Long-term stability of hand-held dynamometric measurements in children who have myelomeningocele.

The measurement of any variable expected to change as the result of treatment or a disease process is vital for clinical documentation and research. Muscle performance is frequently measured. In recent years, the hand-held dynamometer has been used to measure sure forces produced during isometric contractions of muscles, but its clinical application and evaluation has been limited largely to adult populations.[1-4]

Objective methods of monitoring muscle performance are necessary to evaluate treatment effectiveness and patient status. Increasing muscle force production is a common objective of physical therapy programs for children having myelomeningocele. Adequate upper-extremity muscle performance is necessary for the child with myelomeningocele to successfully ambulate using crutches or for efficient wheelchair propulsion. Adequate lower-extremity muscle performance is necessary for ambulation and transfers.

Sensitive, objective measures of muscle performance in children with myelomeningocele are necessary to provide data on deterioration of their already impaired lower-extremity musculature. Recent reports suggest that muscle weakness,[5] especially asymmetric weakness,[6] may indicate tethered cord syndrome. Surgical intervention might be necessary to prevent further deterioration of muscle performance attributable to tethered cord syndrome in children with myelomeningocele. Therefore, physical therapists must detect even slight decrements in muscle performance and refer these children for appropriate intervention.

Measurement tools used to assess muscle performance should be reliable. Several factors are involved in evaluating the reliability, or consistency and stability, of a measure. The hand-held dynamometer has been shown to have relatively high intrarater and test-retest reliability in selected populations. Bohannon[1] found high intrarater reliability for measurements obtained during a single session when the hand-held dynamometer was used on adult patients with neurological impairment. He found lower reliability for measurements of hip and shoulder abduction. Riddle and colleagues'[3] replication of Bohannon's study supported Bohannon's findings of high intrarater reliability when testing sessions were conducted up to 2 days apart. Riddle et al did not test hip or shoulder abduction, however, so a comparison with the Bohannon study for the muscle groups that had low reliability cannot be made.

Stuberg and Metcalf[4] found high intrarater reliability over a 2-day period when testing children without disease or disability and children with muscular dystrophy.[4] They and others[7] reported that children with disease or disability could exceed the upper limit of the dynamometer on hip and knee extension. Stuberg and Metcalf chose to vary Bohannon'[1] shoulder abduction test position by testing shoulder abduction with the subjects' upper extremity at their side and not at 45 degrees of abduction.

Interrater reliability of hand-held dynamometry has been studied by Bohannon and Andrews[2] on 30 adults, the majority of whom were post-cerebrovascular accident. Reliability was high for four of the six muscle groups tested. Lower reliability was found for measurements of the shoulder lateral (external) rotators and the wrist extensors. Shoulder and hip abduction and hip and knee extension, for which lower intrarater reliability have been reported, were not tested. High interrater coefficients (Pearson r=.87-.97) were reported when testing eight muscle groups in 66 subjects with myelodysplasia aged 5 to 21 years.[7]

Long-term intersession reliability or stability[8] has just begun to be studied. Most authors report intrasession results[1-3] or results obtained at 2-day intervals.[3,4] The longest interval studied was 7 days.[7] Riddle et al[3] studied intrasession and intersession reliability or stability using the hand-held dynamometer on 10 muscle groups in two different groups of adults with brain damage. During the intrasession trials, three measurements were taken several seconds apart, and the intersession trial measurements were taken 48 hours apart. Intrasession measurements on both paretic and nonparetic limbs and intersession measurements on paretic limbs were highly reliable. Intersession reliability on the nonparetic limb was poor for the wrist extensors (intraclass correlation coefficient [ICC] = .38, coefficient of determination [r.sup2] = .38) and the knee extensors (ICC = .31, [r.sup.2] = .08).

Reliability, both intrasession and intersession, can be assessed two ways: (1) by measuring the association (correlation) between die repeated measures across time on the same subjects and (2) by measuring the agreement (consistency) of the scores across time.[9] Each statistical approach has its strengths and weaknesses. The Pearson Product-Moment Correlation Coefficient (r) is

... an index of the degree of association

between two [test sessions] or the

extent to which the order of individuals

on one [test session] is similar to

the order of individuals on a second

[test session].[10]

Rothstein has noted that

... a perfect correlation (of 1.00) does

not mean that the paired [measurements]

are identical, but rather that

they are perfectly related.[11(pl7)]

A linear association between the measurements at the two test sessions is the only requirement for a high correlation.

The coefficient of determination ([r.sup.2]) is the proportional measure of the percentage of variation "explained" by the association between two sets of measurements. The coefficient of determination indicates how closely one measurement predicts the other measurement. Although the Pearson correlation coefficient must be used with the sample size to determine the statistical significance of the association between two measurements, the relationship between the two measurements is not due to chance. The coefficient of determination directly indicates the strength of the association.

To determine the consistency of two or more scores between test sessions, an ICC is used. The ICC measures "the extent to which multiple measures agree ... rather than are associated (ie, covary)."[11](p39) The ICC compares the between-subject variance with the within-subject variance. The ICC is the relative amount of variation from the combined mean of the two test sessions explained by differences between the subjects.[9]

The ICC is a comparison of the between-subject and within-subject variance; therefore, it accounts for chance correspondence of the variance between the two measurements. In contrast, the Student's t test for the equivalency of two population means compares the intrasession difference in the means against the average sample variance. This accounting for chance agreement makes the ICC the preferred statistic for comparing the agreement between two test sessions.[9] The ICC does not assess whether the subjects maintain the same relative ranking between the two test sessions. A measure of force production that cannot maintain the same relative ranking between weaker and stronger subjects is not reliable. Therefore, a measure of association (correlation) between the two measurements, such as the Pearson Product-Moment Correlation Coefficient, should also be used. The purpose of this article is to report on the long-term stability of measurements of isometric force production obtained with a hand-held dynamometer for 30 muscle groups of children with myelomeningocele after a 23-day camp experience. The data were collected as part of a larger study of the effect of the camp experience on cardiopulmonary function.

Method

Subjects

Twelve children with myelomeningocele (3 boys, 9 girls), who attended the summer Delaware Valley Variety Club Sunshine Camp (Worchester, Pa), participated in the study. The children, aged 9.8 to 17.4 years (X = 13.65, SD = 2.64), attended one of two 23-day camp sessions. Written informed consent was obtained from both the children and their parents or guardians.

Procedure

A Spark Model 160 hand-held dynamometer(*) was used to test the isometric force production (in kilograms of the children's upper- and lower-extremity muscle groups. Upper-extremity muscle testing included tests of the shoulder abductors, flexors, extensors, and medial (internal) and lateral rotators; the elbow extensors and flexors; and the wrist extensors and flexors. When possible, based on the child's level of spinal lesion, the lower-extremity muscle testing included tests of the hip extensor's, flexors, abductors, and adductors and the knee extensors and flexors.

All testing was performed in a small room on an exercise mat or chair, as required. We used the make-test procedure as described by Bohannon and Andrews.[2] The make test consists of a gradual increase in force for 1 to 2 seconds, followed by an isometric hold for 4 to 5 seconds. The children were given strong verbal encouragement and praise throughout each test session.

The children's position during testing of each muscle group followed Bohannon's protocol.[1] Bohannon did not state whether the subject's leg should be supported on the examiner's arm during the tests of the hip flexors and extensors. We allowed each child's leg to hang freely adjacent to the dynamometer during tests of the hip flexors and extensors. The weight of the dangling leg prevented hip flexion in three children, and two children were unable to obtain the required 90 degrees of hip flexion for testing.

Muscle groups on both sides of the body were tested. Muscle groups were tested in die order shown previously. At the precamp test session, the side of the body to be tested first in both test sessions was selected randomly. Each child's highest force produced for each muscle group was recorded, allowing three trials per position. The same test procedures were repeated 23 days later at the end of the camp session. The same examiner (DAB), who was blind to previous measures, conducted all muscle performance tests at both sessions. In addition, interrater reliability was assessed on randomly selected muscle groups on each day of testing. The dynamometer was calibrated, using known weights, before each day of testing.

Data Analysis

Two researchers independent verified the data entry for accuracy We used the Statistical Analysis System, (dagger) release 6.03, on a microcomputer, (double dagger) to analyze the data. We performed statistical tests on individual muscle group performance scores, total right and left upper-extremity performance scores, total right and left lower-extremity performance scores, and total performance scores for all extremities. The total upper=extremity performance scores were the sum of each child's shoulder, elbow, and wrist performance scores. The total lower=extremity performance scores were the sum of each child's hip and knee muscle group performance scores. Each child's total extremity performance score was the sum of the total upper- and lower-extremity performance scores, as described by Bohannon and Dubuc.[12]

Two statistical tests were used to analyze the stability of our test results: (1) Pearson Product-Moment Correlation Coefficients (r), as a measure of the association between the precamp and postcamp scores (consistency of relative ranking), and (2) ICCs (1,1), as a measure of the agreement between the precamp and postcamp scores (consistency of absolute value). We also determined the corresponding coefficient of determination (r2).[13] We computed the ICC for each measurement from the mean square values of the nested model II analyses of variance for both sets of measurements and calculated the corresponding F-test values for statistical significance.[14]

Results

Individual Upper-Extremity

Muscle Group Tests

The Pearson correlations for the individual upper-extremity muscle group tests are presented in Table 1. The coefficients of determination ranged from .96 for the right shoulder medial rotators to a low of .36 for the right wrist extensors. The mean average for the 18 individual upper-extremity muscle group coefficients of determination was.70 (SD = .20).

[TABULAR DATA OMITTED]

The ICCs for the individual upper-extremity muscle group measurements are presented in Table 2. The ICCs ranged from .99 for the right shoulder abductors to a low of .75 for the right wrist extensors.

[TABULAR DATA OMITTED]

Individual Lower-Extremity

Muscle Group Tests

The Pearson correlations for the individual lower-extremity muscle group tests are shown in Table 3. As expected in individuals with myelomeningocele, as the more distal muscles were tested, fewer children were able to participate in the study, resulting in the smaller sample size. Related coefficients of determination ranged from .87 for right hip flexion to .40 for left hip adduction. The mean average for individual lower-extremity muscle group coefficients of determination was.65 (SD = .16).

[TABULAR DATA OMITTED]

The ICCs for the individual lower-extremity muscle group tests are shown in Table 4. The ICCS ranged from .96 for right hip flexion to .73 for left knee flexion. The mean average for the individual lower-extremity muscle group ICCS was.86 (SD = .07).

[TABULAR DATA OMITTED]

Discussion

This study assessed the long-term stability of isometric force production measurements obtained in muscles of children who have myelomeningocele. Isometric force production was measured using a hand-held dynamometer. Isometric force did not change after 23 days at camp, and there was remarkable stability in the test-retest measurements in the majority of the 30 muscle groups tested. Additionally, the few muscle groups in which reliability was lower were those muscle groups previously reported to have lower reliability when tested within a single session or over a 2-day period.

Some of the children expressed discomfort when pushing against the unpadded test bar with their wrist. Bohannon[15] now recommends adding high-density foam rubber to the end pieces of the dynamometer. We believe a foam-padded end piece would have allowed for greater isometric force production by the wrist muscles and probably higher reliability of the measurements.

In seven of the nine upper-extremity muscles tested precamp, the right upper-extremity muscles produced greater isometric force. This finding was anticipated because 83% of the subjects were right-hand dominant.

Four of the 12 lower-extremity muscle groups had ICC values of .73 to .83. We followed Bohannon's procedure and did not support the leg when testing the hip extensors.[1] The dangling leg might have contributed to the lower reliability of the left hip extensor measurements; however, that assumption is questionable, because right hip extensor measurements had good or excellent (ICC = .94) reliability. Stuberg and Metcalf's procedure for supported leg positioning would appear to result in greater reliability.[4]

We did not find greater stability of measurements in the "paretic limbs" than in the "nonparetic limbs," as did Riddle et al.[3] The children in our study have never experienced normal movement in their legs, unlike the patients with traumatic brain injury in the study by Riddle et al, who had normal movement until the time of their injury. This factor perhaps accounts for the difference in findings.

Strong adolescents would occasionally overpower the tester in our study, and extra stabilization was therefore required. Researchers[2,16,17] have noted that the stabilazation of the dynamometer and the subject's extremity is essential for obtaining reliable measurements when using the hand-head dynamometer. Weak testers experience more reliability problems because of problems with stabilization than do stronger testers.[18] Even with more careful stabilization, the stronger adolescents exceeded the 27.2-kg upper limit of our hand-held dynamometer eight times.

The findings of long-term stability of muscle performance using a hand-held dynamometer in the majority of the muscles tested was a surprise to the examiners. Twenty-three days is an exceptionally long period of time for muscle performance to remain stable. Such long-term stability of muscle performance is especially remarkable in children with neurological impairment who are attending camp. The activities of camp might be expected to have a training effect, which would increase muscle performance, or a fatiguing effect, which would decrease performance. We now realize that our good to excellent reliability probably could easily be improved. The dynamometer we used was large, given the size of our children, the pads were uncomfortable, and the did was difficult to read. We think these correctable mechanical factors are probably the source of the variability in our study, although experiences at summer camp cannot be ruled out. We believe that, with proper stabilization, a strong tester, modification of the hip abductor and extensor testing positions (eg, use of Stuberg and Metcalfs procedure for supported leg positioning[4], padding the dynamometer end pieces, and use of a smaller, digital dynamometer with higher upper limits, highly reliable measurements of isometric force production can be documented over time. Extensive positioning and stabilizing protocols and exact determination of lever arm length might not be warranted in clinical assessments utilizing the hand-held dynamometer.

Future research should focus on the relationship of muscle performance to functional skills. In children with myelomeningocele, the accurate assessment of decreases in muscle performance is critical for early detection of tethered cord syndrome. Well-designed studies, on adequate size populations, are imperative. At present, a growing body of literature[4-7] supports use of the hand-held dynamometers to document change in isometric force production in pediatric populations.

Conclusion

The performance of 30 individual muscle groups in 12 children with myelomeningocele was assessed over a 23-day interval using a hand-held dynamometer. The majority of the muscle groups tested had excellent stability, based on the statistical indicators of association (Person r) and agreement (ICC).[14] Muscle groups having lower, long-term stability were the right and left wrist extensors and flexors, the left hip adductors and extensors, the left knee flexors, and the right and left knee extensors. The individual upper-extremity muscle groups had higher long-term stability than did the lower-extremity muscle groups. Lower reliability was possibly due to the poorer test-retest reliability of the weaker paretic muscle groups, fewer children capable of performing the lower-extremity tests and therefore lower statistical power, and mechanical problems with the dynamometer. Our dynamometer was large compared with the children's limbs and had hard end pads, a difficult to-read dial, and an upper limit that could be exceeded.

Additional study is warranted to determine the long- term stability of isometric force production as measured by a hand-held dynamometer. This study, over a rather long period of time, demonstrated that the dynamometer yielded highly reliable results when used to assess muscle groups in which reliability was documented to be high over short periods of time. Improved reliability probably could be obtained by supporting the lower extremity during hip extensor tests; by padding the dynamometer end pieces, especially when testing over bony prominences; and by using a smaller, digital dynamometer.

(*) Spark Instruments and Academics Inc, PO Box 5123, Coralville, IA 52241. (dagger) SAS Institute Inc, PO Box 8000, Cary, NC 27511. (double dagger) IBM-PC #3286, International Business Machines Corp, PO Box 1328-W, Boca Raton, FL 33429.

References

[1] Bohannon RW. Test-retest reliability of hand-held dynamometry during a single session of strength assessmment. Phys Ther. 1986;66:206-209. [2] Bohannon RW, Andrews AW. Interrater reliability of hand-held dynamometry. Phys Ther. 1987,67:931-933. [3] Riddle DL, Finucane SD, Rothstein JM, Walker ML Intrasession and intersession reliability of hand-held dynamometer measurements taken on brain-damaged patients. Phys Ther. 1989;69:182-189, 192-194. [4] Stuberg WA, Metcalf WK. Reliability of quantitative muscle w" m healthy children and children with Duchenne muscular dystrophy using a hand-held dynamometer. Phys Ther. 1988;68:977-982. [5] Pang D, Wilburger JE. Tethered cord syndrome in adults. J Neurosurg. 1982;57:3247. [6] Alexander MA, Steg NL Myelomeningocele: comprehensive treatment. Arch Phys Med Rebabit 1989;70:6374AI. [7] Hinderer K, Deltz J, McMilla J, Shurtleff D. Reliabilility of the myometer in muscle testing individuals with myelodysplasia. Phys Ther. 1988;68:817. Abstract. [8] Isaac S, Michael WB. Handbook in Research and Evaluation. San Diego, Calif: EdITS: 1981:125. [9] Bartko JJ, Carpenter WT. On the methods and theory of reliability. J Nerv Ment Dis. 1976;163:307-317. [10] Ghiselli EE, Campbell JP, Zedeck S. Measurement Theory for the Behavioral San Francisco, Calif:. W H Freeman & Co Publisher; 1981. [11] Rothstein JM. Measurement and clinical practice: theory and application. In: Rothstein JM, ed. Measurement in Physical Therapy. New York, NY: Churchill Livingstone Inc; 1985;7: 1-46. [12] Bohannon RW, Dubuc WE. Documentation of the resolution of weakness in a patient with Guillain-Barre syndrome. Phys Ther 1984;64: 1388-1389. [13] Kerlinger FN. Foundations of Behavioral Research Educational, Psychological, and Sociological Inquiry 2nd ed. New York, NY: Holt, Reinhart & Winston General Book; 1973. [14] Fleiss JL. The Design and Analysis of Clinical Experiments. New York York, NY: John Wiley & Sons Inc; 1986:1-32. [15] Bohannon RW. Commentary on "Intrasession and intersession rehability of hand-held dynamometer measurements taken on brain-patients." Phys Ther. 1989.,69:190-192. [16] Lamb RL Manual muscle testing. In: Rothstein JM, ed. Measurement in Physical Therapy. New York, NY: Churchill Livingstone Inc; 1985;7:47-55. [17] Mayhew TP, Rothstein JM. Measurement of muscle performance with instruments. In: Rothstein JM, ed. Measurement in Physical Therapy New York, NY: Churhill Livingstone Inc; 1985;7:91-94. [18] Wikholm JB, Bohannon RW. Handheld dynamometer measurements: tester strength makes a difference. Journal of Orthopaedic and Sports Physical Therapy. 1991; 13:191-198.