Physical Therapy Korea

Stroke is the main cause of disability in adults and the most common life-threatening neurological disease [1]. The ability of stroke patients to control their trunk movement plays an important role in balance and postural control and enables selective trunk movement during static and dynamic posture control [2]. In addition, trunk control can be used as an important predictor of functional level in stroke [3-5]. Trunk movements are an important factor in activities of daily living, but they can be impaired by stroke. According to a previous study, it has been reported that chronic stroke patients show impairment in selective movement of the upper and lower trunk [6]. Trunk muscle weakness, abnormal trunk control, and a decrease in position sense have been reported to cause reduced trunk function, resulting in a decreased balance and increased fall risk [7-9].

Measuring range of motion (ROM) is an important part of rehabilitation. Accurate measurements of ROM can help in determining the kind and duration of the treatment to be prescribed [10-13]. To select an appropriate rehabilitation therapy, an examination of joint mobility is required. Thus, measuring the trunk ROM is important for physical therapists. Although spine movement can be measured accurately using diagnostic imaging methods, it is expensive and has the disadvantage of exposure to radiation. Recently, various tools, such as the plumb line, goniometer, and tapeline, have been used to measure spinal movement [14]. Although various methods exist to measure spine movement in clinical practice, there is no effective standardized method [15-18]. The goniometer is mainly used for measuring ROM in the extremities, because it is easy, inexpensive, and reliable [19]. However, goniometer measurements of the spine show lower reliability than those of the extremities [16,20-22]. The goniometer is widely used to measure the ROM of a joint, but it has limitations in that therapists use both hands when measuring ROM, and anatomical landmarks must be accurately identified. It is difficult to stabilize the stationary arm of the goniometer and perform measurements simultaneously; therefore, increasing the risk of errors in positioning and interpretation of results [19]. In addition, owing to the complexity and surrounding tissues of the spine, it can be difficult to palpate anatomical landmarks and set reference points. Because of these characteristics, measurement of spine movement using a goniometer is not accurate [19].

Although a goniometer is commonly used to measure joint ROM, more recently, smartphone (SP) applications have been widely used in clinical practice for measuring ROM. SP has the advantage that it can be measured quickly and easily [23]. According to a previous study, the reliability and validity of ROM measurement by an SP have been confirmed [24]. However, the measured value may differ depending on joint movements or SP applications [13], and it is necessary to determine the reliability and validity of each ROM measuring SP application when evaluating a patient with stroke [25-27].

It is difficult to reliably quantify and measure trunk active range of motion (AROM) [28]. There are studies that measured trunk ROM in healthy people [19,25,29]; however, there are no studies on the reliability and validity of measuring trunk rotation and lateral flexion using a digital goniometer (DG) and an SP application in stroke patients. Thus, this study aimed to determine the reliability and validity of measurements obtained using DG and SP applications for trunk rotation and lateral flexion in stroke patients.

1. Participants

At least 15 to 20 sample sizes are considered appropriate for reliability studies that collect continuous data [30]. Twenty stroke patients agreed to voluntarily participate in the study after experimental procedures were explained to all participants. The study was approved by the Yonsei University Mirae Campus Institutional Review Board (IRB no. 1041849-202108-BM-132-02).

The inclusion criteria were as follows: 1) physical examination confirmed hemiplegia, 2) within six months of stroke onset, 3) had the ability to understand and follow simple instructions, 4) capable of maintaining a sitting position independently for 10 seconds, and 5) no vertebral fractures and no history of orthopedic surgery [31]. Exclusion criteria were as follows: 1) subjects who risk of falling down due to inability to maintaining balance when trunk rotation and lateral flexion and 2) showed a serious pathology, such as surgery on the spine or a tumor on the spine. Table 1 shows the general characteristics of the subjects.

Table 1 . General characteristics of the subjects (N = 20).

Characteristic	Stroke
Age (y)	69.5 ± 12.7
Height (cm)	158.1 ± 7.7
Weight (kg)	58.8 ± 8.2
Gender
Male	7
Female	13
Pathogenesis
Hemorrhage	7
Infarction	13
Time since stroke onset (mo)	2.4 ± 0.8
Paretic side
Left	10
Right	10

Values are presented as mean ± standard deviation or number only..

2. Instruments

DG and SP were used to measure trunk rotation and lateral flexion.

To measure trunk rotation and lateral flexion, DG (iGAGING; iGaging Store, Los Angeles, CA, USA) was used to detect and show measurement values in units of 0.1° on the display. The intrarater reliability of the DG (0.99) showed excellent reliability for the knee joint [32]. The intrarater reliability for thoracic rotation of the goniometer was 0.97 [19].

Galaxy S9 SP (SM-G960N; Samsung, Suwon, Korea) was used to measure trunk AROM. The SP, with an Android operating system, was equipped with an accelerometer and a gyroscope sensor. The intrarater reliability of the SP (0.96–0.98) showed excellent levels of reliability for the thoracic spine in healthy subjects [19]. Applications used for the measurements were Compass (KTW Apps; AppBrain, Seremban, Malaysia) for trunk rotation, and Clinometer (Plaincode; PlainCode, Stephanskirchen, Germany) for trunk lateral flexion.

3. Procedures

Measurements of trunk rotation and lateral flexion were performed by a physical therapist with more than five years of clinical experience with stroke patients. The examiner practiced using the DG and SP applications for familiarization prior to collecting data from patients.

The order of the tests needed to be conducted on each patient were randomized using a website-based application (http://www.randomization.com). However, considering the stroke patients’ movement, the non-hemiplegic side was measured first, and the hemiplegic side was measured next. Once all measurements were completed by the examiner, the measurements were collected again 30 minutes later using the same method [19]. As a result, we provided data on intrarater reliability. Data were recorded by the examiner immediately in a Microsoft Excel (2016 ver.; Microsoft, Redmond, WA, USA) spreadsheet after the measurement for trunk rotation and lateral flexion. To prevent participants from falling down during measurements and to help them understand the movements required for obtaining measurements, the examiner demonstrated rotation and lateral flexion of the trunk, and the participants practiced these movements more than five times. To minimize the change in posture, the participants sat on the treatment bed as straight as possible and maintained the spine in a neutral position. The feet were placed at on the floor, and the knee and hip joints were maintained at 90°. To ensure the consistency of movements, the examiner provided the same verbal commands to each participant.

The DG axis was placed at the thoracic vertebrae 2 (T2) level. The stationary arm was positioned parallel to the ground and the moving arm was pointed at the scapular spine during trunk rotation. The stationary arm was maintained in a straight line with the starting position, and the participant was asked to actively rotate to the end range, and the examiner followed the rotation with the moving arm of the goniometer. When the participant reached the end range, the angle was recorded (Figure 1) [19,33,34].

Figure 1. Measurement of trunk rotation and lateral flexion with the digital goniometer and smart phone. (A) Trunk rotation with digital goniometer, (B) trunk rotation with smart phone (Compass), (C) trunk lateral flexion with digital goniometer, and (D) trunk lateral flexion with smart phone (Clinometer).

The SP axis was placed at the T2 level. The examiner measured the same posture using the same method used with the DG. During trunk rotation, the SP was firmly fixed and followed. The rotation angle was measured using the SP application Compass (Figure 1).

Trunk lateral flexion was performed in a manner similar to trunk rotation. The axis of the DG was placed at the lumbar vertebrae 1 (L1) level. The stationary arm was pointed vertically to the floor and the moving arm was pointed toward the upper thorax. The participant actively reached the end range and the angle was recorded (Figure 1).

The SP axis was placed at the T2 level. Trunk lateral flexion was performed in a method similar to trunk rotation. The lateral flexion angle was measured using the SP application Clinometer (Figure 1).

4. Statistical Analysis

Statistical analyses were performed using SPSS ver. 21.0 software (IBM Co., Armonk, NY, USA). Intrarater (intraclass correlation coefficient, ICC [3,1]) reliability was determined using an intraclass correlation coefficients (ICCs) with 95% confidence intervals (CIs). The ICCs were interpreted as follows: ICC values less than 0.5 (poor), values between 0.50–0.75 (moderate), values between 0.75–0.90 (good), greater than 0.90 (excellent) [35]. The standard error of the measurement (SEM) was calculated to evaluate measurement variability. SEM = standard deviation (SD) × $\sqrt{1-ICC}$ was obtained. Clinically important changes were calculated using the minimal detectable change (MDC). MDC⁹⁵ = SEM × $\sqrt{2}$ × 1.96 was calculated [36]. Pearson correlation coefficient was used to determine the validity of the measurements obtained from the DG and SP. The level of agreement between the DG and SP was shown by the Bland–Altman plot. The 95% limits of agreement (LoA) were calculated as the mean difference between measures ± 1.96 × SD [37].

1. Trunk Active Range of Motion

The mean ± SD were calculated for rotation and lateral flexion. The directions of rotation and lateral flexion were classified into hemiplegic and non-hemiplegic sides (Table 2).

Table 2 . Active range of motion for trunk rotation and lateral flexion (N = 20).

Side		Rotation (°)	Lateral flexion (°)
Hemiplegic	Digital goniometer	18.05 ± 10.15	18.19 ± 18.19
	Smart phone	24.00 ± 14.61	25.20 ± 11.10
Non-hemiplegic	Digital goniometer	23.87 ± 9.91	16.87 ± 8.77
	Smart phone	31.90 ± 15.67	21.75 ± 11.08

2. Reliability Analysis

The intrarater reliability of trunk rotation consistently had high ICC values for both DG and SP application (Compass) (Table 3). The examiner’s ICC value was between 0.96–0.98 for both instruments, indicating excellent reliability.

Table 3 . Intrarater reliability of movement to trunk rotation using a digital goniometer and smart phone.

	Hemiplegic				Non-hemiplegic
	ICC (95% CI)	SEM (°)	MDC95 (°)		ICC (95% CI)	SEM (°)	MDC95 (°)
Digital goniometer	0.98 (0.96–0.99)	1.44	3.99		0.96 (0.90–0.98)	1.98	5.49
Smart phone	0.98 (0.95–0.99)	2.07	5.74		0.98 (0.95–0.99)	2.22	6.15

Intrarater reliability of trunk lateral flexion consistently had high ICC values for both DG and SP applications (Clinometer) (Table 4). The examiner’s ICC value was between 0.96–0.98 for both instruments, indicating excellent reliability.

Table 4 . Intrarater reliability of movement to trunk lateral flexion using a digital goniometer and smart phone.

	Hemiplegic				Non-hemiplegic
	ICC (95% CI)	SEM (°)	MDC95 (°)		ICC (95% CI)	SEM (°)	MDC95 (°)
Digital goniometer	0.98 (0.96–0.99)	2.57	7.12		0.97 (0.94–0.99)	1.52	4.21
Smart phone	0.96 (0.91–0.98)	2.22	6.15		0.96 (0.91–0.98)	2.22	6.15

3. Validity of Measurements

Pearson correlation coefficient showed a strong and significant correlation between DG and SP application (Compass) (hemiplegic side: r = 0.95, p < 0.001; non-hemiplegic side: r = 0.90, p < 0.001).

The Bland–Altman plot was used to graphically represent the level of agreement between the two instruments (Figure 2). The mean difference between the measured values on the hemiplegic side was 5.9° ± 5.87°, and the upper and lower LoAs were 17.41° and –5.61°, respectively. The mean difference between the measured values on the non-hemiplegic side was 8.03° ± 8.08°, and the upper and lower LoAs were 23.87° and –7.81°, respectively.

Figure 2. Bland–Altman plot indicating mean and difference and 95% limits of agreement between measurements from the smart phone (Compass) and digital goniometer for trunk rotation.

Pearson correlation coefficient showed a strong and significant correlation between DG and SP (Clinometer) (hemiplegic side: r = 0.88, p < 0.001; non-hemiplegic side: r = 0.78, p < 0.001).

The Bland–Altman plot was used to graphically represent the level of agreement between the two instruments (Figure 3). The mean difference between the measured values on the hemiplegic side was 7.01° ± 5.34°, and the upper and lower LoAs were 17.48° and –3.46°, respectively. The mean difference between the measured values on the non-hemiplegic side was 4.88° ± 6.92°, and the upper and lower LoAs were 18.44° and –8.68°, respectively.

Figure 3. Bland–Altman plot indicating mean and difference and 95% limits of agreement between measurements from the smart phone (Clinometer) and digital goniometer for trunk lateral flexion.

This study aimed to investigate the intrarater reliability and validity of the DG and SP when measuring trunk AROM (rotation and lateral flexion) in stroke patients. As a result, both the DG and SP showed excellent intrarater reliability. In addition, there is a strong and significant correlation between the two instruments.

The normal ROM of rotation and lateral flexion in the trunk has been reported to be 40° and 45°, respectively [38]. However, in this study, trunk rotation was measured at 18.05° with DG and 24.00° with SP on the hemiplegic side and 23.87° with DG and 31.90° with SP on the non-hemiplegic side. In addition, trunk lateral flexion was measured at 18.19° with DG and 25.20° with SP on the hemiplegic side and 16.87° with DG and 21.75° with SP on the non-hemiplegic side. Consequently, in our study, trunk ROM (rotation and lateral flexion) showed decreased values compared to normal ROM. During rotation, the non-hemiplegic side showed a larger value than the hemiplegic side. However, in lateral flexion, the hemiplegic side showed a larger value than the non-hemiplegic side. The asymmetrical sitting posture of patients with stroke should be considered when determining ROM measurements [39]. In addition, the stiffness and spasticity of the trunk muscles on the hemiplegic side could limit lateral flexion to the non-hemiplegic side [40]. The eccentric antagonist functions of opposite-side lateral flexion of the trunk are extremely important [41]. During lateral flexion to the non-hemiplegic side in stroke patients, the eccentric antagonist functions of the trunk muscles on the hemiplegic side may be insufficient. For these reasons, it is considered that the lateral flexion of the non-hemiplegic side is lower than that of the hemiplegic side. Trunk rotation had higher mean values with the SP than with the DG. The SP showed a higher mean value because it was closely fixed to the spine, and the moving arm of the DG was placed on the scapula. Similarly, the trunk lateral flexion showed higher mean values in the SP than in the DG. The reason for the difference in ROM in lateral flexion is thought to be the difference in the position of the axis between the two measuring instruments. When measuring lateral flexion, the DG axis was placed at L1, which was lower than where the SP axis was placed. These differences could explain why the ROM of SP measurement was greater than that of DG. However, there is a methodological limitation in that the positions of the axis of the two measuring tools were not equal.

In trunk rotation AROM of the hemiplegic and non-hemiplegic sides, DG (ICC = 0.96–098, 95% CI [0.90–0.99]) and SP (ICC = 0.98, 95% CI [0.95–0.99]) showed excellent intrarater reliability for both instruments. In addition, in trunk lateral flexion AROM of the hemiplegic and non-hemiplegic sides, DG (ICC = 0.97–0.98, 95% CI [0.94–0.99]) and SP (ICC = 0.96, 95% CI [0.91–0.98]) showed excellent intrarater reliability for both instruments. According to a study on thoracic rotation in normal subjects, the intrarater reliability of thoracic rotation was universal-goniometer (ICC = 0.97) and an SP (ICC = 0.96–0.97) [19]. In addition, according to a study on lumbar flexion and extension in normal subjects, the intrarater reliability of lumbar flexion was ICC = 0.85–0.89 with an inclinometer and ICC = 0.92 with an SP. The intrarater reliability of lumbar extension was measured using an inclinometer (ICC = 0.88–0.92) and SP (ICC = 0.91–0.92) [29]. The results of this study showed that both the DG and SP had excellent intrarater reliability for trunk rotation and lateral flexion in stroke patients, similar to measurements in normal subjects. Therefore, when measuring trunk AROM in clinical practice, both DG and SP are useful measuring tools available. According to a previous study, the smaller the SEM, the higher the reliability of the measurements [42,43]. In SEM, < 10% of the average score, and in MDC, < 20% of the highest score of the measured values have been reported to be reliable [42,44,45]. The results of this study suggest that SEM and MDC of trunk rotation and lateral flexion were reliable in both the DG and SP application.

Between the DG and SP application (Compass) in trunk rotation, r = 0.95, p < 0.001 on the hemiplegic side and r = 0.90, p < 0.001 on the non-hemiplegic side were observed. Between the DG and SP application (Clinometer) in trunk lateral flexion revealed r = 0.88, p < 0.001 on the hemiplegic side and r = 0.78, p < 0.001 on the non-hemiplegic side. Trunk rotation and lateral flexion showed significant and strong correlations, which might explain the concurrent validity of DG and SP applications. The level of agreement between the two instruments was determined using Bland–Altman plots, and the LoA was calculated. For trunk rotation, the hemiplegic side was 23.02° (LoA –5.61°, 17.41°) and the non-hemiplegic side was 31.68° (LoA –7.81°, 23.87°). In trunk lateral flexion, the hemiplegic side was 20.94° (LoA –3.46°, 17.48°) and the non-hemiplegic side was 27.12° (LoA –8.68°, 18.44°). However, according to a previous study, if there is a high correlation, it does not mean that between the two measurement methods consent [37]. Considering the mean values of trunk rotation and lateral flexion compared with the LoA values, inconsistencies can occur when two measuring devices are used interchangeably; thus, it is recommended to consistently select one measuring device during measurement. However, according to the clinical judgment of the examiner, these differences may vary. This study measured trunk rotation and lateral flexion in stroke patients; it cannot be directly compared to the validity in normal subjects. However, according to a similar study, thoracic rotation ROM was measured using a universal-goniometer and an SP. As a result, concurrent validity for thoracic rotation was strong (r = 0.835) [19]. Lumbar flexion and extension ROM were measured using an inclinometer and an SP. As a result, the concurrent validity for lumbar flexion (r = 0.85) and extension (r = 0.91) was strong [29]. However, radiographic analysis should be considered when measuring the spine since radiography is the gold standard for spinal analyses. Therefore, in future studies, it is necessary to confirm the validity of radiography and both instruments (DG and SP) [46,47].

This study has a few limitations. First, there could be a problem of reliability due to bias incurred during measurement. The study was conducted by a single examiner. Examiner and recorder were not divided separately; the examiner could memorize the approximate ROM value, and bias could occur during the next measurement. In future studies, it will be necessary to divide the examiner and recorder to increase reliability in obtaining measurements. Second, although education and practice were conducted before the measurement, the compensatory movements of the stroke patients could not be completely controlled. In addition, compared to healthy subjects, stroke patients show compensatory movements as their trunk movement increases, which may cause differences of measurement between both instruments [19]. Therefore, one more examiner is needed for safety and limiting the compensatory movements. In future studies, to increase the reliability of the study, additional studies should be performed to eliminate bias and compensatory movements when measuring trunk AROM. Third, in measuring the movement of the spine, measurement values may be different depending on the measuring tools (e.g. Inclinometer); however, only DG and SP were used. Also, in the trunk lateral flexion measurement, the axis positions between the DG and SP are different, there is a limitation in direct comparison. In future studies, it will be necessary to conduct using the standardized clinical methods and various measuring tools.

Reliable methods for measuring trunk AROM are insufficient and come with limitations for use in stroke patients. This study measured trunk rotation and lateral flexion using a DG and SP, which are mainly used in clinical practice. When trunk rotation and lateral flexion were measured, both instruments showed excellent intrarater reliability. Therefore, when measuring trunk AROM in clinical practice, physical therapists can use DG and SP. Although, considering the levels of clinical agreement, DG and SP may not be used interchangeably for measurements.

None.

None to declare.

No potential conflict of interest relevant to this article was reported.

Conceptualization: HP, UH. Data curation: HP. Formal analysis: HP, UH. Investigation: HP. Methodology: HP, UH, OK. Project administration: HP, UH, OK. Resources: HP, UH, OK. Supervision: HP, UH, OK. Validation: HP, UH, OK. Visualization: HP, UH. Writing - original draft: HP. Writing- review & editing: HP, UH, OK.