Intrarater reliability of manual muscle test (Medical Research Council Scale) grades in Duchenne's muscular dystrophy.

Physical therapists freqeuntly use the manual muscle test (MMT) to clinically assess patients with neuromuscular deficits. Manual msucle testing was developed by Lovett and described by Wright [1] in 1912. This technique has been revised, advanced, and promoted, resulting in several methods from which to choose. [2-11] Though each method has differing scales and symbols to represent grading criteria, all methods appear to be based on similar principles with like factors defining the criteria for the various muscle strength grades. These factors include gravity, the extent of arc of movement against gravity, and the amount of force applied by the examiner in opposition to the muscle group being tested. The differences between the methods include positioning, stabilization, application of force, and extent of subdivision among the major categories of grades.

The methods of testing and grading muscle strength described by Kendall and McCreary [9] and Daniels and Worthingham [10] are most often used by physical therapists in the United States. Neurologists appear to most often use the scale proposed by the Medical Research Council (MRC). [11] All three methods have six basic categories for grading of muscle strength. Daniels and Worthingham use words (Normal, good, Fair, Poor, Trace, or Zero) or letters (N, G, F, P, T, O) to symbolize their basic grading categories. They have added the use of a plus or minus sign to the basic grade to denote a greater or lesser amount of resistance or range through the motion. Kendall and McCreary use percentages, a defined by their grading criteria. Traditionally, the MRC scale has used the numeral grades 0 to 5, but, according to the scale's guidelines, use of the plus and minus subdivisions within the grade 4 may be helpful.

Aside from its use with patients with poliomyelitis, little information is available regarding the reliability, validity, or utility of the various MMT techniques in either clinical or research settings or writhin various age groups or patient populations. Minimal attention has been given to documenting the reliability of either MMT grades obtained for individual muscle groups of individual muscle strength grades.

Several studies published during the poliomyelitis era address the role of physical therapists and MMT in drug trials[12-14] and the standardization and reliability [14-16] of MMT grades in the clinical research setting. Gonnella and colleagues, [12] in 1953, discussed in detail the physical therapist's function as a member of the research team with the primary responsibility of muscle evaluation using MMT. The reliability of the testing procedures implemented was not addressed.

In 1954, Lilienfeld and colleagues [15] addressed the interrater reliability of MMT grades and the assignment of a factor describing muscle bulk as used in gamma globulin trials. all examiners had the same orientation to muscle testing procedures for this study, though their educational backgrounds differed (43 physical therapists, 23 physicians, and 8 nurses). A total of 45 individuals with poliomyelitis were examined, and a total of 65 muscles per patient were graded. The average differences in muscle strength scores between examiners ranged from 3.0% to 9.1%. Lilienfeld and co-workers felt these results indicated their system of muscle testing had a high degree of reproducibility among examiners with differing educational backgrounds but similar orientation to the specific methods of testing for their study.

At the Second Congress of the World Confederation for Physical Therapy in 1956, Lucy Blair [13] discussed the role of the physical therapist as an evaluator in the poliomyelitis vaccine field trials. This discussino addressed the interrater reliability of MMT grades, because investigators conducting a nationwide study on muscle testing had analyzed data from 38 physical therapists grading 82 muscle groups per patients, with the total number of patients not being stated. These physical therapists had determined an index of involvement that was based on the MMT grade multiplied by a factor that had been assigned according to muscle bulk. Two examiners grading the same patient agreed with 70% of the grades scored, and, in 95% of these instances, their agreement was within one muscle grade.

In 1961, Iddings, Smith, and colleagues [14,16] described and reported the reliability of grades obtained with a numerical index used in the clinical research of poliomyelitis. This numerical index was based on the MMT, and a factor was assigned according to the bulk of the muscle. The author's description of this numerical index addresses the reliability of the grades in a large-scale research project. The reliability was determined with three studies involving 13 physical therapists. They analyzed the intrarater and interrater reliability of MMT grades obtained in clinical practice. The interrater reliability was reported as 45.3% for complete agreement and as 90.6% for agreement within one muscle grade. Two of the 13 physical therapists retested the same patient, with intrarater reliability being 54% and 65% for complete agreement and 96% and 98% agreement within one muscle grade. Iddings, Smith, and colleagues concluded that, despite differences in training and testing techniques, the MMT grades were reliable in the clinical setting.

The studies during the poliomyelitis era were descriptive in nature and most often addressed the relaibility of a composite score, weighted by a factor that assessed muscle bulk, rather than analyzing grades for individual muscle groups or individual grades within a particular scale. The studies from this era, though informative, do not directly apply to today's clinical or research settings. Other publications have addressed factors that may influence the variability of MMT grades [17,18] or have reviewed the general topic of manual muscle testing. [19-21] The reliability of MMT grades as a measurement tool for analyzing strength as defined by the various methods [2-11] has not been established in regard to individual muscle groups or individual grades within specific patient populations.

Some authors [19,22] have indicated that the criteria for grading muscle strength is relatively specific for the grades for Fair (MRC 3) and below, but question the subjectivity of the grades Good (MRC 4) and Normal (MRC 5). Stuberg and Metcall [23] suggest the variability of the grades Good through Normal may be increased because of the absence of an operational definition of "normal strength." They and other [24,25] suggest that the use of instrumentation may eliminate the subjectivity of grading within these ranges of muscle strength.

In 1970, Silver et al [26] described the MMT for use in the clinical research setting with patients with renal disease. The standardized test was administered to 20 nondisabled subjects by three evaluators who assessed 12 muscle groups per subject using the MMT method of Daniels and Worthingham. [10] There was complete agreement among evaluators for 67% of muscles tested and 97% agreement within one half of a muscle grade.

In 1987, Frese et al [27] examined the interrater reliability of MMT grades obtained by assessing middle trapezius and gluteus medius muscle strength in the clinical setting. Eleven staff physical therapists, with an average of 2.3 [+ or -] 1.2 years of experience, performed the muscle testing on 110 patients referred for physical therapy. The therapists were allowed to use any method of testing with which they felt comfortable, including the methods of Kendall and McCreary [9] and Daniels and Worthingham. [10] Cohen's weighted Kappa was used to determine the interrater reliability, with coefficient ranging from .11 to .58, revealing poor agreement. Their conclusions indicated that the use of the MMT to make accurate clinical assessments of patients status was of questionable value. In this study, the sample was not strictly defined and the positions and procedures for testing were not standardized between examiners. This design probably gives us a realistic idea of the interrater reliability of grades in current clinical practice, but it does not address the reliability of MMT grades as a measurement tool in the research setting.

Ziter et al [28] used the MMT to assess muscle strength in patients with Duchenne's muscular dystrophy (DMD) and to document change in muscle strength over time. The authors described the MMT as a useful measure for documenting disease progression and suggested it be incorporated into clinical studies and therapeutic trials. Reliability of the MMT grades was not addressed.

Florence et al [29] described the intrarater and interrater reliability of a total muscle score used in the assessment of strength in a group of boys with DMD. This composite score was used to define the natural history of strength loss in patients with DMD [30,31] and as an outcome measure for documenting the effectiveness of various pharmacologic agents in the treatment of patients with DMD. [32,36] This composite score served its purpose because systemic effects of various oral medications were being assessed.

The development of myoblast transplant [37] places greater importance on individual muscle group assessment. Myoblast transplant is a potential therapy for genetic muscle diseases in which normal precursor cells are injected into the affected muscle tissue for the potential purpose of integrating with abnormal cells, altering their composition, and regenerating normal muscle cells. [38] Because myoblasts are injected within isolated muscle groups, one must assess individual muscle group strength in order to assess the clinical effect of myoblast transfer. We therefore believe it is imperative to address the reliability of MMT grades of individual muscle groups in a population of boyus with DMD.

Documenting the reliability of measurements is of particular importance in the assessment of children with DMD because this is a population with which a high rate of intellectual impairment and emotional disturbance has been associated. [39] These factors may influence the level of cooperation and hence the reliability of the physical assessment. Based on comparisons of voluntary versus electrically stimulated contractions, variability in muscle force measurements has also been documented and attributed to physiologic factors rather than to motivation or voluntary effort. [40]

The purpose of this study was to document the intrarater reliability of MMT grades, using the MRC scale [11] as a measurement tool, in assessing the strength of individual muscle groups in a sample of boys with DMD. The two research questions were (1) What is the intrarater reliability of individual MMT grades? and (2) What is the intrarater reliability of MMT grades obtained for various muscle groups?

Method

Subjects

Subjects were 102 boys, aged 5 to 15 years,with a diagnosis of DMD. All subjects were participants in the Clinical Investigation of Duchenne Dystropy (CIDD) Group study, a multicenter, collaborative investigation of DMD. As CIDD Group study participants, all subjects fulfilled study entry criteria, with the major inclusion criteria being (1) male, (2) onset of weakness before 5 years of age, (3) proximal weakness, and (4) serum creatine kinase at least 10 times normal at some stage of the disease. Informed consent was obtained from the parents of all boys prior to participation in the study at each of the collaborating centers.

All subjects were participants in a double-blind, placebo-controlled trial examing the effects of prednisone on muscle strength. [41] For the purposes of this trial, muscle strength was operationally defined by the MMT grades given. A previous prednisone trial [36] had reported an increase in muscle strength after 6 months of prednisone treatment in an open therapeutic trial using historical controls. [42] The present study was designed to clarify those results and document the reliability of the evaluation procedures.

All subjects were required to be able to cooperate and perform the MMT. Muscle strength grades ranged from 0 to 5. Subject characteristics and functional abilities are summarized in Table 1.

Examiners

Four examiners representing four institutions participated in the intrarater reliability study. All examiners were physical therapists with 16 to 20 years of experience, including 10 to 15 years of specialty experience in neuromuscular disorders. All examiners have been involved with the CIDD Group as clinical evaluators for the past 10 years and have served as consultants for each center's neuromuscular

Table 1. Patient Characteristics and Functional Abilities Immediately After Treatment (N = 102)

                            No. of
Characteristics             Subjects
Age (y)
 X                           9.06
 SD                          2.69
 Range
   5-8 y                    42
   8-12 y                   44
   2-15 y                   16
Function
  Ambulated independently   72
  Required long leg braces  13
   for ambulation
  Required wheelchair for   17
   ambulation

disease clinic. Examiners were trained in the assessment protocol at the CIDD Group coordinating center prior to data collection. The intrarater and interrater reliability of all assessments have been documented yearly as part of the CIDD Group study protocol at group meetings at which examiners tested the same patients. These reliability values have been described previously. [29]

Procedure

Data were collected as part of a 12-month therapeutic trial to document the effects of prednisone on muscle strength in patients with DMD. The design of the protocol required duplicate visits (two identical assessments performed within 5 days of each other) initially and after 6 and 12 months of treatment in an attempt to ensure a consistent and complete data set in this pediatric population. Examiners were blind to previous testing results.

Muscle strength was assessed and individual MMT grades were assigned using a modified MRC grading scale, with subdivisions of the grades 3, 4, and 5, as follows: 5, 5-, 4+, 4, 4-, 3+, 3, 3-, 2, 1, 0. The modifications to the MRC scale included the addition of the grading subdivisions 5-, 3+, and 3-. Definitions of the individual muscle testing grades are shown in Table 2. All positions and procedures for testing were standardized and strictly defined by the CIDD Group procedures and followed the recommendations of the MRC. [11]

Eighteen muscle groups were assessed in each subject at each session. The muscle groups tested included shoulder abductors and external rotators; elbow and wrist flexors and extensors; thumb abductors; hip flexors, extensors, and abductors; knee flexors and extensors; ankle dorsiflexors, plantar flexors, inverters, and everters, bilaterally; and neck flexors and extensors.

Data Analysis

Data were analyzed using the weighted Kappa, as described by Cohen, [43] to determine the reliability of individual MMT grades and grades obtained for individual muscle groups. Cohen's Kappa is a reliability index used for nominal and categorical data. Kappa is a chance-corrected measure of agreement in which all disagreements are given equal weight. In contrast, the weighted Kappa takes into account the degree of disagreement among raters. Weighted Kappa differentially weighs discrepancies between pairs of scores so that the further apart the two scores, the more effect that observation has on lowering the reliability. The weights used in this analysis were equal to the number of grades separating a pair of scores. Thus, an observation in which a muscle was scored the same on both visits would have a weight of 0, an observation in which the muscle was scored 3 on the first visit and 3+ on the second visit would have a weight of 1, and so on.

In estimating the reliability of grades within individual muscle groups, data were organized by collapsing the data across the sides so that each subject contributed up to six observations for each paired muscle (two

Table 2. Modified Medical Research Council Scale

Grade           Definition
5    Normal strength
5-   Barely detectable weakness
4+   Same as grade 4, but muscle
     holds the joint against moderate to
     maximal resistance
4    Muscle holds the joint against a
     combination of gravity and
     moderate resistance
4-   Same as grade 4, but muscle
     holds the joint only against
     minimal resistance
3+   Muscle moves the joint fully
     against gravity and is capable of
     transient resistance, but collapses
     abruptly
3    Muscle cannot hold the joint
     against resistance, but moves the
     joint fully against gravity
3-   Muscle moves the joint against
     gravity, but not through full
     mechanical range of motion
2    Muscle moves the joint when
     gravity is eliminated
1    A flicker of movement is seen or
     felt in the muscle
0    No movement

sides X three duplicate visits). The data from duplicate visits constituted one observation. Sample sizes after deletion of missing data ranged from 501 for ankle dorsiflexors to 574 for wrist extensors. Neck flexors and neck extensors, not being paired, contributed fewer observations (278 for extensors and 284 for flexors).

Assessment of the reliability of individual muscle strength grades was made by using a modification of Cohen's Kappa, as described by Cicchetti and colleagues. [44] The data were arranged so that each subject contributed up to 102 paired observations on 34 muscles on three separate occasions, resulting in a data set with 9,427 paired observations after removal of missing data.

The intraclass correlation coefficient (ICC,[1,1]) [45] was used to calculate the

Table 3. Intrarater Reliability of Manual Muscle Test Grades for Individual Muscle Groups Obtained Using Modified Medical Research Council Scale

                             Weighted
Muscle Group                 Kappa
Knee                         .93
Hip flexors                  .90
Shoulder external rotators   .89
Hip abductors                .89
Hip extensors                .88
Shoulder abudctors           .87
Elbow extensors              .86
Neck flexors                 .85
Neck extensors               .84
Elbow flexors                .82
Ankle dorsiflexors           .81
Knee flexors                 .79
Ankle evertors               .73
Ankle invertors              .72
Ankle plantar flexors        .71
Thum abductors               .71
Wrist extensors              .69
Wrist flexors                .65

intrarater reliability of the total muscle score for comparison with our previously published results. [29] The total muscle score is determined by transforming individual muscle grades to a 10-point scale (5=10, 5-=9, 4+=8, and so on), adding all converted scores, and using that sum for comparisons. Though muscle scores are ordinarilly scaled, parametic analysis of the overall total muscle score was deemed appropriate because of its linear relationship to other variables previously documented in the DMD population. [46]

Reliability denotes the stability of the measure and whether one can obtain similar measurements of the same variable on separate occasions. There are no universally accepted standards for reliability, but the following criteria have been proposed by Landis and Koch [47] for interpreting agreement of Cohen's Kappa statistics: <.00, poor; .00-.20, slight; .21-.40, fair; .41-.60, moderate; .61-.80, substantial; .81-1.00, almost perfect. In interpretation of the ICC, Fleiss [48] states that >.75 is excellent. Nunnally [49] states that the minimally acceptable reliability for a scale depends on the use of the measurement. Nunnally proposes that reliability of around .80 is sufficient in basic research, but that if decisions are based on individual test scores, one should attempt to attain a reliability of .90.

Results

Intrarater reliability of MMT grades obtained with the modified MRC scale for individual muscle grups, as determined by the weighted Kappa, is shown in Table 3. Grades of proximal muscle groups were more reliable than were grades of muscle groups located distally. The distal upper-extremity musculature was graded less reliably than the distal lower-extremity musculature.

Intrarater reliability of MRC grades 0 to 5, as determined by the weighted Kappa, is shown in Table 4, along with the number of assignments within each grade. The reliability varied among individual grades, with grades in the gravity-eliminated position having the highest reliability values. The ICC (1,1) for the total muscle score was .99, which confirms previously published data. [29]

Discussion

Manual muscle testing, using the MRC scale, provides reliable grades for the assessment of strength of individual muscle groups within a sample of boys with DMD when tests are repeated within 5 days by the same examiner. Intrarater reliability ranged from .65 to .93. The weighted Kappa values for the proximal muscles were more consistent than those for the distal muscles, and the weighted Kappa values for the lower-extremity muscles were generally more reliable than those for the upper-extremity muscles. The range of reliability could be attributed to amount of effort, understanding, or cooperation, particularly

Table 4. Intrarater Reliability of Individual Muscle Strength Grades Obtained Using the Modified Medical Research Council Scales

                 Weighted
Grade  [N.sup.a]  Kappa
5        301     .93
5-     285       .83
4+     1,485     .85
4      3,028     .85
4-     1,838     .83
3+       205     .80
3         92     .84
3-     1,114     .94
2        938     .98
1        132     .99
0          0     .97
[sup.a.N]=number of assignments, first evaluation of
the two evaluations performed within 5 days
of each other initially and after 6 and 12
months of treatment.

in this pedicatric sample. Based on comparisons of voluntary versus electrically stimulated comparisons, however, variability in muscle force measurements has also been attributed to physiological factors. [40]

The less reliable grading in the distal musculature could be attributed to joint contractures such as the equinovarus deformity at the ankles and shortening of the wrist and finger flexor musculature that are often found in patients with DMD. These joint contractures not only limit appropriate positioning for the individual muscle tests, but they also limit the available range of motion (ROM) through which the muscle may work. These contractures and the less available ROM may have a greater effect on the distal than on the proximal musculature. For example, a 40-degree wrist or hip flexion contracture would leave one with 50 degrees of motion at the wrist versus 120 degrees of motion at the hip. The lower reliability values obtained for the upper-extremity muscles in this study confirm the results of a previous study of patients with DMD in which the ICC for the upper-extremity composite score was less thant he ICC for the lower-extremity composite. [29]

Though weighted Kappa values for intrarater reliability varied among individual muscle groups (.65-.93), all had substantial agreement or better. [47] Combining the individual muscle groups to obtain a total muscle score resulted in even better intrarater reliability (ICC=.99). Use of a composite score for patient follow-up may eliminate information regarding individual muscle group assessment, but it also eliminates the variability in individual muscle group analysis and creates a more stable measure when attempting to follow muscle strength changes in individual patients over time when there has been systemic intervention. It appears this finding has been utilized previously in the documentation of therapeutic intervention in patients with poliomyelitis. [12-16]

Intrarater reliability varies among individual muscle grades with those in the gravity-eliminated position (MRC 0-2) grading most reliably ([is greater than or equal to] .97). This finding differs from that of Frese and colleagues, [27] who found poor interrater reliability in grades below Fair; similarly, Beasley [21] found poor differentiation in grades below Fair. These grades were the only categories that had no subdivisions (no plus or minus designations) and were strictly defined, but Frese and colleagues also stated that "compressing the scores by eliminating pluses and minuses did not appreciably change the interrater reliability coefficients." [27](p1074)

The strength grading subdivisions in which gravity and resistance are factors in determining the MMT grade have come under much criticism, [22-24] because they require judgments beyond assigning the original grade on the part of the examiner. The weighted Kappas for the MRC grades 4-- to 5 ranged from .83 to .94, which demonstrates substantial agreement. [47] These grades are less reliable than those given in positions in which the factors of gravity and resistance have been eliminated but, we believe, are still acceptable for measurement in the clinical trial setting.

Stuberg and Metcalf [23] have suggested that the subjectivity inherent in MMT grades in the Good to Normal range (MRC 4-5) can be eliminated with the use of instrumentation. They used a hand-held myometer to measure force in eight muscle groups of 14 boys with DMD and reported reliability coefficients ranging from .83 to .99. The reliability coefficients in our MMT study ranged from .83 to .93 in the Good to Normal range (MRC 4-5). Studies relating to the sensitivity (ie, ability to measure change over time) of these methods of measurement of muscle strength are needed to determine the most useful method for documentation, as both methods appear reliable in the assessment of boys with DMD.

The lowest reliability coefficient for an individual muscle grade in this study was .80, with a grade of 3+. The definition of this grading subdivision in our study required considerable judgment by the examiner because this grade indicates that the muscle "is capable of transient resistance, but collapses abruptly." This grade, however, may represent such a transitional state that there may be real fluctations in performance on a day-to-day basis.

The major differences between our study and various other MMT references in the literature include our documentation of intrarater reliability, our documentation of MMT grades for 18 individual muscle groups, and our use of a total muscle score as compared with other studies that examined only one or two muscles individually or used a composite score that included factors other than muscle strength. Our study was population specific. All examiners had extensive training in working with the DMD population in regard to the specific testing protocol, with all positions and procedures strictly defined. It may be that having one individual in each institution perform all testing and having specific training and strict adherence to the required protocol influenced the level of reliability of the MMT grades, as all measures of muscle strength in this sample had substantial agreement.

One limitation of this study was the fact that all patients could not be tested in all positions, either because of severity of contractures or because of discomfort of the testing position secondary to the severity of the disease. A second limitation is that MMT strength measurements of individual muscle groups are scaled ordinally, thus suggesting that nonparametric statistics should be used. This was the case for using Kappa when analyzing the individual muscle groups and MRC grades, whereas the derived total muscle score appeared to have satisfactory interval properties, deeming parametric statistics appropriate. [50]

Clinical Implications

Controversy exists in the literature over the use of the MMT as a measurement tool in the documentation of muscle strength. Our study suggests that MMT grades obtained with the MRC scale are reliable when recorded by the same trained examiner in a sample of children with DMD. The degree of reliability depends on the muscle group being tested and the specific grade being given. If MMT grades are to be used to make clinical decisions, we recommend that their reliability be documented within the various MMT methods, age groups, and patients populations.

Some authors have suggested that MMT grades below Fair (MRC <3) are not reliable [22,27] and that grades of Good (4) and Normal (5) are subjective. [23] Our study suggests that we need to be most cautious of the grades Fair (3), Fair plus (3+), and Normal minus (5-) and that the most reliable grades are those made with the factor of gravity eliminated, though we believe all grades' reliability coefficients were adequate for the clinical research setting.

The MMT is shown to yield reliable grades within individual muscle groups, but reliability varied proximal to distal within an extremity. The best agreement was shown when MMT grades for individual muscle groups were combined into a total muscle score. This finding suggests the most stable measure for documenting muscle strength in systemic diseases or with systemic interventions is a composite score.

The reliability obtained in a sample of boys with generalized muscle weakness and a high incidence of intellectual impairment and emotional disturbance suggests that the MMT could be adapted and administered to yield reliable results from patients with a variety of diseases.

Though the MMT, using the MRC scale, has been shown to yield reliable grades when administered by the same examiner in the clinical research setting in a sample of boys with DMD, this study has not addressed the validity or sensitivity of the measure in documenting change over time or how applicable or sensitive the measure is as compared with other methods of testing muscle strength and performance.

Conclusion

The MMT grades obtained in this study, using the MRC scale as a measurement tool, were reliable when recorded by the same examiner in the clinical research setting in a population of boys with DMD. High intrarater reliability was found for both individual MMT grades and for grades obtained for individual muscle groups. Though reliable, the range of grades emphasizes the importance of documenting the reliability of various MMT methods within various age groups and patient populations.

Acknowledgments

We are grateful to Anthony Delitto for conceptual and editorial advice, to Patti Nacci for assistance in preparation of the manuscript, and to the CIDD Group, which includes the University of Rochesters (Shree Pandya, Richard T Moxley, Robert C Griggs), The Ohio State University (Wendy M King, Linda C Signore, Jerry Mendell), Vanderbilt University (Jenny D Robison, Gerald M Fenichel), the University of Alberta (Nancy Matheson, Michael H Brooke), and Washington University (Julaine M Florence, Jeanine R Schierbecker, Alan Pestronk, J Philip Miller, Jack Baty, and Brad Wilson).

References

[1] Wright W. Muscle training in the treatment of infantile paralysis. Boston Med Surg J. 1912;167:567.

[2] Lowman CL. A method of recording muscle tests. Am J Surg. 1927;3:588-591.

[3] Legg A, Merrill J. Physical therapy in infantile paralysis. In: Mock HE, Pemberton R, Caulter J, eds. Principles and Practices of Physical Therapy. Maryland: WF Prior; 1932: vol 2.

[4] Kendall H, Kendall F. Care During the Recovery Period in Paralytic Poliomyelitis. Washington, DC: US Public Health Service Bulletin 9.242; 1939.

[5] Brunnstrom S. Muscle group testing. Phys Ther Rev. 1941;21:3-22.

[6] Zausmer E. Evaluation of strength and motor development in infants: part I. Phys Ther Rev. 1953;33:575-581.

[7] Zausmer E. Evaluation of strength and motor development in infants: part II. Phys Ther Rev. 1953;33:621-629.

[8] Hines TF. Manual muscle examination. In: Licht S, ed. Therapeutic Exercise. Baltimore, Md: Waverly Press; 1965:163.

[9] Kendall F, McCreary E. Muscle Testing and Function. 3rd ed. Baltimore, Md: Williams & Wilkins; 1983.

[10] Daniels L, Worthingham C. Muscle Testing: Technique of Manual Examination. 5th ed. Philadelphia, Pa: WB Saunders Co; 1986.

[11] Medical Research Council of the United Kingdom. Aids to Examination of the Peripheral Nervous System: Memorandum No 45. Palo Alto, Calif: Pedragon House; 1978.

[12] Gonnella C, Harmon G, Jacobs M. The role of the physical therapist in the gamma globulin poliomyelitis prevention study. Phys Ther Rev. 1953;33:337-345.

[13] Blair L. The role of the physical therapist in the evaluation studies of the poliomyelitis vaccine field trials. Phys Ther Ref. 1955;37:437-447.

[14] Smith LK, Iddings DM, Spencer WA, Harrington PR. Muscle testing, part I: description of a numerical index for clinical research. Phys Ther Rev. 1961;41:99-105.

[15] Lilienfeld AM, JAcobs M, Willis M. A study of the reproducibility of muscle testing and certain other aspects of muscle scoring. Phys Ther Rev. 1954;34:279-289.

[16] Iddings DM, Smith LK, Spencer WA. Muscle testing, part 2: reliability in clinical use. Phys Ther Rev. 1961;41:249-256.

[17] Nicholas J, Sapega A, Kraus H, Webb J. Factors influencing manual muscle tests in physical therapy. J Bone Joint Surg [Am]. 1978;60:186-190.

[18] Smidt GL, Rogers MW. Factors contributing to the regulation and clinical assessment of muscular strength. Phys Ther. 1982;62:1283-1290.

[19] Lamb RL. Manual muscle testing. In:Rothstein JM, ed: Measurement in Physical Therapy. New York, NY: Churchill Livingstone Inc; 1985:47-55.

[20] Wakim K, Gersten J, Elkins EC, Martin G. Objective recording of muscle strength. Arch Phys Med. 1950;31:90-99.

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

[22] Rothstein JM. Measurement and clinical practice: theory and application. In: Rothstein JM, ed. Measurement in Physical Therapy. New York, NY: Churchill Livingstone Inc; 1985:1-46.

[23] Stuberg WA, Metcalf WK. Reliability of quantitative muscle testing in healthy children and in children with Ducheme muscular dystrophy using a hand-held dynamometer. Phys Ther. 1988;68:977-982.

[24] Andres PL, Skerry LM, Munsat TL. Measurement of strength in neuromuscular diseases. In: Munsat TL, ed. Quantification of Neurologic Deficit. Boston, Mass: Butterworth-Heinemann; 1989:87-100.

[25] Mendell JR, Florence JM. Manual muscle testing. Muscle Nerve. 1990;13:S16-S20.

[26] Silver M, McElroy A, Morrow L, Heafner BK. Further standardization of manual muscle test for clinical study: applied in chronic renal disease. Phys Ther. 1970;50:1456-1466.

[27] Frese E, Brown M, Norton BJ. Clinical reliability of manual muscle testing: middle trapezius and gluteus medius muscles. Phys Ther. 1987;67:1072-1076.

[28] Ziter F, Alsop K, Tyler F. Assessment of muscle strength in Duchenne muscular dystrophy. Neurology. 1977;27:981-984.

[29] 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.

[30] Mendell JR, Province MA, Moxley RT, et al. Clinical investigation of Duchenne muscular dystrophy: methodology for therapuetic trials based on natural history controls. Arch Neurol. 1987;44:808-811.

[31] Brooke MH, Fenichel GM, Griggs RC, et al. Duchenne muscular dystrophy: pattern of clinical progression and effects of supportive therapy. Neurology. 1989;39:475-481.

[32] Mendell JR, Griggs RC, Moxley RT, et al. Clinical investigation in Duchenne muscular dystrophy; IV: double-blind controlled trial of leucine. Muscle Nerve. 1984;7:535-541.

[33] Moxley RT, Brooke MH, Feninchel GM, et al. Clinical investigation in Ducheme dystrophy, VI; double-blind controlled trial of nifedipine. Muscle Nerve. 1987;10:22-33.

[34] Fenichel GM, Brooke MH, Griggs RC, et al. Clinical investigation in Duchenne muscular dystrophy: penicillamine and vitamin E. Muscle Nerve. 1988;11:1164-1168.

[35] Griggs RC, Moxley RT, Mendell JR, et al. Randomized double-blind trial of mazindol in Duchenne dystrophy. Muscle Nerve. 1990;13:1169-1173.

[36] Brooke MH, Fenichel GM, Griggs RC, et al. Clinical investigation of Duchenne muscular dystrophy: interesting results in a trial of prednisone. Arch Neurol. 1987;44:812-817.

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

[38] Partridge TA. Myoblast transfer: possible therapy for inherited myopathies? Muscle Nerve. 1991;14:197-212. Invited review.

[39] Leibowitz D, Dubowitz V. Intellect and behavior in Duchenne muscular dystrophy. Dev Med Child Neurol. 1981;23:577-590.

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

[41] Mendell JR, Moxley RT, Griggs RC, et al. Randomized double-blind controlled trial of prednisone in Duchenne muscular dystrophy. N Engl Med 1989;320:1592-1597.

[42] Mendell JR, Province MA, Moxley RT, et al. Clinical investigation of Duchenne dystrophy: a methodology for therapeutic trials based on natural history controls. Arch Neurol. 1987;44:808-811.

[43] Cohen J. Weighted Kappa: normal scale agreement with provision for scaled disagreement on partial credit. Psychol Bull. 1968;70-213-220.

[44] Cicchetti DV, Lee C, Fontana AF, Dowde A. A computer program for assessing specific category rater agreement for qualitative data. Educational and Psychological Measurement. 1978;38:805-813.

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

[46] Brooke MH, Fenichel GM, Griggs RC, et al. Clinical investigation of Duchenne dystrophy; 2: determination of the "power" of therapeutic trials based on the natural history. Muscle Nerve. 1983;6:91-103.

[47] Landis JR, Koch GG. The measurement of observer agreem,ent for categorical data. Biometrics. 1977;33:159-174.

[48] Fleiss JL. The Design and Analysis of Clinical Experiments. New York, NY:John Wiley & Sons Inc; 1986;1-32.

[49] Nunnally JC. Psychometric Theory. 2nd ed. New York, NY: McGraw-Hill Book Co; 1978.

[50] Miller JP. Statistical consideration for quantitative techniques in clinical neurology. In: Munsat TL, ed. Quantification of Neurologic Deficit. Boston, Mass: Butterworths; 1989:69-84.