A hip test battery for clinical practice: reliability, reference values, and the pitfalls of the Limb Symmetry Index

Abstract

Background

Reliable and clinically interpretable assessments are essential to guide evaluation, monitoring, and return-to-sport decision-making after hip injury or surgery. However, no comprehensive test battery currently integrates intra- and inter-rater reliability, reference values in healthy individuals, and the reliability of the Limb Symmetry Index (LSI). We propose an integrated clinical assessment framework linking measurement reliability (ICC), interpretability (SEM, MDC, reference values), and clinical decision-making.

Methods

This cross-sectional reliability study included 94 healthy recreational adults (mean age 28.4 ± 8.5 years; 39% women). A subsample (n = 20) underwent a retest after 7 ± 2 days to assess intra-rater reliability. Outcome measures included hip range of motion (flexion, internal and external rotation), isometric hip strength (flexors, abductors, adductors), and functional performance tests (medial hop, single-leg squat, modified Star Excursion Balance Test [SEBTm], and endurance squat). The LSI was calculated for relevant outcomes. Reliability was assessed using intraclass correlation coefficients (ICC [2,1]), along with standard error of measurement (SEM) and minimal detectable change at the 95% confidence level (MDC95%).

Results

Most hip strength and range-of-motion measures demonstrated good reliability (ICC > 0.70), with MDC values of 55–79 N for strength and 12–18° for hip flexion. Medial hop performance showed excellent reliability (ICC = 0.89–0.95; MDC = 18–26 cm), and SEBTm demonstrated low measurement error (MDC = 9–14 cm). In contrast, single-leg squat and endurance squat tests showed lower reliability. Medial hop performance was strongly correlated with hip abductor (r = 0.52) and adductor strength (r = 0.51). All LSIs demonstrated poor reliability (ICC < 0.50; MDC range: 11.9–47.9%, mean ≈ 27%).

Conclusions

Range-of-motion measures, particularly active and passive hip flexion, demonstrated the highest reliability, with MDC values ranging from 12° to 18°. Strength tests showed good reliability but substantial measurement error, limiting their ability to detect meaningful individual changes. Among functional assessments, the medial hop demonstrated excellent reliability. In contrast, adductor-to-abductor strength ratios and the Limb Symmetry Index (LSI) exhibited poor reliability and large MDC values (≈ 27%), questioning their usefulness as standalone criteria for clinical decision-making and return-to-sport assessment. These findings support prioritizing absolute measures interpreted against their MDC and reference values within a multidimensional assessment framework.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12891-026-10248-7.

Introduction

Hip injuries account for 10–16% of sport-related musculoskeletal disorders, particularly in disciplines demanding high pelvic–femoral control such as football, hockey, dance, and martial arts [1, 2]. These conditions include groin pain, muscle and tendon injuries, labral tears, and femoroacetabular impingement (FAI), often associated with measurable biomechanical impairments [3]. The Doha consensus provides a clinical framework for hip and groin pain classification, yet the associated impairments and deficits remain challenging to assess reliably [4, 5].

Athletes with hip pain frequently exhibit reduced range of motion (especially flexion and internal rotation) [6, 7], muscle weakness of adductors, abductors, and flexors, and functional deficits in tasks such as single-leg squats and hop tests [3, 8]. Imbalances in adductor–abductor strength ratios have also been linked to injury risk [9]. Previous studies have reported that individuals with hip-related pain may demonstrate specific impairments without consistent alterations in functional performance highlighting the complexity of the relationship between impairments and functional capacity [7]. However, performance in lateral or multidirectional tasks, such as hop tests, may be partially influenced by hip strength, particularly in tasks requiring force generation and dynamic stabilization [10]. Current literature supports a multimodal assessment approach combining mobility, strength, and functional performance to guide clinical decision-making in individuals with hip-related pain [11]. However, hip assessment remains fragmented: studies differ in populations, tools, protocols, and reporting methods, limiting comparability and clinical applicability [12, 13].

In clinical practice, the Limb Symmetry Index (LSI) remains widely used in rehabilitation and return-to-sport decision-making [14–17]. Its popularity is largely explained by its simplicity, ease of calculation, and its ability to provide an individualized comparison when reference values are lacking. In many settings, LSI represents one of the few readily available metrics to monitor recovery over time. When used as part of a broader, multidimensional assessment, as recommended in the literature [3], LSI may still provide clinically meaningful information regarding inter-limb asymmetries. However, its interpretation remains challenging, as it may mask bilateral deficits, reflect compensatory strategies, and is associated with questionable reliability [14–17].

The primary aim was to evaluate the intra- and inter-rater reliability of a structured, hip-specific test battery in healthy active adults. Secondary aims were to (i) establish reference values for recreationally active adults, (ii) explore associations between hip-specific strength and functional performance, and (iii) assess the reliability of the LSI.

Methods

This cross-sectional study adhered to GRRAS, STROBE, and COSMIN guidelines for reliability research and complied with the Declaration of Helsinki. Ethical approval was obtained from the C.H.U. Brugmann Ethics Committee (Brussels, Belgium; Ref. B0772025000042), and all participants provided written informed consent. Data were anonymized.

Participants

Participants were recruited voluntarily from the Faculty of Human Movement Sciences (Université Libre de Bruxelles) and tested between 13 May and 22 June 2025 in a controlled laboratory setting. Eligible participants were healthy, recreationally active adults (≥ 18 years; ≥ 2 training sessions per week) without current or past hip disorders. Inclusion criteria included International Hip Outcome Tool (iHOT-12-fr) score ≥ 90/100 and sufficient French proficiency. In asymptomatic adults younger than 40 years, the mean iHOT score is approximately 94 [18]. Accordingly, a threshold of ≥ 90 was considered appropriate to identify healthy hips while accommodating the normal variability in iHOT scores observed even in symptom-free individuals. Exclusion criteria were previous hip surgery; lower-limb injury within the past 6 months or confirmed/suspected pregnancy.

Testing conditions

All assessments were conducted in a standardized, controlled environment. Prior to data collection, the examiners completed a standardized half-day training session delivered by the physiotherapist responsible for the study (JB), based on a written standardized protocol (Appendix A). The session comprised a 2-hour demonstration of the full test battery followed by 1 h of supervised practice. The examiners then performed four assessments under real testing conditions, during which data were collected, and measurement variability was reviewed under supervision to ensure consistent and adequate test administration before the study began. Assessments were performed by six final-year physiotherapy students, organized into three pairs, with each pair assessing approximately 30 participants under the supervision of a senior physiotherapist throughout. Intra- and inter-rater reliability were assessed using a fixed protocol. For inter-rater analysis, participants performed the full test-battery with one examiner, rested for 10 min, then repeated the tests with a second examiner, with the order of examiners randomized. To ensure blinding, the two examiners worked independently, and the second examiner was blinded to the first examiner’s scores, with no access to any previous results. For intra-rater analysis, the first 20 participants consecutively enrolled were reassessed after 7 ± 2 days by the same examiner, under identical conditions. Tests were administered in a fixed order (active followed by passive range of motion, isometric strength, and finally functional tests) to ensure reproducibility and clinical applicability. For range of motion assessments, a single measurement was recorded for each condition. For strength testing, a familiarization trial at approximately 50% of maximal effort was performed prior to testing, followed by three maximal trials. Functional tests were also performed three times. The mean of the three trials was used for reliability analysis and for the calculation of the (LSI). A standardized rest period of 15 s was applied between trials, except for passive range of motion assessments.

Range of motion

Hip flexion, internal rotation, and external rotation were assessed sequentially in active and then passive modes using a digital inclinometer. Rotational movements were tested in two distinct positions: hip flexed and hip neutral. Protocols were adapted from Nussbaumer et al [19]., ensuring high reproducibility (ICC > 0.90). To ensure measurement consistency and limit compensatory movements, the onset of contralateral limb movement was carefully monitored during testing and used as an indicator of end-range motion. Any observable compensatory strategies or pelvic deviations were noted and used to define the limit of the measurement.

Strength

The handheld dynamometer (MusTec – BioFET, V3) was stabilized with an external belt for all isometric tests, eliminating examiner-dependent counterforce. The lever arm was defined as lower-limb length and measured individually with a tape measure as the distance from the greater trochanter to the apex of the medial malleolus; strength values were then converted to joint moments and normalized to body mass (N·m·kg⁻¹) [7, 20, 21]. Dynamometer placement followed anatomical landmarks, positioned 3 cm above the malleolus or at the midfoot depending on the muscle group [22–25]. Participants performed three 6-second maximal voluntary isometric contractions, separated by 15 s of rest, with standardized verbal encouragement. Adductor-to-abductor strength ratios were calculated. To minimize compensatory movements, specific stabilization procedures were applied for each test. During hip flexion testing, the tested limb was positioned with the hip flexed, the foot flat, and aligned with the limb axis, with participants instructed to maintain alignment throughout the contraction. For hip abduction, similar alignment constraints were applied to avoid pelvic tilt or trunk compensation. During hip adduction testing, a semi-rigid support was placed against the contralateral limb to limit pelvic motion, and the tested foot was maintained parallel to the table without rotation.

Functional tests

The functional battery included the medial hop test (distance), Single Leg Squat (SLS; scored on an 8-point qualitative observation grid [26]), the modified Star Excursion Balance Test (SEBTm; distance), and a squat endurance test at 100° of knee flexion (time in seconds). Participants performed three trials for each test, except squat endurance (one trial) [27]. Performance in the medial hop and SEBTm was normalized to limb length (greater trochanter to the floor [28]). For SEBTm, reach distances in three standard directions, anterior (ANT), posteromedial (PM), and posterolateral (PL), were expressed as a percentage of limb length using the following formula [29] . A composite score (COMP) was calculated as the mean of ANT, PM, and PL normalized values.

Sample size and statistical analysis

The study followed COSMIN and GRRAS guidelines. A priori, ICC(2,1) (two-way random, absolute agreement, single measures) ≥ 0.7 was set as the minimal reliability threshold [30]. Based on Bonett [31] and Mokkink et al,. [30] ~ 100 participants were targeted for inter-rater analysis (expected ICC = 0.70, 95% CI width ≈ 0.30), and 20–25 for intra-rater analysis, allowing for 5–10% attrition. Ninety-four participants completed inter-rater testing; 20 were reassessed for intra-rater reliability. Descriptive statistics (means, SD, 95% CI) were calculated for all collected outcomes using JASP software (version 0.17.2.1). Normality was assessed via Shapiro–Wilk test and visual inspection; paired t-tests or Wilcoxon tests were applied as appropriate. Reliability analyses and reference values were calculated using the right limb only, in order to ensure independence of observations and avoid statistical dependence between limbs. However, both limbs were assessed for all participants, and bilateral data were used for the calculation of the LSI.

Reliability was assessed with ICC (two-way random, absolute agreement) [32], Both single measures ICC(2,1) and average measures ICC(2,k) were interpreted as poor (< 0.50), moderate (0.50–0.69), or good (≥ 0.70) [33]. SEM was calculated as SD × √(1–ICC), and MDC95% as SEM × 1.96 × √2 [34]. The SEM% was derived by normalizing the SEM to the mean value (SEM% = (SEM / mean) × 100), and the MDC% by normalizing the MDC to the mean (MDC% = (MDC / mean) × 100). The SEM and MDC are absolute indices expressed in the unit of each variable and, unlike the ICC, have no consensus classification thresholds. They were therefore expressed relative to the mean (SEM%, MDC%) to allow comparison across parameters, and interpreted through the MDC95 as the smallest individual change exceeding measurement error [34, 35]. For descriptive purposes only, relative error was additionally read against commonly used field-based benchmarks: SEM% ≤ 5% (good), 5–10% (acceptable), > 10% (poor), and MDC% < 10% (good), 10–30% (acceptable), > 30% (poor), which are reporting conventions rather than consensus standards [36]. Pearson correlations examined associations between medial hop and normalized hip strength and were interpreted as very weak (r < 0.20), weak (0.20–0.39), moderate (0.40–0.69), strong (0.70–0.89), or very strong (≥ 0.90) [37]. LSI was computed as (right limb / left limb) × 100 and its intra- and inter-rater reliability were evaluated [38]. Agreement of LSI measures was further examined with Bland–Altman analysis, reporting the mean bias and 95% limits of agreement (bias ± 1.96 SD) with their 95% confidence intervals; proportional bias was tested by regressing the inter-measure difference on the mean.

Results

Participants

A total of 94 healthy volunteers completed the full test battery for inter-rater reliability analysis. Among them, 20 participants also underwent a repeated session 7 ± 2 days later to assess intra-rater reliability. Participant characteristics are detailed in Table 1.

Table 1.

Participants characteristics

Variables	Data (mean ± SD; n = 94)
Age	28.4 ± 8.5
Height (cm)	175.6 ± 8.7
Weight (kg)	75 ± 14.6
Sex	Women (n = 37)
Dominant side	Left (n = 11)
Sportsᵇ	Strength & conditioning (n = 40),
	Running (n = 22),
	Soccer (n = 7),
	Tennis (n = 4),
	Rugby, basketball, cycling (n = 3 each),
	Volleyball, dance (n = 2 each),
	Other sports (n = 8) ª

ª Other sports (n = 1 each): handball, athletics, alpine skiing, yoga, badminton, active walking, gymnastics, grappling

ᵇ Most participants practiced their sport recreationally, without competition. A few took part in running events (without seeking qualification) or in team-sport matches, with occasional or district-level competition. A few participants had practiced sport at a high level during adolescence but were, at the time of the study, regular and active athletes without competitive practice

Abbreviations cm centimeters, kg kilograms, n number, SD standard deviation

Range of Motion (ROM)

Passive hip ROM exceeded active ROM by approximately 12.5° across all movements. Hip flexion demonstrated the highest reliability (ICC 0.76–0.85), whereas external rotation in flexion showed the lowest reliability (ICC ≤ 0.44 for active; ≤ 0.32 for passive). Results are presented in Fig. 1.

Fig. 1.

Active and passive range of motion results : a Flexion ; b Internal Rotation (Hip flexion) ; c External Rotation (Hip flexion) ; d Internal Rotation (Hip extension) ; e External Rotation (Hip extension). Active measurements are represented in blue and passive measurements in purple using violin plots. Mean and standard deviation (SD) values are reported to the right of each plot for both active and passive conditions. Reliability metrics are presented on the right side of the figure. Intraclass correlation coefficients (ICC) are first reported for active measurements (intra- and inter-rater), followed by passive measurements. Associated standard error of measurement (SEM), minimal detectable change (MDC), and their percentage values (SEM%, MDC%) are also provided. Color coding is used to indicate reliability levels: green (ICC ≥ 0.70, strong), orange (ICC 0.50–0.69, moderate), and red (ICC < 0.50, weak). Abbreviations: ROM = range of motion; IR = internal rotation; ER = external rotation; ICC = intraclass correlation coefficient; SEM = standard error of measurement; MDC = minimal detectable change; SD = standard deviation; CI = confidence interval; Intra/Inter = intra- and inter-rater reliability. Active and Passive indicate movement type

Strength

Most isometric strength outcomes demonstrated good reliability (ICC > 0.70). Hip abductor strength showed slightly lower inter-rater reliability (ICC = 0.69; 95% CI: 0.57–0.79), while the adductor/abductor ratio (mean 1.2 ± 0.3) exhibited poor reliability (ICC = 0.46 intra-rater; 0.31 inter-rater). Across isolated strength measures, SEM% values ranged from 14.20% to 17.80%, with MDC% between 40.37% and 48.22%. These metrics complement ICC estimates by quantifying absolute measurement error and the magnitude of change required to exceed measurement variability at the individual level. Results are presented in Fig. 2.

Fig. 2.

Maximal isometric hip strength: raw values (top panels) and normalized values (bottom panels). Panels: a Abduction; b Adduction; c Flexion; d Adduction/abduction ratio; e Abduction; f Adduction; g Flexion. Data distribution is represented using violin plots (purple), illustrating the dispersion of strength values. Raw strength values are displayed in the upper panels, while normalized values (N·m·kg⁻¹) are shown in the lower panels. Mean and standard deviation (SD) are reported to the right of each plot. Reliability metrics are also presented on the right, with the first line corresponding to intra-rater reliability and the second line to inter-rater reliability. For each, intraclass correlation coefficients (ICC) with 95% confidence intervals, standard error of measurement (SEM), minimal detectable change (MDC), and their percentage values (SEM%, MDC%) are provided. Color coding indicates reliability levels: green (ICC ≥ 0.70, strong), orange (ICC 0.50–0.69, moderate), and red (ICC < 0.50, weak). Abbreviations: Abd = abduction; Add = adduction; Add/Abd = adduction-to-abduction ratio; ICC = intraclass correlation coefficient; SEM = standard error of measurement; MDC = minimal detectable change; SD = standard deviation; CI = confidence interval

Functional tests

Medial hop and SEBTm demonstrated good reliability. Medial hop showed ICC = 0.89 (intra) and 0.95 (inter), with SEM% - MDC% of 10.32% − 28.63% (intra) and 6.96% − 19.30% (inter). The normalized hop ratio (hop distance / limb length) also demonstrated high reliability (ICC = 0.84–0.94), with SEM% - MDC% of 12.00% − 33.00% (intra) and 7.00% -20.00% (inter). Mean hop distance was 92.5 ± 28.8 cm, corresponding to a normalized ratio of approximately 1.0. SEBTm reliability was good across directions, with SEM% ranging from 4% to 7% and MDC% from 10% to 15%; normalized reach was 66.9% (ANT), 98.5% (PM), and 103.6% (PL), composite score 89.7%.

In contrast, single-leg squat showed poor reliability (ICC intra = 0.48; ICC inter = 0.50; MDC = 2.6 points), and squat endurance was inconsistent (ICC = 0.36 intra; 0.82 inter), with high intra-rater variability (MDC 121s and 64s, respectively). Results are presented in Figs. 3 and 4.

Fig. 3.

Functional tests results. (a) Medial hop; (b) Medial hop/Lower limb ;(c) Single leg squat ;(d) Squat endurance. Data distribution is represented using violin plots, illustrating the dispersion of performance values for each test. Mean and standard deviation (SD) are reported to the right of each plot. Reliability metrics are also presented on the right, with the first line corresponding to intra-rater reliability and the second line to inter-rater reliability. For each, intraclass correlation coefficients (ICC) with 95% confidence intervals, standard error of measurement (SEM), minimal detectable change (MDC), and their percentage values (SEM%, MDC%) are provided. Color coding indicates reliability levels: green (ICC ≥ 0.70, strong), orange (ICC 0.50–0.69, moderate), and red (ICC < 0.50, weak). Abbreviations: ICC = intraclass correlation coefficient; SEM = standard error of measurement; MDC = minimal detectable change; SD = standard deviation; CI = confidence interval; LL = lower limb length; s = seconds; points = single-leg squat scoring scale; ratio = normalized value (e.g., hop/LL); cm = centimeters

Fig. 4.

Modified Star Excursion Balance Test (SEBT) results. Measurements are expressed in centimeters (top panel) and as a percentage of lower limb length (bottom panel). The composite score represents the mean of normalized reach distances in the three directions: anterior, posteromedial, and posterolateral. Data distribution is represented using violin plots, illustrating the dispersion of reach distances for each direction. Mean and standard deviation (SD) are reported adjacent to each plot. Reliability metrics are presented alongside the distributions, with the first line corresponding to intra-rater reliability and the second line to inter-rater reliability. For each, intraclass correlation coefficients (ICC) with 95% confidence intervals, standard error of measurement (SEM), minimal detectable change (MDC), and their percentage values (SEM%, MDC%) are provided. The lower panel displays values normalized to lower limb length (%), including the composite score. Color coding indicates reliability levels: green (ICC ≥ 0.70, strong), orange (ICC 0.50–0.69, moderate), and red (ICC < 0.50, weak). Abbreviations: SEBT = Star Excursion Balance Test; ICC = intraclass correlation coefficient; SEM = standard error of measurement; MDC = minimal detectable change; SD = standard deviation; CI = confidence interval; Intra/Inter = intra- and inter-rater reliability

Association

Medial hop performance (normalized to limb length) showed moderate correlations with hip strength: r = 0.52 for abductors and r = 0.52 for adductors (p < 0.001). Corresponding R² values (0.27 and 0.27) indicate that strength explains ~ 26–27% of hop performance variance, suggesting additional factors influence functional capacity. Results are presented in Fig. 5.

Fig. 5.

Associations between medial hop performance (normalized to limb length) and hip strength: a abductors, b adductors. Abbreviations: ABD = abduction; ADD = adduction; N/kg = Newtons per kilogram; r = Pearson correlation coefficient; p = significance level

Limb symmetry index

Although LSI values suggested near symmetry (99.0 ± 17.1% for hip flexion and 104.4 ± 10.9% for the medial hop test), reliability ranged from low to moderate. Critically, the MDC of the LSI was consistently large, ranging from 11.9% to 47.9% (mean ≈ 27%): range of motion, isometric strength, and the medial hop all showed an LSI MDC close to 27%, whereas the SEBTm was lower (≈ 14%) (Fig. 6). This indicates that only relatively large changes exceeded measurement error. Bland–Altman analyses revealed limits of agreement of ± 23% and ± 26% for intra- and inter-rater range-of-motion measures, respectively. Isometric strength LSI assessments demonstrated the widest limits of agreement (± 29% and ± 33%), whereas medial hop performance LSI showed comparable variability (± 24%). The narrowest limits of agreement were observed for the SEBT LSI (± 10% and ± 12%). Mean bias remained negligible across all variables (≤ |3.7|%). However, significant proportional bias was identified for hip internal rotation LSI in flexion (inter-rater reliability) as well as for the anterior and posteromedial SEBT LSIs (intra-rater reliability), indicating the presence of heteroscedasticity within these symmetry-based measures.

Fig. 6.

Distribution, reliability, and agreement of Limb Symmetry Index (LSI) measures across the test battery. For each test, the violin plot illustrates the distribution of individual LSI values (expressed as a percentage of the contralateral limb), with the mean ± SD displayed alongside. Relative reliability is given as the intra- and inter-rater intraclass correlation coefficient (ICC) with its 95% confidence interval, accompanied by the standard error of measurement (SEM) and the minimal detectable change (MDC). The colour of each ICC label denotes the reliability level: green, strong (ICC > 0.7); orange, moderate (0.5 ≤ ICC ≤ 0.7); red, weak (ICC < 0.5). Range-of-motion panels report active and passive movements separately: (a) hip flexion and (b) internal rotation in flexion. Strength panels correspond to (c) abductor, (d) adductor, and (e) flexor (rectus femoris) isometric strength. The Star Excursion Balance Test (SEBT) is shown for the anterior, posteromedial, and posterolateral directions, and the medial hop is shown below. For each variable, the right-hand Bland–Altman plots illustrate intra-rater (shaded panels) and inter-rater (white panels) agreement: each point represents one participant (x-axis, mean of the two LSI measures; y-axis, their difference). The solid teal line indicates the mean bias and the dashed violet lines the 95% limits of agreement (bias ± 1.96 SD); shaded bands show the corresponding 95% confidence intervals. The blue line is the regression of the difference on the mean (proportional bias), drawn solid when statistically significant (p < 0.05) and dotted otherwise. Bland–Altman analyses were performed on the available paired data (intra-rater, n = 20; inter-rater, n = 87–90 depending on the variable). Abbreviations: LSI, Limb Symmetry Index; ICC, intraclass correlation coefficient; SEM, standard error of measurement; MDC, minimal detectable change; SD, standard deviation; CI, confidence interval; SEBT, Star Excursion Balance Test

Discussion

The primary aim of this study was to evaluate the intra- and inter-rater reliability of commonly used tests in a return-to-sport test battery after hip injury. Only selected measures (active and passive hip flexion, isometric strength of the abductors, adductors and flexors, and functional tests such as SEBT and hop test) demonstrated sufficient reproducibility for clinical use. The study also provides reference values for healthy active adults to support postinjury interpretation and longitudinal follow-up. Moderate associations were found between hip strength (add/abd ratio) and hop performance. Finally, the Limb Symmetry Index, like the add/abd strength ratio, showed poor reliability, questioning its relevance as a clinical or return-to-sport indicator.

Reliability and clinical relevance of hip range of motion and strength

Our findings indicate that hip flexion, both active and passive, was the only range of motion measures demonstrating high reliability, with clinically meaningful thresholds for real change between 12° and 18°. This is particularly relevant because flexion is frequently restricted and associated with reduced quality of life in individuals with hip pain [39]. The literature also reports considerable heterogeneity in hip range of motion, largely attributable to methodological variability [40]. The observed mean difference of 12.5° between active and passive motion indicates a consistent deficit between active and passive range of motion, a pattern also reported in individuals with hip-related pain such as FAI [41]. In contrast, hip rotations showed excessive variability, limiting their value for isolated clinical decision-making, despite being commonly limited in impingement [7]. This was most pronounced for external rotation in flexion (ICC < 0.44), likely because this movement is particularly prone to pelvic compensation and because end-range was defined by a qualitative firm end-feel, which is difficult to standardize further within a clinically applicable battery.

Strength measures demonstrated high intra- and inter-rater reliability, supporting their clinical utility. MDC values ranged from 55 to 79 N, higher than previously reported thresholds [20], indicating that only substantial changes can be considered clinically meaningful. Thus, adequate relative reliability (ICC) is a necessary prerequisite for the clinical application of a test; however, it does not alone inform its capacity to detect longitudinal change. In our cohort, hip strength assessments demonstrated satisfactory ICC values while remaining associated with substantial MDC values (up to approximately 48%). These elevated MDC values may be partly explained by limited participant familiarization with maximal strength testing, which increases within-subject measurement error; repeated exposure to such assessments, particularly within the context of physiotherapy follow-up, may improve measurement reproducibility by reducing this within-subject variability, and thereby lower the SEM and MDC. Strength levels in our cohort were comparable to those reported in recreationally active adults [42], but remained substantially lower than those observed in trained athletes, where adduction forces can be almost doubled [43]. The adduction-to-abduction strength ratio (≈ 1.0–1.2) aligned with prior studies [9]; however, its poor reliability in our sample limits its clinical relevance, favoring the interpretation of absolute strength values rather than ratios.

Functional tests

The medial hop emerged as the most robust functional test, demonstrating excellent intra- and inter-rater reliability (ICC > 0.90) and an MDC of 18–26 cm, above which change can be interpreted as real [44]. Its normalized performance (1.0 ± 0.3) provides a practical benchmark for comparing patients with healthy adults [10]. The clinical relevance of this test is further supported by the moderate yet significant association between hip abductor and adductor strength (r > 0.51), explaining about 26% of hop performance variance. This suggests that the medial hop reflects both muscular capacity and motor control. These findings contrast with earlier studies reporting weaker associations (~ 7%) [45], likely due to smaller or less standardized samples. Moreover, the medial hop and its variants (e.g., triple hop) have previously been shown to detect deficits in patients with hip pathologies such as femoroacetabular impingement, labral tears, dysplasia, or osteoarthritis compared with controls, reinforcing its value for functional follow-up and return-to-sport decision making [3].

In contrast, the squat endurance test and single-leg squat demonstrated limited reliability, with high MDC values and substantial within-session variability, likely influenced by motivation, fatigue, or learning effects [46]. Finally, the modified SEBT showed good reliability (MDC 9–14 cm) but lower composite scores than those reported in more athletic samples [29], indicating that SEBT performance should be interpreted relative to the patient’s activity level. These results support prioritizing the medial hop and SEBT in hip rehabilitation protocols, while using squat-based tests with caution due to their variability.

Associations

Hip strength was a significant but partial determinant of lateral-task performance (~ 27% of variance), supporting targeted adductor and abductor strengthening. The large unexplained variance (~ 73%) reflects additional factors (neuromuscular control, coordination, balance, movement strategy); strength and functional tests are therefore complementary rather than redundant.

Limb symmetry index

The LSI is widely used to assess recovery and return-to-sport readiness [17, 47], but our findings reinforce existing concerns regarding its reliability and clinical relevance [15, 16]. A fundamental limitation of the LSI is conceptual: symmetry-based metrics assume the contralateral limb is a stable biological reference, yet the uninvolved limb is itself labile after unilateral injury, declining through detraining and compensation, but also adapting through cross-education during rehabilitation. The uninvolved limb may deteriorate, such that athletes meeting a 90% LSI still fail a pre-injury capacity benchmark [15], or conversely rise above preoperative values as a “moving target” [48]. The LSI thus confounds genuine recovery with bidirectional change in an unstable reference. Despite good reliability of several individual tests, the LSI consistently showed low reliability due to its sensitivity to the reference limb, error propagation from both sides, and marked heteroscedasticity, making the ratio increasingly unstable as performance or limb symmetry increases [49]. In our sample, an intra-session variability of ~ 10% was sufficient to alter clinical classification (e.g., LSI ≥ 90% vs. ≤ 80% without any true change). Only changes exceeding 27% (MDC95%) can be considered real, which greatly limits the usefulness of the LSI as a monitoring or decision-making criterion.

Clinical interpretation and functional relevance

These findings encourage clinicians to ground decision-making on the most reliable tests while integrating both their real-change thresholds and their functional meaning. Active hip flexion was the most robust measure (ICC 0.76–0.85; SEM ≈ 4°) and the most clinically relevant, as it correlates most closely with quality of life [50] and directly underpins many activities of daily living [51]; nonetheless, a change exceeding ~ 12° is required to be deemed real. Isometric strength tests were equally reliable but must be interpreted relative to the patient’s profile (age, sport, activity level), since abductor, adductor, and flexor strength underpins pelvifemoral stability and force generation during acceleration, deceleration, and change of direction [52]. The medial hop (ICC > 0.90; MDC > 26 cm) captures lateral force production, neuromuscular control, inter-segmental coordination, dynamic balance, and movement strategy, qualities transferable to cutting and pivoting - while the modified SEBT (MDC > 10 cm) reflects dynamic postural control. Clinicians should therefore favor these robust tests and require changes clearly exceeding the MDC. These MDC values may, however, be conservative, partly inflated by the heterogeneity of our recreational cohort, single-measurement ROM protocols, limited familiarization, and MDC% normalization, and are likely to narrow during longitudinal follow-up. Finally, each LSI measure, the adductor-to-abductor ratio, and squat-based tests (single-leg and endurance squats) show greater variability or limited agreement and should be interpreted with caution, as supporting rather than decisive criteria.

Beyond the findings directly supported by our data, the following points should be read as clinical recommendations extrapolated for practice rather than as study results. The LSI may be most informative in the early stages of rehabilitation, when large asymmetries exceed measurement error and can be reliably detected. As rehabilitation progresses, we would suggest that its interpretation follow a graded logic rather than a fixed cut-off. First, the conventional 10% rule (LSI ≥ 90%) may warrant reconsideration: as noted above, this threshold lies within the measurement error of most parameters, and in our healthy cohort most individuals naturally fell outside the ± 10% band; symmetry might therefore be better read as a continuum supporting clinical reasoning than as a binary pass/fail criterion. Second, the uncertainty of any single LSI value should probably be acknowledged: given the wide limits of agreement, a measured value of 85%, for instance, may correspond to a true value between approximately 70% and 100%, and a value of 60% to a range of roughly 30% to 90%. An LSI of 85% might therefore be interpreted neither as a clear failure nor as a fully reassuring result; rather, it would highlight the inherent uncertainty of interpreting a single symmetry value. More importantly, when symmetry values approach approximately 60–70%, measurement error may progressively outweigh the discriminative capacity of the ratio, which would further limit its usefulness for clinical decision-making. Beyond this range, clinicians could consider moving away from ratio-based monitoring and instead focusing on the longitudinal assessment of absolute measures. These measures could be interpreted in relation to their parameter-specific MDC values and compared against robust reference values expressed as percentiles and derived from targeted populations. Such an approach might provide a more precise and clinically meaningful framework for monitoring recovery and informing return-to-sport decisions. Ultimately, the development of predictive reference equations integrating individual characteristics such as sport, training level, age, height, and sex could offer more individualized benchmarks than symmetry-based metrics alone.

Strengths and limitations

To our knowledge, this study is the first to evaluate the intra- and inter-rater reliability of a comprehensive hip-specific test battery in accordance with COSMIN recommendations, while also providing initial reference values in healthy adults. A major strength of this study lies in its integrative approach, combining range of motion, isometric strength, and functional performance measures within a single standardized protocol. The use of clinically accessible tools and procedures enhances ecological validity and the transferability of findings to routine clinical practice. In addition, this study provides a detailed quantification of measurement properties, including intraclass correlation coefficients (ICC), standard error of measurement (SEM), and minimal detectable change (MDC), enabling clinicians to interpret both reliability and clinically meaningful change. It is also the first study to specifically evaluate the reliability of the LSI within a comprehensive hip-specific test battery, thereby extending current knowledge regarding the measurement properties of this widely used metric.

Several limitations should be acknowledged. First, all tests were administered within a single session in a fixed, non-randomized order. This was deliberate: the battery is intended for clinical use in this same sequence, providing a progressive warm-up, logical task progression, and standardization. Early-battery tests (range of motion, strength) are therefore unlikely to be affected by fatigue, whereas the end-of-battery functional tasks may be slightly affected despite controlled rest periods. Crucially, any such effect would concern only raw (absolute) values, since the identical order at every assessment keeps it constant across sessions and raters and thus leaves reliability estimates unaffected. We accordingly recommend that clinicians apply this same standardized protocol when comparing patient data with the present reference values, bearing in mind that if end-of-battery tests are used individually, outside the battery and its sequence, their raw values may reflect slight fatigue. Second, although the sample size (n = 94) allowed the establishment of preliminary reference values, it remains insufficient to define definitive reference values, and larger studies are required to confirm and refine these values. Furthermore, participants were healthy and physically active adults engaged in recreational activities (≥ 2 sessions per week), with none competing at an elite level. This profile reflects a large proportion of patients encountered in routine clinical practice but may be less representative of high-level athletes. It is also possible that reliability estimates may be higher in populations who are more familiar with maximal performance testing and with greater body awareness.

Conclusion

Range-of-motion measures, particularly active and passive hip flexion, demonstrated the highest levels of reliability. Strength tests showed good reliability but were associated with high measurement error, limiting their sensitivity to detect individual changes. Among functional tests, the medial hop demonstrated excellent reliability. In contrast, adductor-to-abductor strength ratios and the Limb Symmetry Index showed poor reliability and high MDC values (mean ≈ 27%): such error renders symmetry-based ratios unsuitable as standalone decision-making criteria, as a change must exceed roughly one quarter of the index to be considered real. Rather than relying on the LSI, clinicians should combine absolute values, each interpreted against its own MDC, with the normative reference ranges provided here (25th, 50th, and 75th percentiles), within a multidimensional assessment framework. This approach offers a more reliable and individualized basis for monitoring progress and informing return-to-sport decisions than symmetry-based metrics alone.

Abbreviations

ADD

Hip adduction

ABD

Hip abduction

ADD/ABD

Adductor-to-abductor strength ratio

ANT

Anterior

CI

Confidence interval

COMP

Composite score

ER

External rotation

ICC

Intraclass correlation coefficient

IR

Internal rotation

JASP

Jeffreys’s Amazing Statistics Program

LSI

Limb Symmetry Index

MDC

Minimal detectable change

MDC95%

Minimal detectable change at the 95% confidence level

LL

Lower limb

N

Newton

N·m·kg⁻¹

Newton-meter per kilogram

PL

Posterolateral

PM

Posteromedial

ROM

Range of motion

SD

Standard deviation

SEM

Standard error of measurement

SEM%

Standard error of measurement expressed as a percentage

SEBTm

Modified Star Excursion Balance Test

SLS

Single-leg squat

Biographies

Joévin Burnel

is a physiotherapist and PhD candidate at the Université Libre de Bruxelles (ULB), specializing in musculoskeletal assessment and clinical reasoning, with a research focus on hip-related conditions and the development of clinically applicable measurement tools.

Marine Gony

contributed as research assistants, supporting participant recruitment and data collection.

Romane Bourreau

contributed as research assistants, supporting participant recruitment and data collection.

Benoit Pairot de Fontenay

senior researchers and clinicians with expertise in musculoskeletal rehabilitation, biomechanics, and evidence-based physiotherapy.

Joachim Van Cant

senior researchers and clinicians with expertise in musculoskeletal rehabilitation, biomechanics, and evidence-based physiotherapy.

Authors’ contributions

JB: Conceptualization; Methodology; Validation; Formal analysis; Investigation; Resources; Data curation; Writing – original draft; Writing – review & editing; Visualization; Supervision; Project administration.JVC: Conceptualization; Methodology; Validation; Data curation; Writing – original draft; Writing – review & editing; Visualization; Supervision.BPF: Conceptualization; Methodology; Validation; Writing – original draft (contribution to manuscript development); Writing – review & editing (final approval); Supervision.MG: Investigation; Resources (participant recruitment); Data curation; Writing – original draft (initial drafting contribution); Writing – review & editing.RB: Investigation; Resources (participant recruitment); Data curation; Writing – original draft (initial drafting contribution); Writing – review & editing.

Funding

This study was supported by a grant from the Société Française de Masso-Kinésithérapie du Sport (SFMKS), 2025. This grant contributed to the overall support of the project, including publication costs. The funding body had no role in the design of the study, data collection, analysis, interpretation of data, or in writing the manuscript.

Data availability

The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was approved by the Hospital Ethics Committee of CHU Brugmann (Centre Hospitalier Universitaire Brugmann), Brussels, Belgium (Reference CE 2025/63; internal study number B0772025000042). All procedures were conducted in accordance with the ethical standards of the institutional research committee and with the 1964 Declaration of Helsinki and its later amendments. All participants received detailed written and oral information regarding the study objectives, procedures, and potential risks, and provided written informed consent prior to participation. Participants were informed of their right to withdraw from the study at any time without any consequences.

Consent for publication

Written informed consent for the publication of identifiable data and/or images was obtained from the participant.

Competing interests

The authors declare no competing interests.