If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Barden HL, Nott MT, Heard R, Chapparo C, Baguley IJ. Clinical assessment of hand motor performance after acquired brain injury with dynamic computerized hand dynamometry: construct, concurrent, and predictive validity.
To assess the construct, concurrent, and predictive validity of dynamic computerized hand dynamometry.
Prospective correlational study between dynamometry and functional upper limb performance.
Hospital outpatient spasticity clinics.
Adults with upper motor neuron syndrome affecting the upper limb after acquired brain injury (ABI) (n=38; median age, 50y; range, 18–81y) and healthy adult control participants (n=27; median age, 37y; range, 22–62y).
Main Outcome Measures
Dynamic computerized dynamometry elements of hand performance (isometric force, force velocity, isometric grip work, contraction and relaxation duration) and the Action Research Arm Test.
Motor elements of hand performance objectively measured by the dynamic computerized dynamometry protocol achieved moderate to good validity when correlated with standardized measures of functional hand performance. Dynamic computerized dynamometry identified clear differences in hand performance between participants with and without ABI. Within the ABI group, dynamic computerized hand dynamometry achieved fair to moderate predictive validity with regards to whether a participant would be referred for botulinum toxin A injections.
This study provides support for the construct, concurrent, and predictive validity of the dynamic computerized dynamometry protocol.
THE EFFECTS OF UPPER motor neuron (UMN) lesions on hand performance are complex. Current clinical hand use measures after acquired brain injury (ABI) are subjective in nature and often measure a single feature of the multidimensional UMN syndrome. Two such scales, the Modified Ashworth (MAS)
are common measures used by clinicians. These scales attempt to measure spasticity, a positive feature of the UMN syndrome, at the body function and structure level as described by the World Health Organization's International Classification of Functioning, Disability and Health (ICF).
these functional measures evaluate the impact of spasticity on hand function; however, they provide limited information about the underlying physiologic changes that negatively affect hand function.
To provide targeted upper limb (UL) assessment and rehabilitation for adults with ABI, clinicians benefit from an understanding of both the functional impact and the underlying neurophysiologic impairments affecting hand performance in the UMN syndrome. An approach that aims to evaluate both aspects of hand function in UMN syndromes is dynamic computerized hand dynamometry (DCD).
In this protocol, positive UMN features (such as muscle overactivity, spasticity, and clonus) affect the individual's ability to release the dynamometer and are measured by minimum isometric force, minimum force velocity, and involuntary isometric grip work motor elements. In addition, negative UMN features (eg, muscle weakness and reduced motor control) impact on the ability to grasp the device and are measured by maximum isometric force, maximum force velocity, and voluntary isometric grip work motor elements of hand performance.
allowing a different means of investigating the link between clinical presentation and the underlying neurophysiology in terms of its motor output. If DCD can be shown to assess both aspects of UMN syndrome, it has potentially better content validity than existing measures.
To determine the clinical utility of DCD protocols, it is important to demonstrate the validity of this approach for measuring hand performance. Applied in the research context, the concept of validity refers to the “extent to which an instrument measures what it is intended to measure.”
to determine whether DCD can discriminate between healthy controls and individuals after ABI.
Concurrent validity: To explore the relationship between DCD and functional measures of UL performance in adults with ABI.
Predictive validity: To project treatment decisions based on DCD (ie, whether participants will be referred for botulinum toxin A [BTX-A] injection).
This validation study was approved by the local institutional human research ethics committee. Participants provided written informed consent before involvement in the study.
Thirty-eight adults experiencing UL spasticity after ABI (stroke and traumatic brain injury) were consecutively recruited from 3 outpatient spasticity clinics. All participants with ABI were living in the community and met the following inclusion criteria: age greater than 17 years, first onset of ABI, motor overactivity resulting from UMN syndrome of greater than 3 months' duration, and sufficient grip strength in the affected UL to hold the dynamometer (minimum >.75kg). Exclusion criteria included bilateral UL neurologic disease, other causes of UL weakness (including rheumatologic conditions and lower motor neuron lesions), and inadequate ability to understand instructions given in English because of dysphasia or language other than English. Adults in the ABI group had a median age of 50 years with an age range or 18 to 81 years. Twenty-seven healthy adults were recruited as a control group from a convenience sample of hospital staff and adults from the community, with the following inclusion criteria: 17 years of age or older and the ability to understand verbal instructions in English. Control participants were excluded if they self-reported acute or chronic hand impairments or UL neurologic impairment. The control group had a median age of 37 years with an age range of 22 to 62 years. Table 1 shows the demographic details of the ABI and control groups.
below and using a Jamar-style Biometrics G100 Precision Dynamometer.a The dynamometer was calibrated by an independent registered weights and measures company to assess for linearity across the dynamic range of the dynamometer with an accuracy of ±0.5%. The raw dynamometer signal was sampled at 400Hz and amplified through a general purpose amplifier (model ML142b) to a PowerLab 26 data acquisition system (model ML856b) and displayed in real-time on a laptop computer using LabChart 7.0.2 software.b
The ARAT assesses UL performance of individuals in clinical and research settings.
Subtask performance across the 4 subscales (grasp, grip, pinch, gross movement) is graded using a 4-point scale: 3, normal task performance; 2, task completion with great difficulty or in a prolonged timeframe; 1, partial task completion; and 0, inability to perform any part of the task.
Maximum performance of hand function over the 19 subtasks of the ARAT is represented by a maximum total score of 57. The reliability and validity of the ARAT are well established for people poststroke,
Participants were assessed at hospital outpatient facilities by 2 occupational therapists independent of the treating team. Participants were seated during dynamometry assessment using 2 adaptations to the American Society of Hand Therapists
standard testing position. First, the elbow and forearm were supported on the chair or wheelchair arm (fig 1) . Second, the handle setting on the dynamometer was placed in either the second or third handle positions, with the third handle position used to assess power grip for those with larger hands.
Participants self-selected a static wrist position for the performance of grasp and release.
The computerized hand dynamometer was set to 0 before each participant's assessment. Three pretest trials of grasp were conducted to record baseline data and to ensure an adequate understanding of the instructions:
Pretrial 1: A single maximum force grasp and release cycle
Pretrial 2: A single maximum speed grasp and release cycle
Pretrial 3: A single grasp and release cycle integrating maximum force generation at maximum speed Participants viewed real-time force and velocity output during the assessment. After successful completion of the pretest trials, participants completed the assessment protocol
consisting of 10 consecutive grasp and release cycles using the integrated maximum force and maximum speed task from pretrial 3. For all participants, the DCD protocol was undertaken before completion of the ARAT
For the control group, the dominant hand was assessed first, immediately followed by the nondominant hand. For participants in the ABI group, the unaffected hand was assessed immediately before the affected hand. For 16 individuals, hand dominance altered as a result of the ABI.
Participants were then reviewed by an independent physician who was blinded to the research assessment. The physician referred and subsequently injected ABI participants with BTX-A based on an assessment of the presence and severity of muscle spasticity along routine clinical protocols. These protocols suggest that patients who have positive UMN features—that is, muscle overactivity and spasticity—be referred for BTX-A injection if the overactivity adversely affects upper limb function.
Data collected in real-time from the computerized hand dynamometer produced force and force-velocity curves that were analyzed with a previous protocol.
Cycle duration was calculated as the time (in seconds) elapsing from one minimum isometric force marker to the following minimum isometric force marker (fig 2) . This study adopted the concept of isometric grip work previously described by Bautmans et al.
Total isometric grip work is calculated as the area under the force curve and interpreted as having voluntary and involuntary components. Clinically, voluntary isometric grip work represents the amount of motor effort the participant intentionally directs toward the dynamometer assessment. Conversely, involuntary isometric grip work measures the unintentional, non–task-directed work done while holding the dynamometer.
The lower panel in figure 2 contains the force-velocity (Fvel) curve, representing the rate of change in isometric force production during grasp and release (measured in kilograms per second).
The area above the horizontal zero line on the Fvel curve corresponds to the grasp phase, beginning at zero and increasing to maximum Fvel (the point at which peak velocity occurs). Conversely, the release phase falls below the horizontal zero line, also beginning at zero and decreasing to minimum Fvel (the point at which release is being generated fastest).
These data (isometric force, percentage isometric grip work, force velocity, and contraction and relaxation duration) were processed offline using LabChart 7.0.2 and averaged across the central 8 grasp and release cycles for each participant, as these cycles use a consistent motor plan (namely grasp, release, and preparation for regrasp). The first and last cycles were omitted from further analysis because they use different motor plans compared with that of the central 8 cycles, and the resultant data show greater intra- and intersubject variability compared with that of the central 8 cycles.
The resulting DCD values and ARAT data underwent statistical analysis in SPSSc version 19. Descriptive statistics were calculated to report group demographic variables and injury-related variables for the ABI group. Nonparametric statistics were calculated where data were skewed. To address the first aim of the study, to determine whether DCD could discriminate between healthy controls and individuals with ABI, a between-group comparison (ABI and control) was conducted with the Mann-Whitney U test by using the known groups method.
To address the second study aim, and explore the relationship between DCD and functional measures of UL performance in adults with ABI, concurrent validity was examined by correlating the DCD variables with the total ARAT scores of the ABI participants, using the Spearman correlation coefficient (ρ). A correlation of .75 and above was considered a good to excellent relationship; between .50 and .75, moderate to good; .25 to .50, fair correlation; and between .00 and .25, little or no relationship.
Finally, to address the third study aim, whether DCD results are related to the need for BTX-A injection, the predictive validity of the DCD protocol was determined against the treatment decisions of the independent physician. A Mann-Whitney U test was used to identify differences in DCD values between ABI participants referred or not referred for UL BTX-A injections, with significance set at P≤.05. Effect size for between-group differences (ABI and control; referred or not referred for injection) was calculated using nonparametric point biserial correlation (rspb).
The demographic details for the ABI and control groups are shown in table 1. There were no statistically significant differences between the groups for sex and handedness; however, the control group was a median of 13 years younger than the ABI group. All control participants scored maximally on the ARAT (total score, 57). Within the ABI group, 30 participants (79%) scored maximally on their “unaffected” UL compared with 2 (5%) on the “affected” side. Ten ABI participants (26%) scored 0 points on the ARAT. Significant differences in UL function were observed when comparing ARAT scores of the ABI participants' “affected” UL with the nondominant UL of control participants (see table 1); poorer UL function was evident in the ABI group. Statistically significant differences in UL function were also noted when comparing the “unaffected” UL of ABI participants with the dominant UL of controls. While both groups achieved the same median score of 57, a large variation in “unaffected” UL performance by ABI participants (range, 37–57) contributed to the significant difference and challenges the notion that the so-called unaffected UL is “normal” in people with unilateral ABI.
Construct Validity Using the Known Group Method: Differentiation Between Individuals With and Without UMN Syndrome
Significant differences between the ABI group and the control group were identified across each DCD motor element of hand performance for the ABI-affected hand and the control nondominant hand (table 2). Compared with controls, the ABI group demonstrated significantly impaired performance across every DCD measure, with moderate to large effect sizes. The ABI group generated less force (ABI, 8.8kg; control, 31.4kg), more slowly (maximum Fvel: ABI, 20.3kg·s–1; control, 258.3kg·s−1), and required a longer period to perform the grasping task (contraction duration: ABI, 0.5s; control, 0.3s). Dynamometer release was also slower (minimum Fvel: ABI, −16.4kg·s−1; control, −265.9kg·s−1) and less complete (minimum isometric force: ABI, 2.9kg; control, 1.9kg). The greater involuntary isometric grip work in the ABI group indicated a higher proportion of nonpurposeful motor effort (ABI, 29%; control, 3%).
Table 2Motor Performance of the ABI and Control Groups
NOTE. Values are group median (interquartile range), Mann-Whitney U test Z scores, and effect size (rpbs). Effect size magnitudes were considered small at rspb=.10, medium at rspb=.24, and large at rspb=.37.
Concurrent Validity: Relationship Between Hand Dynamometry and Function
Fair to good statistically significant correlations were observed between isometric force, force velocity, isometric grip work, and the total ARAT score in the ABI group (table 3). The associations between DCD measures and the ARAT total score were in the expected direction. Elements of hand performance measuring positive UMN syndrome features (minimum isometric force and minimum force velocity) had a negative relationship with the ARAT, indicating that increasing difficulty in releasing the dynamometer was linked to deterioration in ARAT performance. Conversely, motor elements that primarily measure negative UMN syndrome features (maximum isometric force and maximum force velocity) had a positive relationship with the ARAT total score, signifying that an improved ability to grasp is associated with a higher total ARAT score.
Table 3Relationship Between Computerized Hand Dynamometry Motor Elements and ARAT Total Score
NOTE. Spearman rank order correlation is presented. A correlation of .75 indicates a good to excellent relationship; .50–.75, moderate to good; .25–.50, fair correlation; and .00–.25, little or no relationship.
Predictive Validity: Dynamometry and Treatment Decisions
Twenty-five people within the ABI group (66%) were identified as being appropriate for, and subsequently received, BTX-A injection (the injection group). Dynamometry data for the BTX-A injected and noninjected groups are shown in table 4. Significant between-group differences were evident on force, velocity, and work-related DCD motor elements.
Table 4Motor Performance Differences Between the Injected and the Noninjected Groups
Large effect sizes were achieved for the DCD components of isometric force, force velocity, and isometric grip work. The injection group demonstrated higher involuntary isometric grip work (35% vs 10%) and lower maximum isometric force than those not injected. A higher minimum isometric force, indicating a reduced ability to release, was observed in the injection group compared with the noninjection group. Speed of force generation and speed of release were both significantly slower in the injection group compared with the noninjection group, evidenced by a lower maximum Fvel and a higher minimum Fvel.
This study examined the construct, concurrent, and predictive validity of dynamic computerized hand dynamometry for evaluating specific motor elements of hand function during grasp and release in 38 post-ABI adults with UMN syndrome and 27 healthy adult controls. Construct validity was evident, with dynamometry measures of healthy controls being significantly different from those of post-ABI adults. Concurrent validity was also supported by the fair to good relationship between DCD and UL functional measures measured by the ARAT. DCD exhibited a fair to moderate degree of predictive validity when evaluated against the treatment decisions of an independent physician.
These results support previous research into the use of the DCD protocol in the assessment of UMN syndromes.
the ABI group displayed significant negative UMN features, being weaker than the control group (ABI group, 8.8kg; control group, 31.4kg). These negative features have been self-reported by adults with ABI to be the most significant limiting factor to functional performance.
However, the most significant between-group differences in this study were observed in positive UMN features. In particular, while the control group was able to direct 97% of its voluntary effort toward the grasp and release task, only 70% of the effort exerted by the ABI group was targeted on the grasp and release task, leaving almost 30% of the work undertaken by the ABI group as involuntary. The application of this nontask effort may exacerbate the motor fatigue observed in adults with UMN syndrome. The ABI group was slower to complete the relaxation/release phase of the task, taking twice as long to release the dynamometer (ABI group, 0.6s; control group, 0.3s). Thus, the DCD approach was able to simultaneously identify both positive and negative UMN features with some accuracy, consistent with current research recommendations,
but not evident in current clinical measures such as the MAS and Tardieu. This finding would suggest that the DCD protocol possesses better content validity than the currently used standard measures of muscle spasticity.
Furthermore, the DCD protocol measures performance at both the body structure and function and the activity levels of the ICF, in line with current recommendations.
This study identified statistically significant relationships between isometric force, force velocity, isometric grip work, and UL performance on the ARAT. The strength of these relationships compares favorably with recently reported correlations between isometric force and ARAT total score.
reported a weak, statistically nonsignificant relationship between the MAS and the ARAT. While the MAS has been held as a criterion standard in clinical practice for some time, the desired link to functional performance has not been evident, suggesting that revision of criterion standard status is overdue.
While these findings are promising, it remains to be determined whether the DCD approach will allow for more accurate assessment and enhanced understanding of the interaction between UMN features and their impact on UL function. This study provides preliminary evidence that this may be the case. Within the ABI group, DCD values were moderately predictive for whether a participant was referred for BTX-A injection. Specifically, those injected with BTX-A had a greater severity of positive UMN features compared with the noninjected group. The injection group exhibited significantly greater involuntary isometric grip work compared with the noninjected group (35% vs 10%, respectively, compared with 3% for controls). Further, the slower dynamometer release speed observed in the injected group (.70s vs .54s) provides further evidence that the degree of muscle overactivity is a predictor of BTX-A injection. This substantiates the clinical experience of referring individuals who have dominant positive UMN features and are experiencing difficulty releasing or opening the hand for BTX-A injections.
Although the group referred for injection had more severe positive UMN features, the impact of negative features was also evident. Weakness as one of the negative UMN features has been reported to play a primary role in limiting activity completion in the UL.
and allows simultaneous measurement of positive and negative features of UMN syndrome that is not possible with existing measures of muscle spasticity and UL function such as the Tardieu, MAS, and ARAT. Clinical data collected using DCD also show equivalent or better construct and concurrent validity compared with current measures.
Furthermore, large effect sizes were found for the DCD variables in terms of predictive ability; this contrasts with the MAS, which did not predict BTX-A injection patterns in a recent multicenter trial.
DCD also had no floor/ceiling effects for the affected UL of the ABI group compared with almost a third of participants on the ARAT. While promising, further research is required to assess the specificity and sensitivity of the DCD, and to determine its value as an assessment measure for hand performance after UMN lesions.
The recruitment of the ABI group from outpatient spasticity clinics will have resulted in a greater proportion of participants having clinically significant positive UMN features to be recruited to the study, which may limit the applicability to the overall ABI group seen in other clinical contexts. The heterogeneity of the ABI sample, consisting of both stroke and traumatic brain injury, may limit the application of these results to either specific diagnostic group. However, the heterogeneity reflects current clinical referral patterns in spasticity clinics. Finally, the difference in the median age of the ABI and the control groups has the potential to influence the results. However, the only DCD measure with published normative data is maximum force generation. Published Australian normative data for the 35- to 45-year and 45- to 54-year age groups, coinciding with the median ages of the control and ABI groups, respectively, show effectively identical data ranges (ie, ±2kg).
In contrast, differences between the maximum force of the ABI and control groups exceeded 20kg. In light of this, the effect of a participant's ABI produced a significantly greater impact on force generation than could be explained by the group age differences. Furthermore, the median age of the ABI group is representative of patients seen in clinical practice and should not contribute to statistical issues with the validity findings as presented. Finally, the observed predictive validity may have been adversely affected in situations where BTX-A injections were performed for proximal upper limb spasticity but where the distal arm was relatively spared. These issues should be addressed in future research.
The measures derived from the DCD protocol showed a moderate to good degree of validity. If confirmed in future research, the DCD approach allows for the objective and repeatable simultaneous measurement of positive and negative UMN features as they relate to hand motor performance. Output from the DCD protocol allows the impact of UMN syndrome to be assessed at both the body structure and function, and activity levels of the ICF, demonstrating concurrent validity with UL measures of activity. Further refinements may enable the complex nature of UMN syndrome and the multifaceted impact of positive and negative UMN features on hand performance to be better assessed for the purpose of intervention planning and outcome measurement.
Supported by an Australian Postgraduate Award, The George Burniston Cumberland Foundation Fellowship 2011, and the Helga Pettitt Faculty of Health Sciences Postgraduate Study Award 2011. This article is derived from data collected as part of an investigator-initiated study supported by Ipsen Australia.
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.