Validity and reliability of Veloflex to measure active cervical range of motion in asymptomatic and symptomatic subjects

Subjects

Thirty-eight subjects without cervical mobility disorder (asymptomatic subjects) and twenty subjects with cervical mobility disorder (symptomatic subjects) were recruited. Asymptomatic subjects were selected from volunteers who had no history of head or neck pain (Visual Analog Scale (VAS) = 0) and have more than 75% of normal ROM (Swartz, Floyd & Cendoma, 2005) for all cervical movements. Symptomatic subjects were included if they reported limitation of any cervical movement, either due to an average pain of ≥ 3/10 at rest measured by the VAS or reported vertigo of ≥3/10 measured by the Visual Vertigo Analogue Scale (VVAS) lasting longer than 1 week (Rutledge et al., 2013; Dannenbaum, Chilingarian & Fung, 2019; Carrasco-Uribarren et al., 2021). They were excluded if they presented possible red flags, vascular problems, or a clinical history of previous vestibular disorders different to a benign paroxysmal positional vertigo (for patients with vertigo) (Rushton et al., 2014). All those who initially contacted the researchers to participate in the study were ultimately examined. All subjects provided written informed consent. The institutional review board of the University of Murcia approved the study (CEI-2263), following the ethical criteria established in the protocol.

Measurement instruments

Cervical range of motion device (Performance Attainment Associates, Roseville, MN, USA) consists of a plastic frame mounted on the subject’s head similarly to wearing optical glasses. This measurement system involves several anchors and three inclinometers located in three different areas to measure the ROM in sagittal, frontal and transverse planes. The first anchor consists of an arm placed on the bridge of the nose and ears like glasses, closed with an arch that fixes it to the occipital area. A second arm is placed in C7, which was located by asking subjects to flex, and extend followed by cervical rotation while palpating for the most mobile segment on the cervicothoracic junction region (Palmer & Epler, 1990). The embedded bubble inclinometer on this arm was used to ensure verticality. A magnetic collar is placed over the shoulder to take into account any rotation of the trunk. The two inclinometer dials that measure flexion/extension (sagittal plane) and side bending (frontal plane) are dependent on gravity. The third inclinometer was on the top part of the head measures the cervical rotation (transverse plane), measuring by means of its magnetic dependence to the collar (Fig. 1). This instrument considered the gold standard (Tousignant-Laflamme et al., 2013; Kubas et al., 2017; Rodríguez-Sanz et al., 2018; Feng et al., 2019), as its validity and reliability have been extensively studied (Tousignant et al., 2000, 2006; Lee, Nicholson & Adams, 2004; Audette et al., 2010; Carvalho et al., 2014).

Veloflex (Deportec, Murcia, Spain) is an optoelectronic system designed to measure ROM and feedback from it. As previously described in Medina-Mirapeix et al. (2019), VF includes three elements: an infrared camera on a tripod, a laptop, and markers to place on the skin. The camera is placed on the adjustable tripod at a distance of 1–1.5 m from the participant and is located at the height of the joint to be examined. By tracking the markers, VF can determine the angle formed by two bone segments (taking three reference points: a joint axis and two references), or between a bone segment and the horizontal/vertical line drawn on the joint axis. While for flexion/extension and side bending, the camera captures ROM from a frontal view, for rotation this record is from a superior view (Fig. 2). Two VF markers were used for each movement, one common for all movements located by means of a rigid headband in the center of the head. The second marker was placed on the ear for movements in the sagittal plane, on the chin for movements in the frontal plane, and in the center of the forehead for movements in the transverse plane.

The constant capture of the movement can be used to quantify the ROM or give feedback instantly from it. Depending on the proximity of the camera to the joint, the accuracy of the system has been reported between 0.1° and 1° (Peña García-Orea et al., 2018). It has been previously validated to assess ROM in joints of the upper and lower limb, showing excellent intra- and inter-examiner reliability (Medina-Mirapeix et al., 2019; Martín-San Agustín et al., 2019).

Procedures

For concurrent validity, all CROMs were simultaneously measured by the CROM and VF, except rotations (measured 5 min apart because they were not possible together). Intra-rater reliability was explored for both groups, between-days reliability for the asymptomatic group and within-day reliability for the symptomatic group (Fig. 3). Thus, an examiner (physiotherapist experienced in the VF and CROM) evaluated the participants in two sessions, with a 1-week interval between sessions for asymptomatic participants and with 1-h interval between measurements (to minimize the influence of their changes in symptoms on the device’s reliability) for symptomatic participants. For inter-rater reliability, a second examiner (newly graduated physiotherapist) without experience was instructed in the procedures and performed the same tests in the second session (Fig. 3). A research assistant recorded ROM values in order to avoid bias, reading the C-ROM and recording the evaluation of the VF’s software. Both examiners were unaware of the other’s scores. The order of examiners and tests were randomized.

The testing protocol was strictly identical for the two sessions. All tests required subjects sit on a chair. The VF was placed by physiotherapist (an experienced or inexperienced examiner) on adjustable tripod stand to 1 m distance away from participant and located at the height or above (for rotations) of the head (Fig. 2). Test positions and procedures were chosen based on goniometry manuals (Norkin & White, 2016). To test the movements, the physiotherapist placed the CROM and the markers on the skin and/or a head arch, and trained subjects using standardized instructions for each movement: directions, stabilizations required (examiner’s hands on subjects’ shoulders), pauses, repetitions and sequence. For symptomatic subjects it was asked that the movements did not have to cause pain or dizziness. Afterwards, subjects carried out five full movements to familiarize and warm-up tissues. After warm-up, VF was on and subjects were asked to move from the initial neutral position toward maximal range in a given primary direction, hold it three seconds and return to neutral position. The specific instructions for each movement were: (1) for flexion, “I would like you to tuck your chin then move your chin to your chest without poking your chin forward”, (2) for extension, “I would like you to look up toward the ceiling without letting your back lose contact with the chair”, (3) for side bending, “I would like you to keep your chin tucked and bring your left ear to your left shoulder, and then your right ear to your right shoulder. Please keep your shoulders down”, and (4) for rotation, “I would like you to keep your chin tucked, your eyes level and look over your left shoulder, then your right shoulder” (Kubas et al., 2017). After a 5 s pause this movement pattern was repeated twice for a total of three trials per movement pattern. After five seconds, the opposite direction was measured similarly. There was a 60-s rest period between cardinal planes in order to change camera position and markers.

Statistical analysis

Participant’s characteristics and ROMs were summarized using means and SDs or percentages, as appropriate. Validity is defined as the extent to which the method measures what it is intended to measure (Streiner & Norman, 2008). To evaluate the concurrent validity between the measurements of the two devices (VF and CROM), the Pearson’s product-moment correlation coefficient (r) was used (Lin & Yao, 2014). Validity is defined as the extent to which the method measures what it is intended to measure (Martín-San Agustín, Sánchez-Barbadora & García-Vidal, 2020). To evaluate the agreement between devices, Bland–Altman plots were created and were calculated: upper and lower limits of agreement (LoA), the mean and the standard deviation of the difference between the two devices (these concepts were called ‘bias’ and ‘imprecision,’ respectively) (Giavarina, 2015), and their respective percentage with respect to the CROM values. A priori, we established the acceptable LoA needed to recommend the new measurement system to be substituted for the previously used device. These limits were ±2.5° concerning CROM because this magnitude is the minimal standard error of measurement (SEM) of this device across the six movements (Fletcher & Bandy, 2008).

Relative reliability is the degree to which individuals maintain their position in a sample with repeated measurements, which was assessed using intraclass correlation coefficients (ICCs) (Streiner & Norman, 2008). Absolute reliability is the degree to which repeated measurements vary for individuals, expressed as the SEM and calculated as $SDx\sqrt{1-ICC}$, where SD is the SD of all measures from the subjects. We also calculated the minimal detectable change (MDC) with a 95% confidence interval, calculated as $SEMx1.96x\surd 2$. The reliability was classified as excellent (ICC > 0.90), good (ICC = 0.76–0.90), moderate (ICC = 0.51–0.75) and poor (ICC < 0.50) (Martín-San Agustín et al., 2020). All statistical analyses were performed using SPSS version 24.0.

Sample sizes for the asymptomatic and symptomatic subjects were calculated to determine whether the expected ICCs were significantly higher than 0.5 (with two sessions per participant) assuming a significance level of 5% and power of 90%. For asymptomatic subjects, the expected ICC was set at 0.8, since outcome variables were unknown, and the sample size was 30 participants. For symptomatic subjects, the expected ICC was increased to 0.9 since ICCs from asymptomatic were already known, the required number of subjects was 12 subjects. The WINPEPI program was used for these estimates (Abramson, 2011). However, ultimately, we included 5–10 more symptomatic subjects in the final sample in order to increase the study power.

Subjects

Asymptomatic subjects had an average age of 34.1 years with a BMI of 24.4 kg/m² and symptomatic subjects had an average age of 42.8 years with a BMI of 24.7 kg/m² (Table 1). Thirteen subjects presented limitation of CROM associated with pain (VAS = 4.6) and seven subjects associated with vertigo (VVAS = 4.4).

Concurrent validity

Cervical range of motion and VF showed excellent correlation for flexion/extension (r’s range = 0.993–0.998), side-bending (r’s range = 0.993–0.995) and rotation (r’s range = 0.960–0.974) movements (Table 2). ‘Bias’ (i.e. the differences between CROM and VF) for the six movements were close to 0°, and generally, these differences for each subject were located within the LoAs, as can be seen in the Bland–Altman plots (Fig. 4). Thus, both devices provided acceptable agreement of ROMs across movements with LoA− and LoA+ ≤2.30° and small mean ‘bias’ (≤1.29%) in all movements respect to measures with CROM.

Reliability

Table 3 and 4 show intra- and inter-rater reliability analysis data for the VF in asymptomatic and symptomatic subjects. Intra- and inter-rater reliability was excellent (ICCs > 0.98) in both groups. SEMs ranging from 0.72% to 2.38% in all subjects and the MDC ranging from 1.47° to 2.60° in asymptomatic and from 1.22° to 2.18° in symptomatic subjects.