Interrater Reliability of the Active-Knee-Extension Test for Hamstring Length in School-Aged Children

Rakos, Diane M. MSPT; Shaw, Kelly A. MSPT; Fedor, Robyn L. MSPT; LaManna, Maryalice MSPT; Yocum, Corrie C. MSPT; Lawrence, Kevin J. MS, PT, OCS

Pediatric Physical Therapy:
Research Report

Purpose: The purpose of this study was to determine the interrater reliability of the active knee-extension test (AKET) using a stabilizing apparatus to measure hamstring length.

Methods: One hundred one subjects (53 girls, 48 boys) ranging in age from 10 to 13 years with no known neuromuscular problems participated. The AKET was performed with subjects lying supine with the hip flexed to 90 degrees with a stabilization device attached to a plinth. Next, subjects were instructed to actively extend the knee until the rater detected myoclonus. Then, the rater flexed the knee until myoclonus subsided and the knee angle was measured with a blinded goniometer. This procedure was repeated by each of three raters.

Results: Data were analyzed using ICC (2,1) demonstrating good interrater reliability of 0.79.

Conclusion: Our results suggest that the AKET, when used with the stabilizing apparatus, demonstrates good interrater reliability for children aged 10 to 13 years.

Author Information

Gnaden Huetten Memorial Hospital (D.M.R.), Lehighton, Little Flower Manor (K.A.S.), Wilkes-Barre, Allied Medical and Technical Career (R.L.F), Forty Fort, Allied Services/John Heinz Rehabilitation Institute of Medicine (C.C.Y.), Berwick and Hazleton, and College Misericordia (K.J.L.), Dallas, Pa; and Healthsouth Rehabilitation Hospital of NJ (M.L.), Jackson, NJ

Address correspondence to: Corrie Yocum, 82 Honch Hallow Road, Bloomsburg, PA 17815.

Article Outline
Back to Top | Article Outline


Currently there are numerous methods for assessing hamstring length used by clinicians in various practice settings. 1–7 However, these methods do not necessarily provide adequate isolation of the hamstrings. 1,2 For example, they may include movement of the lumbar vertebrae, pelvis, and/or hips without specifically isolating the hamstring muscles. To accurately measure hamstring muscle length, the testing procedure must provide adequate stabilization of the spine and pelvis. The testing procedure must be standardized so that the measurements represent changes in range of motion and not variation in the testing procedure.

The straight leg raise (SLR) is often used to measure hamstring length; however, this method does not control pelvic movement. Several studies have documented the need to control pelvic motion. 1,6,7 Bohannon 3 used cinematography when measuring straight leg raises performed by nine women and two men, aged 20 to 32 years. The study included three different ways of stabilizing the pelvis and the contralateral leg. The authors concluded that the pelvis must be stabilized or accounted for in the measurement if it is to accurately assess hamstring length. However, the researchers did not mention which of the three positions provided the most stabilization. Furthermore, there is no mention of what “suitable” stabilization would be or how pelvic rotation affects the hamstring measurement. 3

The sit-and-reach test is another test used predominantly in children to measure hamstring length. 1,5,7 The tests consists of having a seated subject reach forward, with knees extended, attempting to bring the fingertips to the great toe or beyond. 1 This technique has been criticized because of the movement that occurs at the pelvis and lumbar vertebrae. Movement of the lumbar vertebrae and pelvis during passive hip flexion was documented by Bohannon, Gadjosik, and LeVeau. 4 In their study, which included 13 female and four male subjects aged 14 to 30 years, passive hip flexion without stabilization of the pelvis produced 27 degrees of pelvic rotation. 4 Hip flexion in the sit-and-reach test can only occur with concurrent movement of the femur, pelvis, hip joint, spine, and shoulder girdle. Therefore, the sit-and-reach test does not completely isolate the hamstrings. Furthermore, Bohannon et al 4 assert that measuring hamstring length to the point of pelvic rotation is of doubtful value.

A study by Kippers and Parker 5 determined that the standing toe-touch test was a good technique for measuring hamstring extensibility, but is not a valid measurement of hamstring length. The subjects consisted of 16 male and 17 female volunteers aged 18.2 to 25 years. The standing toe-touch test consists of the subject standing erect and then flexing hips and trunk to try to touch the floor with the fingertips. 8

The standing toe-touch test assesses gross motion such as hip and vertebral flexion, but does not isolate the components that allow the motion. Because of the concurrent motion at each of the segments involved (ie, lumbar vertebrae, sacroiliac joint, and hip joint) it is impossible to measure the amount of motion that occurs at each segment. 5 Therefore, this test cannot accurately assess hamstring length. 5

Bandy and Irion 8 performed a study in which the subjects (40 men and 17 women, aged 21–37 years) laid supine while one researcher passively positioned the hip to be tested in 90 degrees of flexion and the other hip remained in a neutral position. Then, the other researcher passively extended the knee of the tested leg until the subject complained of pain or the researcher felt tightness in the hamstring. The intratester reliability for their pilot study using this method was 0.98 (ICC 1,1). 8 They reported no attempts to control pelvic movement.

In light of the evidence presented above, we chose to assess the interrater reliability of measuring hamstring length using the active knee-extension test (AKET) as described by Gadjosik and Lusin. 2 This test has been shown to demonstrate good intrarater reliability (.99) in measuring hamstring length for 15 men aged 18 to 26 years old. 2 The advantage of this test is that significant amounts of movement at the hip joint, sacroiliac joint, and lumbar spine are controlled through stabilization. 2

No published studies were found that examined the interrater reliability of the AKET. Therefore, the purpose of our study was to establish interrater reliability for the AKET using a slight modification to the method described by Gadjosik and Lusin. 2 Our hypotheses was that the AKET would have good (>0.75) interrater reliability as described by Portney and Watkins. 9

Back to Top | Article Outline



The subjects consisted of 129 volunteers in fourth through sixth grades from the Leo E. Solomon Educational Complex, Plains, Pa. To be included in the study the subjects needed to return: 1) a signed subject permission slip, 2) a signed parent or legal guardian permission slip, 3) a completed questionnaire (Appendix I). The project was approved by the Internal Review Board at College Misericordia. Any of the following criteria resulted in exclusion of the subject from the study: 1) failure to complete and return the packet to school; 2) positive Thomas Test for hip flexor tightness 3) a pre-existing condition such as cerebral palsy, lower extremity osteosarcoma, fracture, or any disease process that may have affected the lower extremity within the past six months; 4) subject refusal on the day of measurement; or 5) subject not wearing shorts or wind pants on the day of the study. Of the 129 volunteers, 101 (53 girls, 48 boys, aged 10–13 years) were eligible for the study.

Back to Top | Article Outline

Five student physical therapists in their final semester of an entry-level master’s degree program at College Misericordia performed this study. Three of the investigators served as raters measuring the knee range of motion (ROM); the fourth assisted the students with positioning; and the fifth recorded the data. Only the latter had access to the subjects’ identities.

A pilot study was conducted on five student volunteers (aged 7–12 years) who were not participants in the research study to establish intrarater reliability for each of the three raters before the research study. The ICC reliability level was set at 0.90. The raters’ reliabilities for measuring the AKET with the apparatus were 0.99, 0.99, and 0.95, respectively.

Back to Top | Article Outline

The stabilization apparatus (Fig. 1) was constructed out of wood by one investigator. It was placed on a portable table. Clamps were used to attach the apparatus to the table. We added a nylon rope to increase the stability of the apparatus. The apparatus used in our study was a modification of that illustrated by Gadjosik and Lusin 2 and consisted of two vertical bars on either side of the table and two horizontal bars. The top horizontal bar connected the two vertical bars for stabilization. The second horizontal bar also connected the two vertical bars and served as a guideline for maintaining the position of the anterior thigh. Straps were used to stabilize the pelvis and contralateral extremity of the subject.

Additional materials included washable markers and two universal goniometers. The washable markers were used to mark the lateral femoral condyles and the apex of the lateral malleoli to ensure proper placement of the goniometer when obtaining the measurements. One goniometer was used to establish the position of the hip and a second, blinded goniometer was used to measure knee extension. Interrater reliability for a universal goniometer for active knee flexion was found to be 0.87 in a study by Rheault et al. 10 Another study by Rothstein et al 11 on knee flexion and extension found intertester reliability to be 0.91 using a universal goniometer.

Back to Top | Article Outline

In the beginning of physical education class, students who met the criteria were assembled in a room adjacent to the gym where the apparatus was set up. One rater was responsible for using the colored markers to distinguish bony landmarks on each subject’s right leg before placing each subject on the table. Each subject was marked with an “X” at the lateral femoral epicondyle and at the apex of the lateral malleoli on the right leg.

Each subject was instructed to lie supine on the portable table. One of the raters confirmed the angle of the hip of the right leg to be 90 degrees. The subject was then positioned so that the right anterior thigh was resting against the cross bar. One investigator maintained the subject’s right thigh against the horizontal bar. One of the raters positioned a Velcro strap across both anterior superior iliac spines, and a second Velcro strap around the left proximal thigh for stabilization. To measure the right hip position, the fulcrum of the goniometer was placed at the level of the greater trochanter. 12 The stable arm of the goniometer was parallel with the midline of the trunk and the moveable arm was aligned with the shaft of the right femur using the lateral epicondyle of the femur as a reference 12 (Fig. 2). The investigator maintained the subject’s right thigh against the horizontal bar throughout the measurement process. The subject was then instructed to actively extend the right knee until the researcher determined that myoclonus had been detected. Myoclonus is defined as an alternating contract relax pattern between quadriceps and hamstrings. 2 Once myoclonus was detected, the subject was instructed to stop extending the right knee. The rater then flexed the right knee until myoclonus had subsided. The angle of the right knee was measured with the fulcrum at the lateral epicondyle of the femur and the arms of the blinded universal goniometer aligned with the greater trochanter of the femur and the apex of the lateral malleoli 12 (Fig. 3). After the measurement, the right hip was maintained at 90 degrees while the subject was instructed to relax the right knee into a flexed position. The other investigator read the blinded goniometer and recorded the measurement on a data sheet. This procedure was repeated for all three raters, with the raters’ order randomized. Only the third rater rechecked the angle of the right hip at 90 degrees before measuring, due to time constraints. Only the right leg was used for data collection.

Back to Top | Article Outline
Statistics and Results

The interrater reliability was determined for the three measurements of each subject using the ICC (2,1). The ICC (2,1) has greater statistical power than the other forms of the ICC, and is generalizable to other professionals with the same training. 9 The formula for the ICC (2,1) represents changes between raters and not existential circumstances. 9 The result of the ICC (2,1) for right lower extremity AKET was 0.79. According to Portney and Watkins, 9 results greater than 0.75 indicate good reliability (Table 1).

Back to Top | Article Outline


Based on the results of this study, the AKET, when used in conjunction with the apparatus, demonstrates good reliability when performed by multiple raters, although measurements were taken by physical therapist students. Reliability is an important component of clinical practice. In fact, it is crucial in a clinical setting, because it is often unrealistic to assume that the same clinician will measure a patient repeatedly. This procedure standardizes the measurement process so that different measurements accurately reflect a change in patient status, rather than differences in measurement technique. It should not be assumed that the measurement technique would be accurate without the use of the apparatus. Realizing this limitation of the procedure, we recommend additional research to streamline the measurement process. Perhaps studies could include the reliability of the AKET without the apparatus, performed by one person without an assistant, and/or the use of a gravity goniometer. These conditions would more accurately mirror the clinical setting.

Our interrater reliability values were slightly lower than the intrarater results obtained by the Gadjosik and Lusin. 2 This can be attributed to a number of factors. Interrater reliability is generally found to be less than intrarater reliability. 9 The difference could also be because of the use of different statistical procedures. Gadjosik and Lusin 2 used the Pearson-product moment correlation coefficient, whereas we applied the ICC, considered a more stringent statistical procedure. The subject populations were different. Their subjects consisted of 15 men aged 18 to 26 years old, whereas we used 101 children aged 10 to 13 years old. Two variations in procedures could explain the difference as well. Gadjosik and Lusin 2 drew a line from the fibular head to the lateral malleoli. In our study, the lateral femoral condyle and the apex of the lateral malleoli were marked with an “X.” Also, in their study, the subjects were instructed to stop extending the leg once myoclonus was observed. The subjects were then instructed to flex the leg until myoclonus subsided. However, in our study, myoclonus was determined by an investigator, then the leg was flexed by a rater until myoclonus subsided. Furthermore, in our study measurements were taken by student physical therapists, whereas Gadjosik and Lusin 2 are physical therapists.

Limitations of our study include the lack of control of ankle position in dorsiflexion or plantarflexion. Because the gastrocnemius muscle crosses both the knee and ankle joints, ankle position could impact knee extension ROM. Also rotation at the hip was not adequately controlled, thus allowing the measurements to be influenced by the lateral or medial hamstrings. Additionally, the axis of the goniometer could not be accurately aligned with the lateral epicondyle in certain situations. When measuring subjects with smaller leg girths, the universal goniometer’s full circle protractor was too large to have the axis placed at the lateral epicondyle without contacting the crossbar. In addition, the cloth straps secured with Velcro did not completely control the movement of the pelvis. Minimal posterior rotation of the pelvis was observed in certain instances even after the right hip was flexed to 90 degrees. The apparatus, which we constructed as a modification of the original, may have, at times, not allowed the raters to place the pelvis strap directly over the both ASIS. This placement may also be because of the smaller size of our subjects compared with the adult subjects of Gadjosik and Lusin. 2

Back to Top | Article Outline


The AKET demonstrates good interrater reliability when used in conjunction with the apparatus on children ages 10 to 13 years. Suggestions for additional research include studies of different populations, use of a more stable form of the apparatus, additional control for extraneous movement at the hip and ankle, and determination of the reliability of the AKET without the use of the apparatus.

Back to Top | Article Outline


The authors would like to acknowledge the following people for their assistance with this project: College Misericordia Physical Therapy Department; Bobbi Maudsley, EdD, PT, ATC; Lynette Chandler, PhD, PT; students and faculty of the Leo E. Solomon Educational Complex Grades 4 to 6; Maryanne W. Toole, Principal; Paul McGrane, Physical Education Teacher; William Biniek, EdD; Leo and Peggy Fedor; Paul and Sandy Canouse; Marilyn K. Milcavage; Joseph and Maryalice LaManna; and Joseph M. and Maryann F. Rakos.

The research for this study was completed to fulfill requirements at College Misericordia for the degree of Master of Science in Physical Therapy (D.M.R., K.A.S., R.L.F., M.L., C.C.Y.).

Back to Top | Article Outline


1. Kendall FP, McCreary EK, Provance PG. Muscle Testing and Function. 4th ed. Baltimore: Williams & Wilkins; 1993: 44–46.
2. Gadjosik R, Lusin G. Hamstring muscle tightness: reliability of an active-knee-extension test. Phys Ther. 1982; 62; 1269–1274.
3. Bohannon RW. Cinematographic analysis of the passive straight-leg-raising test for hamstring muscle length. Phys Ther. 1982; 62: 1269–1274.
4. Bohannon RW, Gadjosik R, LeVeau BF. Relationship of pelvic and thigh motions during unilateral and bilateral hip flexion. Phys Ther. 1985; 65: 1501–1504.
5. Kippers V, Parker AW. Toe-touch test: a measure of its validity. Phys Ther. 1987; 67: 1680–1684.
6. Gadjosik RL, Rieck MA, Sullivan DK, et al. Comparison of four clinical tests for assessing hamstring muscle length. J Orthop Sports Phys Ther. 1994; 18: 845–852.
7. Cormbleets L, Woolsey NB. Assessment of hamstring muscle length in school-age children using the sit-and-reach test and the inclinometer measure of the hip joint angle. Phys Ther. 1996; 76; 850–855.
8. Bandy WD, Irion JM. The effect of time on static stretch on the flexibility of the hamstring muscle. Phys Ther. 1994; 74; 845–852.
9. Portney LG, Watkins MP. Foundations of Clinical Research: Application to Practice. Norwalk, Conn: Appleton & Lange; 1993.
10. Rheault W, Miller M, Nothnagel P, et al. Intertester reliability and concurrent validity of fluid-based and universal goniometers for active knee flexion. Phys Ther. 1998; 68: 1676–1678.
11. Rothstein JM, Miller PJ, Roettger RF. Goniometric reliability in a clinical setting. Phys Ther. 1983; 63: 1611–1615.
12. Norkin CC, White DJ. Measurement of Joint Motion: A Guide Goniometry. 2nd ed. Philadelphia: FA Davis Company; 1995: 124–125.

reproducibility of results; evaluation studies; child; muscle; biomechanics/skeletal; knee joint/physiology

© 2001 Lippincott Williams & Wilkins, Inc.