Gribble 2003 Star Excursion Balance Test

Considerations for NormalizingMeasures of the Star Excursion Balance

Test

Phillip A. Gribble and Jay HertelDepartment of Kinesiology

Athletic Training Research LabPennsylvania State University

This study was designed to examine the role of foot type, height, leg length, andrange of motion (ROM) measurements on excursion distances while performing theStar Excursion Balance Test (SEBT), a test of dynamic postural control. Participants(n = 30) performed 3 trials of the SEBT in each of the 8 directions while balancing onthe right and left legs. No statistically significant relations were found between foottype or ROM measurements and excursion distances with the SEBT. Significant cor-relations were revealed between height and excursion distance and leg length and ex-cursion distance with leg length having the stronger correlation. Using raw excursionmeasures, males were found to have significantly greater excursion distances than fe-males; however, after normalizing excursion distances to leg length, there were nosignificant differences related to gender. In conclusion, when using the SEBT for ex-perimental or clinical purposes, participants’ excursion distances should be normal-ized to leg length to allow for a more accurate comparison of performance amongparticipants.

Key words: postural control, foot type, dynamic balance

Measurement of postural control is an important tool in the assessment of pediat-ric, geriatric, and athletic populations for establishing levels of neuromuscularfunction for the purposes of injury prevention and rehabilitation. Postural controlis often described as being either static (attempting to maintain a position with

MEASUREMENT IN PHYSICAL EDUCATION AND EXERCISE SCIENCE, 7(2), 89–100Copyright © 2003, Lawrence Erlbaum Associates, Inc.

Requests for reprints should be sent to Phillip Gribble, Penn State University, 266 RecreationBuilding, University Park, PA 16802. E-mail: [email protected]

minimal movement) or dynamic (maintaining a stable base of support while com-pleting a prescribed movement; Winter, Patla, & Frank, 1990).

Static postural control is commonly quantified through instrumented measure-ments of ground reaction forces or less sophisticated non-instrumented means. Aclinician or researcher may assess static postural control by having an individualattempt to maintain a stationary position while standing on either one or both feet.Ground reaction forces may be measured by having a participant balance on aforce plate. Postural control is then typically quantified through various measuresof velocity, area, or variability of the ground reaction forces or a related variable(Guskiewicz & Perrin, 1996). Non-instrumented measures include variables suchas time that a participant can maintain the prescribed stance (Freeman, Dean, &Hanham, 1965) or subjective error scoring systems (Riemann, Caggiano, &Lephart, 1999). A common clinical example of static postural control assessmentis the modified Rhomberg test, first described by Freeman (Freeman et al., 1965).This test is performed by having participants stand as motionless as possible, onone foot, as a series of task demands are added to challenge the postural controlsystem. These task demands include closing the eyes, tilting the head up, andtouching an index finger to the nose. Although this test is commonly used in the as-sessment of cerebral concussion (bilateral stance) and lower extremity joint inju-ries (unilateral stance), it typically does not place strength or movement demandson the participant.

Dynamic postural control often involves completion of a functional task with-out compromising one’s base of support. The advantage of assessing dynamic pos-tural control is that additional demands of proprioception, range of motion (ROM),and strength are required along with the ability to remain upright and steady. Nu-merous tests have been developed to assess dynamic postural control in the pediat-ric (Donahoe, Turner, & Worrell, 1994) and geriatric (Berg et al., 1990;Rossiter-Fornoff, Wolf, Wolfson, & Buchner, 1995; Tinetti 1986) populations, butvery few tests that truly stress the dynamic balance capabilities of the healthy, ath-letic population. The Star Excursion Balance Test (SEBT) is one such test that pro-vides a significant challenge to an athlete’s postural control system (Earl & Hertel,2001; Gray, 1995; Hertel, Miller, & Denegar, 2000; Kinzey & Armstrong, 1998;Miller, 2001; Olmsted, Carcia, Hertel, & Shultz, 2002).

The SEBT involves having a participant maintain a base of support with one legwhile maximally reaching in different directions with the opposite leg, withoutcompromising the base of support of the stance leg. Strong intra-rater reliability ofmeasurements with the SEBT has been demonstrated by Kinzey and Armstrong(1998) [ICC (2,1): 0.67–0.87)] and Hertel et al. (2000) [ICC (2,1): 0.81–0.96]. TheSEBT has shown sensitivity in screening for functional deficits related tomusculoskeletal injuries (Earl, 2002; Miller, 2001; Olmsted et al., 2002). In addi-tion, Earl and Hertel (2001) demonstrated the usefulness of the SEBT for the re-

90 GRIBBLE AND HERTEL

cruitment of lower extremity musculature contraction and discussed its applicationin rehabilitating various lower extremity musculoskeletal injuries.

Researchers have provided evidence that the SEBT is sensitive for screeningmusculoskeletal impairments, such as chronic ankle instability (Olmsted et al.,2002), quadriceps strength deficits (Miller, 2001), and patellofemoral pain syn-drome (Earl, 2002). Olmsted et al. reported decreased reaching distances duringperformance of the SEBT in patients with chronic ankle instability compared tomatched-healthy control participants.

Miller (2001) correlated quadriceps strength deficits with performance of theSEBT among subjects who had undergone anterior cruciate ligament reconstruc-tion (ACLR). The ACLR subjects who demonstrated a quadriceps strength deficitduring isokinetic testing also demonstrated decreased anterior reaching distancewhile performing the SEBT compared to uninjured matched control subjects.

Earl (2002) found that when assessing function among patients withpatellofemoral pain (PFP), the patients with PFP had significantly reduced perfor-mance on the SEBT compared to matched controls (Earl, 2002). Following 6weeks of structured rehabilitation, the PFP subjects had improved their SEBT per-formance such that there no longer existed a deficit between the PFP and controlgroups.

Various predictors of static postural control measures, such as muscle strength,architectural foot type, and mental status, have been studied (Hertel, Gay, &Denegar, 2002; Topp, Estes, Dayhoff, & Suhrheinrich, 1997). The SEBT has beenshown to be a reliable and valid instrument for assessing dynamic postural control(Gray, 1995; Hertel et al., 2000; Kinzey & Armstrong, 1998; Olmsted et al., 2002);however, predictive factors for performance of the SEBT have not been examined.This study was designed to examine the role of gender, foot type, height, leglength, and lower extremity ROM measurements on dynamic postural control asassessed through the SEBT and to determine the need for normalization of perfor-mance data.

METHODS

Participants

Thirty (12 men, 18 women) recreationally active participants volunteered for thisstudy. Descriptive data of the participants is presented in Table 1. All partici-pants self-reported that they were free of vestibular disorders, cerebral concus-sions, and lower extremity injuries during the previous 6 months. The study wasapproved by the university’s institutional review board, and each participant readand signed an approved informed consent form, in concordance with the univer-sity’s Institutional Review Board.

NORMALIZATION CONSIDERATIONS OF THE SEBT 91

Measurements of Predictive Factors

Foot type, leg length, hip internal and external ROM, and ankle dorsiflexion weremeasured bilaterally in all participants by the same investigator. Participants’ feetwere classified into one of three foot-type categories according to the proceduresdescribed by Root, Orien, and Weed (1971) resulting in 22 pes planus, 26 pesrectus, and 12 pes cavus feet.

Height was measured with a standard height chart. Leg length was measured oneach limb with participants lying supine. A tape measure was used to quantify thedistance from the anterior superior iliac spine to the center of the ipsilateral medialmalleolus. Hip internal and external rotation was measured with a standardgoniometer with the participants lying prone. Dorsiflexion at the ankle was alsomeasured with a standard goniometer with the participants standing with one footin front of the other while leaning forward until the heel of the posterior foot beganto lift off the ground (Denegar, Hertel, & Fonseca, 2002).

SEBT Procedures

The SEBT was performed with the participants standing in the middle of a gridformed by eight lines extending out at 45° from each other (see Figure 1). The par-ticipant was asked to reach as far as possible along each of the eight lines, make alight touch on the line, and return the reaching leg back to the center, while main-taining a single-leg stance with the other leg in the center of the grid (see Figure 2).Participants were instructed to make a light touch on the ground with the most dis-tal part of the reaching leg and return to a double-leg stance without allowing thecontact to affect overall balance. The terminology of excursion directions is basedon the direction of reach in relation to the stance leg (see Figure 1). When reachingin the lateral and posterolateral directions, participants must reach behind thestance leg to complete the task.

Participants were allowed to practice reaching in each of the eight directions sixtimes to minimize the learning effect (Hertel et al., 2000). Following a 5-min rest


TABLE 1Means and Standard Deviations Summary Statistics for Participants

GenderAge

(Years) Height (cm) Mass (kg)

LegLength(cm)

Hip ER(degrees)

Hip IR(degrees)

Ankle DF(degrees)

Male(n = 12)

23.2 ± 3.8 69.2 ± 2.1 79.2 ± 11.5 89.9 ± 4.3 41.2 ± 9.2º 34.4 ± 8.4º 27.0 ± 3.0º

Female(n = 18)

22.4 ± 1.4 66.4 ± 3.2 62.0 ± 8.8 87.4 ± 5.0 43.1 ± 11.1º 43.2 ± 10.1º 27.2 ± 5.2º

Note. DF = dorsiflexion; ER = external rotation; IR= internal rotation.

period, participants performed three trials in each of the eight directions. They be-gan with the anterior direction and progressed clockwise around the grid. All par-ticipants began with a right stance leg in the center of the grid. After completion ofthe three trials in the eight directions and another 5-min rest period, the test contin-ued with a left stance leg.

The investigator recorded each reach distance with a mark on the tape as the dis-tance from the center of the grid to point of maximum excursion by the reach leg.At the conclusion of all trials, the investigator measured the distances of each ex-cursion with a standard tape measure. If the investigator felt the participant used


FIGURE 1 Reaching directions for the Star Excursion Balance Test.

FIGURE 2 Performance of the Star Excursion Balance Test with a left stance leg reachinginto the posteromedial direction.

the reaching leg for a substantial amount of support at any time, removed his or herfoot from the center of the grid, or was unable to maintain balance on the supportleg throughout the trial, the trial would be discarded and repeated.

Statistical Analysis

Dependent t tests were performed to compare each of the eight excursion distancesof the right and left limbs of participants. Because no significant differences (p >.05) were identified, data from the right and left limb trials were averaged. A seriesof eight independent t tests was used to examine the differences in normalized ex-cursion distances in the eight directions as a function of gender. Eight separateanalysis of variance (ANOVA), with one between-groups factor (foot type), werecomputed; one for each excursion direction. Pearson product–moment correla-tions were calculated to explore the bivariate relations between excursion distanceand height, leg length, hip internal ROM, hip external rotation ROM, and ankledorsiflexion ROM.

Leg length and height were both found to be significantly correlated, to reachdistance in the majority of the directions (see Table 2), thus the Pearson prod-uct–moment correlation between height and leg length was calculated (r = .89).Because leg length was found to be the most highly correlated factor, and leglength and height were highly correlated, excursion distances were normalized tothe participant’s leg length for further analysis. Normalization was performed bydividing each excursion distance by a participant’s leg length, and then by multi-plying by 100. Normalized values can thus be viewed as a percentage of excursionsdistance in relation to a participant’s leg length.

Eight separate ANOVAs with one between-groups factor (foot type) were thencomputed on the normalized excursion distances. In addition, a series of eight in-dependent samples t tests were utilized to examine the differences in normalized


TABLE 2Relationship Between Height and Excursion and Leg Length and

Excursion (R2 Values)

Reaching Direction Height Leg Length

Ant 0.19*** 0.23***Antlat 0.11* 0.18**Lat 0.01 0.02Postlat 0.04 0.04Post 0.10* 0.10*Postmed 0.13** 0.14**Med 0.14** 0.11**Antmed 0.18*** 0.19***

*p < .05. **p < .01. ***p < .001.

excursion distances as a function of gender. The experiment wise alpha level wasset at p < .05. For the series of t tests (limb, gender) and ANOVAs (foot type),Bonferroni corrections were utilized, resulting in an adjusted significance level ofp < .006 (see note in Table 3).

RESULTS

The dependent t tests revealed no significant differences between the right and leftlimb excursion distances; therefore, data for both limbs were combined for subse-quent analyses. Height, leg length, and ROM measurements are presented in Table1. A significant correlation (p < .05) was found between height and excursion dis-tance, and leg length and excursion (p < .05) distance in six of the eight directions:anterior, anteromedial, medial, posteromedial, posterior, and anterolateral. For thesignificant correlations, r2 values ranged from .10 to .19 between height and excur-sion distances, and from .10 to .23 between leg length and excursion distances (Ta-ble 2). Stronger correlations were found between leg length and excursion dis-tances than between height and excursion distances. Height and leg length werefound to strongly correlate to each other (r2 = .77, p < .05). No significant correla-tions were found between internal rotation of the hip, external rotation of the hip,or dorsiflexion of the ankle and excursion distances (p > .05).

Men were found to have significantly (p < .006) greater raw excursion distancesthan women in three of the eight reaching directions (posterior, posteriomedial,


TABLE 3Differences in Normalized Excursion Distances for Men and Women

ReachingDirections

Raw Scores (cm)* Normalized (% of leg length)

Male Female p Male Female p

ANT 71.2 ± 7.4 67.1 ± 5.4 .014 79.2 ± 7.0 76.9 ± 6.2 .192ALAT 66.4 ± 8.0 65.1 ± 6.5 .493 73.8 ± 7.7 74.7 ± 7.0 .657LAT 71.9 ± 15.5 69.6 ± 12.0 .508 80.0 ± 17.5 79.8 ± 13.7 .962PLAT 81.2 ± 11.9 74.6 ± 11.5 .035 90.4 ± 13.5 85.5 ± 13.2 .176POST 84.4 ± 9.6 74.4 ± 11.4 .001* 93.9 ± 10.5 85.3 ± 12.9 .009PMED 86.0 ± 8.1 77.7 ± 10.1 .001* 95.6 ± 8.3 89.1 ± 11.5 .020MED 87.8 ± 8.5 79.1 ± 9.2 <.0005* 97.7 ± 9.5 90.7 ± 10.7 .012AMED 76.7 ± 7.7 72.5 ± 6.4 .026 85.2 ± 7.5 83.1 ± 7.3 .277

Note. Means of three trials and standard deviation values in each reaching direction. Scores ad-here to a Bonferroni correction level of significance of p < .006 (.05 divided by number of reaching di-rections, 8).

*p < .006.

medial); however, following normalization no significant differences (p > .006)between genders were identified. These results are listed in Table 3.

No significant differences between foot types were found in any of the eight di-rections for raw excursion distances or normalized excursion distances (p > .05).These results are illustrated in Figures 3 and 4.


FIGURE 3 Raw excursion distances for the eight directions by foot type. ALAT =anterolateral; AMED = anteromedial; ANT = anterior; LAT = lateral; MED = medial; PLAT =posterolateral; POST = posterior; PMED = posteromedial.

FIGURE 4 Normalized excursion distances for the eight directions by foot type.

DISCUSSION

Our results demonstrate that height and leg length were positively related to per-formance on the SEBT. The importance of normalizing excursion distance data(by factoring out leg length) was illustrated by the existence of significant differ-ences in excursion distances between genders on raw excursions distance scores;however, a lack of gender differences was found following normalization of excur-sion distances to leg length. When using the SEBT as an assessment tool, consider-ations for normalization should include leg length. This includes either normaliz-ing excursion data to leg length or matching paired participants for leg length. Nosignificant relations were identified between foot type, hip rotation ROM, or ankledorsiflexion ROM and excursion distances.

To contribute to the body of knowledge of the SEBT, we attempted to examineseveral factors that could influence postural control among healthy, physically ac-tive individuals. The SEBT is designed to challenge posture during multiple reach-ing tasks; and the reliability (Hertel et al., 2000; Kinzey & Armstrong, 1998) aswell as sensitivity (Earl, 2002; Miller, 2001; Olmsted et al., 2002) has been previ-ously established. No previous researchers have attempted to examine factors thatmay potentially contribute to performance. While our results do establish correla-tions between leg length, height, and reach distance, the correlation values are low,establishing the need for further investigation into performance contributions dur-ing the SEBT.

The SEBT involves maximizing lower extremity reach distance with one limbwhile maintaining balance on the contralateral limb. Logically leg length wouldcorrelate significantly with excursion distance, as a longer limb would give a par-ticipant an advantage in reaching that limb further. In addition, because height andleg length strongly correlate with each other, it is inherent that height and excur-sion distance would also correlate significantly. While the correlations for heightand leg length to excursion distance were significant, they were not especiallystrong. The highest correlations occurred between leg length and excursion dis-tance in the three anterior directions (anterior, anterolateral, anteromedial), with anr2 value of just .23. This indicates that while leg length is a significant predictor ofperformance on the SEBT, other factors not assessed for in this study account forthe majority of the variance associated with excursion distance.

In a recent study by Hertel et al. (2002), the relation between foot type and staticpostural control was examined. Participants with cavus feet demonstrated signifi-cantly greater center of pressure area measures compared to rectus feet. The resultsof our study show that foot type was not similarly related to dynamic postural con-trol as assessed with the SEBT. During static measurements of postural control, thebody remains relatively fixed over the base of support. Hertel et al. (2002) specu-lated that the reason for impaired postural control in participants with cavus feetwas related to a smaller amount of contact of the plantar surface of the foot with the


ground. The lack of an anatomical block through contact allows the base of supportin a pes cavus foot to shift medially and laterally. In addition, perhaps a lack of con-tact surface limited the amount of cutaneous feedback, which would decreasecompensatory actions leading to increased postural sway.

In a static task of postural control, the goal is to minimize displacement of thecenter of pressure, a derivative of the vertical ground reaction force. In the measureof dynamic postural control used in this study, the goal is to maximize reach dis-tance while maintaining unilateral support. Hertel et al. (2002) demonstrated thatcenter of pressure excursion could be affected by the amount of ground contact re-lated to foot type. Postural sway is not quantified in the SEBT, but it is inherent tothe test that some shifting of the center of pressure will occur to maximize excur-sion distance. Foot type, however, did not significantly affect the performance ofthe SEBT, possibly due to other compensatory motions or reaching strategies, orboth, that allow a subject to overcome a deficit potentially due to foot type.

No relation was found between ROM measurements at the hip and ankle andperformance on the SEBT. We allowed participants ample practice of the SEBT,but we did not dictate their strategy for achieving maximum reaching distance be-yond what was described. Thus, an individual could incorporate a variety of move-ment patterns for positioning of the trunk as well as the joints of the upper andlower extremities. Variations in ROM at the hip and ankle among individuals didnot affect overall dynamic balance performance with the SEBT. Because we didnot require a specific pattern of movement to achieve maximum performance, anindividual conceivably could overcome a deficit in range of motion at one joint byusing more ROM at another joint to achieve the specified goal. Further research iswarranted to explore movement patterns associated with performance of theSEBT.

Another possible predictor of performance that was not investigated in thisstudy was strength. The SEBT requires neuromuscular control though proper jointpositioning as well as strength in surrounding musculature to create and maintainthe necessary positions throughout the test. Future researchers should investigatethe relation of muscle strength and fatigue of various lower extremity musclegroups and performance on the SEBT. Other physical factors that were not exam-ined in this study that may be associated with variations in performance include thefollowing: strength, neuromuscular control, and ROM at additional joints. Byidentifying arthropometric predictors of SEBT performance (height, leg length)and normalizing to these measures, perhaps a more accurate assessment of pos-tural control would be available.

A potential limitation in our study is that the order of trials was not counterbal-anced. The order of testing followed a dominant, non-dominant stance leg order,always in the clockwise direction. Future researchers should include a randomizedorder of testing to avoid potential confounding factors such as learning effect andfatigue.


The SEBT is a promising test of postural control that may be useful in assessingfunctional deficits in those with lower-extremity orthopedic injuries. The results ofthis study suggest that among young, physically active individuals, leg length ofthe participants must be considered in normalizing performance data. We recom-mend that when using the SEBT for experimental purposes, investigators shouldeither normalize excursion distances to participant leg length or use control partici-pants who are matched to experimental participants according to leg length.

ACKNOWLEDGMENT

We thank Anthony Piegaro, ATC, for his assistance in data collection during thisproject.

REFERENCES

Berg, K., Wood-Dauphinee, S., Williams, J. I., & Maki, B. (1992). Measuring balance in the elderly:Validation of an instrument. Canadian Journal of Public Health, 83(Supp-2), S7–S11.

Denegar, C., Hertel, J., & Fonseca, J. (2002). The effect of lateral ankle sprain on dorsiflexion range ofmotion, posterior talar glide, and joint laxity. Journal of Orthopedic and Sport Physical Therapy, 32,166–173.

Donahoe, B., Turner, D., & Worrell, T. (1994). The use of functional reach as a measurement of balancein boys and girls without disabilities age 5 to 15 years. Pediatric Physical Therapy, 6, 189–193.

Duncan, P. W., Weiner, D. K., Chandler, J., & Studenski, S. (1990). Functional reach: A new clinicalmeasure of balance. Journal of Gerontology, 451, 92–197.

Earl, J. (2002). Relationships among dynamic malalignment, neuromuscular rehabilitation, andpatellofemoral pain syndrome. Unpublished doctoral dissertation, Pennsylvania State University,University Park.

Earl, J., & Hertel, J. (2001). Lower-extremity muscle activation during the Star Excursion BalanceTests. Journal of Sport Rehabilitation, 10(2), 93–104.

Freeman, M. A. R., Dean, M. R. E., & Hanham, I. W. F. (1965). The etiology and prevention of func-tional instability of the foot. Journal of Bone and Joint Surgery, 47B, 669–677.

Gray, G. W. (1995) Lower extremity functional profile. Adrian, MI: Wynn Marketing, Inc.Guskiewicz, K. M., & Perrin, D. H. (1996). Research and clinical applications of assessing balance.

Journal of Sport Rehabilitation, 5(1), 45–63.Hertel, J., Gay, M., & Denegar, C. (2002). Differences in postural control during single leg stance

among healthy individuals with different foot types. Journal of Athletic Training, 37(2), 129–132.Hertel, J., Miller, S., & Denegar, C. (2000). Intratester and intertester reliability during the Star Excur-

sion Balance Test. Journal of Sport Rehabilitation, 9(2), 104–116.Kinzey, S., & Armstrong, C. (1998). The reliability of the Star-Excursion test in assessing dynamic bal-

ance. Journal of Orthopedic and Sport Physical Therapy, 27, 356–360.Miller, J. (2001). Biomechanical analysis of the anterior balance reach test. Unpublished doctoral dis-

sertation, Pennsylvania State University, University Park.Olmsted, L. C., Carcia, C. R., Hertel, J., & Shultz, S. J. (2002). Efficacy of the Star Excursion Balance

Test in detecting reach deficits in subjects with chronic ankle instability. Journal of Athletic Training,37(4), 501–506.


Riemann, B. L., Caggiano, N. A., & Lephart, S. M. (1999). Examination of a clinical method of assess-ing postural control during a functional performance task. Journal of Sport Rehabilitation, 8,171–183.

Root, M., Orien, W., & Weed, J. (1971). Biomechanical examination of the foot, Vol. 1. Los Angeles:Clinical Biomechanics Corporation.

Rossiter-Fornoff, J. E., Wolf, S. L., Wolfson, S. L., & Buchner, D. M. (1995). A cross-sectional valida-tion study of the FICSIT common data base static balance measures. Frailty and injuries: Coopera-tive studies of intervention techniques. Journal of Gerontology, 50a, 291–297.

Tinetti, M. E. (1986). Performance-oriented assessment of mobility problems in elderly patients. Jour-nal of the American Geriatric Society, 34, 119–126.

Topp, R., Estes, P. K., Dayhoff, N., & Suhrheinrich, J. (1997). Postural control and strength and moodamong older adults. Applied Nursing Research, 10(1), 11–18.

Winter, D. A., Patla, A. E., & Frank, J. S. (1990). Assessment of balance control in humans. MedicalProgress Through Technology, 16(1), 31–51.


Documents

Gribble 2003 Star Excursion Balance Test