Descriptive statistics (mean and standard deviation) of MIT were calculated for all muscle groups, for all age groups for both the boys and the girls. To determine the effect on MIT of the side tested, of the participant's sex, and of his or her age, a 3-way ANOVA (side, sex and age group) was used. For the adolescent subgroup that had repeated testing, the intraclass correlation coefficient (ICC) and its 95% confidence interval (CI) was used to quantify the intra- and interrater reliability as well as concurrent validity between the measurements obtained with the HHDs and the Cybex.14 For interrater reliability, a 2-way random effects model was used and referred to as an ICC 2, 1, according to the Shrout and Fleiss15 convention. Using the same convention, for the intrarater analysis, an ICC 1, 1 (1-way random effects model) was used and for concurrent validity an ICC 3, 1 (1-way random effect model where systematic differences between the 2 measures is not considered error) was used. For this same subgroup, the measurement error was estimated by calculating the SEM.16 The convention of Rosner17 was used to classify the strength of the ICCs (ICC < 0.40 = poor reliability, 0.40 ≤ ICC < 0.75 = fair to good reliability, and ICC ≥ 0.75 = excellent reliability). Significance was set at α < .05. Statistical analysis was performed using SPSS (SPSS 11.0 for Windows).
Subject characteristics are shown in Tables 2 and 3. All MIT values were obtained using the same HDD protocol. We did not have to adapt the instructions given to the participants, the positions, or the number of tests for a particular muscle group according to the age of the participant or for any other reason. There were also no reports of pain or discomfort during the testing. The mean ± SD for the MIT values of the 11 muscle groups tested for boys and girls for the 9 age groups are shown in Figures 2, 3, and 4. Overall, the MIT varied significantly according to age (P < .05). From the 11.5 to 12.9 year old group onward, for overall muscle strength progression, for all muscle groups, there was a change in strength progression across the older age groups. From that point onward, strength more abruptly increased for each older age group when compared to below that point (Figures 2, 3, and 4) as though the 11.5-12.9 year old group category was a threshold period. Boys were slightly stronger than girls over all (P < .05) and there was no difference between the right and the left sides.
Table 4 summarizes the intra- and interrater reliability statistics, the SEMs, and the concurrent validity statistics for each muscle group. According to a conservative estimate of strength of the ICC statistics (the lower limit of the 95% CI for a given muscle group), and the classification of Rosner,17 4 muscle groups clearly showed excellent inter- and intrarater reliability: the shoulder lateral rotators, the elbow extensors, the hip flexors, and the knee flexors (lower limit for the 95% ICC CI varied from 0.75 to 0.92). Five muscle groups showed fair to good intra- and interrater reliability: the shoulder abductors, the elbow flexors, the hip abductors, the knee extensors, and the ankle plantar flexors (lower limit for the 95% ICC CI varied from 0.41 to 0.71). One muscle group, the ankle dorsiflexors, showed poor intra- and interreliability (lower limit of the 95% ICC CI = 0.36 and 0.11 for intra- and interrater reliability, respectively). The results for one other muscle group, the hip extensors, showed poor interrater reliability, but fair to good intrarater reliability (lower limit of the 95% ICC CI = 0.27 and 0.45, respectively). A conservative estimate of concurrent validity with Cybex values of MIT determined with our protocol showed excellent concurrent validity for 4 muscle groups: the shoulder lateral rotators, the elbow flexors and extensors, and the knee flexors (lower limit for the 95% ICC CI varied from 0.79 to 0.86). Concurrent validity was fair to good for 5 other muscle groups: the shoulder abductors, the hip flexors and extensors, the knee extensors, and the ankle dorsiflexors (lower limit for the 95% ICC CI varied from 0.52 to 0.74). For 2 muscle groups (ankle plantar flexors and hip abductors), concurrent validity was poor (lower limit for the 95% ICC CIs were 0.06 and 0.37, respectively). As shown in Table 5, the SEM, relative to the maximal torque recorded for each corresponding muscle group and gender, was small, varying from 0.5 to 6.0 Nm. The SEM ratios between boys and girls showed that, on average, the SEM was 1.4 times greater in boys than in girls, when including the ankle plantar flexors.
With our HHD protocol, it was feasible to quantify the maximal isometric muscle strength of several upper and lower limb muscle groups in boys and girls over a wide age range. For the adolescent subgroup, for the majority of the muscle groups tested, our MIT values showed acceptable intra- and interrater reliability and concurrent validity with the Cybex. Using a conservative estimate of reliability, the lower limit of the 95% CI for a given muscle group, and the interpretation of Rosner,17 we found that only 1 of the muscle groups tested using our protocol may have poor intra- and interreliability: the ankle dorsiflexors. This was possibly because it is difficult to consistently resist the dorsiflexion force vector using the same dynamometer orientation relative to the plane of movement. Because of the oblique orientation of the surface of the foot relative to the plane of movement and the axis of the tibia when the ankle is maintained at 90° of dorsiflexion, which may slightly change from trial to trial, the force recorded may vary. Placing the body of the dynamometer parallel to the long axis of the tibia (so parallel to the plane of movement) to oppose the force vector developed by the ankle dorsiflexor (instead of at 90° relative to the dorsal surface of the foot) may lead to improved reliability. The results for one other muscle group, the hip extensors, also show that its interrater reliability may be poor. Given that this muscle group showed fair to good intrarater reliability; further research is required to determine whether interrater reliability for this muscle group might be improved with further standardization of the protocol. All of the remaining muscle groups tested showed at least fair to good intrarater and interrater reliability. Given that it may be conservative focusing on the lower limit of the 95% ICC CI to interpret our reliability statistics, we consider our results sufficiently convincing to pursue the continued development and evaluation of our protocol.
Our mean intrarater reliability coefficients (0.67-0.98) for upper and lower limb muscles are similar to those reported in the literature (0.72-0.99) for typically developing, albeit somewhat younger (6- to 10-year-old) children.18–22 Our mean interrater reliability coefficients (0.67-0.96), however, appear to be somewhat higher than those (0.49-0.95) reported in the literature.18 The authors of this previous study, however, cited the different levels of experience of their evaluators as a possible source of variability.18 In our study, on the contrary, the evaluators were both experienced. This difference in interrater reliability may also be related to our participants being somewhat older than the 6- to 10-year-olds in the previous studies. Differing abilities of the evaluators to resist the participant, however, is an obvious source of interrater variability, independent of evaluator experience or the participants' age. In a study with youth with cerebral palsy, Verschuren et al23 reported similar reliability between their make and break tests. While it was not the purpose of this present study to compare make and break tests, there are some notable differences between our make protocol and that of Verschuren et al,23 which may shed some light on why make tests can be more controlled, and thus more reliable than break tests. For example, (1) the hip extensors were evaluated in this latter study using a compression (push) mode as opposed to the distraction (pull) mode used in our study, (2) for the hip abductors, the contralateral, nontested lower limb did not appear to be stabilized in the same manner in the previous study as in our protocol, where we used a strap to stabilize the nontested leg, and (3) for the knee extensors, the other authors used a conventional compression/push technique whereas with our protocol, the resistance was applied through the dynamometer, which was inserted between the leg and a strap, meaning that the examiner did not have to resist, which ensures a true make (isometric) test. When performing make tests with strong muscle groups, such as the 3 mentioned earlier, it may be tempting for the examiner to push hard and “unconsciously” slightly exceed the force of the person being evaluated, which would create a quasi-break test condition, that is, lead to a nonisometric condition. Thus, what is unique, in our opinion, about our protocol and not, to the best of our knowledge, reported in the literature was this use of straps and a hook and a pull technique for the assessment of lower limb muscle groups. For the knee extensors, the HHD was inserted distally between the leg and the strap that was attached to the table in a direction parallel to the floor and perpendicular to the leg. This technique is quick, easy, and safe, and, in our clinical experience, has allowed us to resist muscle forces as high as 900 N. As noted in the “Methods” section, to resist the strong hip flexors and extensors, we also used straps combined with a pull technique. This pull technique, which can be achieved only with a push-pull HHD and lateral grips, offers the evaluator an important mechanical advantage over the participant's force because the evaluator is able to resist using both hands and his or her full body weight, while remaining stable with a wide base of support.
Isometric strength measurements obtained with HHD are strongly correlated to strength measures obtained with isokinetic dynamometers.24,25 In the present study, we had similar results to those previously reported in the literature. Using a conservative estimate of the strength of our ICC values (the lower limit of the 95% confidence interval), the concurrent validity of our HHD protocol with a gold standard, the Cybex, was fair to excellent, varying from 0.52 to 0.86, when excluding the plantar flexors and the hip abductors. The poor conservative estimate of the ICC (0.06) obtained with the plantar flexors was not surprising. We had anticipated some difficulty, on the basis of our pilot work, regarding stabilizing the ankle movement when testing this muscle group with older boys. Indeed, with our stronger participants, it was almost impossible to maintain the ankle joint in the exact same position at 90° of dorsiflexion throughout the tests. In our opinion, this explains the low concurrent validity between the Cybex and the HHD technique for this muscle group. Thus it may not be valid to use our protocol to assess plantar flexor muscles with older subjects, perhaps older than 10 years. As observed in Figure 4, there is a strong and rapid increase in ankle plantar flexor muscle strength after this age group in boys and from 11.5 years for both genders. For the hip abductors, the low value for the lower limit of the 95% confidence interval (0.37) is probably explained by the constraint of using the Cybex dynamometer as the criterion measure, in that the hip abductors can be tested only in the side-lying position when using the Cybex. The side-lying method, compared with the supine position used with the HHD, is an unstable position for the individual in that he or she must actively stabilize the trunk and pelvis while contracting the hip abductors. Since the task therefore differs between the 2 strength-testing protocols, we believe that it is more than likely that muscle fiber recruitment and thus force production would differ somewhat between the 2 methods. As the lower limit of the 95% CI for the ICC value is so close to Rosner's17 “acceptable” criterion of 0.40, however, we would suggest that replication of our work is required before it can be definitively stated that, for the hip abductors of adolescents developing typically, HHD does not show concurrent validity with isokinetic strength as measured by the Cybex.
We found that the strength of all muscle groups progressively increased with age over our age range, with no signs of a plateau. There was no difference between a participant's right and left sides for the muscles of either the upper or the lower limbs. This suggests that the effect of limb dominance may not be particularly important below the age of 17.5 years in healthy subjects. Further work is required to explicitly test this hypothesis. Overall, boys were slightly stronger than girls, and this difference appears to be more important for the vast majority of muscle groups after the age of 14.5 years. Since strength rapidly increases in adolescents, the choice of using a 1.5 years window per age group might not be appropriate. Age groups of 1 year would possibly allow for better discrimination of strength progression with age, especially through the circum-pubertal years. MacFarlane et al,20 who used positions similar to ours, reported mean MIT values (6- to 8-year-olds) of 34.6, 22.8, and 26.9 Nm for the knee extensors, knee flexors, and hip abductors, respectively. These are similar to our respective values of 34.8, 20.3, and 29.0 Nm. Eek et al21 also used positions similar to ours for the hip flexors and abductors, knee extensors and flexors, and ankle dorsiflexors, although they used age groups of 1 year compared with our 1.5 years. When mapping and comparing the mean peak torques of our cohort with identical windows of age, combining girls and boys, the results are very similar. For example, the mean MIT values for knee extension from Eek et al21 (our study) are as follows: 25.7 (26.7) Nm, 39.5 (40.2) Nm, 49.9. (51.5) Nm, 66.3 (64.4) Nm, 75 (79.4) Nm, 98.4 (92.8) Nm, 107 (126.5) Nm for the 5- to 7-, 7- to 9-, 8- to 10-, 10- to 12-, 11- to 13-, 13- to 15-, and 14- to 15-year-old age groups, respectively.
Although we found that our protocol was feasible for younger children, our study is limited in that we did not evaluate the reliability or the concurrent validity of MIT assessed by our HHD protocol for children younger than 13 years. Since the main challenge with the use of HHD is maintaining high reliability and validity when assessing strong muscle groups and since we were under some constraints due to resource availability, we chose to assess validity and reliability only in older youth, who would presumably have stronger muscles. Since younger children may be less apt to concentrate, stay on task, or follow instructions, further work is needed to establish the reliability and concurrent validity of our protocol for younger children. For this protocol to be clinically useful, validity and reliability of the protocol will need to be established with a larger sample size, for the younger children typically developing as well as for different diagnostic groups over childhood and adolescence.
STUDY STRENGTHS AND THE RELATIONSHIP OF THE CURRENT PROTOCOL TO THE PRACTICE OF PEDIATRIC PHYSICAL THERAPY FOR CHILDREN WITH DISABILITIES
Strengths of this study were a well-defined protocol and trained evaluators, which is reflected in the overall acceptable reliability and validity and low SEM values of the MITs reported in our results. Error due to lack of standardization was thus likely minimized. Children and adolescents with physical disabilities commonly seen in pediatric physical therapy practice, such as those with cerebral palsy, spina bifida, or various muscular dystrophies, can show muscle weakness of both the upper and lower limbs at an early age. Thus, our protocol, which is feasible over a wide age range and which allows for the evaluation of muscle strength in both the upper and lower limbs, could aid the clinician in the identification of the extent of muscle strength impairment and its course over overtime. Since youth with these physical disabilities often have bilateral impairments, however, a comparison muscle group “outside of the person” is necessary to identify whether or not muscle weakness exists. This implies the establishment of normative data. Moreover, since muscle strength increases with growth and maturation, normative data may also provide information on the extent of, or changes in, the strength impairment that cannot be detected by simply comparing strength values for the same youth over time. Development of a normative database, however, requires the establishment of reliability and validity in children and adolescents developing typically, as we have begun to do in this study with the adolescent subgroup developing typically. Clearly the next steps in the development of our protocol, to render it of use to clinicians, are the establishment of validity and reliability for the younger age groups of children developing typically and the development of a normative database. After this is done, an evaluation of validity and reliability would then need to be repeated with each physical disability group of interest. This latter work will provide not only reliability and validity data per se, but in comparing the reliability of the disability group in question with that of a reference group developing typically, it may be possible to determine whether there is more variability in the intrinsic physiological conditions from trial to trial and session to session with individuals with muscle strength impairments than with their peers developing typically. Such information would aid in our understanding of muscle strength impairments and in the interpretations of muscle strength results.
The assessment of muscle strength plays a key role in pediatric physical rehabilitation. The feasibility of our protocol over a wide age range and with a variety of upper and lower limb muscle groups and the overall acceptable measurement properties of MIT obtained with our HHD protocol for adolescents are promising. Establishment of reliability and validity of MIT measured using our protocol in younger children and in children and adolescents with various physical disabilities, as well as development of a normative database, is required to ensure clinical utility of this HHD protocol.
The authors thank the children and parents who participated in this study and Ketisa Proulx and Andréanne Marleau who helped to prepare the manuscript. They also thank Thérèse Brousseau from the Neuro-Muscular Disease Program at the Quebec Rehabilitation Centre for her continued support of this work.
1. Pate RR. The evolving definition of physical fitness. Quest. 1988;40:174–179.
3. Damiano DL, Martellotta TL, Sullivan DJ, et al. Muscle force production and functional performance in spastic cerebral palsy: relationship of cocontraction. Arch Phys Med Rehabil. 2000;81:895–900.
4. Wessel J, Kaup C, Fan J, et al. Isometric strength measurements in children with arthritis: reliability and relation to function. Arthritis Care Res. 1999;12:238–246.
5. O'Connell DG, Barnhart R, Parks L. Muscular endurance and wheelchair propulsion in children with cerebral palsy or myelomeningocele. Arch Phys Med Rehabil. 1992;73:709–711.
6. Uchikawa K, Liu M, Hanayama K, et al. Functional status and muscle strength in people with Duchenne muscular dystrophy living in the community. J Rehabil Med. 2004;36:124–129.
7. Mahony K, Hunt A, Daley D, et al. Inter-tester reliability and precision of manual muscle testing and hand-held dynamometry in lower limb muscles of children with spina bifida. Phys Occup Ther Pediatr. 2009;29:44–59.
8. Aufsesser P, Horvat M, Croce R. A critical examination of selected hand-held dynamometers to assess isometric muscle strength. Adapt Phys Activ Q. 1996;13:153–165.
9. Jones MA, Stratton G. Muscle function assessment in children. Acta Paediatr. 2000;89:753–761.
10. Saulnier J, Crête M, Hébert LJ, et al. Protocole d'évaluation de la force musculaire à l'aide du dynamomètre manuel chez les enfants et les adolescents. Québec City, Canada: Institut de réadaptation en déficience physique de Québec; 2006.
11. Blimkie CJR, Sale DG. Strength development and trainability during childhood. In: VanPraagh E, ed. Pediatric Anaerobic Performance. Champaign, IL: Human Kinetics; 1998:193–224.
12. Bohannon RW. Intertester reliability of hand-held dynamometry: a concise summary of published research. Percept Mot Skills. 1999;88:899–902.
13. Seagraves F, Horvat M. Comparison of isometric test procedures to assess muscular strength in elementary school girls. Pediatr Exerc Sci. 1995;7:61–68.
14. Dvir Z. Isokinetics: Muscle Testing, Interpretation, and Clinical Applications. 2nd ed. Edinburgh, NY: Churchill Livingstone; 2004.
15. Shrout PE, Fleiss JL. Interclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420–428.
16. Domholdt E. Rehabilitation Research: Principles and Applications. 3rd ed. Philadelphia, PA: Elsevier Saunders; 2005:245–264.
17. Rosner B. Fundamentals of Biostatistics. 5th ed. Pacific Grove, CA: Duxbury Thomson Learning; 2000.
18. Dawson C, Croce R, Quinn T, et al. Reliability of the Nicholas manual muscle tester on upper body strength in children ages 8–10. Pediatr Exerc Sci. 1992;4:340–350.
19. Horvat M, McManis BG, Seagraves FE. Reliability and objectivity of the Nicholas manual muscle tester with children. Isokinetic Exerc Sci. 1992;2:175–181.
20. Macfarlane TS, Larson CA, Stiller C. Lower extremity muscle strength in 6- to 8-year-old children using hand-held dynamometry. Pediatr Phys Ther. 2008;20:128–136.
21. Eek MN, Kroksmark AK, Beckung E. Isometric muscle torque in children 5 to 15 years of age: normative data. Arch Phys Med Rehabil. 2006;87:1091–1099.
22. 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;68:977–982.
23. Verschuren O, Ketelaar M, Takken T, et al. Reliability of hand-held dynamometry and functional strength tests for the lower extremity in children with cerebral palsy. Disabil Rehabil. 2008;30:1358–1366.
24. Bohannon RW. Testing isometric limb muscle strength with dynamometers. Crit Rev Phys Med Rehabil. 1990;2:75–76.
25. Hill C, Croce R, Miller J, et al. Muscle torque relationships between hand-held dynamometry and isokinetic measurements in children aged 9 to 11. J Strength Cond Res. 1996;10:77–82.
adolescent; child; muscle strength; muscle strength dynamometer; reliability and validity© 2011 Lippincott Williams & Wilkins, Inc.