
broom and the ruler tasks were characterized by weaker correla-
tion with the other tasks, with correlation coefficients ranging It should be noted that the participants included in this study
from 0.55 to 0.68. were relatively homogeneous in terms of functional capacities
The Cronbach’s a coefficient for the MFS was 0.97. (see table 2/fig 1), the majority being rated between 4 and 6 for
the overall score. This relative homogeneity likely provides
Intra-rater reliability greater value to the ICCs reported here, which might have been
even higher with a more heterogeneous sample of participants.
Regarding intra-rater reliability for the 4 raters (table 2), ICCs Interestingly, the scores obtained in this sample of subjects
were excellent for each task, ranging from 0.91[0.85;0.95] to 1.00 seem representative of typical patients with chronic hemipare-
[0.99;1.0]), and for the overall MFS score, with individual ICCs sis enrolled in previous clinical studies using the MFS and also
ranging from 0.98[0.97;0.99] to 1.00[1.0;1.0]). The mean intra- of routine inpatients and outpatients with hemiparesis followed
rater differences between ratings over the 2 sessions for each task in neurorehabilitation centers.31-33. Indeed, in clinical practice,
were all <0.3 (table 3). The mean intra-rater differences for the MFS is often administered to patients with moderate upper limb
overall MFS scores ranged from 0.06§0.04 to 0.16§0.15 for the 4 paresis, which would typically correspond to MFS overall
raters (mean, 0.10§0.04). scores between 3 and 7. Patients who would have an overall
For the overall MFS score, SEM and MDC were 0.06 and 0.17, MFS score below 3 are often too impaired in their upper limb
respectively. function to usefully undergo a functional test, which might then
mostly have a negative psychological effect. At the other end of
Inter-rater reliability the spectrum, patients with an overall MFS score above 7
(which corresponds to an overall sense of smoothness in the
ICCs ranged from 0.82[0.69-0.9] to 0.95[0.92-0.97] for individ- performance of tasks) have recovered sufficient upper limb use
ual tasks, with higher values for unilateral tasks (table 2). The in daily life and have then often exited the rehabilitation

www.archives-pmr.org
1600 Z. Laclergue et al

Fig 1 MFS scores for each task and overall (Rating 1). Each bar indicates the mean; Error bars indicate SDs.

Fig 2 Correlation matrix between MFS tasks. The color code on the right indicates the values of Pearson’s coefficients.

www.archives-pmr.org
Reliability of Modified Frenchay Scale in hemiparesis 1601

Table 2 Intra- and inter-rater ICCs for each task and for the overall MFS scores with 95% confidence intervals
Intra-rater ICC Inter-rater ICC

Tasks Rater-1 Rater-2 Rater-3 Rater-4 Mean ICC Rating 1 Rating 2

Jam jar 0.97 0.97 0.95 1.0 0.97 0.90 0.89
(0.95;0.98) (0.95;0.98) (0.91;0.97) (0.99;1.0) (0.95;0.98) (0.85;0.94) (0.83;0.93)
Big bottle 0.99 0.99 0.97 1.0 0.98 0.93 0.92
(0.97;0.99) (0.98;0.99) (0.94;0.98) (0.99;1.0) (0.97;0.99) (0.89;0.96) (0.87;0.95)
Small bottle 0.98 0.99 0.97 0.99 0.98 0.92 0.93
(0.96;0.99) (0.99;1.0) (0.94;0.98) (0.99;1.0) (0.97;0.99) (0.89;0.95) (0.89;0.96)
Cup 0.96 0.99 0.97 0.99 0.98 0.92 0.92
(0.93;0.98) (0.98;0.99) (0.95;0.98) (0.99;1.0) (0.96;0.99) (0.88;0.95) (0.89;0.95)
Ruler 0.95 0.97 0.91 0.98 0.95 0.87 0.89
(0.91;0.97) (0.94;0.98) (0.85;0.95) (0.96;0.99) (0.92;0.97) (0.8;0.92) (0.83;0.93)
Clothespins 0.96 0.98 0.95 0.99 0.97 0.84 0.82
(0.92;0.97) (0.96;0.99) (0.92;0.97) (0.98;0.99) (0.95;0.98) (0.76;0.9) (0.72;0.89)
Comb 0.97 0.99 0.97 0.99 0.98 0.95 0.94
(0.95;0.98) (0.98;0.99) (0.94;0.98) (0.99;1.0) (0.96;0.99) (0.92;0.97) (0.9;0.96)
Toothbrush 0.95 0.96 0.92 0.96 0.95 0.87 0.84
(0.91;0.97) (0.94;0.98) (0.87;0.95) (0.93;0.98) (0.91;0.97) (0.77;0.92) (0.75;0.91)
Fork 0.97 0.97 0.93 0.99 0.97 0.89 0.89
(0.95;0.98) (0.95;0.98) (0.89;0.96) (0.98;0.99) (0.94;0.98) (0.84;0.93) (0.83;0.93)
Broom 0.96 0.97 0.91 0.96 0.95 0.84 0.82
(0.93;0.98) (0.95;0.98) (0.85;0.95) (0.94;0.98) (0.92;0.97) (0.75;0.91) (0.69;0.9)
Overall MFS score 0.99 1.0 0.98 1.0 0.99 0.97 0.97
(0.99;1.0) (0.99;1.0) (0.97;0.99) (1.0;1.0) (0.99;1.0) (0.95;0.98) (0.95;0.98)

process, having gained enough satisfaction with their upper though the patient may have achieved a change in scores above
limb function. the MDC value, such improvement may not be perceived as mean-
ingful to the patient. To further promote the use of the MFS, future
Internal consistency specific study may be needed to estimate the MCID of the MFS.

The analysis of internal consistency showed that there were 3

apparently redundant tasks with a correlation coefficient close to 1 Practicality and potential for dissemination of the
(little bottle, big bottle, and cup), but these tasks are qualitatively MFS
relevant with a different degree of hand opening between the 2
bottle tasks (patients often recover grasping of the small bottle The MFS may be used in acute, subacute, and chronic stages after
before the big bottle) and with an additional shoulder rotation task a central nervous system injury, in various disorders involving the
for the cup. The broom and ruler tasks were characterized by corticospinal tract such as brain trauma, multiple sclerosis, cere-
lower inter-rater reliability compared with the other tasks. In clini- bral palsy, and spinal cord injuries, although psychometric proper-
cal practice, these are not simple tasks to assess; more precise ties have yet to be specifically studied in these populations. To be
instructions in rater training might help improve the reliability of validated and used by extension, an assessment scale must have
these 2 tasks. appropriate psychometric qualities. According to the COSMIN
checklist,24 several criteria are required, such as validity, reliabil-
MDC vs minimal clinically important difference ity, and responsiveness. The MFS has good face validity because
most tasks are derived from an already validated scale (Frenchay
MDC is the smallest real difference score that is not due to random Arm Test) and because they all measure various components of
variations in measurement, to differentiate real change from ran- daily upper limb functioning. Convergent validity of the MFS was
dom measurement error. In this study, we can assume that a varia- previously reported in 2 studies, with strong correlations with the
tion between 2 evaluations in the MFS overall score >0.17/10 for total Fugl-Meyer Assessment for the upper extremity score (UE-
the same rater, and >0.37/10 between different raters can be inter- FMA) in 22 post-stroke patients (Spearman correlation=0.89)31
preted as a real change. and good correlations with the composite score of the maximal
The MDC is conceptually different from the minimal clinically range of active motion XA (CXA) for the upper limb in the Five-
important difference (MCID), which is a clinical measurement. Step Assessment in 161 patients (Pearson correlation=0.62).32
The MCID defines the smallest amount an outcome must change These prior results thus support the validity of the MFS as a mea-
to be meaningful to patients, so this is a patient-centered concept. sure of upper extremity function in patients with hemiparesis.
This value helps clinicians determine whether a change in score Regarding responsiveness, an international study on the effects of
can be considered important from patient perspective. Thus, even repeated botulinum toxin injections in the paretic upper limb

www.archives-pmr.org
1602 Z. Laclergue et al

Table 3 Intra- and inter-rater differences between ratings for each task and for the overall MFS score
Intra-rater Inter-rater

Tasks Rater-1 Rater-2 Rater-3 Rater-4 Mean Rating 1 Rating 2

Jam jar 0.16 (0.25) 0.14 (0.23) 0.25 (0.35) 0.04 (0.14) 0.14 (0.13) 0.45 (0.20) 0.45 (0.23)
Big bottle 0.08 (0.21) 0.08 (0.18) 0.24 (0.35) 0.06 (0.16) 0.11 (0.12) 0.44 (0.22) 0.50 (0.21)
Small bottle 0.14 (0.25) 0.07 (0.17) 0.26 (0.32) 0.10 (0.20) 0.14 (0.14) 0.44 (0.24) 0.44 (0.21)
Cup 0.22 (0.32) 0.12 (0.21) 0.25 (0.34) 0.12 (0.21) 0.18 (0.14) 0.50 (0.23) 0.51 (0.19)
Ruler 0.21 (0.29) 0.15 (0.23) 0.25 (0.40) 0.12 (0.21) 0.18 (0.17) 0.42 (0.24) 0.37 (0.24)
Clothespins 0.17 (0.29) 0.07 (0.17) 0.23 (0.27) 0.09 (0.19) 0.14 (0.12) 0.46 (0.29) 0.48 (0.32)
Comb 0.17 (0.31) 0.12 (0.21) 0.30 (0.30) 0.07 (0.17) 0.16 (0.15) 0.34 (0.25) 0.39 (0.23)
Toothbrush 0.18 (0.26) 0.13 (0.22) 0.30 (0.30) 0.21 (0.25) 0.20 (0.13) 0.39 (0.24) 0.41 (0.24)
Fork 0.18 (0.24) 0.17 (0.24) 0.25 (0.37) 0.13 (0.22) 0.18 (0.15) 0.47 (0.22) 0.45 (0.24)
Broom 0.18 (0.30) 0.17 (0.24) 0.34 (0.44) 0.21 (0.27) 0.22 (0.15) 0.56 (0.25) 0.58 (0.27)
Overall MFS score 0.10 (0.08) 0.08 (0.06) 0.16 (0.15) 0.06 (0.04) 0.10 (0.04) 0.23 (0.12) 0.24 (0.11)
NOTE. Results expressed as mean (SD).

showed an average improvement in the MFS overall score of mobility. Additionally, the MFS tests a relatively complete palette
+0.43 after 3 injections of 1000 U abobotulinumtoxinA and of of human hand use, with tasks of overall hand opening and elbow
+0.76 after 3 injections of 1500 U abobotulinumtoxinA.33,34 supination for vertical objects of large caliber (jar with the shoul-
der in neutral position, bottles and cup with flexed shoulder), a
Potential advantages of the MFS over existing task of opening with the elbow pronated (ruler), forceful hand
scales closing tasks for small caliber objects (tube, clothespins, broom),
light index-thumb pinching tasks for thin objects (comb), a mixed
Compared with other classic upper limb functional scales,9-17 reli- hand opening and closing task (fork), and shoulder external rota-
ability appears similar, but the MFS takes about 10 minutes to tion tasks (comb, cup). Finally, the MFS can be clinically helpful
complete, and uses daily and low-cost equipment. The scale can using qualitative observation to figure out which antagonistic
thus be used in routine care for in- or outpatients. All tasks are muscles may limit movements, to guide rehabilitation prescrip-
simple, uni- and bimanual, and involve proximal as well as distal tions or botulinum toxin injections.

Fig 3 Bland-Altman plot of the inter-rater reliability for overall MFS scores. Differences between two overall MFS scores are displayed as a func-
tion of the mean score. Solid line, mean inter-rater difference between 2 overall scores; dashed lines, 95% limits of agreement (mean§1.96£the
SD of the differences between raters).

www.archives-pmr.org
Reliability of Modified Frenchay Scale in hemiparesis 1603

Study limitations - 20 cm ruler and pen

The patient sample size was moderate but similar or larger than in - Three clothes pins
previous comparable reliability studies.25−27 The study was mono- - Comb
centric and the 4 raters received a group training session from the - Tube of toothpaste and toothbrush
same neurorehabilitation school prior to the study assessments. - Knife and fork
Also, the present study was not designed to assess the role played - Broom
by experience in using the scale, for which a larger sample of
raters would have been required. To mitigate potential bias related
to the experience of raters, we recruited a junior resident with only
2 months of experience using the scale; we still observed excellent
intra-rater reliability for that investigator. In comparable literature,
it should be noted that evaluations were performed by experienced
therapists in most reliability studies for upper limb functional
scales14,25 and that the efficacy of training sessions was also not
evaluated. However, we would recommend that users of the scale
be given prior training similar to the one used in the present study
to optimize their rating skills.

Conclusions
The present findings provide evidence that the MFS is an internally
consistent and reliable scale in hemiparesis to evaluate upper extrem-
ity function through 10 daily tasks. The MFS is a simple tool that can
be used in routine care for patient evaluation and follow-up, as well
as in clinical research. It allows for both quantitative and qualitative
analysis of the paretic upper limb. Today, the MFS constitutes an Guidelines to raters:
integral part of the Five-Step Assessment for the evaluation of Place the objects in front of the patient in the order of tasks,
deforming spastic paresis.22 Further studies are needed in other cen- from left to right in an arc shape, approximately 1 arm length
ters35 in comparison with other functional scales to improve its valid- away from the patient.
ity, with patient samples of various conditions. Ask the patient to pull up the sleeve on the paretic upper limb.
Perform the tasks 1 by 1, always in the same order.
Unimanual tasks should be performed with the paretic hand
Keywords only, without help from the other hand.
If the patient cannot reach the object, try again after placing it
Hemiplegia; Muscle spasticity; Neurologic rehabilitation; Reha- next to him close to the edge of the table (task rating will necessar-
bilitation; Reproducibility of results; Upper extremity ily be <5).
Overactive muscles of the paretic hand may be self-stretched
before each task.
Corresponding author The patient should not stand up to perform the task.
Zo
e Laclergue, Service de R
e
education Neurolocomotrice, Suggestions to facilitate rating: Half points are allowed
H^
opitaux Universitaires Henri Mondor, 1, rue Gustave Eiffel,
94000 Cr
eteil, France. E-mail address: zlaclergue@gmail.com.
0: No movement
1: Movement initiated with ≥1 movement component
Acknowledgments 2: Task still closer to not being achieved than to being achieved
2.5: Task midway between being achieved and not
We are grateful to the patients who helped us carry out the study. 3: Four movement components still incomplete (eg, for a
unimanual task: shoulder flexion, elbow extension, elbow
supination, and finger extension, which deducts 4 half points
Appendix 1 Guidelines for the Modified from 5)
Frenchay Scale 4: Task close to being achieved but 2 movements components
still incomplete (two half points deducted from 5)
Required equipment: 5: Task barely accomplished
6: Sense of security: no doubt that the task was going to be
achieved but task completed with difficulty and slowness
- Jam jar of about 8 cm diameter 7: Sense of smoothness: task accomplished with some smooth-
- Big bottle of about 8 cm diameter, filled up one-third ness but still slow
- Small bottle of about 6 cm diameter, filled up one-third 8: Sense of speed: task completed fast but not normal
- Empty solid cup about 7 cm diameter 9: Almost normal
- Notepad or slate 10: Normal performance

www.archives-pmr.org
1604 Z. Laclergue et al

Appendix 2 MFS Tasks and Scoring 4. Baer G, Smith M. The recovery of walking ability and subclassifica-
tion of stroke. Physiother Res Int 2001;6:135–44.
5. Kwakkel G, Kollen BJ, van der Grond J, Prevo AJ. Probability of
regaining dexterity in the flaccid upper limb: impact of severity of
Tasks Score
paresis and time since onset in acute stroke. Stroke 2003;34:2181–6.
1. Open and close the jam jar using both hands 6. Suzuki M, Omori Y, Sugimura S, et al. Predicting recovery of bilateral
(paretic hand holds the jar) upper extremity muscle strength after stroke. J Rehabil Med
Specific comment: Hold the jar above the table (not in 2011;43:935–43.
contact with the table), otherwise rate <5. 7. Alt Murphy M, Resteghini C, Feys P, Lamers I. An overview of sys-
2. Reach, pick up and release the large bottle using tematic reviews on upper extremity outcome measures after stroke.
paretic hand only BMC Neurol 2015;15:29.
8. Santisteban L, T
er
emetz M, Bleton JP, Baron JC, Maier MA, Lindberg
Specific comment: Grasp the bottle by the side from its
PG. Upper limb outcome measures used in stroke rehabilitation studies:
initial position at arm length; if bottle is grasped from
a systematic literature review. PLoS One 2016;11:e0154792.
the tip, or from table edge only, rate <5. 9. Fugl-Meyer AR, J€a€ask€o L, Leyman I, Olsson S, Steglind S. The post-
3. Reach, pick up, and release the small bottle using stroke hemiplegic patient. 1. A method for evaluation of physical per-
paretic hand only formance. Scand J Rehabil Med 1975;7:13–31.
Specific comment: Grasp the bottle by the side from its 10. Mathiowetz V, Volland G, Kashman N, Weber K. Adult norms for the
initial position at arm length; if bottle is grasped from Box and Block Test of manual dexterity. Am J Occup Ther
the tip or from table edge only, rate <5. 1985;39:386–91.
4. Reach, pick up the cup, and bring to mouth using 11. Mathiowetz V, Weber K, Kashman N, Volland G. Adult norms for
paretic hand only the nine hole peg test of finger dexterity. Occup Ther J Res
1985;5:24–38.
Specific comment: Grasp the cup by the side from its
12. Rapin I, Tourk LM. Evaluation of the Purdue Pegboard as a screening
initial position at arm length; all components of the
test for brain damage. Dev Med Child Neurol 1966;8:45–54.
movement must be barely completed to achieve 5, 13. Lyle RC. A performance test for assessment of upper limb function in
including external rotation of the shoulder to bring cup physical rehabilitation treatment and research. Int J Rehabil Res
to mouth. 1981;4:483–92.
5. Rule a line with the ruler using both hands (the 14. Wolf SL, Catlin PA, Ellis M, Archer AL, Morgan B, Piacentino A.
paretic hand holds the ruler) Assessing Wolf motor function test as outcome measure for research
Specific comment: If ruler is stabilized by paretic hand in patients after stroke. Stroke 2001;32:1635–9.
closed, rate 5. If ruler is barely stabilized by paretic 15. Uswatte G, Taub E, Morris D, Light K, Thompson PA. The Motor
hand almost open, rate 6. Activity Log-28: assessing daily use of the hemiparetic arm after
stroke. Neurology 2006;67:1189–94.
6. Clip 3 clothespins to the notepad using both hands
16. Penta M, Tesio L, Arnould C, Zancan A, Thonnard JL. The ABIL-
(paretic hand holds and clips pins)
HAND questionnaire as a measure of manual ability in chronic stroke
Specific comment: If paretic hand clips only 1 out of the 3 patients: Rasch-based validation and relationship to upper limb
clothespins, rate 5. impairment. Stroke 2001;32:1627–34.
7. Reach, pick up the comb, and mimic combing 17. Dehmiyani A, Taghizadeh G, Azad A, et al. Psychometric properties
using the paretic hand only of dexterity questionnaire-24 in Iranian chronic stroke survivors. Top
Specific comment: Pick up comb from initial position at Stroke Rehabil 2022;29:201–7.
arm length, if comb is picked up from table edge only, 18. Wade DT, Langton-Hewer R, Wood VA, Skilbeck CE, Ismail HM.
rate <5. The hemiplegic arm after stroke: measurement and recovery. J Neurol
8. Put toothpaste on toothbrush using both hands Neurosurg Psychiatry 1983;46:521–4.
19. Parker VM, Wade DT, Langton Hewer R. Loss of arm function after
(paretic hand holds and presses on tube)
stroke: measurement, frequency, and recovery. Int Rehabil Med
9. Pick up knife and fork using both hands and mimic
1986;8:69–73.
cutting (paretic hand holds the fork) 20. Heller A, Wade DT, Wood VA, Sunderland A, Hewer RL, Ward E.
Specific comment: To reach 5, the paretic hand must both Arm function after stroke: measurement and recovery over the first
grip the fork firmly enough (with Digits 3, 4, 5 and three months. J Neurol Neurosurg Psychiatry 1987;50:714–9.
thumb), point the index finger along the fork axis, and 21. Gracies JM, Hefter H, Simpson D, Moore P. Botulinum toxin in
have sufficient elbow pronation and shoulder spasticity. In: Moore P, Naumann M, eds. Handbook of botulinum
abduction/internal rotation to be able to mimic toxin, Hoboken, New Jersey: Blackwell Science; 2002:221–74.
planting fork. 22. Gracies JM, Bayle N, Vinti M, et al. Five-step clinical assessment in
10. Sweep the floor with the broom using both hands spastic paresis. Eur J Phys Rehabil Med 2010;46:411–21.
23. Kottner J, Audig
e L, Brorson S, et al. Guidelines for Reporting Reli-
TOTAL SCORE
ability and Agreement Studies (GRRAS) were proposed. J Clin Epide-
miol 2011;64:96–106.
24. Mokkink LB, Terwee CB, Knol DL, et al. The COSMIN checklist for
References evaluating the methodological quality of studies on measurement proper-
ties: a clarification of its content. BMC Med Res Methodol 2010;10:22.
1. Gracies JM. Pathophysiology of spastic paresis. I: paresis and soft tis- 25. Nijland R, van Wegen E, Verbunt J, van Wijk R, van Kordelaar J,
sue changes. Muscle Nerve 2005;31:535–51. Kwakkel G. A comparison of two validated tests for upper limb func-
2. Gracies JM. Pathophysiology of spastic paresis. II: emergence of mus- tion after stroke: the Wolf Motor Function Test and the Action
cle overactivity. Muscle Nerve 2005;31:552–71. Research Arm Test. J Rehabil Med 2010;42:694–6.
3. GBD 2016 Stroke Collaborators. Global, regional, and national burden 26. Lin JH, Hsu MJ, Sheu CF, et al. Psychometric comparisons of 4 meas-
of stroke, 1990-2016: a systematic analysis for the Global Burden of ures for assessing upper-extremity function in people with stroke.
Disease Study 2016. Lancet Neurol 2019;18:439–58. Phys Ther 2009;89:840–50.

www.archives-pmr.org
Reliability of Modified Frenchay Scale in hemiparesis 1605

27. Platz T, Pinkowski C, van Wijck F, Kim IH, di Bella P, Johnson G. 32. Bayle N, Maisonobe P, Raymond R, Balcaitiene J, Gracies JM.
Reliability and validity of arm function assessment with standardized Composite active range of motion (CXA) and relationship with active
guidelines for the Fugl-Meyer Test, Action Research Arm Test and function in upper and lower limb spastic paresis. Clin Rehabil
Box and Block Test: a multicentre study. Clin Rehabil 2005;19:404–11. 2020;34:803–11.
28. Cronbach LJ. Coefficient alpha and the internal structure of tests. 33. Gracies JM, Brashear A, Jech R, et al. Safety and efficacy of abobotu-
Psychometrika 1951;16:297–333. linumtoxinA for hemiparesis in adults with upper limb spasticity after
29. Bartko JJ. The intraclass correlation coefficient as a measure of stroke or traumatic brain injury: a double-blind randomised controlled
reliability. Psychol Rep 1966;19:3–11. trial. Lancet Neurol 2015;14:992–1001.
30. Koo TK, Li MY. A guideline of selecting and reporting intraclass 34. Gracies JM, O’Dell M, Vecchio M, et al. Effects of repeated abobotuli-
correlation coefficients for reliability research [published correction numtoxinA injections in upper limb spasticity. Muscle Nerve 2018;57:
appears in J Chiropr Med. 2017 Dec;16(4):346]. J Chiropr Med 245–54.
2016;15:155–63. 35. Ghroubi S, Alila S, Elleuch W, Ayed HB, Mhiri C, Elleuch MH.
31. Pila O, Duret C, Gracies JM, Francisco GE, Bayle N, Hutin E.
Evolu- Efficacy of botulinum toxin A for the treatment of hemiparesis
tion of upper limb kinematics four years after subacute robot-assisted in adults with chronic upper limb spasticity. Pan Afr Med J
rehabilitation in stroke patients. Int J Neurosci 2018;128:1030–9. 2020;35:55.

www.archives-pmr.org