Measurement of isometric force in children with and without Duchenne's muscular dystrophy.

Duchenne's muscular dystrophy (DMD) is a progressive, inherited disease of muscle that affects one of every 3,500 male births. It has recently been demonstrated that DMD is the result of a defective gene that encodes a protein termed "dystrophin." [1] Despite this major breakthrough in medical research, an effective treatment or cure for this disease does not exist.

New interventions such as myoblast transfer [2] have increased the demand for better clinical documentation of the natural history of DMD and the effects of experimental procedures on its clinical course. At present, however, no reliable clinical markers that accurately describe disease progression in individuals with DMD have been reported. Levels of serum creatine phosphokinase and 3-methylhistidine have been monitored in clinical trials, but they have been ineffective in judging therapeutic responses to new interventions. [3,4]

The major clinical characteristic of DMD is a gradual deterioration of muscle force. A practical and potentially effective means of measuring muscle deterioration is by quantifying changes in the production of isometric force. Previous studies on the natural history and effects of therapeutic interventions in patients with DMD have shown the importance of force measurements in defining the clinical course of these patients. [5,6] These studies, however, have relied on measurement of force by manual muscle testing (MMT) and other clinical measures that have questionable reproductibility, sensitivity, or scaling properties. [5,8] The purpose of this article are to describe an electronic strain-gauge system adopted specifically for children for measuring isometric strength and to examine technical measurement qualities of the system when applied to children with DMD and to an age-matched comparison group.

Until the recent thrust for quantitative instrumentation in physical therapy clinics, MMT had been the most widely used system of muscle force testing. [5,9] Standard MMT techniques and recent derivations are based on the patient's ability to move against gravity and hold against the examiner's resistance [10-12] Investigators [9,13] have reported acceptable reliability using MMT in patients with DMD. Manual muscle testing, however, produces ordinal-scale data, which limits its usefulness in clinical trials. [8,14-16] Additionally, MMT depends on the examiner's strength in the Good to Normal range [17] and on the weight of the limb within the Fair grade. [18]

Isometric force measurement provides the most direct method of assessing the amount of contractile activity in a particular muscle group. [19] Whereas isotonic and isokinetic muscle testing are influenced by biomechanical and neurophysiological factors, isometric force testing maintains constancy in muscle length, joint angle, and velocity. [20] By simplifying the complex function of human muscle performance into discrete segments via isometric force testing, one can minimize inference in estimating disease progression by changes in muscle force. [21] The use of isometric force testing, however, de-emphasis functionally based muscle characteristics that are obtained with isotonic and isokinetic testing.

Isometric muscle force recordings in both childern and adults have been documented by use of cable tensionmeters, [23] hand-held myomters, [16,23-26] and strain-gauge tensiometers. [26-28] The hand-held myometers have been reported to objectively record muscle force in patients with DMD. [23] Because the therapist must provide resistance equal to that given by the child, handheld dynamometers have hot been effective in detecting early muscle force loss, [29] or in measuring strong muscle groups in children without DMD. [23]

The use of an electronic strain gauge for the measurement of isometric force has been described as a component of the Tufts Quantitative Neuromuscular Examination (TQNE). [30] The TQNE consists of four motor component tests: (1) pulmonary function, (2) oropharyngeal function, (3) timed dexterity, and (4) maximal isometric force production. The isometric force protocol, developed for quantitative motor assessment in patients with amyotrophic lateral sclerosis (AIS), samples maximal isometric force of nine muscle groups. Test-retest reliability has been established for non-disabled adults (r=.82-.99) and for patients with ALS (r=.92-.99). [30] Because it is connected to immobile uprights, an advantage of the strain-gauge system is that it permits testing of weak as well as very strong muscle groups. Natural history and clinical trial investigations of DMD require a force evaluation system that is applicable for both clinical and normative samples of children with varying degree of force capabilities. The objectives of this study were to modify the TQNE force-gauge protocol for children and to determine selected technical measurement characteristics in matched groups of subjects with and without DMD. Although the strain-gauge, system has been shown to have acceptable test-retest reliability in patients with ALS, replication of the quality of measurements of the force-gauge system should be demonstrated for children with and without DMD before it can be applied to these populations with confidence. [31] Additionally, adaptations in testing positions and administration procedures were necessary secondary to population differences such as contractures in children with DMD, smaller body proportions, and general motivational and attentional considerations for testing children.

The quality of clinical measurements of the strain gauge can be estimated and presented within a generalizability framework. [32] A generalizability approach allows one to examine simultaneously a number of sources of variation in measurement that may be of clinical interest. In this study, we sought to examine the extent of isometric force variation resulting from testing different populations (ie, children with and without DMD), extremity sides, and repeated occasions. In a subsample of children we also examined variation secondary to different raters. The strain-gauge measurement system would be considered optimal for clinical research applications if it demonstrated the ability to discriminate between populations with and without DMD (construct validity for criterion groups) and if it demonstrated relatively small amounts of variation in test scores as a result of (1) repeated testing (test-retest reliability), (2) different sides, and (3) independent raters (intertester reliability). The specific aims of the study were (1) to determine the ability of the measurement system to discriminate isometric force capabilities between children with and without DMD; (2) to identify sources of variation in the testing of children with and without DMD, including factors such as repeated testing and variation

Table 1. Demographic Characteristics of Subjects with and without Duchenne's Muscular Dystrophy (DMD)

                 DMD          Comparison
Characteristic   Group        Group        p (a)
Age (y)
 [unkeyable]     10.2         10.4         .42
 SD               2.8          2.8
 Range            6.2-14.0     4.8-13.4
Weight (lb (b)
 [unkeyable]     83.9         79.4         .725
 SD              32.3         24.5
 Range           43.8-127.0   47.4-118.0
Height ([in.sup.c])
 [unkeyable]     53.5         54.5         .73
 SD               6.9         6.3
 Range           43.2-63.0    43.5-62.0
Sex
  M              10            8
  F               0            2
Handedness
  R               9            9
  L               9            1
  (a) P<.05, paired t test.
  (b) 1 lb=0.4536 kg.
  (c) 1 in=2.54 cm.

between side; and (3) to estimate intertester reliability for subgroups of children with and without DMD.

Method

Subjects

All children with DMD were recruited from the Neuromuscular and Muscular Dystrophy Clinics at New England Medical Center (Boston, Mass) and local area hospitals. Any child with DMD with cardiovascular, orthopedic, or unstable medical problems; decreased attention span; or marked cognitive deficits was excluded from the study. The children without DMD who volunteered for the study were friends of patients and employees of New England Medical Center and had no history of medical or neurological problems. Prior to participation in this investigation, all subjects and their parents signed informed consent forms. A convenience sample of 10 children with a diagnosis of DMD of varying stages and 10 age-matched control children between the ages of 4 and 14 years were recruited for the study. We decided to exclude children under the age of 4 years because of the difficulty in testing young children. Additionally, we chose not to include children with DMD over the age of 16 years, as their forced capabilities are often zero to minimal on the muscle groups tested in our protocol. Although this was a convenience sample, we believe that it represents the spectrum of children we see in our clinic.

One control child was unable to complete the entire battery of force tests, and scores on three of the muscle groups from a control child of the same age were substituted for the missing data in order to maintain a balanced design for statistical analyses. Isometric force readings for one child with DMD for left knee flexion were not interpretable, so data analyzed for one muscle group using only nine paired subjects.

Because of the problems of fatigue in children with DMD and time constraints within the clinic, it was not possible to collect data from the same rater on two ocassions (within one clinic visit) as well as collect intertester tester reliability data from two different raters on all the children. A subsample of six children with DMD was recruited for intertester reliability during a separate clinical visit. Because of omitted tests or fatigue problems, interster reliability data across muslce groups were available for up to six children with DMD and five comparison children.

Table 1 describes the demographic characteristics of both groups of subjects. No significant differences in mean age were present between groups. All subjects with DMD were male, but the comparison group included two female subjects. Although seemingly problematic, there is evidence that suggests that strength differences between nondisabled boys and girls under the age of 13 years are minimal. [33] No significant differences existed between weight and height in the children with DMD and the control children.

Procedure

The height, weight, and age of each subject were recorded. In addition, Vignos classification scales [34] were completed on all children with DMD (Tab. 2, Appendix). Isometric force testing of seven muscle groups of the right and left extremities was performed in the Neuromuscular Research Unit at New England Medical Center. The following muslce groups were tested using a standard test procedure: knee flexor and extensors, hip flexors, ankle dorsiflexors, shoulder abductors, and elbow flexors and extensors (Tab. 3). The muscle groups tested were chosen based on their representation of proximal and distal musculature and their importance

Table 2. Upper- and lower-Extremity Functional Grades for Children with Duchenne's Muscular Dystrophy (DMD) (a)

             Functional Grade
Subject No.  Upper Extremity    Lower Extremity   Age (y/mo)
 1           3                  9                    14/0
 2           1                  2                   8/5
 3          3                   9                 13/2
 4           4                  9                 11/10
 5           2                  2                 11/8
 6           2                  9                 10/10
 7           1                  1                   6/3
 8           1                  2                   6/4
 9           2                  8                 11/0
10           1                  3                   7/9
  (a) Based on Vignos classification scales for children with
DMD. [34]  See Appendix for Vignos classification
scales descriptions.

in charting disease progression (S Pandya, W King, Clinical Investigation of Duchence Dystrophy [CIDD] Group; personal communication; 1989). [23.35] All testing was completed on the same day with approximately 1 1/2-hour intervals between testing sessions. Initial pilot testing of children with DMD included measurements of hip extension abduction and adduction. We did not measure hip extension because many of the children were unable to generate consistent force. Stuberg and Metcalf [23] also reported difficulty in testing hip extension, as measurements of hip extension on only 3 of 14 children could be obtained. Additionally, we were unable to isolate hip adduction and abduction using the testing device described, as many of the children substituted with their body weight, rendering the test results invalid. Peak force generated during maximal voluntary isometric contraction (MVIC) was measured using an SM-250 electronic strain gauge. (*) The strain gauge was connected to a strap that was looped around the extremity being tested. Pulling of the limb segment creates distortion of a metal coil within the strain gauge, leading to a change in the electrical resistance. This current change was amplified and recorded on an LM-24 strip chart recorder. (+) The strip chart recorder was calibrated with the strain gauge using a 10-kg weight. Forces were recorded to the nearest 0.05 kg.

Test-retest reliability was assessed by one physical therapist (CMB). Inter-tester reliability was assessed by the same physical therapist and a physical therapist assistant who had experience with force testing of patients who have neuromuscular disorders. Subjects were positioned on an adjustable table to which aluminum uprights and clamps were attached (Fig. 1). Modification of the testing equipment, as described by Andres and colleagues, [27] included a back support and stabilization straps for testing the lower-extremity muscle groups (Fig. 2). During ankle testing, a bolster was placed under the knees of all subjects to accommodate for hip and knee flexion contractures in the children with DMD. Forty percent of children with DMD had knee flexion contractures, 60% had hip flexion contractures, 70% had ankle plantar-flexion contractures, and 40% had shoulder and elbow range-of-motion limitations.

Prior to reliability testing all subjects participated in a training session consisting of one to three practice trials for each muscle group tested. This training session allowed the subjects to become familiar with the examiner and the task and to reduce the effects of learning on the test results. For each testing session, the subject was requested to perform two MVICs for each muscle group. Each MVIC was maintained until the examiner was satisfied that the force produced was no longer increasing. Continual verbal encouragement was given thoughout the contraction. If the tester was not confident that a maximal effort was made, the measurement was repeated. In addition, any jerky movements that produced a force were eliminated. A rest period of approximately 10 seconds followed each muscle contraction. Each testing session was 30 minutes in duration, with approximately 90 minutes between testing sessions. Testing was the same for each session and done in an order that required minimal position change for the subjects and instrumentation. The best of two contractions, rather than the average, was selected for analysis because of the difficulty in obtaining two good contractions from a child. Although some have argued that reliability is improved when averaging trial scores, the best of two MVICs was determined to be optimal in recording maximum force capabilities in adults with neuromuscular disease. [27] Additionally, Riddle et al [25] found no difference in reliability between the first trial and the average of repeated trials for measurements of force production obtained with a hand-held dynamometer in adults with brain damage.

Peak force measurements were taken directly from the strip chart recorder graph paper following the completion of the two tests. Maximum force measurements were obtained by using a measurement scale that corresponded to the set sensitivity of the strip chart

[TABULAR DATA OMITTED]

recorder. All upper-extremity measurements were taken from the zero baseline to the peak of the better trial. Knee flexion and extension measurements were taken from the resting tension to the peak of the better trial, and hip flexion and ankle dorsiflexion measurements were taken from the resting tension following each contraction to the peak of the better trial.

Data Analysis

A series of three-way analyses of variance (ANOVAs) were computed to analyze main effects and interactions between subject groups, sides, and tests for each muscle group. Using variance components derived from the ANOVAs, intraclass correlation coefficients (ICC[2,1]) [36] were calculated to develop test-retest and intertester estimates for individual muscle groups for each side. [37] Although various methods have been used to determine reliability, the ICC is the preferred statistic to estimate the reliability of measurements made on the same person over several trials. [37,38]

Because ICC values are a function of the variation among individuals, they may be artificially high when the range of scores among subjects is large or spuriously low ith small between-subject variability. Thus, to help interpret predicted variation for individual scores, the standard error of measurement (SEM) was also calculated. [39] The SEM is an estimate of measurement error in the metric unit of the measurement system. The SEM provides a frame of reference for interpretation

[TABULAR DATA OMITTED]

of data obtained from subjects undergoing intervention, and it provides outer bounds of the potential variation of individual scores at the 95% confidence level.

Results

A summary of the results of the series of three-way ANOVAs, showing only significant effects for each muscle group, is presented in Table 4. Significant main effects for force were found between subjects with DMD and control subjects for all muscle groups tested. Significant main effects were found between sides for elbow flexion (F=4.7, P=.04) and ankle dorsiflexion (F=9.1, P=.007). These findings indicate that one extremity side (right) consistently demnstrated greater ismetric frce readings acrss bth subject grups in elbow flexin and ankle dorsiflexion. A significant interactin was found between groups, sides, and tests for elbow extension only. This finding indicates that the control children produced greater peak force in elbow extension on the first trial on the right side, whereas children with DMD were consistently more effective on the second trial. As expected, no significant main effect was found for test-retest for either subject group.

Descriptive statistics, ICCs, and SEMs for test-retest reliability of peak force measurements for each muscle group and side are presented in Table 5. Although no single reliability coefficient is accepted as a gold standard, [40] test-retest ICCs were relatively high for all muscle groups tested in both groups of subjects. The ICCs for test-retest reliability ranged from .88 to .99 for the subjects with DMD and from .85 to .98 for the control subjects (Tab. 5). The SEMs ranged from 0.52 kg for right shoulder abduction to 2.38 kg for left hip flexion for the subjects with DMD. For the comparison group, the SEMs ranged from 1.28 kg for right shoulder abduction to 4.70 kg for left knee extension.

Intertester ICCs were somewhat lower than those found for test-retest, ranging from .74 to .97 for the children with DMD and from .71 to .98 for the control children (Tab. 6). The corresponding SEMs for intertester reliability ranged from 0.24 kg for left elobw extension to 2.14 kg for right hip flexion in the subjects with DMD. In the control subjects, the SEMs ranged from 1.00 kg for left elbow extension to 6.22 kg for right hip flexion.

Discussion

The results of this study support the use of the adapted protocol of the electronic force gauge for testing muscle force in children with and without DMD. The force-gauge system demonstrated the capability of measuring (1) small as well as large forces across subject groups, (2) relatively small amounts of variation across sides (although significant differences for two muscle groups were found), and (3) very small variation across repeated tests. In addition, data on intertester reliability on a small subsample of children indicated relatively small amounts of variation ttributable to independent raters.

Examination of the ANOVA results (Tab. 4) and the descriptive statistics for test-retest and intertester reliability (Tabs. 5 and 6, respectively) revealed large force differences in the mean values for all muscle groups between the children with DMD and the control children. This finding indicates a high degree of construct validity of this system because it distinguishes muscle force abnormalities in diseased muscle from the muscle force of the controls and thus argues for the system's usefulness in future comparison studies.

Although a small amount of measurement variation was attributable to different sides, a significant main effect for sides was noted for two of the seven muscle groups (elbow flexon and ankle dorsiflexion). For both muscle groups, mean force production was larger on the right side. Fowler and Gardiner [22] also dmonstrated greater strength on the right side of children with DMD (n=43) and children without DMD (n=45) in six muscle groups. Actually, more variation between sides might have been expected, because 9 children with DMD and 9 children without DMD were right-handed. Other studies reporting force measurements of children with DMD have not reported side differences; however, they did not examine the data separately between sides, as was done in our study. This issue may need to be examined in future studies in children, particularly if comparisons between right and left sides are made in therapeutic trials. One three-way interaction was found to be significant for elbow extension. As this occurred for only one muscle group and cannot be meaningfully attributed to testing or order effects, this isolated finding appears to be random and should not be overinterpreted.

This high degree of test-retest reliability, as evidenced by the small variation attributed to repeated testing, provides evidence that the test protocol is appropriately standardized for

[TABULAR DATA OMITTED]

children and will produce consistent results. Similar test-retest results using a force gauge have been demonstrated in adults with ALS [27] and in the knee extensor strength of 12 boys with DMD. [28] Comparative test-retest coefficients in adults reveal slightly higher correlation coefficients than in children. Because test-retest variation depends on measurement variables as well as human muscle performance, reliability would be expected to be slightly less for children. Although different muscle groups and statistical analyses were used in this study, we agree with Stuberg and Metcall [23] that factors such as cooperation, motivation, attention, and understanding of instructions are potential difficulties that make force testing more variable in children than in adults. Procedurally, it appeared that shoulder abduction and ankle dorsiflexion tests were the most difficult tests for the children to complete. This observation was partially corroborated by the data, as these muscle groups were among the most variable across the repeated tests. Tournvall [41] also noted a more extensive trial variation in isometric tests that involved stabilization of two or more joints.

The test-retest correlations in this study were slightly higher than those reported for a recent study using a hand-held dynamometer. [23] In addition, we were able to demonstrate relatively high reliability for all muscle groups in both the children with DMD and the control children. A significant advantage of using the force gauge over the hand-held dynamometer is the abiity to test muscle groups that produce large forces in nondisabled children. Researchers using hand-held instrumentation were unable to test the quadriceps femoris and hip muscles in the majority of the nondisabled children studied. [23,33] Although previous reports [33,42] have suggested that reliability of measurement is inversely proportional to the magnitude of force, our data suggest measurements of both large and small forces can be accurately recorded

[TABULAR DATA OMITTED]

with the force-gauge protocol used in this study.

The SEMs are reported for each muscle group in Tables 5 and 6 to provide an estimte of the bounds of error that could occur on one test occasion. The SEM is an important measure by which clinical change or deterioration can be properly interpreted. For example, if a force of 2.5 kg is obtained in right shoulder abduction in a child with DMD, the force in shoulder abduction would need to change at least 0.52 kg before the difference could be attributed to a changing clinical situation and not to measurement error.

As expected, the intertester reliability coefficients were not as high as the test-retest reliability coefficients. The control subjects demonstrated intertester reliability results that were the most variable. Reliability was lowest for right hip flexion (ICC=.71), right and left dorsiflexion (ICCs=.79 and .78, respectively), and right elbow extension (ICC=.78) in the control subjects. Although reliability was generally acceptable in light of relatively small between-subject variance in both groups of children, the small sample size limits the stability of these estimates.

Clinical implications regarding the possible correlation between MVIC and joint contractures cannot be easily drawn from the results of this study. In general, the children with DMD who had knee joint contractures generated less force than did those who did not have contractures. We believe an interpretation of these data will require analysis of force in various age groups, with a comparison with nondisabled children and a larger sample. Also required are an examination of functional level and of degree and amount of exercise and an understanding of the lever system on which the muscle acts when contractures are present. A previous study on children with DMD did not find a strong correlation between amount of contracture and amount of force generation at the knee. [22] Future studies should reexamine possible relationships between contractures and MVIC at the knee and other joints.

Conclusions

The results of this study suggest that the use of an electronic strain gauge with a standardized administrative protocol is a reliable and valid method for measuring isometric force in children with and without DMD. Quantitative measurements obtained with the force gauge will improve the quality of dependent measures in natural history and clinical trial studies of DMD. the high levels of variability obtained across subject groups and the relatively small amount of variation detected across sides, over repeated tests, and between raters in seven muscle groups support the assertion that the adapted protocol provides objective measurements in children with varying force capabilities.

The described procedure is currently limited for use in a physical therapy clinic, because it requires a setup of the equipment as depicted in Figures 1 and 2. A potential advantage of the system is its relatively low cost as compared with currently available isokinetic dynamometers, although the force-gauge system is restricted to isometric force measurements. Results obtained from the testing procedure, however, will provide the clinician with potentially valuble information with regard to treatment planning, effectiveness of therapeutic exercise programs, and the nature of disease progression.

Further research should focus on intertester reliability with a larger sample to obtain more stable estimates of variability among independent testers. Additionally, force measurements should be repeated separately on homogeneous clinical and normative samples to further examine the issue of measurement error in different age groups and with different magnitudes of force. Future research should also focus on obtaining the moment arm length in order to calculate torque. Considerations of the growth of the limb in serial force measures may be important in the longitudinal documentation of force changes in children.

Acknowledgments

We thank the staff of the Neuromuscular Research Unit, New England Medical Center, and especially Brenda Thornell, PTA, for their support and assistance with data collection. We also thank Melvin Meister for his technical assistance, the therapists of the CIDD Group for their valuable comments with protocol design, and David E Krebs, PhD, PT, for his assistance with statistical analyses.

References

[1] Koenig m, Hoffman EP, Betelson CFJ, et al. Complete cloning of the Duchenne muscular dystrophy (DMD) cDNA and preliminary genomic organization of the DMD gene in normal and affected individuals. Cell. 1987;50:509-517.

[2] Partridge TA, Morgan JE, Coulton GR, et al. Conversion of mdx myofibres from dystrophin-negative to -positive by injection of normal myoblasts. Nature. 1989;337:176-179.

[3] Munsat TL, Balch R, Pearson CM, et al. Serum enzyme alterations in neuromuscular disorders. JAMA. 1973;226:1536-1543.

[4] Thomas FM, Ballard FJ, Pope LM. Age-dependent changes in the rate of myofibrillar protein degradation in humans as assessed by 3-methylhistidine and creatine excretion. Clin Sci. 1979;56-341-346.

[5] Brook MH, Griggs RC, Mendell JR, et al, Clinical trial in [DD.sub.1]: the design of the protocol. Muscle Nerve. 1981;4:186-197.

[6] Madsent KS, Miller JP, Province MA, et al. The use of an extended baseline period in the evaluation of treatment in a longitudinal DMD trial. Stat Med. 1986;5:231-241.

[7] Mendell JR, Province MA, Moxley RT, et al. Clinical investigation of DMD: a methodology for therapeutic trials based on natural history controls. Arch Neurol. 1971;44:808-811.

[8] Brooke MH, Fenichel GM, Griggs RC, et al. Clinical investigation in Duchenne dystrophy: determination of the "power" of therapeutic trials based on the natural history. Muscle Nerve. 1983;6:91-103.

[9] Florence JM, Pandya S, King WM, et al. Clinical trials in Duchenne dystrophy: standardization and reliability of evaluation procedures. Phys Ther. 1984;64:41-45.

[10] Daniels L, Worthingham C. Muscle Testing: Techniques of Manual Examination. Philadelphia, Pa: WB Saunders Co; 1972.

[11] Kendall H, Kendall F, Wadsworth G. Muscle Testing and Function. 2nd ed. Baltimore, Md: Williams & Wilkins; 1971.

[12] Medical Research Council. Aids to Investigation of Peripberal Nerve Injury: War Memorandum No.7. 2nd rev ed. London, England: His Majesty's Stationary Office; 1943.

[13] Ziter FA, Allsop KG, Tyler FH. Assessment of strength in DMD. Neurology. 1977;27:981-984.

[14] Aitkens S. Lord AF, Bernauer E, et al. Relationship of manual muscle testing to objective strength measurements. Muscle Nerve. 1989;12:173-177.

[15] Heckmatt JZ, Hyde SA, Gabain A, et al. Therapeutic trial of isaxonine in Duchenne muscular dystrophy. Muscle Nerve. 1988;11:836-847.

[16] Rubinstein RY. Simulation and Monte Carlo Method. New York, NY: John Wiley & Sons Inc; 1981.

[17] Beasley WC. Quantitative muscle testing: principles and applications to research and clinical services. Arch Phys Med Rebabil. 1961;42:398-425.

[18] van der Ploeg PJO, Oosterhuis HJ, Reuvekamp J. Measuring muscle strength. J Neurol. 1984;231:200-203.

[19] Edwards RHT, Hyde S. Methods of measuring muscle strength fatigue. Physiotherapy. 1977;63:51-55.

[20] Cook JD, Glass DS. Strength evaluation in neuromuscular disease. Neurol Clin. 1987; 5:101-123.

[21] LaRocca NG. Statistical land methodologic considerations in scale construction. In: Munsat TL, ed. Quantification of Neurologic Deficit. Boston, Mass: Butterworth-Heinemann; 1989:50.

[22] Fowler WM, Gardiner GW. Quantitative strength measurements in DMD. Arch Phys Med Rebabil. 1967;48:629-644.

[23] Stuberg WA, Metcalf WK. Reliability of quantitative muscle testing in healthy children and in children with Duchenne muscular dystrophy using a hand-held dynamometer. Phys Ther. 1988;67-977-982.

[24] Bohannon RW. Test-retest reliability of hand-held dynamometry during a single session of strength assessment. Phys Ther. 1986;66:206-209.

[25] Riddle DL, Finucane SD. Rothstein JR, Walker ML. Intrasession and intersession reliability of hand-held dynamometer measurements taken on brain-damaged patients. Phys Ther. 1989;69:182-194.

[26] Scott OM, Hyde SA, Goddard C, et al. Quantitative muscle function in chicken: a prospective study of DMD. Muscle Nerve. 1982;5-291-301.

[27] Andres PL, Hedlund W, Finison L, et al. Quantitation of motor assessment in ALS. Neurology. 1986;36:937-941.

[28] Wiles CM, Karni Y. The measurement of muscle strength in patients with peripheral neuromuscular disorders. J Neurol Neurosurg Psychiatry. 1983;46:1006-1013.

[29] Burnett CN, Betts EF, King WM. Reliability of isokinetic measurements of hip muscle torque in young boys. Phys Ther. 1990;70:244-249.

[30] Andres PL, Thibodeau LM, Finison LJ, et al. Quantitative assessment of neuromuscular deficit in ALS. Neurol Clin. 1987;5:125-143.

[31] Rothstein JM. Measurement in clinical practice: theory and application. In: Rothstein JM, ed. Measurement in Physica; Therapy. New York, NY: Churhchill Livingstone Inc; 1985;7:8-10.

[32] Mitchell SK. Interobserver agreement, reliability and generalizability of data collected in observational studies. Psychol Bull. 1979;86:376-390.

[33] Hosking GP, Bhat US, Dubowitz V, et al. Measurement of muscle strength and performance in children with normal and diseased muscle. Arch Dis Child. 1976;51:957-963.

[34] Vignos PJ, Spencer GE, Archibald KC. Management of progressive muscular dystrophy in childhood. JAMA. 1963;184-89-96.

[35] Edwards RHT, Chapman SJ, Newham DJ, et al. Practical analysis of variability of muscle function measurements in Duchenne muscular dystrophy. Muscle Nerve. 1987;10:6-14.

[36] Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428.

[37] Krebs DA. Computer communication: intraclass correlation coefficients--use and calculation. Phys Ther. 1984;64:1581-1589.

[38] Bartko JJ. The intraclass correlation coefficient as a measure of reliability. Psychol Rep. 1966;19:3-11.

[39] Ghiselli EE, Campbell JP, Zedck S. Measurement Theory for the Behavioral Sciences. San Francisco, Calif: WH Freeman & Co Publishers; 1981.

[40] Cronbach LJ. Test "reliability": its meaning and determination. Psychometrika. 1947;12:1-16.

[41] Tournvall G. Assessment of physical capabilities with special reference to the evaluation of maximal voluntary isometric muscle strength and maximal working capacity. Acta Physiol Scand [Suppl]. 1963:201.

[42] Bohannon RW, Andrews AW. Interrater reliability of hand-held dynamometry. Phys Ther. 1987;67:931-933.